📚┃suno-school

[cyan jewel]

The song was made on suno

I just also made a lyric sheet for the song

[floral python]

but its not about it...hmmm

[cyan jewel]

i... sure fine whatever.

[floral python]

is what it is

[cyan jewel]

whatever you say your highness. my b for wanting to do something extra besides prompt and print...

[final hazel]

It's fine @cyan jewel ... It's suno related 🙂

[floral python]

I kinda figured that

yup thats all I do..ya got me

Im pretty damn good at it too

[humble verge]

there are times where i want to take a motherflipping flamethrower to this editor. I swear it was designed for people to torch all their credits and get zip in return.

[polar spade]

Can I show you something better than I showed you?

I dont necessarily think of something working better. So, no, not really -- to answer your question. There are alternatives, of course, which anyone picks up after awhile from experimenting.

[cyan dagger]

Ooo aight aight thanks a lot really

[hexed venture]

Heya - When i create my ambient music, it usually comes with a good soft and haunting female vocal. Relaxing to listen to, with a feeling of solitude. With the new song im creating the lady singing keeps being like opera inspired. She sings WAY to high notes and sustain them for to long. Any way i can make sure to avoid this?
I've tried putting this before the lyrics; [soft haunting female vocal]

[cyan dagger]

Si I tried your idea and frankly if I don’t put any lyrics it works insanely well tho the only issue is that it can’t do two different pitch one so it’ll only be some kind of solo

[polar spade]

I did not offer an idea of how to do duets, did I ? My samples were showing its difficult to get anything more than a vocoder effect, so far that Ive tried. Im glad something works for you ,in one way. Build from that, I guess.

You can search "duet" in this channel , if looking for ways to have two people talking/ singing and there are over 400 conversations of different ways to do duets. They all say theirs works best, so try a few recent ones. Setting pitch, is not consistent, no, you're right.

Using edit UI and extensions , along with adding personas of different vocals [thus different pitches] within each extension, is a better means of controlling the conversation.

[near flume]

use [Spoken] in front of each line you want spoken.

Long embellishments depend on your punctuation, genre, and cadence of the lyrics. If your cadence isn't the same Suno tries to balance the length.

[near flume]

try airy, haunting, alto female vocals.

[long pivot]

Same. I was writing a lullaby and the woman kept hitting high notes and doing runs. Which is not very soothing.

Male vocals have more range. All of the women are alto or soprano. No tenors or bass vocals for women.

[polar spade]

You try contra-saprano? Even "kids song" might raise the pitch.

[tranquil leaf]

And try tags like mellow, calm, etc... or even specify pitch-shifted down.

[polar spade]

Different lyrics? Ytou may need to tweak the style prompts each time.

Add "ambient" twice instead of once in the prompts.

How about "dreamy" ?

Even haunting could mean silently dreamy like a soft voice, or loud and frightening like a banshee. Suno might pick a new definition with a new song.

[long pivot]

I was using "melodic, calm, soft, nursery rhymes"

[wispy laurel]

I run into this a lot too
[Low, soft, relaxed vocals] helps, [gentle] helps
[Soft rap] can give a more laid-back, matter-of-fact vocal delivery with less "showing off", usually it won't actually result in rap but more of a half-sung, half-spoken style
Exclude "sustained vocal notes", "high pitched vocals" and "powerful vocals" too, Exclude in general doesn't work very well but it can help to reinforce the above
If you're using a persona who usually sings in the way that you're trying to achieve, you also can put [use the persona's voice and vocal style] right at the start of the style prompt

[polar spade]

By the way, I should mention "vocaloid" is another prompt that goes in the direction you're looking for.

[cyan dagger]

Oh right

[cyan jewel]

if you want lower register for fem you need to go with mezzo-soprano, or contralto. also use additional vocal technique tags

[cyan dagger]

Man vocaloid is the trick !

https://suno.com/song/27a4e490-37b4-4a9d-a3a1-17764cde397e

[tidal hare]

I’m working on improving a song I created by first recording with acoustic guitar, then using the cover function to add lyrics and instruments. I’ve managed to create a version I’m almost happy with, but I’m wondering: is there any way to prompt the vocal melody to shift certain words slightly back or forward, or to adjust note lengths (longer/shorter) to create more natural pauses?

[wispy laurel]

Punctuation and spacing in the lyrics can influence this, but it's hard to control it precisely, especially when using Cover on an already-established melody
Commas, ..., -, even just adding extra spaces between words can sometimes do what you want them to
Moving or adding line breaks, or even running words together without spaces, also might or might not help

[cyan dagger]

Dasvon idk what to tell u u really really REALLY made it so better than before

Thank you so much much much

[polar spade]

Now when I set display to full acreen and fonts to medium, im only showing Discord in the upper left quadrant of my screen. All other sites are fine. But I think its my cover on my phone screen. One more thing to fix.

[tidal hare]

Thanks for good answer! I should also use that more to try to control how the vocal melody will turn out straight from the guitar recording instead of trying to make a cover of the result i'm almost happy with?

[turbid pagoda]

im on this page and im trying to download my song with the new chances but cant seem to do that can someone please help me. thank you

[wispy laurel]

Try both ways. You could cover the current version and alter the lyrics to add/remove spaces and punctuation, and also try making a new version with that stuff in place from the start, and see which works better.
Another option is, record yourself singing it with the exact melody that you want. Doesn't matter if you can't sing well or if your voice just isn't right, you can potentially cover your recording using a persona that has the right voice. There are ways of changing the voice while keeping the same melody and timing. That's a bit more involved and can sometimes burn through a lot of credits though so I wouldn't recommend that unless other methods fail

[tidal hare]

Great advice, thanks again😁 😁

[cyan jewel]

?

[polar spade]

Im mostly just talking out loud. 😆 but its annoying, until I figure out why.

[cyan jewel]

oh ok. was just weird to get pinged by that lol

[small violet]

I find it works sometimes, or maybe that's just because I explain it further in the positive prompt anyway.

[wispy laurel]

I find it can help to reinforce positive instructions, if you were to exclude saxophone and you elsewhere have a clear and specific list of instruments that doesn't include saxophones then it can help to make sure that instruction is followed
It does seem to heavily rely on being supported by positive prompts though. And even then, Suno tends to just straight up ignore exclusions if you try to exclude something that it thinks is "supposed to" be in that style of music (try getting it to generate a rock song without a bass guitar)
I'm hoping that in future updates the Exclude prompt might be weighted a bit more heavily to make it more useful but as it is, it doesn't really seem to do much most of the time

[granite pawn]

How do you get suno to start the song without an instrumental intro

[sweet siren]

I know this is offtopic, but had to say it. what you said: "Im mostly just talking out loud. 😆 but its annoying, until I figure out why." That is actually genius for use in a song!
And now I am here, thanks for all your helpful comments and info that you are sharing!

[stuck dust]

This seems to be work. You can change it to reflect what you are going for
[Cold Start, starts right away with powerful vocals immediately when song starts]

[quiet solar]

If you do go the route of singing over the top, you may want to turn up audio influence to at least 60%-70% to make it more likely that Suno doesn't simply overwrite your cadence and timing with what it thinks would suit better.

If you get the side-effect that Suno then only wants to replicate the guitar and not add other instruments, be clear about what instruments you want added and nudge up the style prompt to offset it.

Good luck

[granite pawn]

Thanks i will try it

[granite pawn]

Yeah that did not work

[feral skiff]

Hey. Im trying to cover a folk music to death metal, but it seems like i cant get growling to work. Anyone else had this problem?

[scarlet bison]

Not sure about growling, but I did "Three Little Pigs" metal tune, LOL. I am linking, if you would like to see the closest I have come? But, it might make you laugh a little. HaHa, Let me know what you think of the little feller. HaHa

https://suno.com/s/raGbhI9NhptWogFW

[warm prairie]

Hi everyone 👋,
I’m working on prompts to create 1980s funky music and I’d love to refine them with the help of experts. I’m aiming for that classic vibe from bands like Change, The Cool Notes, D Train, SOS Band or Central Line, with strong bass lines and vintage synths (for example, Yamaha DX7 or similar).

👉 Who here has experience in this style and could advise me or share more polished prompts to really capture that authentic groove?

Thanks in advance! 🎶✨

[hidden yarrow]

You can always cut the intro in editor, if the verse has a good start.

[flint kraken]

Hey, I'm not sure if this has been covered, but with duets is there a way in the lyrics section to force a line to be sung by one singer and not the other? i've tried [Singer 1] [Singer 2] etc but it still seems to just alternate between them.

[fathom siren]

Not really, but you can really increase the odds by both labeling, and using a descriptor that is highly weighted towards one gender or the other.

[flint kraken]

ahh bugger, I was hoping that wasn't going to be the answer.. i might have two do two versions one male, one female, and then slice them together in audacity.

[fathom siren]

[Singer 1: Sultry flow]

Something like that is going to be highly weighted towards a female vocal, more so then just singer 1.

[flint kraken]

ahhh... that's a good top!

tip rather.

[fathom siren]

Also never hurts to throw [Singer 1: female vocals]

[flint kraken]

It is kinda funny hearing a the male voice sing "You're not the boy that i once knew" to the female part

Very progressive though.

[fathom siren]

haha ya I have had some bangers that I wish I didn't use some identifying gender phrasing in the song

[devout sinew]

Hi everyone, I need help with pronunciation control in Suno.

I want the acronym L.Sp.L to be sung as alphabet names:
• “L”
• then “S” and “P” back-to-back with almost no pause (said as “ess pee,” not the blended consonant “sp”)
• then “L”

When I type “Sp,” Suno blends it into “sp.” When I type “S-P” or “S P,” Suno inserts too much of a pause. Is there a prompt trick or phonetic spelling that forces letter-name pronunciation with a very small gap between S and P, without fusing them into a syllable? Examples of punctuation, IPA, or markup that worked for you would help a lot.

Optional test lyric line to illustrate:
“We carry the name L.Sp.L tonight.”

One-liner version
How can I get Suno to sing “L.Sp.L” as letter names “L, S, P, L,” with S and P said back-to-back as “ess pee” with almost no pause, not the blended “sp”? Any phonetic or punctuation tricks to force alphabet-name pronunciation?

[flint kraken]

Have you tried dropping the tonight part onto the next line?

[wispy laurel]

Ell-espee-ell
Then capitalise whichever syllable you want the emphasis on

[flint kraken]

oh? it takes that into account? Awesome!

[wispy laurel]

Sometimes it does. Not 100% reliable but mostly it works (in my experience anyway)

[devout sinew]

This is actually quite helpful-thanks.

[honest heath]

I hate it when suno pusts the em-PHA-sis on the wrong sy-LA-ble

[wispy laurel]

Yeah that's almost as bad as when it mispronoonsies things
In English it just sounds a bit weird but I also write in Arabic sometimes, and getting one syllable wrong can change the meaning completely
Although I might just leave it that way rather than fiddling around trying to fix it, on the basis that 95% of listeners probably don't speak Arabic anyway 😂

[echo crescent]

I was having trouble getting it to say complicated. lickated....

[honest heath]

I spent 300 credits on the word defiance the other night 😄

[wispy laurel]

"Spectres"
Damn AI gave me every single possible way of pronouncing it, except right 😂
Speck-trees. Speak-ters. Sceptres. Spuckties. Speesters.
Oh my god fine I give up, I'll spell it SPECK-ters instead
"OHH you mean 'spectres'"
F U Suno 🤣

[sweet siren]

That's funny. I made a song with some Arabic words in it 2 weeks ago.haven't released it, but I like it. I am pretty certain it pronounces it wrong. 😀

[humble verge]

does SUNO have something against Hammond organ? I've tried like 200 songs now with Hammond organ featured prominently in lyrics and styles prompts. SUNO hasn't given me Hammond even once. Also tried endless variations on 'gospel style electric keyboard', 'electric organ' etc etc.

[urban zephyr]

I'm trying to write a musical and I upload my musical sketches using the add audio prompt which works great for getting the theme turned into a song. My question is how can I prompt suno and give it multiple audio themes that could be interwoven into a song with different singers singing the different themes?

[slow edge]

I managed it with this one, https://suno.com/s/umWRHEPsnk6Kol4l

but it did require, multiple versions until I got the start version right, plus a few extensions, noting where it consistently got the wrong singer, then adding an extra line in before the error. For the alternate singer to sing, or just reordering the lines.

so if you want him singing first. But constantly got her. Then write a line or verse for her here, then his. I had to do that a few times on the extensions. But it did cost quite a few credits and time to get the desired effect. And correct Singer, singing. There still was one voice error, where the male voice says two words of her line, before she takes over. which decided to leave in.

Why...because the editor refused to get right any of the many re-generate sections it produced.
it repeatedly skipped to the end verse, but the video is messed up, even trying to redo/realign the lyrics in editor, it just regenerated the repeated end lyrics error many times, so video just purely shows those that lines and none of what it should do.

Even doing a cover or remaster, didn't change the fact this lyric issue repeated, that wasn't an error on the lyrics shown on the song, but is on the video.

[rain lance]

Anyone have tips on how to use Extend properly? I'm getting a lot of "Credits Refunded"
I had a generation that had the perfect sound, but the 4 minute song just ended at 1:30 so I'm trying to finish it.

[slow edge]

I usually try to extend from a bit further back from the very end.

So if you have a great version generated but it ends well before it should. Let's say at 1:30, then try to extend from say 1:25 or if the final verse it starts singing, starts at 1:21. You can either, start the extension from 1:21 or try 1:20 if enough space between previous verse and last it sung.

If it's purely instrumental tune or at an instrumental break in the song, then pick a point at least a few seconds from the end to extend from.

If you extend from 1:30 where the song exactly ends, it will often just create a couple of seconds of ending, that's why the credit refund, or you get two 8- 10 second clips of nothing much more than a fade out or couple of notes then silence and it eats your credits...

Sometimes you have to extend from further back, say at 1.14. It can be just a second difference to extend from that makes it work how you expect.

[slow edge]

Maybe it doesn't know specifically what a Hammond organ is... You could always ask an LLM that you prefer to use, how to describe the sound it makes and try that.

[modest venture]

Can you use real voice samples to train a unique AI vocal on Suno?

[sweet siren]

That is really well done. I like this!

[stuck dust]

Where do you want to Extend from? Select the timestamp you want to Extend from. If it's too close to the end where it is already fading out it may fail to generate a long enough section so if you are getting "Credits Refunded" it's usually because it's too close to the end of the song.

[rain lance]

Thanks. Removing the last chunk helped. I was also trying to use an Perosona and I've been having better luck without it.
Now I'm having the issue of the song not following my prompts as well as they were before, and the extension's tone not matching the original's too well after a bit.

[sweet siren]

I am confused now. I have done a remix before. But when I add my persona the name of the song on the top disappears. Will it still take is as reference? Maybe it was always this way and I am becoming insecure now 🙂

[tidal hare]

Wow, this really helped! Actually gives me almost full control on melody and vocals 🤯

[stuck dust]

If you are Extending, you shouldn't need to apply a Persona as it will now be able to go off of the context of the song it's extending from.

[humble verge]

I've tried. Lots of lots of alternate ways to describe an electric organ. SUNO simply doesn't seem to know what that is. At least not in the context of blues music. Maybe if I was doing a gospel.

[rain lance]

Yeah, that's the tough part. It's always very close, but only a few out of all my generations are picking up on some key themes. I guess I just have to keep regenerating until I can get it just right.

[quick sonnet]

It's a gamble, it can sometimes be difficult to get the whole song performed by a specific voice.

[quiet solar]

There's no straightforward way to do what you're attempting - Suno can only work with a single audio upload, and there's no way to assign part or all of the upload to a specific part of the song.

If you know what those musical themes are already, and they're sutiable to blend together, then your best bet is recording them all into a single audio upload, handing over to Suno and crossing your fingers. If you're trying to create multiple separate parts in Suno that need to ultimately be converged into a single unified and coherent whole in the end, but you don't know what that whole looks like yourself to start with...whew, good luck.

[quick sonnet]

Is there an instruction to make a chorus really a chorus and not just the refrain?

[quiet solar]

Do you have an example of the problem?

[urban zephyr]

Yeah I figured I'd have to just write the song with all the themes and then upload. But if like a way to assign different singers to each theme. I'll try creating an instrumental first and then adding lyrics and trying to assign or rewrite lyrics to march the right singer/theme section.

[quick sonnet]

I don't have a problem, I would just like to give instructions that the chorus is sung by several voices, which is also called a choir, while the rest is sung by a single voice.

[quiet solar]

Advance warning; working with multiple singers in the same song with Suno is pain personified. Expect to have to do a lot of reloads if you need to switch singers mid-song, or have them duet. Good luck.

[urban zephyr]

Yeah I'm dealing with that now on a duet and some songs with "Greek chorus". My fav Suno mistake was one song where Suno just added lyrics like "Greek chorus sings" to replace actual lyrics in that section. Lol.

[quiet solar]

I don't think there's an easy way to instruct Suno to do that, but you could play around with prompts in the lyrics at the start of the chorus like "multiple harmonies", "layered harmony", "multiple singers" until it gives you something that works.

I'd also suggest telling Suno in the style prompt that the chorus should have multiple singers, or a choir, or whatever descriptions will help it look for references in other types of music that have done the same thing you're after.

[quiet solar]

Gotta love it when it sings your prompts. xD

[quick sonnet]

I don't want the entire song to be sung by a choir. The choir is simply meant to highlight the chorus from the rest.

[quiet solar]

I get that - what I'm saying is that because that's fairly complicated to get Suno to do, you want to reinforce that's what your asking for as both a style prompt and a lyric prompt.

[pine island]

Ciao! Is possible to change the key of a generated song in zone way?

[quiet solar]

After it's been generated? Not easily. If you want that specific version of the song, you'd be better off taking it out of Suno into a DAW or audio editor like Audacity.

It is possible to shift the key if you cover the song using the same style prompt, specify the new key (might take a few tries to get right) and turn up audio influence and reduce weirdness so that it doesn't change the song too much. But expect that it will change the song, for better or worse.

[storm pawn]

Guys if I have a song that turned out quite well, but I would just like to add a Mexican trumpet to the main melody loop, what would be your approach? Creating a cover and in style section just mentioning the trumpet, or copying the old style, adding trumpet to it and covering, or making new track from scratch?

[wispy laurel]

This is hard to do without changing the original song more than you wanted.
The way I'd go about it would be, cover the original song, add VERY specific instructions about the trumpet in both the lyrics and the style prompt, turn Weirdness way down to probably about 3%, Audio influence up to the high 90s and Style influence, at a guess, around 80
In theory that should give you something very close to the original, with a strict (invisible and internal) limit on how much it can differ from the original. And since you've specified an extra instrument that's not in the original, and given very clear and repeated instructions reinforcing that, hopefully Suno will pick up on that, add the trumpet, and that should take up most of its internal "quota" of variation from the old version.
This is not even close to 100% reliable, you might need to make quite a few attempts. And be prepared, you might just have to accept one that just about managed to put a trumpet in the right place, then use Replace Section to alter that specific bit to be closer to the way you want it. I've done this myself to crowbar a guitar solo into a song that didn't previously have one, it does work but, not easily

If it kind of works, you can try fiddling with the Weirdness/Style/Audio numbers to refine it. But be prepared to spend a lot of time and credits on it

[storm pawn]

that sounds like a plan, thank you!

[vague kraken]

Hello everyone,
I’m working at a radio station (RADIO ASWAT) and I’d like to know if Suno can be used to create radio trailers / station IDs with a natural, clear, non-singing voice (voice-over style).
So far, I only see music/singing voices, but not plain speech.
Does anyone know if this is possible, or if it’s planned for the future?

Thanks a lot for your help!

[polar spade]

Short answer seems to be that depends on the genre. A gospel might build to have the lead singer on the first choruses, and a choir on the last. A Christmas song may start immediately with a chorus and ensemble singing, a Whitney Houston ballad could be all her or build to a chorale chorus. A sing-along could start in on the second chorus. A musical song can build from one or two, then 4 and 10 , then the whole ensemble by the end chorus.

