📚┃suno-school

[near flume]

Spend $10 and try it out. At least you'd have ownership of those songs.

[hard beacon]

i think we need 2 more sliders... "screaming level" ... and "long singing" 😄

and yes.. i know.. a well known "feature" 😉 just wanna say hi ^^

[hard beacon]

I can't follow the facts at all, but thank you 😄

[polar spade]

I deleted the video. General room said I was bricking, so they can find a mod. Im hungry.

[hard beacon]

btw... I know you're familiar with DAWs... Are Suno stems easier to edit than stems generated by DAWs?

I would say “yes” at first, since Suno stems are generated from the output... but before I use up 50 credits, I wanted to ask someone who knows about this.

[wispy laurel]

Suno stems are weird sometimes
Sometimes it's absolutely perfect, WAY cleaner split than I get from DAW. But other times it gets confused and splits one instrument across multiple stems or combines unrelated and non-similar-sounding things into one
It seems to particularly struggle with some instruments. I find it'll consistently file a violin under FX instead of Strings, it identifies a harp as Guitar but then if there's an electric guitar in the same song that gets identified as Synth or FX instead
With a simple, standard rock band setup it works well, though. I'd say try it and see, because it's hard to predict what it will or won't get right

[fierce token]

OK hoping for help. When I do a COVER of a song I wrote and created in SUNO it often cuts off before the end of the song. SO I lode lyrics and music. Is there any resolution for this? My song currently in process is around 6 min and keeps cutting back to 4:58 everytime I use the COVER option

[wheat iron]

How do I go about releasing a track where I keep the exact same beat but change the lyrics? Like, if I already have a version of the song out and I want to drop another one with different or edited lyrics over that same instrumental, what’s the right way to do that? Do I need to register it differently, call it a remix, or just release it as an alternate version?

[wispy laurel]

Covers, or any song using a persona, almost never run longer than 4:50-5:00
The data from the original song and/or persona, takes up some of the 'space' that the AI has to work in, so it can't generate as much new output as it otherwise would
Best fix that I know of is, use the editor to crop a section from the end of your original song, including whatever parts don't fit into the 5-minute length. Then use that section to make a persona.
Then, cover the full version of the original song. It won't all fit BUT, use Extend starting from around the point where your persona-fied clip begins. Set Weirdness 1%, Style 0%, Audio 99%, and use the persona that you made from the ending of the original song
In theory, if the Suno Gods are willing, this should give you a very close recreation of the original, but tweaked to better match the style of the cover version
If you don't want it to match that closely, turn Weirdness up a bit and Audio down, and use tags in the Lyrics (of the extension) to tell it what it should change

[polar spade]

Was it your original track (song) to begin with? You didnt quite say.

[wheat iron]

https://suno.com/s/iGWpLlozuLUe2RDT

[tranquil leaf]

Even if you base a persona on an 8 minute track, songs you base on it will still come out at <5 minutes.

[polar spade]

Bouncing off your great idea, Would it be more exact to make an additional persona also of the original 6 min output. Then Cover with adding the full song persona. This would keep the new Cover more in line with the song tahn a fresh cover without persona? Assuming it will cut off somewhere. ThEN use short persona and extend. Maybe overkill. 🤷

[polar spade]

I assume that means yes.

So you can separate out the instrumental and the "add vocals to instrumental" . Thats one option. Your melody would change to completely different, but the music stays the same.

[wheat iron]

I'm on phone

[polar spade]

Im at lunch. . . . And driving. 😁

[wheat iron]

Oh..

Sorry

[wispy laurel]

I find that if I drop Weirdness and style all the way down, audio all the way up, and a blank style prompt, I get a very close replica of the original anyway without needing a persona for that bit
Which then allows me to use a persona to identify which specific bits I do want to change
Been doing this a lot lately because I'm remaking a lot of my songs as close as possible to the original but with different (better) vocals
So if I set minimal weirdness and style, no style instruction, and maximum audio influence, I get an almost exact recreation with very small changes
And by using a persona with a different voice, it forces Suno to make that particular change to fill what I imagine as its quota of variations
I haven't tried but I think that for example if I wanted to keep the same voice but change a guitar solo from hair-metal to blues, I probably could get the same result using a persona that's nothing but a bluesy guitar solo
Experiments are ongoing

[polar spade]

Good to know. Thanks.

[quick sonnet]

Not everybody here is from USA to use $ and in some places 10$ is huge money.

I am wondering if it's posible to explicitly use the theme from an uploaded audio in the chorus and let the voice sing it in some defined key, for example minor key?

[wispy laurel]

You could try [tagging] the desired key signature in the lyrics (not the style instruction)
So it'd look something like

[Chorus]
[Vocal in key of A minor]
Insert lyrics here
And here
And probably also here

Suno generally isn't too great at taking precise direction on very specific details. It should at least understand the instruction if you do it that way, but whether it follows it is anybody's guess

[quiet solar]

If you've completely changed the lyrics then the traditional musical term for it is a contrafact, and the short answer is that provided you own both the music and lyrics you can do whatever you like when it comes to releasing it. Generally speaking, most writers simply release the new song under an entirely different name, with no explicit reference to the original work.