So, from my view, its about massaging in the right prompt at the right time to bring out the choir/ensemble. This, again, could be best by building the song from extensions, not from prompting to get a one-shot song. Certainly, experiment with both ways though.

Choose genres that might ggo in that direction. These days, just ask chatgpt or an AI which genres do that the best.

[odd epoch]

why can't i paste a prompt into the "describe your song" field? ctrl+v does nothing

[polar spade]

Oh. You want non-singing. Hold on. Let's see.

[vague kraken]

Think uuuu

[polar spade]

Just showing its easily done. Just may need lots of tweaking to get something to your liking. Spoken verses and sentences are commonly done.

[wispy laurel]

It can do spoken word as a section of a song
Sometimes it doesn't want to just talk without also having some musical content too
But you could possibly frame it as a spoken-word intro to a song, then just throw away the bit that you don't need

[Intro - spoken word, voice only] [natural speech]
Hi I'm Bob!
You should all tune into my radio show, RIGHT NOW
Because I'm awesome
And if you don't, I'll burn your house down

[Verse 1] [voice, piano, drums]
(Literally any lyrics can go here)
(It doesn't matter what you write)
(Because you're not going to use this bit anyway)
(Oh no, here comes Bob with a can of petrol)

[polar spade]

OK. Im done. Haha

[polar spade]

Suno likely uses their own interfacing. Right click and paste not appear?

[vague kraken]

Thinks friends !

[clear pebble]

I would also say make sure you are on Pro or Premier so you can use the output commercially.

[proper condor]

Question. Persona captures the essence of the vocals and melodies of the song you made from the persona from. But for some reason my rap persona wants to get into screamo vocals when there aren't any in from the original song i made the persona from. Is that a bug?

I even put it in exclude it and still does it.

[honest heath]

Have you tried defining the vocal quality within the style prompt?

[ebon sand]

If it's 4.5+, the model is just biased to scream a lot and persona isn't working well on the 4.5 models on average
You could play with the sliders or use inspo instead, inspo performed a margin better for me with capturing the vocals

[proper condor]

Interesting, still trying to figure out the sliders and how they work. Very inconsistent so far. I'll try an inspire next. Thanks.

I make a lot of fantasy music and it works fine for that, but for rap its broken lol

[odd epoch]

right & paste doesn't work either

[quick sonnet]

Examples with chorus with many voices can be found i quite every genre I know.

[carmine yacht]

Definitely needs an ability when music is ready to download the ability for it to generate a full action music video

[shell trout]

hey what are stems for, just downloading?

[wispy laurel]

Mostly
You can manipulate them within Suno but you can't currently recombine the modified versions within Suno so the usefulness is limited
If you get a song where the instrumental is perfect but the voice isn't quite right, or vice versa, you can split the stems and then cover the good stem, using the same prompt and lyrics that originally produced it and with a very high Audio influence. That way you can keep at least some attributes of the good stem while replacing the bad one with a new version that still fits in terms of structure, tempo etc.
That works sometimes but it's not an exact science. If you want to directly edit a stem without regenerating any part of it, or if you want to combine altered or edited versions, you have to do that offline in a standard DAW instead

[shell trout]

Thats exactly the information I was looking for. thank you.

doing a cover of a stem is a great idea

[polar spade]

I'll make an assumption you dont wish to type in the lyrics as they are many. -- and are you certain youre not past the limit? -- So what happens of you copy only a few words? Do you have the copy/paste capability then?

[autumn rock]

I was about to post the same thing, the intro's generally work flawlessly like this... but then legato come'n in hot 🙂

[odd epoch]

oh i see, there are too many words

[polar spade]

I go under the assumption we use prompts to nuance and nudge for a response we want, not to force one. So it may be a group of prompts which steer it into a final ensemble chorus.

Are you saying you are trying the extend approach? Adding in something like [ensemble] as a prompt in the last chorus.

But heres one by just using extensions on the last chorus. It came out well to share the last two choruses as a ensemble chorus [2:55] then followed by a lead vocal

[polar spade]

That video seemed as if it would not paste either way, few words or many words. Though you seem to have the paste option.

I think you have 200 characters as a limit for describing a song in the simple version where suno writes the lyrics.

I dont know, but maybe that plays into not being able to paste lyrics in. (?)

[polar spade]

Is this idea from something youve done from somewhere else? Even with the best of AI sites focused on text or audio to video, it is quite a feat to get this right at this point.

You can mention it in the suggestion channel if you wish. https://discord.com/channels/1069381916492562582/1154587937611190354

[carmine yacht]

No no idea if there even is something like this

[quick sonnet]

Would like have in every chorus that [ensemble].

[polar spade]

Well, with just a few prompts I got that within 10 generations. So have fun playing around with prompts like "gospel","ensemble" and even "isicanthamiya" [African chant]. See what you get.

[polar spade]

Ahhh.. OK. That would revolutionize music videos then. Certainly.

[hidden yarrow]

Even more fun when you use english words in other languages. Could be that the other language is suddenly pronounced like an english-man would do...

[half coral]

I think I have pushed Sona to its edge. Several times changing information in style box has failed, title doesn't change, song is almost identical to the last song before my changes. I would like my musical instrument choices and how I restrict that choice to be moved into + persona. I thought I saw a thread that covered persona really well. I am not much of a discord contributor. I get frustrated scrolling threads trying to find info

[narrow ridge]

Anyone know what song this is?

[quiet solar]

@narrow ridge I don't recognise the video, but the song is Gentleman by Psy.

[narrow ridge]

Thanks

But the rock remix isn't on Spotify

[quiet solar]

Seems like the creator might have done the remix themselves. It's not clear if a full song actually exists.

[narrow ridge]

Lol

[stuck dust]

There are full video generators but the quality is usually not very good or coherent. It is much better to create individual clips yourself and edit it yourself

[sweet siren]

What are your experiences. I am really curious.
When working on a new song, do you use a persona more or a playlist with (vocal) stems? I find the more songs I am happy with the more bits of them I make stems of, tryinv to refine/fine tune my style.
Thanks in advance for the reply!

[fathom crypt]

I don't use any of them 😔

[calm condor]

Notification and Activity bar is not getting closed and the moment i click on create it shows a stop sign can someone help

[sweet siren]

Do you define everything in the style box and at the lyrics input?

[fathom crypt]

I use both.
I am using it in a limited environment because of 3.5.

[fathom crypt]

How about you?

[sweet siren]

I switch between using the Persona and an Inspo playlist. At the start it was just Persona, but now I have added more vocal stems in a playlist, it is balanced. I define the way should be sung in brackets where the lyrics go and in the style prompt,
I want to do various styles of music, but sung by one persona, be that through Persona or an Inspo playlist.

[fathom crypt]

Great 👍

Your thoughts are clear 💪

[frail grail]

Any of you prompt and actually get sound effects?

I got some coughing when I use cough

in asterisks

This is with Suno 3.5

[polar spade]

Some FX are easier to get: giggle, laugh, cough, thunder amd a few more. If you think about it, these are found in several recorded pieces in history and would make sense they are trained on, but still would need to be taught or defined what they are, like any style or genre.

[sage jetty]

I find I get laughing effects the easiest when I actually type out "hahaha, or HAH HAH HAH"

trying Currently doing a trance/daub step song and struggling to get it to do deeper voices.

Like I'll have a verse, then before the beat drops I have it say a word. but it struggles to do anything with that word. In fact suno has entirely ignored the word in some cases XD

[fathom crypt]

Is suno able to make atmospheric sound?

[sage jetty]

like a room full of people talking? birds chirping? I know I've seen crowd cheering in the wild.

[fathom crypt]

Like atmospheric rain sound

[sage jetty]

a quick google found this one:
https://suno.com/song/80aa36dc-3aac-43a6-a418-27904c504e8d

and this link it mentions sounds here:
https://suno-ai.notion.site/Using-Metatags-34944efe09ec41d693e314f13af44695

that link might work for you too @frail grail

[polar spade]

Kind of. You can prompt for it and it may appear randomly. But not on command, where youd want it, really.

And these had to be stemmed out.....

https://suno.com/s/p5yZMrGrxbRYq1bR

https://suno.com/s/NzwaUWbbx0yM4SZG

[azure hawk]

Hi everyone
Since we are working with one of the best tool in music industry
I have a quick check with you for a resolution
Problem 1: The output is missing the whole verse. I am writing the lyrics
Trials: I tried extending but its not taking the new verse still, if it does, it changes the whole music arrangement
I tried, changing the lyrics 1. just adding what I want, it wont take lyrics just add just matching music 2. replacing the whole lyrics, its changing the music arrangement

I tried using the cover, remaster and other direct options as well. none seems to be working without changing other things

Anyone know some hack or do i have to just keep trying till i get lucky?

Problem 2: There are few words mis pronounced in hindi language only at 2 places , the same words are pronounced correctly in the rest of the song.
Any quick hack for this or is there way to let the suno know about the wrong instances?

[fathom crypt]

Oh man. Terrible issue 😕

[sage jetty]

Looking for some feedback on these dubstep/trance songs, suno has jorked the lyrics on both, but I'm trying to just figure out which one sounds better.

Been trying to get the AI to drop to a deeper voice when it as "Gravity" without too much luck. I made this song with "Red" in mind, but "blue" sounds more in tune with previous songs I've done.

❤️ Red:
https://suno.com/s/VNAuoUARH5es6kEX

💙 Blue:
https://suno.com/s/7W2UdYKwDx7pnkOk

[fathom crypt]

I see. I think I have to find a solution for it.

[polar spade]

#2: Hindi Language ( and others including English)

If you can spell some word phonetically, that may help. This is a common issue with every AI generative program on the internet including audio chatboxes. So its not specific to Suno. Its universal. You may wish to join the Hindi language channel to keep updated of others and any solutions. https://discord.com/channels/1069381916492562582/1154779664565403768

#1: extensions.

Extensions do work. It might take tweaking on the right timemark to get it to take, or adding in a linespace or a prompt. As said, with practice comes skill. Thats to say, we start to recognize the nuances and start to recognize where it "may have gone wrong". So, so far, there isnt one way to do something that works for the thousand of varied songs prompted for.

As you said: ".... its not taking it, but if it does..." so it does insert the verse occassionally. Sunos , as well as any other AI generative music software excels on the "If's" , not always the "When's". So, see below, slide the audio up and hope it adds the verse AND gets the melody at the same time.

As for melody, remember to slide Audio between 78% and 98% to get the closest to retaining the melody.

[azure hawk]

Thank you so much, You gave me another boost of confidence

[polar spade]

To answer your other question: No, we dont get credits back for the generation not meeting our expectation. Generative AI is strong on the "generative" and makes its best guess. Sometimes its trying to fit 500 grams of flour into a thimble and so we dont get what we ask for. That includes not adding in lyrics we clearly write in the lyrics box, or make that guitar we clearly prompted for.

If something sounds right or has all the lyrics, we may think it "worked" , but it can vary on the next generation, even when no prompts or lyrics are changed. So the expectations should be geared towards "success" at getting what we want as being somewhere around 30%.

That said, Ive gotten far better results often while NOT getting what I was going for, on that other 70%.

[fathom crypt]

Yeah. Suno struggles with clean pronunciation in some language

[polar spade]

Ooo.... so my post was originally fine. I added Sunos official rules page and Woody deleted my whole post. 🤣 sounds suspicious to me. 🤔

@smoky dagger You might consider also, the judging is based on:

Creativity
Originality
Number of likes
Number of plays
Number of shares

Make sure you set to "Publish" so it will show on the original song's stats list as a remix. Im sure they'll use that as part of the entry.

[azure hawk]

Thanks @polar spade for the elaborate answer. Just to confirm, even though i am trying, i just have to add a section and write the missing lyrics

for second I can try with english phonetics which will be the transliteration of the hindi word wrongly prounounced ?

[polar spade]

Yes to the English phonetics. Its a plaster on the wound, but its about the best way so far. You won't get it perfectly as Hindi and English are not the same sounds. But do your best. Yes, spell it out in English as best you think it would sound in Hindi.

Using IPA is not as helpful, just to say, as Suno tries to phonetically pronounce IPA symbols.

[polar spade]

Do you have other lyrics AFTER this missing verse? Since I write in English, I often just add a verse, then extend again and add the chorus, then extend again and add an outro. So its about building onto the previous.

With 4.0 to 4.5+ its seems to do pretty well, to just click extension and paste in the new lyrics to add. There should be some consideration to the timemark it is extending from. Listen to the end of your output you will be extending from. Choose a timemark that seems appropriate in the music to begin the next verse. So, if the output ends or fades, you should back the timemark in a few seconds . This means manually typing in the time.

[azure hawk]

Yes, thats exactly i was trying

[azure hawk]

Yes, actually its part of the verse, AI generation skipped 4 lines in between so from verse2 3rd line skipped to verse3 4th line. so the meaning just got disrupted.
I am trying to replace the lines to completing the verse or merging the meaning of botht he verse to retain teh identity of each verse. there is small bell sound drop at the end of verse2 4th line, but when i add verse/lyrics, the AI retains(somewhat) the music continuation but completely misses the lyrics
Nevertheless, I am giving another few shorts , hoping it works ... Thanks for your help @polar spade

[outer token]

Any idea what would cause a remix origin to not show? It's a remix of someone else's uploads so it doesn't even give me the option to turn it off. The song's not cropped or faded. At first I was like wait did I forget to use the upload, but it's showing me audio influence so no that's not it

[native estuary]

I'm having a big trouble making suno to not add gipsy folk to my hip hop tries. Every time it get to the chorus, baaam gipsy like singing. what I want is like a storytelling hip hop rap witohut super emotional choruses.

[weary glen]

hey guys so how do i make all the vocal is the same like this and maintain for each lyrics paragraph?

do i need to add [verse]? [chorus] or what?

[polar spade]

@mossy zealot Uploads for video artwork can be MP4 and MOV file extentions only. Not OGG for instance.

Upload a video
Videos that do not follow Suno policies will be removed
File requirements:
9x16 aspect ratio
at least 720px tall
MP4, MOV files
Videos can be at most 10 seconds long

[polar spade]

You might add a metatag before each stanza and repeat in styles prompts. Heres some sample prompts used in these two sample outputs

Ensemble choir, shapenote, exalting, powerful tabernacle, deathmetal, power choral, majestic , epic, cinematic, triumphant

https://suno.com/s/3rVIjc52XJx8s2ZT

https://suno.com/s/fqLZF6bn2x8WIqmM

[graceful condor]

All my prompts say "NO Singing" in each possible way and still the freakin Chorus is sang all the time but this is a epic cinematic Rap song like NF and I do not want any dump opera singing all the freakin time.

[polar spade]

Suno, as I understand, doesnt understand negating a prompt. It ignores "no" and "not". So any voids must be in the affirmative.

Exclude a number of vocal styled songs, like, ballad, aria, requiem, a cappella.

Then use [spoken] as well as [rap] and dont indicate any metering, such as four or six lined stanzas with end rhymes. Write the rap out as long paragraphs .

You can try putting them in quotes as well.

Any line breaks and Suno will see that as an indication to sing the next bit so ytou would need to re-mark your metatags [prompts] before each new paragraph.

[graceful condor]

This is not what I meant. I know that "no" is not working. I meant that my whole Lyric and Style prompt tells in all ways possible that I do not want any singing and it still does it

[polar spade]

Got it. Regardless of my not reading between the lines, the suggestions still are valid if not already tried. Otherwise, let's hope others may give some insight better than I have, as well, my friend.

[quick sonnet]

Posibly try to write something like [chorus - rap].

[quick sonnet]

Sorry, but are you tallking now about videos used instead of the pic for the cover?

[queen basalt]

Hi, is there a way to tell Suno 4.5+ to not go over 3 minutes ? it's doing song way too long, even if I balery have lyrics

[wispy laurel]

You could try:

★ Change the tag for the chorus to something else other than [chorus]

★ Exclude "chorus" and "melodic vocals"

★ Really hammer the fact that it's rap into the AI's tiny little silicon brain
[Refrain - fast-paced rap]
[hard-hitting rap vocals] [young male rapper]
Etc
Repetition is your friend.

★ If there's ONLY meant to be rap and no singing at any point, make sure you specifically refer to the vocalist as "rapper" or "MC" rather than "vocalist" or anything else

If you already did all this and it's still not working, I have no idea. Sometimes Suno just really doesn't want to do something (currently whatever I prompt it for, it's giving me jazz instead 😂 )

[queen basalt]

it's not rap, it's 70's progressive rock

the song are perfect, just too long. I will trie to remove the intro

[wispy laurel]

It's hard to directly control the length of the output. You can try tagging a length, it seems to work better as a number of bars rather than time, and also make sure that you have an [ending] tag in the lyrics field at the right point

[queen basalt]

it really is. I hope someday they'll add a length handle

thank you for the tips

[polar spade]

Yes. The OP asked about uploading videos for the artwork.

[quick sonnet]

Since you use v4.5, you can simply edit it later with the editor.

[hard beacon]

The Editor is bugged as hell

[unreal narwhal]

Id like to know how to add a video loop or GIF files for song art also if anyone has any tips.

[polar spade]

For Suno artwork, this is all we can do.

Upload a video
Videos that do not follow Suno policies will be removed
File requirements:
9x16 aspect ratio
at least 720px tall
MP4, MOV files
Videos can be at most 10 seconds long

It would loop automatically

[unreal narwhal]

Than you

Thank you

[kindred jetty]

is there a good way to enforce two different singers? i tried a lot already. i'm writing a black metal song and I want Singer 1 to have more high pitched screams and singer 2 to have deep growls... most of the time i only get singer1 doing all parts. I have things like [Verse 1][Vocals: High Black Metal Screams] and [Verse 2][Vocals: Deep Brutal Death Growls] in the lyrics. and this is part of my prompt "Symphonic Black Metal with two distinct vocalists, one delivering high tortured screams and one delivering deep brutal growls, alternating and overlapping as written in the lyrics."

[polar spade]

Im a broken record to say, its been easier to set up vocal personas with those qualities, then extend with the preferred persona at those points of change.

How far/comfortable are you into using extensions and making personas?

[kindred jetty]

i've not worked with it... i tried to make a persona yesterday, but only got an error whenever i tried to save the description... and personas seem to be more a blueprint for the whole band if I understand it correctly

[polar spade]

Well, we can stem out the vocals only, of something we like. And even chop those vocals down by removing extraneous bits, so we have a good 1 minute of special vocals, as an output, then make a persona from that, calling it " deep growls" or whatever makes sense. From there we can use them whenever we wish.

Yes. Initially about a 5 step process, but its with the easy tools in the edit UI. Stem vocals & instrumental . Remove and crop. The make persona. Not too hard.

[kindred jetty]

and you can use multiple personas in a song?

so basically create two songs for each voice and create personas from the vocal stem. then use both personas to create the song with both voices

[polar spade]

Yes. Thats it.

[kindred jetty]

sounds good. and in the lyrics probably use [persona name] right?

[polar spade]

Your Audio sliders set about 75-78% control the personas to embue the vocal. But the styles slider set around 78% will give weight to sing whatever lyrics are in your lyric box so that any persona lyrics dont take over.

One reason for segmenting the best bits of the pre-persona output is to hopefully remove any lyrical patterns. More superstition probably, but I feel it helps.

Works best as verse by verse , than line per line.

But at some point after a couple of extentions you might also get away with a line per line prompting of

[Growl vvocal) "This will be the way"

[Soft vocal] "I will not follow"

[Growl vocal] "you must go"

. . . . And it may intuitively alternate the lines anyway.

Lots of ifs and maybes, but if just extending verse by verse to change vvocals , its pretty easy.