If you've just tweaked the lyrics a little and it still largely resembles the original work, then to avoid confusion amongst listeners it's more common to keep the name of the song but identify it as an alternate version or remix in some way.

But there are really no hard and fast rules on this.

[trail canyon]

This sound some very valuable tip. Will try it. 🙏

[polar aurora]

So, I have been using Suno for about 9 months, just casually, for turning my poems into songs for my YouTube channel. Is there a way to tell Suno to only create a 1-minute or 2-minute song? Some of the songs are amazing with 4.5, but I don't always want a 3 or 4-minute song.

[sharp moth]

I’m having an issue where no matter what I do—refresh, reset, reinstall, redo changes, switch between app and web—it won’t fix the style setting. I’m stuck with this trash pop style and can’t switch back to my usual music style. Even when I clear everything and start fresh, it still defaults to the same one. Nothing works.

[lament shale]

is there a place where you can find all the options you can use for [ ] in lyrics section?

[cinder grotto]

Any suggestions for how to get a song with just a single guitar and vocal? Everything I’ve tried incorporates other instruments and layers. I’m trying to approximate the sound of a single singer with a single acoustic guitar. Thanks!

[half dragon]

Yeah, google will help you find most of places that maintained by users as Suno doesn't have one.
For example, https://sunoaiwiki.com/ and some others...

[half dragon]

Like this?
https://suno.com/s/25OLXmPZiqSOmXME

"A solo acoustic guitar in Bb minor leads with delicate fingerpicking, allowing open, expressive voicings to resonate, The male vocal sits forward, intimately woven with the guitar’s subtle dynamics, The minimal arrangement highlights nuanced pauses and gentle harmonic movement throughout"

[quiet solar]

I don't believe so, since there's an infinite variety of things you could include.

Suno recognises anything in square brackets as a style or song structure prompt, rather than lyrics to be sung. So yes, there are classics like [Verse] and [Chorus] which I'm sure you've seen, but there's nothing stopping you putting in [Spoken like a cheesy 1940s used car salesman from Michigan], for example.

However, bear in mind they're guidance prompts rather than commands. Just because you ask it to do something doesn't mean it definitely will, or that it will execute in the way that you'd hoped.

[graceful condor]

Why is it lately singing parts of the prompts? This is is so annoying. All parts of the prompts are as always in [....] So why is the AI singing parts of the prompts? Anyone else experiencing this lately.

[quick sonnet]

Of course, there's no 100% guarantee that it will work the first time, but it increases the likelihood.
And what about music instructions for the chorus so that the audio I uploaded is used?

[polar spade]

Didn't it cover your audio in the same key? And cover as uploaded, so the arrangement, including chorus, was the same? On average, my uploads are more often covered in the same key.

[quick sonnet]

Well, actually I would like to generate a rap, hip-hop track where the chorus is sung to the uploaded melody, the rest is then rap and the AI ​​is free to sample the melody.

[polar spade]

So your audio is only the rap part(s) and you want to edit in a chorus?

[polar spade]

If i were to try what you describe, I would simply rap the rap parts and , if I cant even sing well enough, still also sing the chorus part , all at one time, and in the key you want, and get close enough for Suno to pitch correct anyway.

So you can literally get it done just by covering your upload.

Molding a clay pot from soft clay can be easier than making a vase from broken pieces.

What has been your starting point then?

[near flume]

For Covers this is generally; true, you will normally get the same length or a little less than the original. For Personas it's a crap shoot sometimes. Personas I created in v4 are less than 4 minutes in length but I get them to produce songs of up to 7:59 without using extensions (typically 5:30 - 6:15).

[quick sonnet]

No, my audio is intended for the chorus part, let's say a classical piece by Beethoven, I know it might sound weird, but because of copyright it should work and there are well-known examples where something like that has been used.

[torpid haven]

Is there already something like music to text which can describe music ? So music to prompt

And can I give specific tokens a different weight like in stable diffusion ?

[quick sonnet]

Do you mean something like, you upload your music and then get the corresponding description?

[torpid haven]

Yes

Music to prompt as use case

It feels like a hard problem to solve but I would have said the same in the image and video area and we have it already

[quick sonnet]

Should work, at least it worked for me. Have uploaded my audio and suno generated a realy good description.

[torpid haven]

Yes I tried it at Suno and the description is very generic for me. Not good enough to create something that comes close to the input. So I was asking for alternative providers for this functionality at least. Maybe some describe the sound of specific synths erc

[quick sonnet]

Are you a musician yourself?

[torpid haven]

I know musicians and I'm definitely not on their level so I would not put that title on me

[quick sonnet]

But your level here is probably still higher than mine, I'm an armateur when it comes to all the theory concepts.

[graceful condor]

Is there a way to keep the voiced consistent, because sometimes it changes slightly from verse to verse.

[left sapphire]

what is suno's obsession with giving me drums/percussion? I don't mention drums/beat/percussion anything of that nature. I even literally say NO DRUMS, NO PERCUSSION ... 2 minutes of it sounding fine, then throws in a groovy beat ... 20 songs later, no matter what I say, Suno decides it needs some drum action halfway through

[wispy laurel]

Suno is really bad at leaving things out, that would usually be present in that genre of music (or, in whatever genre it thinks it's making)
It's also basically impossible to generate any form of rock music without a bass guitar
You can try specifically identifying which instruments are to be used, right at the start of the style prompt

[Acoustic guitar, piano, violin, flute]

And it will sometimes stick to the list but not always, especially if you're leaving out something that's usually 'standard issue' in that genre.
But you might find that you have to just split the stems, download and use offline software to rebuild the track without the unwanted drum and percussion stems

[left sapphire]

yeah, thats pretty annoying to have to spend 50 credits just because suno keeps wanting to add drums to ambient atmospheric music

[wispy laurel]

You could try using 'arhythmic' in the style prompt, if you didn't already.
Telling it "no drums" doesn't work because the style instruction just doesn't understand negative statements. It'll ignore the word 'no' and think that you're asking for drums. So you need to either just leave that out, or find a positive instruction that has the same practical meaning.
The Exclude Style box, annoyingly, isn't very reliable either at present. You can try putting drums in there (do not say NO drums there because that'll just confuse it), but it doesn't always put much weight on exclusion prompts
This is one of my pet hates in Suno too but unfortunately, so far splitting the stems and rebuilding seems to be the only really reliable way to keep specific instruments out

[left sapphire]

looks like a bunch of credits are going to get used. I thought I finally cracked it bust sure enough 2:30 comes along and the beat kicks in. My only other method was keep extending every 2 min of non-beat music ... but I start making 7-10 min songs.

stems it is

[knotty stag]

Can Suno do translations of songs? I mean, I created a song in English, all finished and good. Now I want to do a version in Spanish, same music, same meaning, but with Spanish lyrics. Can Suno do that? I have the Spanish lyrics I want to use already, is there an easy way to do a new version with the new lyrics without changing the music?
At the moment, I can do it line by line, but the timing does not always fit perfectly.

[lost holly]

With exactly the same music is hard. What you can try is working with the cover option here, means create a cover of the song you like the music off and change the lyrics to Spanish.
Personal i would recommend to make sure that the syllables match with the English lyrics or you might have a hard time getting the result you want.

[dreamy sparrow]

??

[copper cedar]

Can't believe a music program is flagging the word skank when used to describe guitar skank in ska music. SIlly children

[polar spade]

Just prompt "ska stroke" or even just "ska guitar". Life will be happier in Suno-land.

[olive moon]

Does anyone know how to remove a random “yeah” that shows up at the start of a song even though it’s not in my lyrics? I was able to mute a similar section later in the track because it matched an actual lyric, but these “yeah” and “smoke” ad-libs were never part of my lyrics. What’s the best way to replace or remove them?

[devout sinew]

Question adding chord progression does suno ai really going to execute chord progression when added to the prompt?

[chilly wyvern]

In general case, no. I find some genres have better probability of doing so. Jazzy things and the like but it's not like your Dmaj9 is actually obeyed.

[devout sinew]

Thank you.

[thick estuary]

If you can use a DAW, download the stems and edit them in there

[wispy laurel]

You'd need an actual recording of the chord sequence. Could be played on any instrument, although it'll work better if it's the same one that you want Suno to play it on. Upload that recording, and cover it with added lyrics and clear instructions for what else you want along with those chords.
Or, the more involved way - use the upload as Inspiration, if you can find a clear enough prompt you should be able to get Suno to produce something that includes that specific progression in a nice clear form. Split the stems (if it used other instruments too) and crop out just a single instance of that chord sequence. Use the cropped section to create a persona, then use that persona to create the song that you're actually trying to make.
This sometimes takes quite a few tries but it does work, it's heavy on credits though so if you're a free user I wouldn't really recommend it

[hollow trellis]

hey can someone point me in the right direction im trying to setup bark ai in n8n for my local ai

[scenic coyote]

I have gotten it to do chord progressions successfully multiple times. it's hit or miss. If you remix this song and change the chord progression to a I-IV-V-I or something You should hear the difference. This is the format that I got it to work with most successfully.
[Chord Progression: Imaj7(#11)–♭III7(b9)–VI13–♭VImaj7(#5)–II7alt–V13sus4–Imaj7(13) (Cmaj7(#11)–Eb7(b9)–A13–Abmaj7(#5)–D7alt–G13sus4–Cmaj7(13))]

https://suno.com/s/YqtwzMvSQuK5Nuu8

I didn't validate that this song actually is those chords, but it's a starter. I remixed that same song to a 1-5-4 progression and it took.

[scenic coyote]

This works as well and is more consistent. you can go to beepbox and create the cord progression and export it for free. then upload and extend. although this is still Hit or Miss. it might not follow the chord progression and go off on its tangent. I've actually had more success with the prompt I posted above.