[polar spade]

Sorry, i'm busy that I cant whip a sample up,

Heres a sample of 6 different personas in one song. First persona I think comes in about 1:20. I didnt stem out the vocals, so each segmemt change also includes the musical style of the persoas, yet the whole song and lyrics are still cohesive

[kindred jetty]

all good, i'm trying myself... thanks for the tips and ideas

[queen linden]

any way to lower the octives of female singers.... I'm looking for Nina Simone sound and end up with some screechy disney princess girl power song.... shudder

[wispy laurel]

[contralto female vocals - low, husky, soulful]

Or if it doesn't like 'contralto' try 'low alto' instead.
You also might need to go through a couple of cycles of covering a song and prompting for a low voice each time, then covering the cover and again prompting for a lower voice, and so on. In theory you can get the voice to drop a bit lower each time, then make a persona when you get to the sound you want
From personal experience this is a pain in the backside, you're right that Suno doesn't really like low female voices, but it can do it if you're persistent (and find a prompt that consistently works)

[queen linden]

thanks I'll keep trying .... really appreciate the info.... I tried contralto I got a weird belting , soaring that is so not what I was attempting lol thnx again

[wispy laurel]

Another thing you could try is, make a completely different song, in a genre where that particular voice would be expected (ie, write a prompt that really obviously is code for "sounds like a Nina Simone song"). Do everything you can think of to make sure that Suno will naturally think of that particular artist as a reference point.
Doesn't matter if you don't actually want that song. Because what you're going to do with your fake Nina Simone song, is split out the vocal stem, crop out whichever part of it is closest to the way you want the voice to sound, and then use that little section to make a persona.
And you can then use that persona to help you use that voice, or one close to it, in other genres and styles too

[deft girder]

not just u

i just get 5-10 seconds

[proud temple]

rip my boring lyrics. I'll keep them in japanese then

For context ¯_(ツ)_/¯ https://suno.com/s/ohej95TxpHaMiaah
(don't mind me, I'm just venting)

[wispy laurel]

You might be able to get around it by spelling it "foss-fate" instead
It can pronounce phonetic spellings as long as they're clear, I don't honestly know I the copyright filter would still flag it though

[proud temple]

yea, no worries. I posted mostly to expose the absurdity of copyrighting english words. I'm just happy we can still talk about apples

[mossy zealot]

The video loops I tried to give it were first one straight out of comfyui I rendered with wan2.1 and saved with h.264 MP4, yuv420p with a typical video combine node, and it looked good playing video, but the thumbnail was squashed in playlists to fit the tall ratio (gave it a square video). Then I tried encoding it with handbrake to make sure it was "streamable" and I just used the default fast 30fps setting at whatever resolution, which fixed the squashed thumbnail, now is just black on my android and plays in desktop browser fine (I have no other platforms to test on). Videos from Veo3 and Kling also work fine (no issues whatsoever for those!), but out of my pc the needed format for universal playback on all platforms is yet to be figured out. Never tried ogg.

[quiet solar]

Seems Dasvon has you covered, but just to note that while your prompts are absolutely on the right lines, you have to remember that you're talking to an AI rather than a producer/performer. So for example, "alternating and overlapping as written in the lyrics" probably won't be something it can work with, as the style prompt doesn't explicitly cross reference the lyric prompts (and vice versa).

Some people have had some success with opening the lyrics prompting to specify two vocalists, e.g.
[Vocalist 1: Male tenor, high black metal screams]
[Vocalist 2: Male bass, deep brutal death metal growls]

And then using those same prompts to indicate where Suno should switch between singers during the song. You can also name the singers and Suno will typically understand that means they're to be performed by their respective parts - although the execution can be a bit hit and miss to say the least.

[rigid folio]

hey guys can i get some advice? the remix i made sounds almost exactly the same as another person even through they are completetly different style prompts.. what do i do?
Others https://suno.com/s/B5Vk0AKQAluh1rMQ
mine
https://suno.com/s/leerFv40V6ffYZ9G

[proud temple]

my 2 cents:

• try using a personna when you remix (it will affect the style a lot)
• try describing the style in the title (you can change the title afterward. The title has a very strong impact on the generated audio)

[crude mountain]

Hi ya'll, I'm very new to suno but have been using it for a bit with interesting results. I am attempting to use it to superdrive my creative/production process. Is there a good way to get suno to simply create subtle melodic and rhythmic variations from my source audio (I own all the instrumentals/voice etc.). I usually find remaster makes it essentially identical with a different voice, I attempt to remix with my own wip album as inspiration and audio influence 100% and decreased weirdness and style influence to no avail. I can't seem to generate something "close" but different. Only things that have no resemblance of the original composition and melodic structure/vibe or essentially a copy. Any advice would be great.

[honest heath]

crank up the audio influence slider in the advanced settings

err wait you did that,

no, you will always have subtle variations.

[crude mountain]

I put it to 100% and the track it spits out is still nothing like the reference or inspiration

Should I remove the auto-generated style prompt from my audio upload

[honest heath]

are you giving it a strong style prompt?

[crude mountain]

I have the prompt that is automatically generated with my upload but I also turn down style influence to like 20%

[honest heath]

try giving it your own style prompt, i suggest using music theory and production terms to describe it, then a bit of vibe
so something like "4/4 time, a minor, dorian mode, smooth jazz"

[crude mountain]

Should I be using the cover feature or the use style and lyrics feature for subtle variations?

[honest heath]

I use cover

[crude mountain]

Ah, maybe that's also my issue

[honest heath]

are you making sure to select the audio you uploaded as an influence?

[feral surge]

Anyone have any success in replicating the factory industrial ambience style that permeated the first Silent Hill? I've tried "industrial ambience" and "dark ambience" and "horror ambience" and "factory ambience" and other terms and even added into the exclude styles things like "drums" and "melody" and "percussion" and "studio mastering" but generally what will happen is it will start out kind of okay albiet a bit Boards of Canada esque, sometimes even envoking that industrial factory sound I'm looking for, and then it will suddenly break into a wild piano melody or a hardstyle techno track or a funky beat not unlike lofi trap, even when I'm generating out of nothing vs doing a remix. Any ideas?

[crude mountain]

Does selecting an inspiration remove the audio reference?

Right now I have a playlist of my uploaded demo tracks as an inspiration

[honest heath]

yes

[crude mountain]

Ah, that is probably why.

[honest heath]

if you select a playlist as inspiration it will over ride the audio influence iirc

[crude mountain]

Perfect, thank you much! It is working much better now, especially being more direct with the prompt.

[sage jetty]

Has anyone had any experience with the musical instrument: Theremin

Me and my BF Been trying to make some songs with it without any luck.

[ebon sand]

Jen

[sage jetty]

[ebon sand]

I'm vishanka

[quick dome]

so suno can have a comp with explicit lyrics but i cant make a track with explicit

[sage jetty]

I have incredibly bad memory, feel free to DM me tho if you reconize me form somewhere else!

[polar spade]

Its noted in the Terms of Service that due to Sunos attempt at optimizing for the best and most effective music forms, outputs may contain similar or repetitive elements found in other users outputs.

Its the traits of the beast. Not much can be done but attempt to avoid it by trying another direction. Finders keepers, "first come first owned" kind of thing. Moreso, we should be diligent to try to find similarity to our songs to those in the real world, if we plan to distribute ours publicly.

[polar spade]

Many songs prompting for "halloween" " spooky" " horror" often end up having theramins in them. Heck, just prompt "halloween" and nothing else and youll likely get one.

[rigid folio]

i ended up remaking it, saves the whole issue, new one is actually way better lol. yeah im very aware how risky it can be using this stuff negligently. im just glad i caught it myself

[sage jetty]

Nice! good to know!, asking around in another group, they mentioned that Theramins are not a registered instrument in suno but you can encorage the sounds with prompts.

[rigid folio]

https://suno.com/s/pjrEg7mF1FqkgGai if your interested

dang it.

[polar spade]

Yea. Its a quandary. Not so much though about whether I should be allowed to do something "banned" if someone else can. I could go a few rounds debating this one. Respectfully, I won't.

[polar spade]

Well. Just try one.
https://suno.com/s/tMRqSd3GMtSJMPW1
https://suno.com/s/MCUiACfSTPwVEvpm

Stem out the theramin and throw a bunch of them in a Playlist and click inspire.... see if you get a theramin song out of it.

[polar spade]

I can clearly hear melody lines of my older songs from last year in my new songs this year and some rhythms (which seems obvious) and intros. I always wonder though is this that prevalent between users, or are my new songs influenced specifically by my previous works. We'll never know.

[potent agate]

I'm not sure if anyone else has been getting this, but over the past few days, the very very start of a song seems to be clipped out, much more than it used to be. Like, all the tracks are starting about half a second in, or something.

[polar spade]

Is it every song of different creations? Open edit ui and see if it's there as well, or can you manually slide the marker a bit more left to include more of the output. Zoom in to see if its physically looks like something is missing. It may be the player itself, then needing to report that as a bug.

[sage jetty]

Ooo that second one cook'n there abit!

Thank you for your time, care and credits for testing this!

[polar spade]

No problem. Those were old outputs from earlier this year.

[sweet siren]

I did actually, but didn't think it stood out so I was not impressed and changed it.

[brisk flint]

Yes. Here's the link so you can get some inspiration of my prompt: https://suno.com/s/0YDI4vgdXBaNOj4d

[limpid spade]

About 1 in 3 of my audio upload gens are starting in odd places which just ruin the song. The image below is about 2 bars too late into the song start. There's no amount of stretching or manipulating in Editor to fix it unless it's exactly duplicated later and you can do some wav surgery. I've reported it in the Testers chat.
I even had one song start halfway thru the 2nd chorus tonight. 😮
@potent agate

[pastel sluice]

Hi Dasvon, thanks for the reply. Now with another song, I tried this way and it worked perfectly!

[flat violet]

I can't get the best result for sound effects in my lirycs any guide pleasE?

[potent agate]

I've tried the "edit trick" on a few, but none of them seem to be that as the cause. It's definitely Suno not starting tracks from the start. At first I wondered if I'd clicked Extend instead of Cover, but .. nope..

[pastel scaffold]

How to change male vocal song to female vocal completely without changing instrumental at all?

[sage jetty]

Looking for some feedback on this space Themed Ballard!

Emoji your fav, or send me some feedback, ether is great!

1️⃣
https://suno.com/s/GgQZoVUJ4SYwOlFs

2️⃣
https://suno.com/s/BJB06BsXEBaqNsy7

[feral surge]

Hey I'm still wondering if anyone has had any luck with this?

[polar spade]

Yea. 👍 Thats what i was wondering. If the song was all there but the player is starting somewhere other than the beginning. I may try a couple. Thanks.

[polar spade]

Oh cool. Glad it worked this time. SUNO is fickle, but this has worked fairly decently for keeping in your toolbox.

[polar spade]

@pastel sluice , would you like to suggest any technique for him?😀

[pastel sluice]

@pastel scaffold

I was lazy and not in the mood to change the previous one,
In the edit view, try to change the lyrics and add [Male/Female vocal] before the lyrics you want.

but with the new one I just made
[Male vocal]
Lyrics
[Female vocal]
Lyrics

And create as much as songs until I like the vocals and music

[humble verge]

there are no words in the English language that could accurately describe just how deeply I detest SUNO's editor.

[polar spade]

Ahhh.. OK. I thought you had used the edit UI and replace to switch the vocal genders. Ill just paste my previous post then. Maybe that can help.. Thanks.

[polar spade]

Here a suggestion that seems to help in that direction.

Its two posts which hopefully make sense.

#📚┃suno-school message

#📚┃suno-school message

The replacement -- done in the Edit UI -- though helpful to change the gender, will also generate a new melody or even a new rhythm. Thats not what you want, I know. But you might get something that is unexpectedly pleasing instead. It may even randomly and unpredictably change the vocal genre but keep the melody as you want.

But other than the "add vocals" feature for the whole song, Suno doesnt have a means of adding new vocals (and therefore adding a the same lyrics and melody but change the gender) yet, to just a segment or extension.

[crude mountain]

Anyone have any success with generating organic electronic pop like Nurture? What keywords/prompts do people use?

[halcyon sundial]

is there a way to translate songs?

[half dragon]

Keeping melody and instrumental arrangements? Not on Suno, not possible.

[humble verge]

What do you mean exactly?

[halcyon sundial]

i dont want to betray suno hehe

[halcyon sundial]

i want to made multilingual versions of my songs

[half dragon]

I did this a few times, start with Suno then finish in DAW. Basically you added both lyrics into the lyrics, in both language, have full structure, Intro, verses, chorus...etc..
Then generate normally, once you like the version of gens... then you have to download it and chop and edit into 2 songs.

[halcyon sundial]

i tried this...

i made songs in portuguese with parts in english

[halcyon sundial]

the worst problem is t

that

[humble verge]

Can't you just edit the lyrics in Song Details and then do a remix?

[halcyon sundial]

some words are identical with different pronounces

as an example the word "trace" and "trace"

if there is a part of song saying trace

it will say in portuguese

i have some cheats to solve this

but annoys a lot

[half dragon]

I think that happens in any language... even in just English songs sometimes...
You can fix them in Editor.

[honest heath]

suno hates heteronyms

[halcyon sundial]

a forgot a word in a phrase past week

it costed 90 replaces to fix

regen features just ignored all times the word i tried to include lmao

[quick sonnet]

Have asked the questin i help chat, but was tolld there, that a good place to ask it is here, so Irepeat. I have a track that's actually pretty good, except for the chorus. I'd like to change the chorus a bit but keep the rest of it the same. What would be the best way to achieve this? I'm a free user, so I can't use features like the editor. I also have a track where the chorus is good, but there are problems with the verses. Would the same solution work there?

[half dragon]

For the song that verses are ok but Chorus is bad, I think Extend would be your only choice to fix it with some sacrifce.
Basically regen the rest of the song by regening from the Chorus on.
For the song that issues with verses but Chorus is good - it'd be difficult to remedy - I'm not sure if possible for Free Tier.
My advice is just don't get to attached to it and generate a new one 🙂 jk

[sage jetty]

I am looking to capture a good mature Japanese woman voice for a song I have in mind, Most of the ones so far are fairly high pitched.

Think classic anime villainess. anyone familiar enough with the tags that could generate something like that?

I know for capturing accents it's a good idea to try and capture in another language but any other advice would be great!

[wispy laurel]

It sounds obvious but did you try "classic anime villainess"? 😂

It's surprising sometimes, how well this thing can interpret stuff that theoretically shouldn't work

[neat igloo]

Does anyone know if I can merge 2 files? For instance, I recorded 1 vocal track on my own (WAV) and 1 instrumental track (wave). I want suno to merge the 2 tracks. Is this possible?

[wispy laurel]

You can't (currently)directly combine them in Suno like you could in a normal DAW program
What you can do, is upload them both, put them both into a playlist and use that as inspiration. Write lyrics and style prompts that clearly and exactly describe what the song would sound like with both tracks together
Set Weirdness 1, style and audio influence high (75-95, you might need to experiment)
In theory that will prompt Suno to generate something that sounds very close to the combination of your two uploads. It will recognisably not be an exact replica, but if your prompts are good it'll be the right song, with vocals and instrumentation that closely match your two tracks
This is basically the same as if you mixed the two tracks together yourself, then asked a real band of real musicians to cover it as close to the original as possible
That all assumes you have access to inspiration and the influence sliders. If you're a free user, this doesn't work. Even with the paid functions it's not quite the same as directly combining the two files, but it's about the closest that Suno can currently do without using other software too

Also though, there are dozens of free apps that you can download, either for desktop or mobile, to directly combine the two files yourself. Which is honestly probably a lot easier that using Suno

[neat igloo]

Awesome! Thank you for this. How do I get SUNO to not change my voice too much?

[neat igloo]

I'm having the toughtest time finding one. Garage band is too advanced for me lol

[wispy laurel]

Very low Weirdness, very high Audio influence, and make sure the prompts very clearly and accurately describe the actual sound of your recorded vocal

[graceful condor]

Why in the freakin heck is this dude screaming this line for 300 credits already. Regardless sliders up and down

"[Arrangement:Music breaks down dramatically.Clean,reverb guitar plays beautiful,sparse melody.Drums simple or drop out,leaving a sparse electronic beat or reversed cymbal swell.Atmospheric synth pads create vast,introspective soundscape]
[Vocals:Vulnerable,almost whispered, melismatic approach on end-words,wave-like up-down phrasing,and breathy transitions]
And perhaps that is enough…
…for us."

Nothin says screaming. Even my exclude prompt says "Screaming" for the whole freakin song

[wispy laurel]

I use a free Android app called AudioLab for this
It has a bunch of more technical stuff to experiment with but the basic function to combine two (up to four) tracks, is as simple as selecting the two files, setting the relative volume, and clicking Mix

[neat igloo]

Youre awesome. I appreciate it!!

[polar spade]

Audacity, perhaps?

[wispy laurel]

If you're using model 4.5+, try switching it to 4.5
The more advanced model is known for over the top vocals, the previous version isn't as fancy but I find it takes direction better and shows off less

[neat igloo]

I'll give it a try. Thank you!

[sage jetty]

you know.. maybe I will try this 🤔

[queen linden]

I used the Teremin in a song or two I loved how Suno added it in just the right spot... I needed a sound... not a voice or instrument... worked great

[tranquil orbit]

Is there a way to expand on or add things to a track Suno made? This one track I've got, Suno put in these awesome choir "ahhhs" in one part. I'd like to have them in the chorus, too. I'd also like to expand the guitar solo and add a chorus reprise at the end. Is that possible with the studio/track editor, or is that a bit too much to ask of the AI?

[tranquil leaf]

I haven't gotten access to Suno Studio yet, but theoretically from what they've said it should be possible in that. For now, the way you'd do is it download the stems and cut and duplicate the tracks as you want them in an external DAW.

[hollow pine]

I haven't gotten access either 😭

[wispy laurel]

Currently you'd have to do this by regenerating a new version of the track which hopefully comes out very close to the original, but with the new bits added in.
Cover the original song. No persona, unless the "singer" has a really unusual voice that you can't replicate without using one.
Lyrics should include all the original lyrics, plus add any new parts in there at the right point. When you're doing this, ideally you have a ton of [tags] in the lyrics, labelling and defining everything precisely and accurately - this includes any new additions, described in detail in tags, but also you may need to add tags that weren't already there, to fully describe everything in the song that you want to keep the same.
You need all that in the Lyrics prompt because the style prompt is completely blank.
Set Weirdness to 1% and Style influence to 1%, Audio influence to 95%.
Keeping the style instruction blank means you don't need style influence, meaning that the original song and the lyrics are the only influences on the output (this is also why you want to avoid a persona if possible - if it only has one source to pull from, it won't randomly introduce new bits without being told to)
Theoretically this should basically recreate the original song with the changes you wanted. It may make small other changes too, if it breaks something just try again.
When you have one you're happy with, if necessary you can then cover that with no changes to the prompt but with a persona (clear the style instruction after adding the persona). Sliders at 1%, 1%, 99% for this one. Should once again recreate the amended version of the song, but now using the persona's voice. If the voice was already close enough to the original, you can skip this whole section
Suno gonna Suno so it can do weird stuff during this process, be prepared to spend some credits trying this more than once, and editing to fix things

[tranquil orbit]

Thanks a bunch for the detailed answer! I'll try that, thank you!!

[lament marsh]

Hi I am new to Suno. Can someone recommend a good resource to learn what all the tags do? i want to learn how to exercise more control over what is produced.

[polar spade]

This is a good start to read the depths of basic prompting , including tags for vocals. Peruse through or just Google "Suno prompts" and many sites or links pop up.https://howtopromptsuno.com/making-music/instrumental-tags

Otherwise, click on outputs and see what other do. From there, find patterns or similarities.

[lament marsh]

TY Dasvon!! I appreciate it!

[devout sinew]

Hello. I am trying to get Suno AI to generate a children’s choir for my duet song, but it keeps giving me adult vocal choirs instead. Does anyone know the best way to prompt for a true children’s choir sound?

[devout sinew]

What prompt wordings did you used?

so what prompt did you used to make that persona? I am confused, you just provided me a link to click. But not giving me an advice what technique to use for prompt. Thanks for your help. 🙂

[quiet solar]

Let me save you a lot of headaches by telling you that tags are not commands that are programmed in, nor is there a fixed list of them.

Suno understands that anything in [] is a style or structural prompt.

So you can put [Epic Ukulele Solo] if you like, or [Vocalist shouts every word like their life depended upon it].

There is always a fair chance Suno will disregard any prompt you make based on randomness and genre fit, but prompts that are most successful are ones which the AI can use keywords from to reference music that contains what you’re looking for it to do.