[Chord Progression: I–III7–VI7–ii7–V7–I (C–E7–A7–Dm7–G7–C)]
I'm pretty sure this worked
https://suno.com/s/BMBFRpMWevra70MC

[wispy laurel]

Interesting. I've never been able to get it to follow specific chords from a text prompt only. Although I'll admit, I gave up pretty quickly when I realised that I could use the persona/inspiration method instead

[scenic coyote]

If you give it a shot let me know the results!

I just copied the prompt in the On Cue that song I just posted and changed the chord progression (A new song, not a remix) And it followed it. maybe it's more reliable than I've thought

[wispy laurel]

Hm. Might have fixed it in recent updates then, I couldn't get it to work a few months ago.
Although it also might have been because I was telling it to put a traditionally jazz sequence into heavy metal

[quick sonnet]

Is there a way when I cover a track to force using the melody from that track?

[wispy laurel]

Set low Weirdness and high Audio influence
You can't force it to do anything but you can make it more likely to keep the major parts of the original

[quick sonnet]

That will not work with v3.5. In previous results it has used that melody, but now I geting results which has nothing common with the original.

[wispy laurel]

In 3.5 you can try [match the melody of the original song] in the style prompt, I had some success with that but no guarantees
Otherwise all you can really do is either keep trying, or give up and do something different. I can't think of any other way to achieve it in 3.5

[quick sonnet]

Then I guess I got really lucky with my first attempts. The melody was always included, but the vocals in the chorus didn't work the way I wanted. Probably for promotional purposes, I got two very shortened versions, exactly 1 minute, with v4.5, but they were just as far from what I wanted.

[cinder grotto]

Yes -- thank you! I'll adapt this. Much appreciated!

[quick sonnet]

Since [] used, is it a prompt?

[wispy laurel]

Yes

[quick sonnet]

But it's styles prompt, right?

[wispy laurel]

Yep. Disclaimer, I am not a programmer or AI specialist so I don't know the specifics of how Suno interprets a prompt
But, putting something in [] within the style prompt seems to improve the chance of the AI actually following that bit (as long as it understands the instruction]
Things at the beginning of the prompt also generally seem to have a higher priority than stuff that comes later

[quick sonnet]

While I am not programming AI, as a programmer I know that certain things are marked like that. In the lyrics, these markers are supposed to ensure that these instructions aren't interpreted as part of the lyrics, even though some people omit them there without consequences. In version 3.5, the styles prompt is limited to 199 characters, so every additional character you use can be problematic.

[wispy laurel]

Yeah. I started out with 3.5 only, I remember having to pick and choose my words really carefully to fit everything in. Suno does recognise [] in the style prompt as marking a specific and important instruction, but I realise you can only fit a few such instructions in

[viral pond]

trying to extend an old v4 song, but it just keeps failing eveytime, hallucination and makes horrible weird result.. is this a known issue ?

I wanted to give it a new proper ending, but i ended up fading it out instead

[wispy laurel]

Never extend from the very end of the song. Go from at least a couple of bars before the end, or earlier if the structure of the song allows it
Empty the Lyrics box (unless you had to extend from before the last vocal section, in which case include those lyrics only).
Then add in [instrumental outro], followed by [tags] explaining what you want in there. So for example it could be something like

[Instrumental outro]
[Drums, bass, piano, guitar]
[Long melodic guitar solo - heavy rock]

[Ending]

Obviously change the list of instruments and the guitar solo to whatever works for the song that you're trying to create, but make sure you put something there so Suno has a clear instruction to add music rather than just ending the song right away.
Do not switch it to Instrumental mode, even if there aren't meant to be any vocals - at least from my own experience, that tends to make it less likely to follow any instructions
All of that goes in the Lyrics field.
If the extension you want is in the same style and musically similar to the rest of the song, you can empty the Style field. If you want the extension to go in a different direction, explain that in the Style instruction.

[viral pond]

yeah i tried to go back 10 seconds, but it seems there was something weird going on in the background, it is okay with the fade out, but it was weird

should the style be removed, will it just use the current one ? never really tested this..

also from the newer editor, i trying to insert some empty space to remove some hallucinated lyrics, but it seem slike this wont work,, i might do something wrong when I remove a part, then insert a new one in place adding new fresh lyrics

[wispy laurel]

Extensions or covers with a blank style instruction will just default to trying to match the original style, influenced by the persona if you're using one
I can't help with the new editor, I hate it and it hates me 😂

[viral pond]

hqhq

haha

yeah it does hate me sometimes as well

but i figured if you commit, you can actually drag and try to match and blend them together manually, thats awesome

and i hate davinci resolve that have to do some weird hickups everytime i start it... must be a bug in that version'

[wispy laurel]

I haven't really used it much, to be fair. I use the browser version but via an Android tablet, in Desktop mode everything technically works but the new editor lags so much that it's unbearable to use. Legacy editor works fine for me though so I just use that instead and deal with not having some of the newer functions

[viral pond]

yes i sometimes have to revert to legacy

[humble verge]

I don't know why SUNO has to put so much echo in songs.

[sullen sage]

Hello everyone, I hope this message finds you well. My name is Stacey. I'm not new to Suno per se, I first started trying suno in late 2023, after a friend recommended it to me. But I'm new to the community and I was actually recommended to post on this channel by someone who replied to me in the help channel earlier. As I said in the help channel, I'm really impressed with how far Suno has grown in the almost 2 years since I've used it. Especially with the 4.5 plus model. My question that I posed to the help channel was has anyone been successful in intentionally telling suno to do key changes, like in the middle of a song? For example, telling Suno to change the key up or down a half step, or a minor third or a major third? If so, what kinds of tags did you use and what instructions did you give? It would be really cool to be able to do that intentionally in some of my songs. When I posted this in The help channel a really nice person pointed me over here so I thought I would write and see if any of you have had luck with it and if so, could you tell me how to do it? Thanks so much and have a great one.

Stacey

[cerulean mesa]

Testing at rock... https://suno.com/s/3avKXiyxsxuXyXoV

[fresh atlas]

Something I notice is Suno isn't the best when it comes to choir/chorus/choral music.

I'm find it difficult to find the right prompts compared to other genres

[viral pond]

hmm, can someone tell me what to do, I use massively amount of meta tags througout the songs i make, and this seems to mess up the editor, it cannot show me the lyrics, and this make it hard to replace lyrics or sections

[gloomy shore]

I tried to create in 3.5v and it didn't work. I don't think, children song prompt, would even existed in 3.5v 😔

[viral pond]

this is one of the most genuine new features of the new editor, is that when you commit, you can expand the replaced portions and manually find where it vibes best

the light orange part there is where the original replacement was set, but expanding it further back to seamlessly make it match, it's so awesome ! THANKS !

[polar spade]

Most dont know about that, I would think. Definitely something to make a video how-to. It helps decide where to make splice points as well as more options for which segment os favored -- slide the left segment to the right, or slide the right segment over the left. Kind of a DAW feature, and useful with Suno.

[viral pond]

yeah, I disvovered this purelly by accident. but you had to commit it first, and also if you want to check the other render you can ctrl + Z to undo and commit the other one

but it should be ossible to drag and test without commiting first

I always tought this was a hit or miss before, so I discarded many renders and regenerated many times unecessary because of this, but its awesome, and i can understand that it generates much more than just that small part since when I listen with the pap i got lots of generations that was quite long and i did not really understand why it was like that

[graceful condor]

What is working best for Spoken word () or "" or ("")?

[viral pond]

that's adlibs

[spoken word]i think work best

( ) and (" ") might give slight different results tho `?

[graceful condor]

yeah I know that [Spoken word], "but as for the" (spoken part)

[viral pond]

I mostly use ( and ) for adlibs, backup singer or even backup spoken words.

but mostly I use them for non-speech vocals

like uhh.. yeah... oooh... and so on

[graceful condor]

Thanks @viral pond

[viral pond]

i think you can also do like this... [intro: Spoken words] .... [verse 1] but it might be a dice roll

@polar spade I know you been here long time and have vastly knowledge, but i have one question, you might not know it tho but i'll try anyways

the style of music has a cap limit of 1000, but I realized that since it is an LLM, it does understand typos, and other things that might be unrecognizable for most people, so i intensionally shorten words and compressses it pretty much to squeeze those 1000 character caps

hold on.. i have to find that example

here it is

I set this as a style of music, but suno starts singing it, and it sounds like hallucinations
"E Phryg(E, F, G, A, B, C, D):frag/glch mel, root E–F b2 drag, phr unrl, nt stut/drp/frx, gliss warp, det¢ micro, slid microt, ln swy/crk/bnd, tp drift, seq rpd clps→silc→rejag, Pls unstab, brth hiss bleed, irr acc brst, Mel frag+venm undertn, Bkg incoh whsp L↔R, dual tone(mat+sin), ovlp hum/rasp, wet ASM, detun layr, syll blur, ½bt offs ech, gsp+crkl mic, Surr whsp swl/rtr vol rnd, swarm orb, Gibbr vox rear, mutt syl, grgl+grn, pch ±oct rng, del offs, dbldcy→stat, chor faint gstv glch, Txt bld tail, rad hiss, met drn bed, rev brth swell, stat brst tear phr, Gls=tp warp, note resis res, E–F pull stres rpt, Dying mch motif:mel→cut/fail/clps, noi gap, frag mem bleed, Fin sec:narr fq→thin, sib hiss, whsp coil rpt, low gibb chnt rotout, All mono, raw, detun impf, no perc, no harm, no fx, mel+glch+brth+whsp+gibb+clps"

it's a pretty complicated prompt

it might misunderstand some of it but it still weird that it starts to sing it

maybe should post this to suno support and get theire take on it, might be a bug

[polar spade]

You got me. Thats interesting if it can interpolate it perfectly, consistently and throughly. But it can also just be a way to produce a certain vibe. Its your experiment, so I would not know. Cool thought though to try things in that direction. Thanks. 🙏. I may play around with the idea.