As a result, something like [Male rock vocalist, Tenor, Intense] will typically be more likely to influence towards what you want than [Dude with a scratchy voice] which may be interpreted as [Male - Voice - Scratching effect] - so not an especially detailed steer, and with a keyword that might give you some interesting and unexpected side effects.

[lament marsh]

TY for explaining that, I was wondering why some people used terms that seemed way too vague. When I tried to use myself ... sometimes they don't work. Thanks again for letting me know what to expect.

[pine ingot]

In Future Bass, the drop usually switches to halftime, but when I generate Kawaii Future Bass, the drop doesn’t consistently go halftime. Does anyone know how to make the drop halftime?
I’m on model v4.5+, instrumental (no vocals).

[quick sonnet]

How can I create an intro without music, where, for example, the names of the artists involved are mentioned, not sung?

[pine ingot]

I have used the prompt "spoken word" before.

[native bloom]

Is there anyone that can help me with adding music to vocals ? The vocals being originals ?

[spiral frost]

Yo guys, now, baking a 30sec stupid recording can bring way more "pro" production into suno. It's crazy how it changed the way we can produce... just crazy

[spiral frost]

Yep

[spiral frost]

Look what i posted its just the invert lol

i think you can just upload your lyric file, and type keyword for a genre? do you have any test you did? or maybe just send me your recording

[native bloom]

Actually its a group song by my children, and I just want to add music to it. if you just give me the steps , it will be awesome.

Here is the rough recording.

After generating the music, I just want them to sing on the music , like karaoke.

[polar spade]

Is that a Manipuri folk song? Is that sung in a minor or major key (using western music terms) the kids sound like they are singing it in a major key, almost like an Irish tune. But it seems likely it could be more minor?

[polar spade]

@spring thicket [from need-help channel]

Most sound FX are not a common usage in music so likely not as familiar to Suno in its training. A few, like cough, clap, laugh, whir, crackle may work, but even then, its hard to force them by a metatag to hit in the right spot if they are produced at all.

[queen basalt]

I see that some voice come more often than other, I'm not ready to make a persona because I haven't found what i'm looking for, but is there voice tag to try to suggeste voices ? for example [ALEX] [DREW] etc...

[quick sonnet]

Which prompt?

[polar spade]

I really dont think so. Technically we might use a name like Yomato and get an anime styled voice, but thats stretching that Suno would make many associations between names and voices, or if popular enough, would be included on a censor list, anyway.

If you called one [Growler] and one [Whisperer] , also, I dont know if it would recognize them as names, but as metatag styles instead.

[hard beacon]

wooosh... layout switch 😄

was confused for some seconds 😉

but i like it.. good one @steady kelp and Team 🙂

[native bloom]

Can anyone help me add music to my vocals ? Please tell me the steps ??

[small bridge]

are personas not working today? I can't get it to apply to my cover

[hard beacon]

Unfortunately, I can't help you with that, as I have no experience with it.

[eager compass]

Do you mean it's just not showing up above your lyrics? It should reappear when you click create/remix. You will also be able to swap personas freely.

[eager compass]

Only help I can be is to click on the 3 dots on your original audio, hover over remix/edit and select the add instrumental. You can deselect the instrumental on the lyric section to the left and have the lyrics visible. It will try to recreate the voices but wont use the original. The voice quality will probably be the same though. My experience has been 10 minutes ago as I do not sing XD. You could do that and then extract the stems to have just the instrumentals.

[native bloom]

Thanks a lot

[eager compass]

You're welcome.

[polar spade]

Had previously asked:

Is that a Manipuri folk song? Is that sung in a minor or major key (using western music terms) the kids sound like they are singing it in a major key, almost like an Irish tune. But it seems likely it could be more minor?

So you're wanting an Indian vibe or just maybe a guitar .

First, your quickest route can be to just make a Cover of it. Just upload, click Cover and done.

Anyway, if you add an instrumental, Suno will guess whether its major or minor key and if you make a few, may get a bit of both.

You may then need to separate out again the vocals with Get Stems under the outputs menu. Just the two stems vocals and instrumental.

Possibly, needing to also add a small intro, which to Cover that stemmed instrumental can do that fairly well.

There's a couple of good ways to do this. And about 200-500 credits to get something you like.

[graceful condor]

Did your UI also changed from one sec to the next on Web? the whole interface is totally different and the "Get Full Song" button is gone.

[faint pier]

is there any way to "add instrument" to my uploaded vocal track without altering or regenerating the vocal track itself? Currently, this option regenerates the uploaded track, which I want to avoid.

[eager compass]

Nope. All normal here.

[polar spade]

Well, at times we can get lucky to "Cover" the upload as an instrumental only. And hopefully that one of the outputs dont contain vocals. Expectedly, it should turn any vocals into an integrated melody line. That doesnt always happen.

Its suggested to turn off 'instrumental' toggle perhaps add some [metatags] in the lyrics box to emphasize a no-lyric result but spaced out to keep the song's length.

[graceful condor]

Man it seems I am cursed

[quick dome]

suno might be changing ui to get ready for the studio

[eager compass]

You do seem to be running into a few interesting issues. I do like the layout you have. Maybe it's a soft release? Some get it while others don't.

[graceful condor]

I would rather have it not if it turns everything worse. Even the publish button is gone in most cases only on "Full Song" it is shown. At least in this case right mouse click helps but with "Get Full Song" I have to work with my smarthphone now. Create songs on PC and then put them together on the phone. This is ridiculous. Why change something that worked well before? I do not get it.

[polar spade]

And if you add an instrumental, then separate out the vocals again through Get Stems (vocals/instrumental), you can have an instrumental that way, usually without any melody line.

[near flume]

clear your cache, there's been a stylesheet change.

[graceful condor]

I just did that yesterday and spend 30 login into every damn acount again because suno Persona was not shown and it did nothing. I do not wanna spend again 30 minutes doing so. And who guarantees it is even changing anything.

[woeful crystal]

Hey guys I just got into Suno few days ago. I found a song on there and I really like the style, but when I try the same prompt on my song the style are just different. My song just always have that synthwave style in there, but if I remove that the result are feels dry.
This is the style I'm looking for:
https://suno.com/s/4wIAOc1c01pu5aAs
And this is what I got:
https://suno.com/s/4tHlEUwdLiVjoUnv

How do I match that style?

[polar spade]

This is what I mean. #📚┃suno-school message whether a minor key sound or major key?
https://suno.com/s/mXQgxSG1ZzM2w1dE

https://suno.com/s/pc2wLwTjnk820XsT

[polar spade]

With about 60 million combinations of styles it could make, the "vibe" is still there. Though the first prompts in the list are usually weighted more, Suno will often pick a style in a random position in your list and emphasize it.

You can take that output, find the prompt ytou feel it lifted and put it later in the prompts or even put in the exclude list.

J-pop won't help, uplifting, soft beats, actually quite a few you could probably do without. Try just the ethereal sounding ones up to Nocturnal and see what happens. Ill do one.... hold on.

https://suno.com/s/kjMyl6xEwA3AmuW4. So here's a slower one. The vocal is not as ethereal.or Dreamy with that angelic choral sound, so it just takes tweaking. . A prompt like "massage spa" might help also.

Ill delete this song in an hour as its not mine.

[woeful crystal]

Thanks, I think that might help. I will do some experiment.

[thin nebula]

I’m trying to accomplish a girl group sound with distinct different singers harmonizing and taking turns. Any tips?

[lusty juniper]

Already restarted my computer. I am unable to use any Personas. Anyone else having this issue?

[eager compass]

Do you mean it's not showing up when you select it? A lot of people are having that bug. You can still select a persona and it will apply it. Once you click create/remix the persona will reappear.

[lusty juniper]

ok i will try

[polar spade]

And remember, that user posted just one, the best one. The other 30 they did that you dont see may have all failed. 😄

[woeful crystal]

Agree, I just got 10 songs down and none got that style, but there are some that is good in it's own way

tho another question, do suno change your song after it's crated or am I just imagining things? I had this one song I think that sounds good, but then when I come back to it later I felt like the start sounds different.

[lusty juniper]

cover the song if possible and select high sound quality and then maybe put weirdness at 10-15% or audio at 50%. i havent tried it though. 😛

[hallow crow]

Hello everyone.
I tried to make a song with first two verses sung in solo and the last two verses sung by a choir.
It was fine until the fourth verse were it was sung solo again instead of a choir.

Whatever I try to make it two solo's followed by two choirs it never seems to get it.
Is it even possible? The AI seems confused on how to do it.

[lusty juniper]

as for me im having a problem with a persona. its a new one i got that covered a song of mine CRAZY GOOD. so i wanted him to cover another of mine, and... he sounds different -.- and he didnt have a texas country cowboy accent... but now he does

[dusky gorge]

Also working with Personas and looking to create a sort of personal anthem for this character and continue to get male voices in the mix for whatever reason, despite explicitly choosing female vocals, spell out female vocals in the prompts, etc. yet still some choruses (also called out no chorus and/or choir) still include male parts. very frustrating. Some of them work great, but it is so inconsistent. Any ideas?

[lusty juniper]

we need Suno v5 ahaha

[woeful crystal]

my song sounds differnt, what just happened? I heard it once when it's first generated, I think it's good so I left it there, now I come back and the few notes I reconized are all gone, it's like a new song now.

[dusky gorge]

burning 1000 credits to try to get the song to have the right gender voice is so tedious

[eager compass]

Yep lol. I have made some duets and a couple with choirs. It's a pain to get it right. Spent so much on just single songs.

[lusty juniper]

are you using a persona?

[dusky gorge]

yeah, it worked great for 8-9 songs, now having a LOT of issues

[polar spade]

Maybe the one you heard was an edit, it showed a bit then got hidden? So clear [un-tick] your filter in Library? So edits and stems are not hidden?

[hallow crow]

Found out the answer to my own question. I made an extension after verse 3, in that way verse 4 became a choir aswell.

[woeful crystal]

I unchecked all filters, it is the correct song, the length is right the name is right. I don't know maybe the AI remastered it after the song is generated? the first time I heard it was when the length hasn't showed up, is it possible suno came back and edit to smooth things out after done generated the whole song?

[polar spade]

No. It doesnt re-do a song to "fix" anything. It lays it down, beginning to end. So im not sure. You can check your Library History to see if it shows the one you played with the melody line in it.

updated info: #📚┃suno-school message

[woeful crystal]

Maybe I'm just imagining things.

Too many song with the same name, maybe I get confused

[near flume]

happens all the time, Suno doesn't keep an audio "hard copy" of a generation, it reproduces it but sometimes errors and creates a new version.

[near flume]

Even when using a Persona you need to select either Male or Female for it to come out right. Didn't have this problem before they added the gender selector.

[woeful crystal]

so you mean everytime I replay it there is a chance the old one is lost?

[unique stone]

Once a song has finished generating, the MP3 file won't get changed.
What can happen on very rare occasions is that a bug can cause the song generation to fail right at the end, so it will start generating the song again. But, that won't replace the MP3 that the website plays.
So a song generated today should still sound the same whenever you play it in future.

[woeful crystal]

oh that make sense, because the first time I heard the song it hasn't finish generating, maybe it was replaced.

[dusky gorge]

oh it's selected female

[quick sonnet]

Which UI was changed?

[quick dome]

the one on the screen , not version like 4.0 or 4.5 pro

[white quail]

Do you also receive muffled quality with 4.5+ right now?

https://suno.com/song/992672e7-a8e8-4559-bab8-cb00b4306697 Lots of artifacts and noise

[primal lotus]

Is anyone else having issues with song generations having pops and skips?

[quick dome]

no , ssometimes some bad noise

[placid spindle]

Hi kids, One of generations on this one song came out pretty good. Is there a way in Suno to say - hey - give that version another try, but don't stray too far. Kinda like Remix or Subtle varaition in an image generator?

[wispy laurel]

Cover it, make sure the text in the Lyrics box exactly matches what's actually in the song, leave the style instruction blank.
Set Weirdness slider down to about 20-30%, Style to 0, Audio to around 70-85 (you'll probably make a few attempts at this so experiment with varying the numbers a little)

[neat nexus]

Try Variations?

THERE IS NOTHING HERE.

[hearty moat]

Hey team! Has anyone ever accidentally created a song and forgot to hit generate the lyrics first, then the song came out amazing and the lyrics were the prompt you used.

HOW do I keep the music and get Suno to generate new lyrics to match, or do I just have to commit to screwing it up haha

I've two amazing songs and frustratingly the lyrics are the same one sentence over and over that I was about to use to generate the lyrics... Just forgot to hit generate lyrics 🙁

[neat nexus]

You can try covering it with new lyrics and adjusting the sliders to try to get it closer.

I actually have some luck on this with hallucinations sometimes.

[wispy laurel]

?
Why do I not have this in my menu?
Web version, on tablet in Desktop mode, everything else is where it should be but that option doesn't exist

[velvet granite]

Hi! I'm trying to cover a song and I love everything about it except I want to add some noises like "hollow wooden clunks, distant metallic clinks, wind, leaves rustling, or faint animal sounds" But for some reason suno is interpreting this as really weird vocalizations, especially in the intro and no matter what I do, I can't get rid of them. Does anyone have any insight?? Ive tried

1. In the text of the actual song by intro no Vocalizations
2. Since this is a cover, there's nothing in the style box and usually that would be fine, but I have specified the additional sounds in there. So now in front of that I've added no vocalization in all caps. I've also included no human sounds. moans screams or any other vocalizations.
3. I've also added this on its own line.

I don't know what else to do. I've stripped any descriptors because at first it said primal or tribal. I've removed all of that. I've simply stated "noise."

I'm open to all suggestions

[rain lance]

If you want to add sound effects, my best suggestion would be to use an editing software and add in the sound effects manually. Relying on Suno for them is a massive gamble, and while it can sometimes churn out some cool ones, keeping that all consistent also with the best sounding generation of the cover is incredibly rare.
If you have a version you like right now but it is plagued with weirdness like weird vocalizations, you can get the stems of the song (50 credits for all stems, 10 for voice+instrument) and, again, use an editing software to cut out the weird vocal bits. The stem version of the instruments don't always sound as good as the original, however, so it depends on the stem generation.
Suno also tends to stick with any ideas it's had and I haven't found a way around this. So if it starts putting weird human sounds in your song, it'll keep doing it because it thinks it's a good idea, even if you throw that into the negative prompts.

[woeful grove]

technical question... if I export the tracks, chop the parts I like and import them back into a playlisti to be used as inspo. will it get the lyrics-correlation properly? or is it better to chop them in suno's editor? ( I find it very choppy , this is why I would preffer to do it on my pc)

[lost nexus]

I have to say, Suno is amazing but unfortunately the cover department still needs work. I have a bunch of songs I originally made myself , so I'm like , let's give them a "better sound" by remastering/covering them. I have to say it's a bit underwhelming. The audio quality is either worse, instruments are missing or the voice is completely changed and giving the song the wrong vibe, not to mention it sometimes simply skips a beat or glitches and repeats it like a record scratch. I've reported these things, I hope they look into this technical bit.

[polar spade]

There is probably no known answer. Certainly you would be the one to test this theory and let us know how well it works. I would guess it doesnt matter which way you choose, the results would be about the same.

Maybe to say, will Suno be good to follow the lyrics, as normally pasted in one create form, but this time have Suno follow each entry in the Playlist sequentially by clicking Inspo to result in one complete song of all the individual lyric posts.

No. I absolutely doubt that. But you can be the tester and find out. Id love to know.

[tepid moth]

I did notice a UI change today, and the quality of the generated songs have been so bad since then. No idea if it`s linked or what. Maybe I need to switch back from the 4.5+ to 4.5.....or maybe I just have bad luck right now.

[graceful condor]

The UI is terrible but honestly songs turn out better than before. Which is the only good about it so far. Normally if it cuts off a songs just near the end and I extend it generates way too long part 2s. But today all extend are perfectly fit outros. BTW if you use it on WEB too are you missing the "Get Full Song" button too after extend? Since the new UI it happens to me and others too I could confirm.

[unique stone]

From the "..." menu, go to "Create" and then "Get Full Song"

(I think it would be good to add it to the Remix/Edit options since that's slightly more convenient)

[graceful condor]

I think it would be good to add it where it was before directy at the track to easily just click it because going into an extra menu is far from convenient. Because why is there no button anymore it was so much easier and more userfriendly like so many things before this UI happened? At least there is something now. Also each time you have to arrange all windows new. Lyrics Prompt, Style prompt window and advanced is closed. Before it remembered your selection, what you opened, which size those windows had and now each new creation I have to do it over and over again. Shouldn't a new UI be more convenient and QoL instead this is making everything worse and complicated.

[wispy laurel]

I've tried this and it can work but it isn't very reliable. You'd have more luck combining the parts back into a single track offline, even if the joins and edits are messy, and then uploading the whole song as one thing. Then use that one single song as inspiration, if Audio influence is high and the style prompt and lyrics are accurate to what your upload sounds like, it should come out at least close (but it'll usually patch any gaps or bad edits, because it wants a song that flows like it thinks it's supposed to)
Then you can make adjustments to the output, either by covering it with altered prompts or using Replace or Extend

[graceful condor]

Forget what I said. I jinxed it. Now every Extend is all of a sudden terrible.

[polar spade]

You've seen Inspire combine the lyrics as separated verses, set as individual outputs in a Playlist, to have combined as one intended contiguous song? I haven't heard that happening.

I was curious if someone wanted to test, let's find out. ----->> @woeful grove

[hollow ridge]

I made a song and I want to insert a single verse. When I click create it doesn't return the exact verse I entered. I want it to return those exact words.
How should I do it? thanks

[quick sonnet]

@stuck junco Looks like suno sometimes ignores the style prompt when making a cover, just some examples:
https://suno.com/s/WrlJiln2KcyfN106
https://suno.com/s/FP59udQBapntz4t1
https://suno.com/s/BB8QzB5e955VPF5q
Not saing it's a bug, sometimes it could be a great feture.

[unborn star]

That's known =)

[polar spade]

Just to point out: If the tag to a user isnt blue the recipient isnt pinged. Just to let you know. Edit your post and try to edit the tag until its blue (hyperlinked).

[quick sonnet]

Thanks.

[quick sonnet]

Sure, but 10 times.

[wispy laurel]

I tried it quite a few times and only had it work once. I think the key was I put instrumentals with different lead instruments between each verse and then signposted the hell out of it in lyrical tags. So it knew that the bit with the guitar solo should be before the violin solo, and the harp goes at the end. And even with that in place it still only got it right once
Hence being better to stitch it together yourself and let Suno fix the bad edits, it's way more reliable than trusting the robot to solve the puzzle correctly. Possibly will work better in future if it gets smarter about complex instructions

[polar spade]

Once means it possible. Thats a good sign.

[tepid moth]

Okay, so not only me then with that impression. Hopes it only AI having a bad moment, and I`ll try again later.

[woeful grove]

OK , I may even try also to mix manually (with rough blends) and try to use inspo (filled with same sections, wether from extenarl chopping or internal using editor) and see if it manages to blend wether by using a remaster or so

worth testing imo

would be nice if suno devs drop a couple thousand opints for heavy testing on this 😄

I've already expanded twice this month

I reset creditso on 18... so I can go chopping meantime

I'll also try to have multiple layers with mixing different sections, wonder if suno may detect the layering... I really hope suno brings a decent editor because this is too raw

[placid spindle]

Thank you!

[placid spindle]

No I have never tried that because I never saw it! LOL How do I access that? Right Click?

[neat nexus]

Should be in the hamburger menu (three dots) then the Create option, it should be in that menu listing.

There is no war in Ba Sing Se.

[placid spindle]

I don't see it? 🙁

trying to upload a screengrab

[placid spindle]

Create Variation is not avail for me. Is that only on certain subs?

[neat nexus]

https://tenor.com/view/tom-and-jerry-tom-worried-scared-nervous-gif-14995822

YOU SAW NOTHING.

THE NOTHING THAT NO ONE SAW HAS BEEN REMOVED.

[placid spindle]

LOL

NOPE Variation is not avail for me

May I ask, what is ReMaster OPtion? DOes that clean up the mixes?

[neat nexus]

I don't want to accidentally speak out of turn so I'll wait to respond to this until I get an exact wording on it first.

[placid spindle]

from me? LOL

why can't I see
Variation
and
What is Remaster?

[neat nexus]

It is supposed to assist with cleaning up tracks, yes.

https://help.suno.com/en/articles/3591617

[placid spindle]

YES!!!!!!!!!!!!! Thats amazing

Thank you!! THere are so many tracks that I have that are so noisey !! Thanks!

[daring bay]

maan the covers getting cut off 4-5 minutes in bug still hasn't been fixed Q_Q

[hollow hare]

Believe it or not, it is not a bug. Some of the resources dedicated for the generation is reserved when using certain functions, like persona, which reduces the max time.
https://discord.com/channels/1069381916492562582/1404828447695569037

[daring bay]

ahhh. I hope it gets fixed nonetheless…

[neat nexus]

Just in case, you do know that you can extend those and complete the song, right? Even better, you can choose the time at which you will extend from. So you can extend from an earlier point and get the best possible entry moment.

[daring bay]

yeah I know

[neat nexus]

Just wanted to make sure. I know a lot of people started after that bit was added and not everyone knows the older work flows and just expect to get a full gen single shot.

[storm pawn]

I have a question - how to use consistent vocals with newly sounding tracks? I have a persona, for weeks I was creating different tracks not moving any sliders, just using styles and this persona. The songs were unique, vocal was consistent.

Today I just spent 500 credits and ALL the songs sound like a remix of the original song from which the persona was created. What happened? Am I doing something wrong there?

[daring bay]

the one problem with extending is it stops being a cover at that point

[polar spade]

We had a 4 minute limit. Everyone had issues with it cutting off early once Covers and Edit tools were introduced . Others demanded 5 minutes songs. So....The 3:12 / 3:24 cut-offs (mostly for edits of songs already longer than 6 minutes) were "fixed" when 4 minute limits went to 8 minutes, allocating the space needed to seed the edits and have more cleaner and less glitchy issues on the outputs.

The 4-5 minute cut-offs (mostly for song already about 5-8 minutes) will be fixed when limits go to, what, 10 minutes (?) probably.

Any song going over 20 minutes . . . (composing a symphony?) . . . . you should just download that and learn DAW for now 😉

Otherwise, writing songs under about 4 minutes tend to do pretty good.

Just explaining. Not resolving.

[polar spade]

Is that 4.5+ beta? or same issue with 4.5 or 4.0

[storm pawn]

4.5 beta

never tried other ones

[polar spade]

There could be change ups while in beta. Work in progress. Often something to improve one thing that might affect something else. You can point it out in the bugs channels to bring awareness to the change of the feel/vibe /character of the vocals from before. They appreciate knowing those things.

If everything is generally consistent and no prompts changes, the personas, with sliders set to "neutral" should, I'd think, have the same weights and emulate the vocal as before. But I cant say how personas adhere to their own "prompting" and data gatherings from the seed output.

At least , grab a pen and pad and try some new settings on the sliders to see if another spot hits closer to what it was doing before. Write down the settings for furture use. Periodically, check the neutral/default settings in a couple of weeks.

[polar spade]

Oh... sorry.. almost forgot... one save that might work is to go to the original song and make an updated persona. Even though both are personas of 4.5+ beta songs (I assume), if there's been any tweaks by the devs, an updated version then would be a better chance that it works with the default settings , rather than having to fiddle with the sliders each time.

How long before another tweak puts it "out of alignment" , if thats happened, is anyones guess. But updating a persona may work and then you have two persona versions to volley if the other doesnt work well.

[storm pawn]

The vocal is ok, the melody is the problem - in all these tracks I could hear the same piano melody, same chords progression, just different scale / instrument, tempo, no matter what I put in style (alt hiphop, hiphop-rock, funky r&b etc.)

and it was the same melody as in the first track where I saved the persona from

I have two personas created around 3 weeks ago - I used either of them and the pattern repeated - I heard the melody of the original song from where I saved the persona, in various styles (depening what I put in styles)

[graceful condor]

I am not sure if it is because of the new crap UI but it makes me nuts if literally all I kept in my prompt section is this

[FADE OUT]
[END]

Which freakin means: END THIS SONG FOR F SAKE. And it starts to sing again. But there are no lyrics. but it starts to sing each freakin create.

[polar spade]

I misunderstood when you said "...consistent vocals.."

It reads now that you WANT to have the same musical/instrumental motifs (or riffs and melody lines) of the original persona's output showing up in the new outputs?

Thats kind of in the realm of "lietmotifs". Something popular in Classical music as well as concept albums, prog rock or jazz.

That would be cool to have that. I'm envious that you got it on lots of outputs before. Personally I havent done enough in that direction as I would wish to give any insight. And its certainly on my list as I'm writing some concept pieces.

Maybe someone else can step in who's more practiced in this area with the sliders.

Otherwise, sorry to say, try variations in the sliders. I know ... more credit burning. Turn up Audio and turn down Styles, is my first guess. I'd hope whatever the change, it wont be too far left or right from the default settings.

[storm pawn]

Hah no 😄 I want to create new melodies. I provided new styles for the prompt. I just want to use the same vocal. But instead of having completely new melodies, it uses the melody of the original song from the persona

[ancient stratus]

Anyone else get blips in sound levels? Anyone know how to prevent it happening?

[storm pawn]

Imagine I have song A from which I saved the persona. In the span of last weeks I created like dozen of tracks - all sound different, very similar vocal - that's what I expect. However today, I wanted to create a new song, and somehow whatever I put in style, it always results in a melody very similar to song A.

[polar spade]

Adding prompts (optional) and sliding styles up will affect more of the NEW feel of the prompts, while slightly kicking up Audio slider (in the 60-78% range) hopfully will pull in and maintain only the vocals of the persona.

The guess as to why the defaults not working now, I've already mentioned, so report the issue and do a bit of slider experimenting until something clicks. I'd send you credits if I could. Sorry, mate.

[polar spade]

create the song first in any vocal suno gives you. Cover and add persona. Then sliding Audio up will give more weight to the newly created song, while still pulling out only the vocal qualities of the persona and usually the nuances of instruments used. Thats been my results. I use that to cover songs in my Fake Artist's vocals. without imparting any or too much of the original personas melody lines....

I use a persona of a Country band over , lets say, some Christmas songs for a new x-mas album. So the lyrics are same, the styles are same as persona, but I often add about 20% more prompts like, Christmas bells, chunka beat, boys choir, whatever. Style sliders about 75% and Audio about 68% ... weirdness optional.

I get a full album now with steel guitar, banjos, a country beat and the persona vocals, but with the melody line of the christmas song I had previously made, along with a few bells and whistles for the Christmas feel.

[storm pawn]

Thanks! That's exactly what I tried and it works - the only problem I always have tha the first verison sounds best, all the rest always miss a detail here or there 😉

[graceful condor]

I am sure they freakin messed up the AI again because of a secret silent Update or being a test subject for a new crappy AI because they never give us a heads up. Because now the AI is unusable. They can't finish a song. Even if you put all into a prompt or in this case nothing expect "FADE OUT; END" Etc and it still goes on singing without lyrics addin another 3 minutes to an already 6 min long song. And regardless what you do it never gives you a simple fade out of the last instrumental music section. At this points it is just eating up our credits and delivering nothing and we do not even get compensations. They are just messing around, not informing us. Because this is ridiculous at this point. Just ridiculous. I wasted around 400+ coins for nothing.

[tepid moth]

I know! After the update the quality has been real bad, so I`ll wait a few days.

[graceful condor]

Waiting a few days totally messes up my schedule. I am used to do certain stuff because routine is important to me. I have autism. And now my head is full of creative ideas. I really do not get it why they force something onto people which is unstable and not tested enough? And not even offering an option like "Use the old but if you want test the Beta new UI here you go" And you can switch back and forth if you like. But they are not even giving a heads up or any insides what they are doing.

[wispy laurel]

Find the song that you made the persona from. Split the vocal and instrumental stems. Use Edit mode (Crop and Remove - usable in the legacy editor, which works without the bugs that apparently are affecting the newer editor)
What you need to do is firstly, take just the bit of the vocal stem that best represents the voice you want. Then use Remove to eliminate silence within that section. This will break the rhythm and melody of the stem - that's what we want.
When you've trimmed the vocal stem down to just the best bits without pauses or silences, use that to make a new persona. Breaking the original rhythm means that songs made using this new persona, theoretically won't be able to extract rhythm or melody from it, and there won't be any instruments there either - but it has to use a certain amount of influence from the persona, and the only thing it can use now is the sound of the voice itself

This is kind of awkward and annoying, but it really works

[glad mesa]

Is there prompts for making lengthy songs past 6 minutes?

[glad mesa]

Distracted by cuteness

[near flume]

Are you using [Outro: Fade Out] [End] or [Fade Out] [End]? Using [Outro] helps prep the song to end.

[slate trail]

For some strange reason it will not let me position the time stamps to 0 to the ending

I'm on mobile

[trim adder]

Anybody know how to get ambient without any drums? Prompting “no drums” doesn’t work. Drums still come in at some point in the finished generation.

[glad remnant]

Hi, so I’m trying to help my dad with SUNO. We tried doing the record audio earlier because he had a tune in mind. But when we hit the record button it seems like it’s recording, then we hit the stop button and there’s no other prompt. The only thing left is aX in the right hand corner to close out. I’d really appreciate any help because I’ve looked a lot of different places and been unable to figure out what to do next

And we are just using the free version, 3.5

I’ve seen on some help stuff that when you record a clip it goes to your library… But I’ve never been able to find it in the library

[sly citrus]

Does anyone know a smart approch to just creating a simple motif - no frills, just a theme structure that can be used for remixes ?

[opal orbit]

It's better to put "drums" in the exclude styles section, rather than in the style prompt, though Suno tends to add some kind of drum or processed beat anyway.
What's worked best for me is putting in-line descriptors where I want the beat to 'calm down' or stay 'ambient', but that is with specific sections of a song, not the whole thing. What genre of music are you aiming for?

[opal orbit]

I just tested it on my end and it seems to be working ok, so it might be something with your computer or recording input. If your mic isn't actually picking up the sound, then Suno wouldn't register anything and give you the option to save. You should get something that looks like this after recording;

Double check that your mic input is set correctly, or, if its still not working, you can record it through your computer and upload it to Suno instead. Most PC's have some kind of simple sound recorder app built in.

Also check by selecting "Uploads" under the filter option to verify if you've successfully uploaded or recorded.

[storm pawn]

Thanks! I’ll give it a try

[proper owl]

Help me please.
Bug Report: My persona description is being flagged as if it has an “artist name,” even though it’s only descriptive (no names/titles). Please check — seems like a false positive in the filter.

[tepid moth]

After yesterdays horrible experience in the sound quality, today has been a blast! Never heard Suno generate so high quality tracks as for now. So maybe they had to adjust to the new UI and had something magic on their end! Nice work!

[polar spade]

Taking a page from @wispy laurel , one possible solution is you could use her "vocal persona" suggestion on the instrumental side to make an instrumental persona of the best of riffs or motifs. Adjust sliders to gauge how much leitmotif shows up in the new outcome.

[polar spade]

Glad thats working for you now. I like to wait issues out a few days, do some real world stuff, like you say, to then decide whether its a bug or just an update.

[graceful condor]

tried both, same outcome. Literally all you do is the same outcome like if the Ai spliced in right now is resistent against everything.

[graceful condor]

Okay I am done with this. They ruined a great song. Really this is no fun anymore. I tried everything but the new AI model they are testing on us since yesterday it seems is absolutely learning resistent. I wasted easily 500 credits for nothing only the get an easy short fade out of the last instrumental part. Because as you can see the song it is extended from is A Fleeting Dream 1.2 which is already over 6 min long. And it still beyond all logic is creating 4 min tracks and even those are freakin cut off again at the end. I am sick of this. I even let a night pass sat down this morning and still the same.

I am so aggressively already forcing the AI not to use voices again because as you can see there is no FREAKIN LYRIC part in it anymore and it still starts to sing at the same spot over and over and over again regardless when I chose the extended cut it produces the same structure regardless what you add into the fields. And do we get compensated? No...of course not

[normal lagoon]

I need help! The sample rate of the WAV file I downloaded from Suno is 44100 Hz, but when I publish the work on other music platforms, the platform automatically changes the sample rate to 48000 Hz, which causes a problem: there is a distorted hissing sound when listening to the work on the platform. Is there a good way to solve this problem?

[ebon sand]

Can you crop it, fade it and then replace? I usually create ends like that

[graceful condor]

WOW Suno putting out this crap morning? I used credits for this garbage?

[graceful condor]

If the editor would work but it doesn't. And yeah I could fade it in Audacity but that sounds cheap

[ebon sand]

Try the legacy one, I'm using it a lot

[graceful condor]

I made a lot of Feedback in Bugs about it. The new Editor is not loading at all stuck on Black loading screen and the old one crashes if I wanna cut a song. Suno became really a hell lately.

[graceful condor]

Welcome to the NEW SUNO where fun was a thing of yesterday...😑

[ebon sand]

The new editor never worked for me either but the old one always gets the job done for me
Even if cropping throws an error the song is getting cropped

[digital kraken]

Do you have any tips for managing to use only one instrument?
I only manage it once in 40 times, I think. Yet I put a lot of prompts and instructions between [].

[wispy laurel]

Can confirm, this does work
I have a persona made from a guitar solo. With low influence it produces a similar 'vibe' and causes any guitars to play in a similar style. With influence turned all the way up, it will literally generate almost the exact same solo again, if there's any possible way of it fitting into whatever I'm generating
This also works as a partial fix for the "amazing ending but I hate the chorus" problem, you can persona-fy the ending of a song, extend from an earlier point to fix problems in the middle, and then somewhat reliably bring back the original ending instead of whatever the extension put there

[wispy laurel]

It depends what instrument, to an extent. Suno is really resistant to making anything play unaccompanied that normally doesn't
You possibly could split stems from a song that includes that instrument, then use just that specific stem either as a basis for a cover, or to make a persona that's more inclined to solo that particular instrument, or both

This was the only way that I could persuade it to use a harp as main melodic lead in non-classical genres, it took me like a week of arguing with the AI but I eventually managed to get a persona that will happily play a blues harp solo 😂

[unborn star]

I got a question about Mastering prompts

What is stronger [Mastering] or Mastered]? What do you guys think?

[brazen jasper]

hi, im trying to remix song, and make it piano only.

i use Remix/edit > Cover, and set prompt like this

but it doesnot work

i use 10USD/month version, would like to know how to adjust it? thx~

[wooden sapphire]

tried using the adv. options?

[main saffron]

is it possible to recreate a vintage style, that covers the old shellac sound?

[wispy laurel]

Lyrics -
[Instrumental]
[Piano only]
[Rhythm and melody both played by piano alone]

Style prompt - be more detailed and more descriptive, and really hammer home the fact that the piano does everything -
"An instrumental piece performed by a solo piano, unaccompanied.
Begin with xxxxxxxx, building to xxxxxxx, before xxxxxxxx in the final section of the song. Throughout the song, all rhythmic and melodic parts are played by the piano alone."

Use the xxxxxxx's to describe what the piano should be playing at each stage of the piece.

You can potentially use more detailed structures in your lyrics, it can understand [instrumental verse] [instrumental refrain] [instrumental bridge] and you can then give more specific details in each of those sections too. Just make sure everything in the Lyrics field is clear, concise, and enclosed in [square brackets] so the AI won't start singing the instructions

[polar spade]

Whether [master] works or not, I think of single prompts more often as nouns. Not instructing Suno to do something, but the described result as a I want it. Mastered i would say [master] (not capitalized) as in the result is THE master.... not a mastered result.

I dont want it rocked, or folked, or do some hardmetaling. I want rock, folk and hardmetal.

I want a [whisper] not whispered or whispering.

Strictly my opinion and taste. I have not done 1000 comparative gens to test the theory. I may be going down the wrong path.

Exception for -ing words is I dont say [audience] [applause] [clap] , I would say [clapping] or (smattering) as a gerund nouns. Still nouns.

[unborn star]

Hmm means, when using Mastering means it build in nuons, which can be interesting as well 🤔
When it is possible, I will test it later.

[polar spade]

Sure. Users have used: crackle, vintage, 1940s, 1930s, Mid-Atlantic, gramaphone, nostalgia

Mid-atlantic is a western bias. You could use "Shanghai nostalgia" and probably get similar results.

Though the fidelity of shellac records [78s rpm] was warm and muffled, Lo-fi is possibly a help, but may hurt as lo-fi is a current genre for a current sound and so you'll get more hip hop results.

[wispy laurel]

Yeah you'd probably want 'low-fidelity' rather than 'lo-fi'

[brazen jasper]

it's more close now, i'll keep to try, thx for your direction ❤️

[wispy laurel]

The more details you give it, the better. It helps if you have decent knowledge of music theory, but even if you don't, if you can clearly explain what you want, the AI can usually do it.
Sometimes you run into problems when a word means one thing to you but means something different to the AI. The more you experiment, the more you'll learn Suno's "language" and be more able to explain stuff in terms that it will interpret correctly

[polar spade]

I know you meant noun, but if it was built in "muons" (where my mind went when I first read that), we would have a negatively charged output! 😅 morning Dad humor. Sorry.

[main saffron]

i'll give it a try, thank you

[unborn star]

Good thing you cleared that up. I was starting to worry my grammar might decay like a muon! ;)

[sleek stump]

This should be very simple. Put the track in Audicity or some other DAW and save it out as a wav at 48Khz.

[glad remnant]

Thanks I appreciate the advice. I will have to mess around with permissions and see if that does anything, it should have access to the microphone but I really appreciate seeing that screenshot because I don’t get the save option. This is what I get after I have hit the stop button. I also looked under the filter for uploads and I don’t have anything so it definitely isn’t saving automatically. But at least this gives me a better idea than I had before of what needs to happen

Maybe I need to get the app, I’ve been using the browser on my phone and my dad’s iPad

[upper escarp]

I need some help fixing vocals to a few of my songs that will be going on my album.. Is there anybody here that would be willing to help me? I would be open to putting another name on the credits for my songs if I can get sufficient help

[mellow moat]

Hey guys is there anyway i could upload an instrumental track and add vocals to it and keep the instrumental exactly the same?

[livid summit]

Anyone have problem with female vocals - no matter prompt you give, it regardless sings with veeeeeerry looong vowels and lous, shouts.?

[polar spade]

This has been popping up in general. Prolonged singing of words. So far, with my folk and pop, not so much. If I can sing the same line in one breath, I give it a pass. But it does seem to be across the board, perhaps. And not male/female specific.

[livid summit]

is there other than male/female? a robot?

[polar spade]

....if they prolong their vowels. Sure. Why not. I have an androgenous chipmunk singer.

[raven turtle]

I’ve been using a negative prompt of “vocal runs” which seems to help

[hybrid pagoda]

Having issues with shortened song generations now 6 times today

[hollow hare]

What prompts are you using? How long are your lyric? Are you using persona, cover, etc.? Those affect the generation's time

[hybrid pagoda]

Ver 4.5 songs are 5 mins long and use about the same amount of lyrics each time

Create Custom profile in suno

Last credit 6 times looking to see why this is happening now after months

[hard beacon]

new creation ? Cover ?

[hybrid pagoda]

new creation

Everytime new creation

[hard beacon]

can you copy paste your lyrics (song structure)?

[hybrid pagoda]

yes

[hard beacon]

forget about the time stamps 😉 and set it into [ ] instead of ( )

like [Verse 1]

not (Verse 1)

[hybrid pagoda]

ok got it

[hard beacon]

test it

[hybrid pagoda]

will do need to go line by line let you know

[hard beacon]

btw... version 4.5+ isn't able to create mixable dj intros for electronic dance music... covers in 4.5 with specific promts can realize it 🙂

btw²: that was an offtopic comment 😉

[hybrid pagoda]

Still a hit and miss, but I made other changes to see if I get better results. However, it is better than before, thanks to the change from ( ) to [ ]. I will keep working on it until I get what I am looking for in this song. I was able to get the last two sets to work. Thanks

[quick sonnet]

@stuck junco Have noticed there are at least 10 remixes of that "happy" track which sound very similar to the original, while the styles they use a very diferent. The only common thing, all that remixes are made with v3.5. While here it sounds more like bug, generating more or les exact the same result, could be also used as a feature.