[viral pond]

yes it can understand allot of typos and shortend words, but i think this was toomuch for it, not sure if it understands the symbols like the back abd forth arrows etc 😄

yeah I am highly experimental when it comes ot my methods

[polar spade]

You should have been around in '23-'24 in ver 2.0 thru 3.0 when everyone was trying all sorts of wild stuff. Now everyone is already too serious. 😅

[polar spade]

Good stuff and nice ideas. I always like finding out more techniques others discover so I dont have to. 😀

[viral pond]

ah yeah i came to 4.0 , februar this year, and i think my music really evolved thorugh this time, I even intentionally use hallucinations as a part the artistic creativity here

I'll show you an example

here is a track i cranket weirdness up to about 85% , and near the end you can hear how I use hallucinations as a part of the creativity process
https://suno.com/s/ZZqmmJVUuZJIGQNC

9:15

I surely hope Suno will never fix this "feature" 😉

it would be awesome to get 30 seconds video clips tho, but 10 seconds work as well, I really love to make them loop seamlessly, so many people upload clips that just jitter loop... no seamless

[quick sonnet]

Can you simply use keywords in the prompt instead of spelling it out in sentences? For example, instead of "rock track in an E minor key," simply use "rock E minor."

[viral pond]

yes, if you do that, you lock the genre

[quick sonnet]

And what about the key?

[viral pond]

"Rock track", "E Minor"

you can also just tell example chatgpt to prompt... "Write a suno AI Style as a narrative "A Rock track in E Minor Key"

and it will spit out a nice prompt that suno loves

[quick sonnet]

But why lock the genre?

[viral pond]

you don't have to, but if you want it strickt like that it will try to follow it as good as it can, but it's always a dice roll

[quick sonnet]

Well, I'm actually trying to reduce the prompt to as few characters as possible so that it doesn't exceed the limit. In v3.5, the limit is 199 characters.

[viral pond]

https://suno.com/s/YXRP8TL86FJzuMhA this song is genre locked to Eurodance, Bubblegoth, Dance Macabre

oh, i have never used 3.5 so i'm not very familiar with that one, but it's still an LLM, you might write Rock Track, or what kind of rick this is, and put the other stuff as a meta tag in the verse... i.e.

[Verse 1: Play guitars in E Minor mode]

3.5 you using the free plans then i guess ?

[quick sonnet]

Exactly.

[viral pond]

after seeing some guy creating AI music on youtube, i instantly found a new hobby, and bought in instantly, so i always used V4, and now 4.5+, it's worth it if you can afford it, its just $10 per month

I paid for the $30 plan, but i bought earlier many small pacakges, so this month i downgraded so I can spend those credits

I am not familiar with any of the minor or major notes, but https://suno.com/s/DQq3GtVxzJmYYqTE just did a test, have no clue if this work with 3.5

tried a few of these major / minor keys

second generation here. https://suno.com/s/GPJU2tqSd4kYwuge maybe there are some difference, but i am not capable of hearing the difference on this kind of stuff 😄 I just know what sounds nice or not

I'll try to make it with 4.5 see if there is a difference

https://suno.com/s/zCEcZH650J6OaKCs
https://suno.com/s/6mmKvzk3NZb8cHVT

You decide

@polar spadewill they ever add more genres to this discord ? 😄

[quick sonnet]

I'm wondering if I select the instruments in the styles prompt, suno will automatically understand which one should be used for what?

[polar spade]

More Genres to the discord. As in what? Oh nevermind. I know what . The genre channels... yea... I dont know. I peruse them about twice a year. And post a few in 80s. So I dont keep up there. They are created as a moderator thinks up one, so maybe tap one of the creators names you see , top left of the channel and suggest a new one. Dont squirrel them so much, but one mention would be enough.

They may also want to limit sub-genres as well as a policy to give more songs to one channel than to be inclusive too much, so I dont know how far they'd go.

I notice many just use these channels to link their youtube (non-Suno) vids, which I then lose interest, personally.

[jagged narwhal]

there is a way to download my song without vocals?

[torpid haven]

Stem separation

You can do it also local for free

[quick sonnet]

If you are a Pro user posibly, I don't know, but else you can remove vocals later, that's what I am doing.

[torpid haven]

Suno stem separation is 50 credits

[quick sonnet]

But you need to have Pro, to use it.

[viral pond]

aha, yeah i'll try to grab one, heck i could administer it if they dont have time 😊

[torpid haven]

Sure, it's worth it

[viral pond]

download it as a .wav file, and use UVR 😊

[torpid haven]

Else that for free and try around different models

[viral pond]

aha yeah i guess only reason i would link a youtube is if i made some visuals, but you can add two links afaik so no excuse not to include a suno link

[jagged narwhal]

not a good solution

[viral pond]

why ?

[jagged narwhal]

yes I am

[viral pond]

you need to keep them on the platform for persona ?

[polar spade]

Im just trying to be a purist as much possible. Once a link is NOT Suno, its basically unknown what other DAW features were added of its not even Suno created. Then why bother listening to be impressed at ones Suno skills prowess when its a YouTube link. 🤔

[viral pond]

yeah thats true

one thing that i wish suno would implement was to upload custom sound effects to give songs more atmosphere,, I never had much luck using *glass break* or similar

[polar spade]

Other sites have "add fx" features, but then each has their better features than the other. Its partly up the what to market that will get the most users to join.