[tepid moth]

Is the add instruments to vocals and otherwise sometimes buggy? I am a premier member, but today I got that "buy pro or premier too access" to use that function. Not a big deal since that is not often within my workflow, but anyhows, just wondering if anyone else experienced that. Everything else as a Premier member works.

[polar spade]

A glitch may be.

To compare, I clicked add instrumental to a vocal stem of a remix. No problem. No pop-up to upgrade.

[tepid moth]

Okay. Thanks for checking. Then I`m glitched out on that part. Everything else works as a paying member, but since I rarely use that I will just wait and see if Im beeing unglitched.

I sent support a message just in case I need it, so they can unglitch me

[static river]

Hey guys, is there a way where can I connect these 2 parts of the song? I'm trying to shorten the song duration

there was a verse part between them, I deleted it and now I want to connect these 2 parts

[ebon sand]

do you mean visually removing the cut or in terms of replacing the transition? if the latter then, well, replacing over the cut and hope it works

[boreal nest]

I'm continuing to consume credits to get results that are the opposite of what I write in the style prompt. I also reinforce the instructions by inserting in the text between square brackets what must be done and how, especially the vocal refrain, but there's no way, I have all the songs with street-shrieking singers, excessive high-pitched vocalizations like someone dismembered alive, how should I get whispered, whispered, spoken voices without having to scream like crazy?

[silver turtle]

I had started to use Suno to make new cover versions of some Fallout 4 oldies (cover in the traditional sense, NOT the Suno 'Cover' feature). I did three songs with good results, at least I think so, but I'm not a musician or anything. Two of them Suno allowed the original lyrics to be used; one of them Suno did not. I used Grok to "rewrite the song just enough so that it's the same but legally-distinct" (that works really nice btw lol). I also tried converting those original lyrics to phonetics but it just didn't sound right and I didn't want to spend a lot of time making it sound right lol.

Anyway, here's a couple sets of tags that worked well for me:

sexy female,lounge singer,1930s,vintage,oldies,1940s,big band,jazz,analog,lo-fi,vinyl,phonograph,broadway,musical,brassy,classical,ragtime,marching band,patriotic,theater,opera

sexy female,lounge singer,1930s,1940s,1950s,broadway,musical,big band,jazz,quartet,vintage,oldies,analog,lo-fi,vinyl,theater,opera,brassy,classical,ragtime,marching band,patriotic,radio,

Legally-distinct "Dear Hearts and Gentle People": https://suno.com/s/zvGgoURuTrcbhNkc

[rigid folio]

why am i not getting the 50 credits a day now? how do i get this seen by support?

[split tundra]

Did my message get automatically deleted or did a mod delete it? Is it not ok to post links to custom GPTs? Just want to know so I don't get myself banned 😊

[split tundra]

Any tips/ideas on how to tell v4.5+ to never generate mono samples/vocals? Trying to figure out if there's a way using tags in lyrics or add something in the styles field.

[silver turtle]

My post in workflows just got deleted for some goddamn reason. For a split second I saw a "NO NO NO NO NO reply from a mod I guess, and then it was deleted. WTF.

[bronze helm]

@split tundra @silver turtle It was Woodhouse most likely. About the workflows post I looked at the members and Woodhouse had joined that channel. Bot's the only account with elevated privileges there.

Now as for who controls Woodhouse (most of its dumb filters are automated) I have no idea. Sometimes Woodhouse does send random messages. I have always assumed that it's someone with control over Woodhouse trolling people. But in this case they literally deleted an entire post. <@&1082792303976001547>

[silver turtle]

yeah that was it, Woodhouse. I was editing the post, adding to it, reorganizing it, thanks for the info

[bronze helm]

I hate Woodhouse with a passion

And that's an understatement, don't get me started

So it would be nice for this sorta bs to be sorted out.

[silver turtle]

Yeah I dunno if it was flagged cause I made several edits in a short time or something? All I saw was the bot message at the bottom say "no no no no no-" and I couldn't read the rest before it disappeared when it deleted my post. Oh well I'll sort this later I guess lol

[bronze helm]

Yeah you shouldn't care about your edits no matter how rapid they were. Woodhouse regularly deletes entire posts and entire paragraphs on a random whim because the mods set up a non-AI filter that can't even distinguish between fanaticism and genuine religious motifs for Suno songs

[livid summit]

it's got confused by the devil himself!

[viscid lagoon]

ambience help.... please

[polar spade]

Using too many dashes, spaces, repetitions like hashtags. Adding any outside links. Could be a number of things.

Always copy anything over about 4 lines long, just in case, before you post. Or just draft somewhere else to paste in the post.

Probably if your posting a workflow, it shouldnt be a link to an outside workflow. If needed, copy and paste as your own post , and reference the source.

[untold briar]

I can check for you, but it's probably just some random punctuation that triggers something in the auto moderation. Especially things that are used in Markdown syntax seem to trigger it.

[untold briar]

The logs don't specifically say what triggered it, but the reason was "Duplicated characters/words".

[solemn mantle]

Do you guys know you can really compose music by adding [ ] and ( ) in between your verses/chorus to guide suno in making whatever intrumental solo you'd like?

example:
[Intro – 8 Bars Instrumental]
(Sax drifts in slow, like a midnight train callin’. Guitar picks a lazy rhythm, each note heavy with dust. Organ hums steady, low and warm, while bass and drums roll soft like worn-out wheels on a lonesome road.)

or:
[Outro – Sax & Guitar Fade]
(Sax sings the final melody, long and aching, fading into the distance. Guitar answers in soft, bending whispers. Organ hums low, fading slow, as the bass and drums gently dissolve into silence, leaving only the echo of the journey.)

Try it out! 🔥

[main saffron]

is it possible to set the music style for each verse, something like this? [Intro: dark electronic pulse]
[vocal-style: whispered, ghostly, detached]
Die Straßen flimmern, als wären sie aus Rauch,
ein leises Zittern liegt über der Nacht.

[Verse 1: glitchy beat]
[vocal-style: spoken, glitchy, robotic]
Ich zähle die Schatten wie kaputte Zahlen,
meine Stimme bricht im Rhythmus der Drähte.

[Chorus: wide synth pad]
[vocal-style: sung, soaring, emotive]
Lass mich fallen, höher als der Morgen,
zwischen den Lichtern brennt ein neuer Anfang.

[Verse 2: funk groove]
[vocal-style: playful, crooning, scat]
Die Wörter tanzen wie Münzen auf Asphalt,
ein stolperndes Lachen springt über die Takte.

[quick sonnet]

You can try. But do you really want it in German, or the English translation?

[main saffron]

the first style seems to be very dominant, extreme style switches do not seem to work

[quick sonnet]

Suno sometimes just ignores instructions.

[main saffron]

it's just a test song, for science 😄

[quick sonnet]

But since suno tries to sing it and rhymes mising, could sound strange.

[main saffron]

you don't need rhymes, most of my songs have no rhymes 🙂 it's more sung storytelling

[neat field]

when did v4.5+ remaster become available ??

[opal orbit]

Suno is not going to do extreme style switches like that between verses. Certainly not that often at least.

HOWEVER If you regenerate the same song a couple times in different styles, splice various peices together in DAW, then re-upload to SUNO and cover, you might get something interesting. For cleanest merging, try to gen all the songs with the same BPM, (put 120 BPM, or whatever you want to use, in your style description and in the beginning of the in-line lyrics for each generation).

If you do this, pretty please post the results? I'm really curious to see what you'd get.

[polar spade]

About two weeks ago it was noted in the menu.

[hard beacon]

It's available since 2 weeks. but only on the Browser and not in the app

[tranquil leaf]

I've hadn't been able to get it to do drastic style shifts like that, but I have been able to get it to start soft/slow and build over the course of the song. I have one where it starts off essentially a slow ballad and ends as a power metal song.

[thorny cypress]

Ok I'm not sure if this is the best place to put this but I didn't know where else to put it.

So I was trying to pay with PayPal but the web version doesn't do that, and when I tried to set it up in Google Play so I could do it through the app, it told me I need to attach a card to my PayPal account. Like wtf I thought PayPal existed so you didn't HAVE to use your actual card and stuff lol

[brisk flint]

Just last week, I noticed that within a genre, a song can change drastically in its arrangement when I turn up the “weirdness” slider. Around 65% to 75% is the range that sometimes produces quite strong results for me. If you like, I can post an example where, after about 2 minutes, the break becomes completely different from the otherwise minimal track.

[tranquil leaf]

I would try putting the transition into the style field though, instead of inline. They really seem to be leaning towards the style field giving more weight. So if you tell it "slow and whispered in the first verse, transitioning to glitchy and robotic in the second verse, to a funk groove at the end", etc.

You might have to play around with the exact wording.

[tranquil leaf]

I haven't messed around with the weirdness slider much. Feels like adding an unpredictability I don't want. 🙂

[brisk flint]

Yes, that's right, but because it's easier with an example, pay attention to what happens after about 2 minutes, and I'm convinced that it was due to the high value of the weirdness slider (I think it was 75% here): https://suno.com/s/7gaoQf0RAfpTANrI
Nevertheless, the rest of the track follows my detailed prompt pretty closely.

[opal orbit]

Have you been able to navigate it to a style you prompted? Or is it making a complete change to a random one?

I haven't messed around with the weirdness slider much recently.

[brisk flint]

Very good question. With values up to a maximum of 75%, at least in my experience and with my preferred genres, Suno continues to adhere pretty strictly to the style prompt and the information in the lyrics section, which I also fill with prompts 😉

[tranquil leaf]

It's an interesting transition, but it works for the song.

[opal orbit]

Huh. Nice, that's an improvement from my previous experiences with it. I guess multi-genre songs are more attainable than I'd thought, haha.

Have you messed around with doing drastic style changes more often, like every verse/chorus? What's been your success rate with stuff like this is general?

Curious to hear if you've got anything with even more sharp style swaps mid-song.

[brisk flint]

@opal orbit I don't have any real experience with that, because although I mix styles, I do so consistently for an entire track, i.e. I don't try to change genres within a track:
I've tried it from time to time, e.g. adding a garage or 2-step break to a 4/4 house track, but I'd guess that it only worked in about 10% or even less of cases. I had prompted this in the style prompt, at the beginning of the lyrics prompt and at the corresponding point in the lyrics, but was rather disappointed at how rarely it worked 😉

[opal orbit]

Ahh, gotcha.
Still, good info. I'll keep that in mind if I mess around with it.

[wheat cipher]

Question for all: I’m finding the Inspo button + a Playlist to be the most reliable way to get a song I like. The problem I’m facing is that I can’t seem to add a song to a Playlist unless I Publish it, making it Public. If I don’t publish it the Playlist says it has 0 songs in it. Is this a bug or do I really have to make my unfinished songs Public. (Currently my workflow is to Publish the song, use the Inspo button until I get something I like and then Unpublish the songs.)

[opal orbit]

That sounds like a bug. I would post it on the #bugs channels depending on what you are using. Are you using the app, desktop version, website on mobile? And are you free or pro-premium? Want to see if I can replicate.

Also, is it showing 0 songs after refreshing inside the playlist itself? Or just while adding it as audio influence?

If your referring to it showing 0 as audio influence, that's just a visual bug. It is still taking the audio influence from your playlist even if the songs aren't public.

[sterile yacht]

gonna paste this here, any idea how to resolve this issue:

#🪲┃bugs-web message

is on mobile and web.
TL;DR: I generate a song and i like majority of it but want to use cover to for instance change the vocals from female to male or change the lyric language… but the new generation finishes after 1 second and is exactly the same as the original with maybe a little bit of noise at the start or end.

is on free tier cause i’m a poor guy that is not allowed monthy or yearly subscriptions

[opal orbit]

What works best for me is turning the weirdness down below 10-15%, style influence to about 20-30%, and audio influence at least 75% or higher (depending on how much you need to change).

In the style box, only give the gender and vocal description, nothing else. (Assuming its just the vocals your changing)

If you're genre type does not have a lot of compression heavy instruments (synth, electric guitar, metal, ect...) you're safe to turn down weirdness even more.

[quick sonnet]

Thanks to a possible bug that seems to be present in v3.5, I was able to partially fix a broken track. Here's the original:
https://suno.com/s/yGIf63ZvZWR9eo4Q
And here is the improved version:
https://suno.com/s/5TuyFu79BHsSkzGl
There are still some issues there, especially with voices, but I hope this potential bug will become a feature, even if I probably won't be able to benefit from it myself.

[sterile yacht]

don’t have access to advanced settings like weirdness (at least not on the mobile app what i use most often)

[quick sonnet]

As a free user you can't manipulate weirdness, style influence and audio influence.

[opal orbit]

mmmm shoot. I forgot about that. In that case it's kinda difficult to get it to replicate exactly. Adding alot of inline instructions for [male vocals] or [female vocals] helps, as well as emphasizing it in the description box is the best bet.
Suno still tends to swap vocals genders partway through when covering a song in the opposite gender though.

There are some other AI tools online for covering tracks, not sure how well it'd work with SUNO but another possibility.

[quick sonnet]

That not really helps, suno doing what it wants, it just ignores [male vocals] or [female vocals]. My link above is the best example of this.

[sterile yacht]

yea, the only cover like change i am able to do is turn a regular vocal into vocaloid.. but that’s all it is able to do in the free tier.

[opal orbit]

Yeah it ignores it most of the time. But it's more effective at getting segments of consistent voice, rather than changing pitch and tone mid-way through. It's useful for covering a cover occasionally.

But yeah, getting vocals correct is an issue for everyone

[fluid copper]

If I have a song in the library and put it in EDIT, If I am just trying to change one word or one line, it's always changing the voice and sometimes the music. Is there a way to do this I don't know (I'm only using Suno for a drop over a week.)

[sterile yacht]

sadly that level of fine tuning is not possible without it basically making a complete new song.

hopefully suno 5+ will be 10x smarter and granular

[fluid copper]

It sorta worked ONCE. Probably by accident 🙂 . Other times it changes the inflection of some words which throw off the line. I tried replacing the whole piece ([PreChorus] or [Verse 2]), but it changes the music and voice. It's too bad because I like the original except for some words that need changing

[sterile yacht]

same issue i have… the song really needs to be sung by a male as lyrics don’t make sense if by a female 🤣

[fluid copper]

I guess I can play around with it and if I get one where the vocal is good, but the music is too different, I can get the stems and edit them. I'm new here, but I've been using stuff like Adobe Premiere Pro since the early 90s.

[opal orbit]

When you regen the segment, how long is it and have you tried completely emptying the style box first?

[quick sonnet]

If you don't have a pro, it won't do you any good personally.

[fluid copper]

Actually, I left the style box the same as the original.

[sterile yacht]

if 5 releases free might get access to 4.2+ 🙂

[quick sonnet]

You mean 4?

[opal orbit]

If you leave it empty, the only influence it has to go off is the audio you have. It's gotten the closest results for me.

Also select a segment slightly larger than what you need, with the start/end points being in the middle of when the singer is singing. This gives SUNO a better reference to how the audio and vocals should be matched

You can shrink the edited segment down after to where you want it once you've inserted it

[fluid copper]

That's exactly one problem. The original was a male vocal and it gave me a female vocal version too. I actually liked it a drop better but two or three lines had to be changed. And It never really gave me what I Really want. But what I got was okay

I'm at Pro membership

[fluid copper]

Good idea. I'll try that this evening.

I wish there was a way to download my Library with one click so I can clear it out (my next thing)

I was just reading in "updates" about migrating v3 to v4 (4.5). How do you do that?

[storm pawn]

Just a hint to you guys - I recently had a song, that had poor vocal quality, but great composition. I've spent more than 1000 credits to recreate it using cover and different personas - always there was something wrong (either the chorus sung differently, or some words cut that couldn't be fixed with replace section, or some significant artifacts) - today I decided to try with values <5% weirdness and <5% audio (nothing in style) - to my surprise, all 10 covers that I created sounded really close to the original, but none of them had any weird artifacts.

It seems that turning audio down very much, but leaving style empty, is helping to avoid a lot of issues keeping the composition of the original song.

[polar spade]

Did you happen to click Cover and not touch the sliders as a test comparison, to let it make a cover with default settings. I think turning sliders to near 0% (but not 0%) is similar to not doing anything at all (default 50% settings).

Before slider were provide, just doing a cover with no styles was just a way to get the same song but hopefully clean up some issues. Similar to remastering. So, thats what you've discovered. 😊

If you have hiss or clicks or issues, Covers (at just default settings) will often either eliminate the noise or attempt to change the noise to a more instrumental sound while keeping melody, on and the general sound about the same -- with some changes still, like gender changes on vocals.

[winged gyro]

I was having a great time creating music. But, suddenly the create button is no more longer orange. Now, It is black, and the button inside Style section disappeared. So, I am no more able to generate additional suggestions of styles.

What happened?

[polar spade]

Add something, a word , a space, a period, in the style or lyric box to see buttons or the Create button to turn orange.

[winged gyro]

I did that. It remains black.

The button still works. So, I can create songs. But, the real problem is the button in the style section. It no more longer there. So, I have to guess the styles...

[polar spade]

Do you mean one of these buttons? You can always log out and back in, or close and reopen browser just to start things fresh, see if that helps.

[quick sonnet]

How about just refreshing the browser, sometimes it helps.

[winged gyro]

See the pic

The first thing in this line used to be a circular button. So, when you clicked on in, it used to create more styles suggestions....

[quick sonnet]

In the world of the woke ideology it makes sence. 🙂 🙂

[polar spade]

I see the "archive" button of prompts [library icon] and the suggested prompts. This is NOT what you see after login in or restarting browser?

[winged gyro]

I did log out log in and nothing changed. I will restart the browser now

[polar spade]

Can you screenshot the "circular" button? I think issue resolved?

[winged gyro]

It disapeared. In Its place you have now the "archive" button

[polar spade]

Thats sounds like a good sign.

[quick sonnet]

And what about pip tones, not talking about explicit lyrics, can they be eliminated?

[polar spade]

Im stupid. What's a pip tone?

[winged gyro]

Nothing changed. See the black create button even with text in the lyrics section. Also the archive button is there... It makes my life more difficult...

[quick sonnet]

Most of the time it's just very loud sounds that often drown out explicit lyrics, as example. https://www.youtube.com/watch?v=TDykSetMKog

[polar spade]

First. Yes. That looks glitchy, but not a bug.

Make mention of that in the bug channel so they can investigate. It could be a bit of a bleed over, probably from an update. At any point, the Create button should at least be a greyed out orange colour. 🍊 [ or blue]

Clear cache as well.

[winged gyro]

Tks. It used to be orange. It changed to black today. I will talk to them there

[polar spade]

Got it. I dont know how it treats a "beep" sound. It might keep it, assuming its not so much noise but part of the song. You just have to test.

Sometimes Cover will make an instrument out of a cough, audience noise, clapping, wind, hiss, click... I dont know about beeps/pip noises.

Stick [beep] in a few places to get the sound... then Cover it. 🙂

[quick sonnet]

I think its really called beep in English, my fault.

[polar spade]

Oh. No problem. I just need to keep learning. Thanks for the new word. 🙏

[quick sonnet]

Not often having such tones. To be honest only in a single track, have noticed 3 such short beep tones, while the rest is realy great.

[polar spade]

It may be a hallucinated tone though. So remaster or cover and see. Even a strange split-second guitar lick in the middle of a verse can be a hallucination.

[fresh wave]

even if the stems aren't club quality they are useful for making new songs with a good vocals generation or adding your own vocals to a track or making a kareoke version

i think it is good when you want to lock in a vocal style or feed something back into suno

[flint kraken]

Has something changed? the covers i'm getting now... are really... just.. odd... to the point that i'd like to check in on my AI Singer.. i think he may have had a stroke...

[fresh wave]

Did you accidentally escalate the crazy slider?

[flint kraken]

Nah.. I've jumped back to 4.5 rather than 4.5+ it seems to be more reliable.

which is sad coz i think 4.5+ gives a much nicer sound.