Sound effects and spoken parts are coming back into pop music (last really popular in the 70s) , so it may gain more prevalence as a new feature to offer. Who knows.

[viral pond]

yeah if i ever upload to distrokid i do some post editing to enhance that

hopefully this new daw coming can have some features like this

I have made a few nice 80s style songs, It's so awesome that this kind of genre can come back now today with a fresh twist

[jagged narwhal]

I guess it has not that option 😄

sad

[quick sonnet]

But mostly released by people who didn't experience the 80s themselves. 🙂

[viral pond]

yeah I am so happy that i experienced that era

remi made Rick rolled lyrics to this song, so i had to rewrite it, did not take any chances 😄

#1262592742945456148 message

[quick sonnet]

But you mean only musically, right?

[jagged narwhal]

ah I guess I found

[viral pond]

yes, both musically and the era, of course the cold war in the shadows was always the scary part,.. i remember i played "It came from the desert" back in 89 on the amiga 😄

oh... wtf.. I was not aware that there was a remake 😶

[jagged narwhal]

stems 😛

[viral pond]

that music actually still haunts me from that game

[quick sonnet]

The scary part was what I actually meant. Although it's currently experiencing a resurgence, and the lyrics from back then sound really current. Shame, I can't do anything with it due to copyright reasons.

[viral pond]

can always rephrase it, personally i feel its really weird to have copyright on speech, or phrases... to me it makes just as much sense as copyrighting instruments abd how to play.. like,,. there is a limit to what you can make without touching someone elses music

[quick sonnet]

With the twelve-tone technique there aren't that many combinations if you leave out all the dissonances.

[polar spade]

Hey if Rick Springfield can try to make a comeback [as hes doing , at what, 72(?) ] . . . We all have hope to try to revive that genre.

[viral pond]

well...

Mel brooks is 98 and will play as yogurt in spaceballs 2

😊😊

can't wait for this, I hope it wont dissapoint

[quick sonnet]

You seem to have misunderstood me. I meant that the whole 80s sound is being brought back today by people who were, at best, born in the 90s.

[polar spade]

I understood that. Just making an inferential random comment of my own. 😁 And it does relate, people born in the 90s making the effort should be played that Rick is heading the march. He's already released his new song on a 45rpm.

[quick sonnet]

That fits, it's supposed to be a parody of the character from Star Wars, which is much older. 🙂 🙂

[polar spade]

When his album is on 8-track, ill be ready!

[viral pond]

I remember this

My dad had this kind of stereo equipment, I remember the buzzing sound it made when warming up

[winged peak]

Hi there, maybe i'm too late to the party but did Suno over night change the limit of characters for the prompts or am I just confused?

(Ah, me stupid, was in Model 4 and not 4.5+ ) - Thanks 😉

[wispy laurel]

I was born in 1998 and most of my favourite music, and the stuff that influences me the most, is from the 80s or earlier. So this maybe isn't so surprising

We absorb whatever our parents were listening to when we were small

[viral pond]

@wispy laurel I most definetely did NOT listen to what my parents were listening to :DDD

but the 80s and some from the 70s is special, I was born in '78

[wispy laurel]

Not by choice
But you heard it, a lot, and it still will have made an impression, even if it was a negative one that defines what you don't like rather than what you do

[viral pond]

yeah ofcourse, I heard it, but unfortunately it never gave me any good feelings, it was more or less felt like old people music, but I grew up with kraftwerk, technotronic.. pump up the jam and so on... so i think those genres pre 70s cannot be compared, there was a crossroad 🙂

[wispy laurel]

"Old people music" 😂

[viral pond]

i just felt my parents music was just boring 😄

no bad feeling 😄

just my neurodivergent brain spitting out words

[wispy laurel]

That's fair. I felt that about a lot of it too. The majority of 70s rock and pretty much the whole of 80s pop never really did much for me, but there definitely were bits that lodged in my head, that later steered my own writing and composition in directions that don't at all fit with what was current when I was growing up. Leonard Cohen, Bowie, Led Zeppelin, and pretty much the entire "goth" movement of the 80s found their way into my head and took up residence there 😂

[viral pond]

there were ofcourse some gold back then also, led zeppelin, pink floyd, david bowie and so on tho I liked those

I am really horrible about music theory, but I just know what I like and whats good, that's why AI is perfect, cause you can just guide it to what you want, and it makes wholesome music, It's cringe but i really just lke to listen to all the stuff I make driving car, Infact i made so much ,that I can listen and hear something awesome, and it turns out its one of the random generations I made but totally ignored, never listened to

[wispy laurel]

Yeah I've had that a few times too. My phone refuses to separate music, any downloaded audio just goes into one big playlist and every so often something comes on that I didn't remember having made

[viral pond]

wish there was a better player tho, more like winamp style with custom playlist, random functions and such, i might need to make my own one day

[tranquil leaf]

But if you make your own, will it whup the llama?

[lost forge]

Ok, not sure what to search for, so here goes: Anyone has any idea how to do overlays of background vocals reliable? As a special case, I want childrens laughter during a particular part of the pre-verse. Not singled out, but in the background when the the sung vocals flow. Any ideas?