[storm pawn]

Yes, keeping default sliders often had issues with structure - since the first verse is slightly irregular (is missing the last line for a buildup), leaving default sliders often resulted in extending the lines to cover the gap at the end of the verse

[polar spade]

Good you have it to work out your own way, then.

[lusty juniper]

any way to add static or radio station changing channel effect to the track? i am starting the song with an RnB vibe and i want it to suddenly cut off

[glacial rampart]

Anyone notices when trying to remix to a cover, the track's quality suddenly decreases and then back to normal?
Also, tracks get chopped like it's getting off beat on a dusty CD player.

[manic vapor]

New to Suno, LOVE what it does but have hit it only playing the first minutes of the two 4.5 versions it provided (earlier versions it's fine) https://suno.com/s/RASQkYgwOLxtHEnV

I did pay to upgrade from free and 'unlock' the songs, got to hear them once or twice, now - sign out/in, refresh etc. - Gemini said it's a common, frustrating bug in Suno. Said I could download the song but it's about sharing ya know?

Oh figures, HERE it plays the whole thing!;-)

And I copied the link to play in Firefox (mainly I use Chrome) - same first minute dangit;-/

[hexed iris]

New to Suno Discord. Suggested by Suno CUST service to join as I could not get answer to my issue on reddit.
For a street jam kinda song how to get group/crowd vocals/singalong. I tried to test with just one verse or one chorus separately with commands inside bracket in lyrics and/or in style and it works. But when I am trying to use the same for songs it doesn't work etal.
Please guide.
Alongside, also help with prompts for echo effects in lyrics and/or style.

[hexed iris]

Yes. Everytime. Better to try a new composition unless you want the exactly same vocals /music/instrumental/ style or if you are an expert in mastering!

[glossy plaza]

Any suggestions on suno ignoring gender prompts? It drove me THIS crazy (no gender changes attempted was just a way to laugh instead of throwing the keyboard) https://suno.com/s/7Ex1KgAt2Py7hoeK

[polar spade]

Last bit first:

The simplest "echo" effects is to repeat the word or line in parentheses.

Such as:

This is the way (way) (way) (way)

Or This is the way (This is the way) <<--- more a repeat or call and response feel

[polar spade]

Referencing the[ crowd ] as a metatag at each point you want participation is helpful. Promtping same styles as well in the styles prompt. " Audience" will get a larger audience, so you could try [street crowd] which i throw in my folk songs to keep the audience smaller. Overall that's for my smaller auditorium or café feel.

You could forego a full one-shot attempt for creating the song in segments and extend for each segment with prompting, tweaking until you get the vibe and sound tyou want, then move on.

Anything to get Suno to believe, overall, that this is a grungy, hip, street crowd that wants to sing along, so adding prompts of styles that have had sing alongs in that style, will be helpful. From "hip hop" , "edgy", punchy, in-your-face, free form, ad-lip, dynamic, chant-like, and even "musical" and "ensemble", [think West Side Story. ]

As always, if you get 1 out of 10 gen tries a success or close, then thats a win in my book.

[polar spade]

I assume this song is no sample referencing to do with your issue? As there are no tags encouraging gender references in the lyrics.

[polar spade]

Congrats on your first starts with Suno.

That IS unusual these days to have it just sing the first half of what are a modicum of lyrics. It would normally get through about 4 minutes of all of those lyrics.

It could be a matter of just "going out and coming back in". So delete cache, clear anything you can, just keep one suno site open, log out, log back in, close browser open browser and refresh page to get a clean Create page and try again. Rule out anything stuck in cache or local storage that might limit it

. . . And try another. .

[polar spade]

Reliably and purposefully, I haven't gotten it to do that yet. Ive gotten some radio station changing from hallucinations and even some similar effects from sliding weird slider to near 95%, but thats not much help for you, I think.

Since you mentioned that. Im going to find that Christmas song that had the hallucination and try making s persona of it. Ill see what that can do.

[hexed iris]

Thanks a lot for both the tips 🙏 Totally agree on the 1/10 part.
But for my next release I’m kinda stuck after the 50 mark...feels like something’s missing, maybe that crowd/audience singalong vibe. It’s so close, yet feels far. I’m looking for something like yours....that street jam / folk / café vibe with a small audience feel.

Quick Q...when you add lyrics, do you write it like:
[Verse/Chorus - Crowd] ?

And do you also mention that in the style/description section?

For me, I don’t just jump into Suno with random AI lyrics + prompts. My lyrics are original (mostly from known writers), and I usually imagine the whole composition in my head first...that’s what takes the most time. Only after that I move to Suno, and it never gets done in one go. One song at a time - one in a month.

[glossy plaza]

Ahhh so yeah the 'song' just summed up my frustration ( in a humorous way) with different song that had male and female call and response-esque that kept switching genders and got worse and worse, but the 'clean' / refreshed page might work that never even crossed my mind as I assumed the 'work' was all done server-side so I'd have never thought something might linger from client side

Hehe re read and realized you answered someone else but it still might help me so 2 for 1 Thank you! Even if it was on accident 😀

[polar spade]

Im a single prompt at a time person from starting with version 3.0, but whatever works for you is fine too.

My lyrics would look more like :

[Verse 1]
[street crowd] [Ad-lib] (clapping)
Take me to the top (yea)
I ain't gonna stop (no)

[Crowd] "yea"
We gone make it work
(Work) (work) (yea)
Show me what you got

And even then, id likely "fail" and so slowly tweak them around depending on the previous outcome.

Id even click the magic wand, get a full paragraph and try one, then use any ideas it gives and ggo back to single prompts. Usually, once we do one or two outputs, and make one or two changes, we can see the direction a change makes and keep tweaking to get closer and closer.

"street crowd, audience, hip hop , edgy ensemble, punchy, in-your-face, free form, ad-lib ensemble, dynamic, chant-like, musical, improvisation, ensemble"

[hexed iris]

Thanks so much! If you've made any song on Suno public, would you mind sharing one?

[polar spade]

Hope you got what info you needed. This message now deleted.

[quick sonnet]

It's common practice for many other AIs to generate two different versions at once, but I've noticed that the few AIs that consistently generate only one version generally deliver better quality results, but then potentially have problems elsewhere. Wouldn't it be more beneficial to generate one higher quality version for Suno, too, rather than two versions that each often have specific traits?

[main saffron]

now it works, it seemed to be a to "complex" style prompt, a shortened version did the trick https://suno.com/s/by6qTj8qsY2Z6SSG

[plucky bloom]

Not a fan of the new track listing in Create, the tracks don't stay selected when you select them!

[cunning mulch]

How does Suno link lyrics to the audio for a video? Can this be converted into a .lrc file we can download?

[polar spade]

Its hard to debate a subjective opinion. But that helps that there are many AI sites out there to choose from.

[lucid drum]

Are there any tools for a 'reverse' prompt, ie link or upload a song then it gives you a generic description you could use in a prompt?

[quick sonnet]

If you uploading an audio to suno, it generates some prompt.

[lucid drum]

Ah nice, I'll try, thanks!

[quick sonnet]

Possibly you could just explain an advantages of generating 2 tracks instead of 1? Sure such explanation could be called subjective, but it's still interesting to hear.

[hollow hare]

I cannot find the message but, mentioned a LONG time ago by a Team member, I believe they said it was more efficient resource-wise to have two tracks generated at once compared to having the track generated twice individually.

As people often generate multiple tracks of the same input, they decided to go for the two at once

[quick sonnet]

But if the first track is good, why generating the second one?

[hollow hare]

Because how would the AI know if it is good or not? That is subjective

[hollow hare]

Think of it this way: your time and effort is better spent putting down two railroad tracks at once (left & right sides together) rather than doing left-side-than-right-side

[quick sonnet]

If it's not good you would ask the AI to generate an other one.

[hollow hare]

That is the point. The team member said it basically is cheaper (on their side) to make two at once.

Imagine it like this: to them, it is like spending $2.00 for having a 2-for-1 pizza. If it is done individually, it would cost them $3.50 for two pizza

It is about using the resources efficiently. Music generations uses a LOT of GPU (based on the figures given by YuE), we are talking about usually >>24 GB VRAM for timely generations

[quick sonnet]

Sure, but the resorces are shared by 2 tracks, using them just for a single one would posibly give it more quality. Now about the price, the generation of 2 traks takes you 10 credits, I personaly would not car to give the same 10 credits for a single track, but beter quality.

[polar spade]

No thanks. Respectfully, I wouldnt want to discuss opinions and get into a debate. Thanks mate.

[quick sonnet]

I see your point, but possibly having a chance to select between 2 models, one model generating 2 track and an other just 1, but as yo already said it's a subjective opinion.

[hexed iris]

Yes, thank you so much! Yet to try.
I like the idea of one prompt at a time. But AI is so unpredictable. Would luv to see one of your composition.

[polar spade]

Yes. Profits are stronger when giving someone a 2 for 1 deal.

[polar spade]

Ill repost the sample i did that moment when we chatted yesterday.

[hexed iris]

I did save one of the sample. But I would luv to listen to the final composition out of one prompt at a time.

[polar spade]

There won't be a "final" composition. This was a quick sample demonstration of putting prompts and brackets in lyrical form to get something back. mate.

Its probably 1 of 8 ways to do this.

Frankly much of what is learned applies to the AI level we had previously, as AI gets better month by month, these techniques could become obsolete.

[polar spade]

Ahhh... then ill delete again. If you already copied. Thanks. 😁

[hexed iris]

In AI there is nothing called a perfect prompt. Even the creator can't ensure a perfect prompt. So there won't be any perfect output, especially in creativity. Inputs for a song are not that simple, even if just mention verse, hook, interlude.... etc
And for people who compose the song in their imagination first before going to Suno, 2 is the best thing.

[quick sonnet]

I was not talking about perfect prompt, something like this doesn't possibly exist, at the end it's a gamble and I am talking just about increasing the chaces.

[honest tendon]

Pardon my lack of language on the matter but im hoping maybe someone can tell me what exclusion to use to prevent or remove the ticking sound that comes in around 53 secs? I get it alot in my track mixes. It does generate versions without it but I'd like not to waste so many credits 😅
https://suno.com/s/VDeSyVPwJagG0AqR

[timber robin]

When making an album with the same persona, what percentage do you recommend for audio influence?

[manic vapor]

I did most if not all of that, even went to Firefox and it only played the first minute! BUT on the Android APP, the whole thing plays! What the...?

[polar spade]

Do you mean the high-hat taps? The ts-ts-ts-ts of tapping on the high-hat to get a crisp "chick" sound?

You could try to exclude "staccatto" or "staccatto percussion" or both with a comma "staccatto, percussion" which your song does not seem to have otherwise.

If its clicking noise like it shouldnt belong I didnt hear anything on my playback.

[honest tendon]

Hight hat exclusion seemed to tone it down at least..ill try that staccato etc..I don't know music terminology..tysm! It is ok on some tracks but doesnt fit when remixed down to soul or rnb

Like its there but not in the forefront with high hat exclusion..I tried to remove Male vocal but the original style seems too strong

[polar spade]

Yea. Just unusual. Worth report as a bug as feedback.

You could try to force a longer output by putting maybe "dreamy" or "ethereal" which Suno seems to like to make extended outputs from. Just to see if it will go past the halfway point in the lyrics. If it works, then tweak from there to get a vibe back to what youre wanting.

And if plays on Android app, then definitely a bug ! https://discord.com/channels/1069381916492562582/1082775596767117312

[polar spade]

Cool. Hope one of those works out. Its always a bit of roulette to what will work.

[honest tendon]

Always. 😆 🤣 if it can't remove..ill take toned down..im on mobile so I can't pop in and remove or reduce it

Should increase my ratio of exclusion at any rate. 😆

[quick sonnet]

I already mentioned this yesterday. An instrumental track was generated for me quite randomly. It actually turned out quite well, but it also generated at least three beeps. To get rid of these annoying sounds, I generated a cover of the track, and the result was one that's very similar to the original, but unfortunately, the last minute and a few seconds were cut off. I'm wondering if I should use Extend here, or try a new cover and hope that something very similar comes out, but without the end being cut off?

[analog kettle]

Try both. I use Advanced Options to say 100% Audio Influence. Sometimes you have to use Extend. Sometimes it's even better than the original.

[restive radish]

Every time I type a prompt and then press the magic wand it's doing this ... Why is this happening?

[honest heath]

that ususally happens if it thinks the prompt is blank. idk whats going on

[restive radish]

Well I had several things written in like I would normally do and haven't experienced that before. Perhaps it's just an app thing. I usually use the desktop site

[ornate hull]

Hi can anyone give me prompts to make doo wop's similar to this song? https://youtu.be/RBj2HN2uuNA?si=w7d-LEM85tiyWW5y

[shell kraken]

Does PRO plan allow LONGER prompts?

FYI: Free Plan caps it at 200 characters

[lost holly]

Yes, lyrics is limited to 5000 and style to 1000 with pro.

[solar badger]

For a free user v3.5, anyone know the key word in prompt style if i need solo vocal no harmony or fullband

[whole crescent]

I've run into an issue while creating a remix. The original song was generated in Suno 3.5, and the female vocal was velvety and deep.

Now I'm trying to create a remix of this song in Suno 4.5. Despite my instructions, the new version of Suno consistently changes the voice to a high, "soprano" sound in the choruses.

I've tried different tags (no shouting, low voice, Female very low contralto vocal, in the style of...), but the problem persists.

How can I create a remix without this vocal change? Thank you.

[keen quiver]

Here's how:

Open ChatGPT

Ask "Describe the song "The Great Pretender" by The Platters as Suno v3.5 style prompts. Max 200 characters"

[quick sonnet]

Unfortunatelly I can't use Advanced Options.

[shell kraken]

thanks mate

[echo prawn]

How to make AI pause at a specified place when covering a song?

[sweet siren]

Probably been answered but generally very high. I have between 75-95% generally. Although I am working on a song now for an album where it is lower. Hope it helps. Drop me the link when the album is done. I am curious.

[shadow plaza]

Guys I'm gonna need little help. I plan to make AI Artist. But don't know how to give him same voice for the songs he supposed to sing?

[wispy laurel]

Signpost it in the lyrics

[Break - all instruments stop playing]
...
[Dramatic pause, 4 beats]
[Silence]
...
[Music resumes]

This isn't 100% reliable because literally the whole point of Suno is that it can improvise and do its own thing, it may take quite a few attempts and you might need to experiment with the wording but I have known this to work sometimes

[lusty juniper]

It did it. I put it on weird and it started glitching and making like a radio station channel change sound. However i had to cut it cause the whole next 3 minutes was just the radio station changing channel sound

[graceful condor]

The Audio Cover function became absolutely terrible hardly ever does it sound like the audios melody anymore. Anyone else noticed?

[ionic turret]

what i do to not have a limit in the style is [style] and put well the instruments the names for me TR909 or 808, MOOG, Virus TR, Nordlead etc in there and also the track structure and i do it all in parts part 1 is intro - verse - pre drop - drop depending a bit what track it is, and then part 2 breakdown etc etc each does have its instruments and styles and in the style window i put the audio quality so that keeps going with the song

example : Part 1 – Intro / Build / Breakdown

Lyrics
[Intro]
(No lyrics)

[Build 1]
The night is calling,
I follow the sound.
Through endless horizons,
I’m lost but I’m found.

[Breakdown]
Echoes are rising,
Like waves in my mind.
I reach for tomorrow,
The future I’ll find.

[style]
(Progressive Trance, Melodic Trance. 128 BPM, G minor. Arrangement: deep rolling intro → dreamy melodic build → atmospheric breakdown. Analog-inspired sound: TR-909 kick, warm Moog Sub37 sub bass, evolving Nord Lead airy pads, analog-style percussion. Build with Virus TI plucks, JP-8080 filtered supersaws. Breakdown: Korg M1 piano, lush pads, subtle analog drift. Ambience: hypnotic, warm, emotional, wide and organic.)

etc etc

Styles (Audio Quality Only)

studio quality, fully mastered, warm analog feel, balanced highs/mids/lows, smooth highs no hiss, airy but controlled treble, centered sub bass, no muddy low end, subtle tape-like saturation, gentle harmonic richness, reduce noise oscillator levels, shorter reverb tails on highs, no white-noise wash, vocals upfront and natural, clear diction, polished and dry with subtle delay/reverb, sharp transient definition, transparent limiter −1 dBTP ceiling, no pumping, cinematic wide stereo field, hypnotic progressive trance analog-inspired club mix

thats in the styles window "_

you get something like this ":) https://soundcloud.com/sube-el-volume/we-are-infinite-tonight

[worthy marlin]

Why is this unreadable?

[ionic turret]

?

[worthy marlin]

The create button is unreadable for me

It appears white

I think it's Suno's CSS glitch

[ionic turret]

could be

refresh?

[worthy marlin]

I done it, but still happens when remixing/extending it

[ionic turret]

hmm

[viral pond]

Suno eating all my data again

When will they fix this problem

On wifi is even worse, 40gb

[ionic turret]

huh

how

by playing the track?

[tribal bobcat]

Hi everyone, I am facing massive lag as of late? Anyone else have this problem or know how to resolve. I have cleared Cashe but still no luck there... updates seem to be causing massive delays/lag

[magic nest]

why am i getting so many indian samples for my intros?? ehhh annoying.

[hollow hare]

What are your prompts? There is likely a keyword causing that

[magic nest]

epic drillbeats

with melodic cinematics🤦‍♂️

[hollow hare]

Ah, that is quite unusual indeed. Hmm, try to be more specific with the subgenre, like "UK drill beats".

In addition, you can try adding a location (e.g. England, New York, etc.) at the end. This may influence to get a regional style - or that used to be the case back in 3.0 days, haven't tried it lately

[wispy laurel]

It still does that
Any word that can be taken as the name of a place, culture or musical genre, will skew generation in that direction even if it's not intended to
It's like the Monkey's Paw, it feels like it actively looks for any possible way that it can misunderstand you 🤣

[glossy scaffold]

Hi everyone. I'm currently working with the editor and the transitions are not smooth at all, even after exporting. Any advice?

[wispy laurel]

The new editor has a Heal Edits function which I think is designed to help with this
But if that's not helpful, you also could try doing Replace Section in the legacy editor (NOT the new one) and replacing a very short section around the transition. I find this usually gives a fairly smooth join

[glossy scaffold]

Okay, I'll give that a try. Hadn't noticed this function.

[fluid copper]

Have a quick question: If I click on a song in my Library, it opens a page with the lyrics and other info. There is a Play button, some with a number. What is that number?

[glossy scaffold]

Amount of plays

[fluid copper]

Does that include by me or just others?

[glossy scaffold]

I imagine it depends on whether you allow your songs to be heard by others. I keep mine private.

[fluid copper]

Oh, okay. Thanks.

[wispy laurel]

Plays by anyone

[ocean timber]

Did they change the editor or did I screw up a setting? Last month I was able to go into the editor and songs were separated by Verse, Pre-Choprus, Chorus etc... which made editing so much easier. Now it's just one long audio file.

[polar spade]

A vocal question, regarding all versions.

Suno tends to sing the same melody in verses, change up if its a bridge, tag, hook, etc, and a "chorus" like melody for choruses.

Prompting for inprovisation, impromptu and freestyle helps, but it does seem like a built in go-to for Suno on vocals tp keep the melody the same throughout.

As a storyteller/songwriter its usually fine to repeat a melody on the second verse, but if there is a third following after, some changes or vamping works better than singing the same melody. Same for the second to last chorus if there are repeating at the end.

for everyone else, who writes in other genres, any thoughts of whether you like a bit of change up, within the framework of the melody line, but by a more vamping feel by the third verse?

I often go in Edit UI and just replace the verse with the same, to get some choices of a livelier melody line.

[wispy laurel]

Few of my songs have more than two verses or choruses
But if they do, yeah I like some variation
Usually I force it by adding or removing syllables or messing with the spacing, though (although I usually still end up having to replace bits when Suno mispronounces things or just randomly starts yelling in the middle of a gentle folk ballad 😂 )
A lot of my poetry wasn't originally written to be sung, though. Suno can usually find a melody but I don't necessarily keep the same rhyming and scansion from one verse to another. So I guess there's a kind of built-in anti-repetition there. When I write in a more standard pop/rock format with a consistent structure, it definitely does prefer to keep the same tune if I let it

[viral pond]

im not sure.. it just uses massive amounts of data

[small bridge]

Suno is ignoring me in the replace lyrics function doing whatever it wants and even going to a totally new song vibe direction while i am trying to rewrite my chorus. Anyone else having this tonight?

I even changed browsers to try and give it a chance. Bro is being glitchy tonight for me

[velvet granite]

So I have almost 100 versions of a song and I'm so frustrated with suno right now because I keep redoing it due to Suno errors. It seems every version I like Suno glitches which there doesn't seem to be any fix for according to the support email. Last night it kept returning songs that were 58 seconds or a minute and 45 seconds. I gave up. Why is this a thing?? It's never done this before.

[brazen kraken]

Using 4.5+ and I just can't get it to end a song, keeps giving me 4 minute extensions in a rock song.

Any tricks and hints?

[timber robin]

Are you using a persona?

[solar badger]

Quick question, if i need suno to play instruments from beginning do i need to put lyric structure just like your ? Does it work in every type of music? I’m using version3.5 by the way
[intro]
(No lyrics)

[verse 1]
Lyrics…

[wide flame]

I've been getting indian or almost Goa trance type patterns doing jungle,drum n bass too lol

[long spindle]

Does anyone know how to get it down to one instrument. Im trying to get fully acoustic country music and it keep throwing drums in the back round even know there in exclusions

[shadow plaza]

Guys I'm gonna need little help. I plan to make AI Artist. But don't know how to give him same voice for the songs he supposed to sing?

Is it persona?
Can the artist sing different styles music with same voice?

[feral ermine]

Yes and yes

[shadow plaza]

I mean I want only the voice not whole music vibe and melody

[feral ermine]

If you use persona, the voice will be mostly the same throughout any style description you enter

[shadow plaza]

OK thanks

[halcyon sundial]

any tip to make suno sing exactly how i am asking in this metatags?

it always invert or just ignore it

[prime prairie]

Having issues with the Android app right now. It won't play a d@mn thing!

[autumn vault]

Really? Might be minor update? Cuz somehow some song now have fx...

[graceful condor]

Suno is a mess lately and I noticed that some stuff is highly interfering with each other. SO PLEASE EXPLAIN ( @wispy laurel , @polar spade, @young geode or Someone)

If I have as follow:

1. Persona
2. AUDIO Track
3. Detailed complex Lyric Prompt (includes description of vocals too)
4. Detailed complex Style Prompt (includes description of vocals too)
5. Exclude
6. Sliders used mostly around this Weird: 20 / Style: 85 / Audio between 90-100 (especially if I use a Persona)

So if I use all of that at once. What does Suno prioritize the most and what interferes with each other and what kinda cancels the effect of the other one?
This is what I really would like to know.
Thank you

[halcyon sundial]

in my case suno prioritize nothing lmao

he just dont do what i ask

[graceful condor]

xDD yeah lately that is sadly the case.

[halcyon sundial]

as im trying to do here

i changed this prompt at least 30 times

and 300+ tries

no one obeyed it

male singing at female verse

unasked metal scream everytime in every end of chorus

[graceful condor]

I also use specific Personas and maybe 1 out of 10 tries it might sound a bit like the Persona I saved but then loses its tone already within the same song. It is so annoying. I do not know what they have done lately because they are not communicating with us but they really f-d up this AI.

[halcyon sundial]

i know my cause its a very specific song and prompt so its harder

but i had a lot of trouble with more simple ones

[graceful condor]

Have you also encountered that after extend even if you only need a small instrumental outro with fade out it still drags on the song for minutes and even sings again and again. Regardless how many times you direct him not to sing, remove all that has to do with singing.

[halcyon sundial]

yeah, a lot of times

in my case, i do songs in portuguese, so this extensions sing some cursed unexistent words

it looks like a demon summoning chant

its scary

[graceful condor]

Yeah I know. I do a lot Japanese stuff and it sounds like total gibberish.

[halcyon sundial]

are you japanese?

[graceful condor]

No 😅 Austrian

[halcyon sundial]

i give up for today

i said to my subscribers that the song was experimental and im not sure i will post it

so im prepared to say that i will not do it

suno isnt smart enough to do this now

maybe in 5.0

[ionic turret]

sorry for the late response nah you dont have to you can also put instrumental if you like but i just did it as an extra

[polar spade]

Typically metatags are in [square brackets]. That may get a better response.

As for inverted, nothing says you cant run one male /female and another female/male and take the one that works. 🙂 sometimes the easiest way without knowing why is better.

[polar spade]

Keep in mind that Suno, though generative, or makes attempts to be generative, needs to feel some "vibe" in the direction you're taking it. If there are not many hardcore fast-paced duets in the real world, which you may know as I wouldnt, suno may have to just veer a bit towards what occurs more often within all the prompts given, so you end up no better than just putting in a couple of prompts.

Have you come across any samples songs in here that have emulated a similar sound? And looked at their prompts?

A back and forth accompaniment like this could be thought as occurring more in Opera, or a Musical. Even adding those prompts may get more of a tug in the direction you want.

If not overdone, you can also separate the lines , and further , indicate them as [spoken] per each line or at maximum just at the beginning line of the verse. . This time, you hope Suno DOESNT speak them but does give a more demanding voice while hopefully still singing them. So do NOT put any lines in quotes.

[peak veldt]

Hey guys, my friends suggested my this app on the website and I loved what it had shown that I immediately bought the pro plan. However, I am missing features that my friend has on his side. I tried different web browsers and cleared my browsing data but I am still missing features. This still applies in the mobile app too.

Just an idea in this image that my friend screenshotted on his side, he has a switch between auto and write lyrics. He has '+ By line' & '+ Full song'. I have put it on 4.5+ and tried switching between but I am not having any luck. Can I get some help please.

[zinc cloud]

Hi, a question about pitch, I find that even when I specify "low pitch", most of Suno songs still tend to go very high-pitch for chorus. How do you guys use certain prompt words to make sure the whole song stay in the lower pitch range?

[half dragon]

Even with older model version selected - that still shows, are you in Custom mode like the screenshot?

[peak veldt]

yes I am my friend and I screenshared we were dumbfounded.

[half dragon]

Can you share your screenshot?

[polar spade]

Do you have "Desktop mode" or "desktop site" clicked off?

[velvet granite]

Sometimes. Both yes and no. The persona was created from a full song. How does this affect it?

[peak veldt]

im on my desktop/pc

[polar spade]

Is it not showing when you click Cover on an output, rather than just a NEW create?

[peak veldt]

sorry can you please show me what you mean

[peak veldt]

this is the whole view

[polar spade]

I was curious if you go to any output, click the 3-dot menu, click Remix/edit and then Cover... what does the Create window look like then? Is it also missing the by line / full lyrics tab

Certainly, also, try just logging out, clear cache, close browser. And open browser and logg back in. Could also be a glitch. Vwrsion 4.5+ , top right, should also be indicated by magenta colour, not just white lettering.

[peak veldt]

Yeah done all that too 🙁

[polar spade]

All that = 1) Cover panel looks same. 2) And logging out / login not help?

[polar spade]

When you click the version drop down.... is 4.5+ in magenta colour and the rest 3.0, 4.0, and 3.5 in white? If not, then that seems glitchy. I cant repeat BOTH the by line/ full lyrics tab AND 4.5+ not magenta colour.

[peak veldt]

yeah logging out/in, cleared broswing data, incognito, different browser you name it sadly not the same

[polar spade]

OK. 🤔

[peak veldt]

yeah im letting one of the tech guys know of this but hes telling me and I quote "When you click the Creative Prompt Booster button (looks like a magic wand) at the top right of your lyrical field you should see a drop down with additional options. The link I shared with you in the prior chat explains a bit about what's currently happening with the shift toward a new Create experience."

[polar spade]

Yea.... ill take a photo of that..... hold on.

[polar spade]

Sorry.... suno is really slow today.... hold on.

[peak veldt]

all good i really appreciate your time just to help me, thank you 🙂

[polar spade]

Well... hmmm... I cant log in at the moment. 😢

The next two weeks, they are phasing out versions for Basic level. In turn, they may also be slow-rolling out a new format/layout on the Create panel . The fact that you have a white 4.5+ indicates you may have a more current updated panel, or its a glitch still 😳

As always, too, check your app store for updates.

Paste some lyrics in the box and then click the magic wand and choose one of the options which now drop down. See if that now gives you options to edit the lyrics...... or click Cover on an output that already has lyrics and test the same.

I can NOT see my account at the moment. These next two weeks may be glitchy for everyone as they make these deprecating / updating changes.

[polar spade]

In all candor, I dont use by line/full lyrics or the prompt booster button with a drop-down menu enough to notice either one missing. As soon as I can log in again, though, ill check further.

[polar spade]

Many are having login / logout issues. So im putting all on the back burner and going to touch grass for awhile. So i cant do much now. Be back later.

[ionic turret]

small thing 🙂 try to add this to your track style, either in the part where you have the lyrics but use [style] and put the prompt between ( ) as the space is limited in the standard style field below the lyrics field 🙂 you get real clear audio tried it few tracks

studio quality, fully mastered
well-balanced highs, mids, and lows
tight sub-bass, centered and controlled
punchy low-end with clear kick separation
warm midrange presence, no boxiness
crystal-clear leads, minimal reverb, no muddy mix
vocals upfront and natural, crisp presence with smooth air
light de-ess on sibilance, gentle saturation only
short plate reverb with subtle pre-delay, tight slapback delay for space
double/triple vocal stacks low in mix, clear diction, no over-tuning
sidechain preserves vocal transients; sub energy never masks vocals
carved 200–400 Hz for clarity, added 8–12 kHz sheen for sparkle
dry and polished vocal sound, controlled reverb, sharp transient definition, no distortion
festival-ready, cinematic, wide stereo field, energetic club mix
transparent mastering with -1 dBTP ceiling, no pumping on mix bus
spotless noise floor, no hiss, no artifacts

perhaps it helps 0

07

or you can do it the other way around also 🙂 for the style parts

[polar spade]

Could you tag the users name you're replying to?

[ionic turret]

all

🙂

its a prompt setting which everyone can use :_)

[polar spade]

Maybe if you have a "without" and a "with" examples, from your tests, that would be helpful to see the significance. Looks like a lot of work coming up with this.

[ionic turret]

sure

lemme put somethingup

[polar spade]

One without. One with. Im always interested if I can find something thats effective. I am limited by lyric space though. So a test on just a couple of verses and one or two style prompts is best. Ill do that as well. Once I can log in... 😳

[ionic turret]

one with https://suno.com/song/8558f389-898d-4b6b-86f3-3696c52ac6ba

now one without 1 sec

https://suno.com/song/94b5bafe-e28d-449b-a8ff-aa64d3e6afdb

you hear how the second one is over doing it

same prompt

well almost heh

now ill add the same settin for the audio to this prompt and redo it wiht it

https://suno.com/song/a74e4eb2-809f-4472-9758-564bdb521188

clear 🙂

balanced high mid low

🙂

[half dragon]

Might also want to consider posting it in #1252267444773326848 - for easy to find for others - it might be helpful for others.

[ionic turret]

uhmmm

oki

[hexed iris]

So finally, I’ve got an almost perfect song after so many attempts, complete with perfect crowd vocals and cheers. Big thanks to @polar spade!
Just one small glitch: a single word is pronounced incorrectly. Should I go for an edit or a cover to best replicate the song with the correction? I don’t want to use third-party tools yet, since I’m still learning those techniques.

[ionic turret]

il set something up 🙂 but try and see if it works for you also

[half dragon]

If it's just a single word - go for Edit
If too much trouble with editors - Cover would be next thing to try
Each tool has their quirks - be patient and work through it.

[polar spade]

Yes! Very obvious clarity and brightness. Ill play around with it in different genre settings. The limited space and ability to Cover later might be an issue things to deal with, but it looks promising, especially in mastering a cover for older 4.0 songs . Thans for the time you put in and for posting up. 🩵 🙏 🤘

[polar spade]

Meh. 🤷‍♂️. Covers are working pretty good these days.... but do both. I usually go to Legacy Editor and try first to literally replace one word..... then if not good... the whole verse with the one-word change.... and lastly try new Edit UI for the same routine.

Keep in mind, these next two weeks there may be some glitchiness from a few change overs as mentioned in update announcements. So, put things on back burner if you need to.

[polar spade]

Yes.... what he said. ☝️

[hexed iris]

I've had bad exp with editors in the past. I never tried after that. Cover is fine but this time I don't want even a minute change/variation in the song

[ionic turret]

https://discord.com/channels/1069381916492562582/1417162963709857913

[ionic turret]

https://discord.com/channels/1069381916492562582/1417162963709857913

[polar spade]

Thats what I demand in my relationships. I dont want even a minute change, darn it! . Thats why im single and just write songs now. 🤣

Yea, even with edit, it has to do its best to re-tokenize and blend whats to the left and write of that word. Typically if you can go from a longer word to a shorter one, then its easier to keep the melody, and nothing change, but thats rare.

[keen quiver]

Hey, how do I correctly extend an instrumental song? I am using v3.5 and extend function only gave me 2-3 seconds extension.

[hexed iris]

This is great. Equally good for lyrical songs?

[ionic turret]

yea should be

it balances the audio

id say try hehe also tried on the other AI gen and also works there 🙂

you know who 🙂

and well id say people should try to tweak the settings in the prompt see what happens

instead of sliders and knobs we use text lol perhaps SUNO dev can make something with this like a few settings and a window for audio quality styles that would be kewl

[ionic turret]

https://suno.com/song/2a8d5d94-7ab0-4bea-8080-4935f8e490e1

little extra test 🙂

nice dynamics

what i also did ... is not selecting any male or female vocals this way they both showed up lol

[hexed iris]

Nice. This style in lyrics and SQ in style is a new thing for me. Will Def try.

[tight hatch]

Is there a way to limit a song length to 10 Seconds? I only need a jingle for new followers on twitch

[half dragon]

Something like this?
https://suno.com/s/HrA0uE1YQkJVZTJc

[tight hatch]

Yes

How do limit the length?

[half dragon]

You can try to put that in the style prompt like this one from 3Daizy
https://suno.com/s/0dq3qnTfHAmGGZmv

[tight hatch]

Didnt work - songs are much longer

[half dragon]

Oh ok - while specifically controll the length might not be reliable maybe "news/short jingle" or "radio jingle" would help to keep the length down

[tight hatch]

Ok, thanks - i give it a try

That worked! Thanks!

[fresh oxide]

The Suno UI was just updated and nothing is creating

The generator started working with the new UI, but it's not giving the option for 'get full song' once extended.

[halcyon sundial]

do you have any ideia how i can write better this prompts then

[polar spade]

@halcyon sundial I had mentioned a couple of things already in the last post.

But I'm one now to get close to something thats OK, then tweak 100 times until something works better to what i like, rather than spend so much time to think each line should be tagged and prompted to get it in one try.

But start with what you have.

Just put things in [square brackets ] NOT (parentheses).

For fun, add "opera" as an end stytle prompts, see if that gives any nice change.

[halcyon sundial]

my biggest question is if [Chorus] then [Prompts] for each verse affects smething

a prompt inside a prompt

[polar spade]

Mori is typing something exceptional. Let's see if they are expounding on this....

[wispy laurel]

Something you could try perhaps, if you don't mind burning some credits on an experiment that might not work out
Like Dasvon said, identify a genre where this kind of duet would be common, I'm guessing musical theatre might be a good place to start. Then throw out all the metalcore and other styles, generate the song with the correct lyrics but as a Broadway-style duet (or whatever style will allow you to get the right vocal arrangement)
Once you have the duet properly executed in any genre, split the stems. Cover the vocal stem, with audio influence set very high, and this time use whatever style prompt you were originally going to for the 'proper' version that you're trying to make
Hopefully this will preserve the duet vocal, but build the song you were trying to create around it, with the style and genre that you wanted - but because the vocal is already established and is being covered with high influence, in theory that part shouldn't be changed too much

[polar spade]

Prompts beside prompts is usually OK. The suggestion is no more than three bracketed prompts of any emotion or action.

[Chorus] is more a placemaker and doesnt have much affect.

Oh.. wait.... let's read.....Mori post.

[halcyon sundial]

lets try it

[low laurel]

Hi there, How would you make a male and a female dialog: what tag, what style?
Using this [Female voice speaking: Hello?], but remains instrumental or it sings and does not respect male and female. Thanks

[polar spade]

One sample can be:

[chorus]
[Male vocal] [male] (excited)
This is the way

[spoken] [female] (whisper)
Tell me the way. I want to learn.

[Male] You will learn by your actions

[Female vocal] oh
How shall I know if I succeed.

Style prompts: Hardcore, metal, speedmetal, opera, musical theatre

---

So. I dont know what youre going for.... a little growling too?

Anyway, here's one just from using the above lyric tags and style prompt. Anything to get us another 20% further, then keep tweaking.

https://suno.com/s/Xoy1YnaIP2cEOHQT

So, it seems though, the line spacing helps switch gender vocals, but still, once Suno starts into a pattern, they may get switched, so again, we often have to move something up or down or even we can add a line earlier to force Suno to start with a gender to get the pattern correctly.

[polar spade]

Ytou can see I used the tag [spoken] but it didnt stop the female from singing her line. Yet I think it can benefit for giving some emotion and directness to the vocalization.

[polar spade]

I took out "musical theatre" which gives more weight to hardcore and metal, and left Opera.... and it switched the pattern, so the BUILD of oh...oh...oh.. would likely need to be removed... but how about that operatic vocal of the female and growl of the male..... pretty cool huh?

https://suno.com/s/BiY6vq6YQwEQ3grr

https://suno.com/s/hsnAMDtbmHLOMXHa

[low laurel]

not sure iif it was a reply for me

[lost holly]

Suno can do funny things, modified it a little 😄 https://suno.com/s/iMnmpRAlb9fRBarl

[graceful condor]

Suno is a mess lately and I noticed that some stuff is highly interfering with each other. SO PLEASE EXPLAIN ( @wispy laurel , @polar spade, @young geode or Someone)

If I have as follow:

Persona
AUDIO Track
Detailed complex Lyric Prompt (includes description of vocals too)
Detailed complex Style Prompt (includes description of vocals too)
Exclude
Sliders used mostly around this Weird: 20 / Style: 85 / Audio between 90-100 (especially if I use a Persona)

So if I use all of that at once. What does Suno prioritize the most and what interferes with each other and what kinda cancels the effect of the other one?
This is what I really would like to know.
Thank you

[lusty juniper]

what in the world? the interface suddenly change mid creating

it's throwing me off.

this is horrible ahahaha

[half dragon]

You mean, it changed to Simple mode out of no where?

[lusty juniper]

both look weird

i am on custom and it still looks different

[half dragon]

yeah I just got updated too - looks like they removed "by line" and "whole song" on Lyric panel

[lusty juniper]

nah i dont need that part. just the whole thing is different

[half dragon]

all panels are still there - looks a bit different, maybe different fonts or what not

icons rearranged in the main listing panel

Looks cleaner imo

[lusty juniper]

it sucks

its like going from windows to apple

its so "clean" its too in my face. right side is empty which helped me navigate. i can see the songs i liked and planned to work with. now everything is in the middle the whole right side is empty

[half dragon]

Apparently this is the new interface - I just got updated as well.
So your friend will probably see the same thing as you do really soon, if not already.

[half dragon]

If you select a song, it'll open up a right panel for the song, while I wish I can resize that panel to be bigger, it'd help to elimiate the emptiness on the right side

[lusty juniper]

i pressed on a song and the button of GET FULL SONG appeared out of nowhere as i clicked. LOL

and typing in styles is delayed like crazy.

[half dragon]

Because you clicked on the title of the song instead of the "Row"

[lusty juniper]

well... this aint good at all.. not for me

[half dragon]

I do not experience any delay there.

[lusty juniper]

it opens and closes as well for me. when i click styles. dunno anyway -100 points