[polar spade]

Getting any children's voices is tough enough...

But strategically placing..... (laughing) or (giggle) .... among tags that might invoke a child's laugh can be helpful to blend laughing and singing together.

I dont think I would be successful to force it in a first try, and if I got it, I would then change the whole song to fit around it, rather than the other way round.

[lost forge]

Thanks. Figures. I was planning on wasting LOTS of credits. I was also trying to get countervocals to work, but that is also hit and miss with [Countevocal] or [Harmony] stuff.
Generally any overlay of the vocals seems to be hard.

[polar spade]

Being experimental can help. Since we can add vocals to vocals now, we can experiment stemming out the vocals, then click Add vocals to the vocal stem by adding parenthetical lines peppered into the lyrics , stem out THAT as a vocal and add instrumental to that last double-vocal output.

Or maybe 6 other ways about it to experiment.

[polar spade]

We still have this "Suno Studio" coming out sometime. So if that means we can overlay tracks, things can get easier. Fingers crossed.

[lost forge]

Will look into stems, haven t done much with them. The editor interface is currently so unusable, I only do crops and extends mostly.

[autumn rock]

Well done! to satisfy the kids I just had to give them a kpop song about their school 😛

[hard beacon]

i would love to see the prompts on the "listen and rank" page.. 🙂

[wispy laurel]

"A hackneyed, uncreative faux-country ballad, compressing every cliché of the genre into the first 8 bars" seems to be a popular one 😂

[hard beacon]

😄

i have an uplifting trance track right now 🙂

[radiant breach]

4.5+ is back to giving the same melody every time. Have a new song, new lyrics. I made a persona of a different song, and with this new one, using the persona, every song starts the same, as the persona song. Suno.... I didn't pay for this.

[ripe galleon]

why does suno process my voice so weirdly and make it vaguely robotic whenever i generated a song using a recording of my voice as the clip for it to extend from and then generated the full song from?

does it do this for other models other than v3.5?

[wispy laurel]

Try "clean, unprocessed vocals" in the style prompt for the extension
Or I think the more technical terminology is something like "dry signal chain"?
But also, 3.5 does kind of suck in a few ways. I haven't tried extending from uploaded vocals in 3.5, but I've done it in both 4.5 and 4.5+ and never had anything like this happen. I find it tends to add an American accent if the upload doesn't already have one, but otherwise I find it usually works like it ought to

[ripe galleon]

dunno if that will work on v3.5 properly if just the base recording morphs into that lolo

but ill try

[wispy laurel]

Always worth trying anything once

[reef tree]

i think Suno needs to add a tickbox to indicate additional vocals, male/female duets. Naming male/female within the prompts, or using male//female vocal, does not work. But separating the lyric section for different people, to help the system identify who is who.

[ripe galleon]

did that

nothing changed 😭

[grizzled scroll]

Is there a trick for getting Suno to switch vocal genders accurately in multi singer songs? I've been having so much trouble getting it to follow gender tags for different verses. It will do the complete opposite sometimes, sometimes switches genders inaccurately in the middle of verses, etc...

[reef tree]

costs me a lot of credits to switch back to the correct vocals. Need a more streamlined measure for the system to know. As it seems go guess.

[wispy laurel]

Reliable duets seem to be the Holy Grail of Suno creation 😂
There are a few different methods that various people say have worked for them
To be honest your best option would be to search this channel for posts with the word "duet", try out any suggestions you find, and see if any of them work for you
And, be prepared to spend a lot of credits trying to gt it to work

[grizzled scroll]

appreciate that, wasn't sure if it was something obvious I was doing wrong. I've tried so many tricks I've found online and done hundreds of generations and it still gets it wrong. I tried just using the editor but it screwed up the melody and rhythm of the song.

[umbral steppe]

I made a track in Suno that I really like, but I’m running into a weird problem. About a minute into the song, the EQ shifts dramatically. the intro has a great balance of highs, mids, and lows, but then it suddenly feels off and doesn’t match the vibe of the first section. I’ve already tried different covers with personas and inspiration options, and I’ve also run it through Remaster. The issue is, those approaches change the instrumentation and vocals too much, so the song ends up feeling like a different version instead of the one I want to keep. I’ve also tried working with the editor tool and regenerating the instrumental, but honestly, it doesn’t seem to do much to fix the EQ shift.

Does anyone have suggestions on how to keep the mix consistent without losing the core sound? Are there certain prompt strategies or tricks you’ve found that help smooth out EQ or tonal balance issues across different parts of a track?

[wispy laurel]

Yeah, unfortunately all the current models are just simply bad at duets. I've heard successful duets so it definitely is possible but there doesn't seem to be a consistent or reliable (or cheap) way to do it

[crisp slate]

Hey guys, my first message here
I've noticed recent few weeks persona are not being consistent at all, is that a known issue?

[polar spade]

As for duets, i always scatter bullets (prompts test) everywhere and whereever the most duets are, I draw a big red circle and dot around them and say I hit my target. Then all I do is post the one link to the one best song and everyone is amazed and no ones knows any wiser.

[wispy laurel]

That's pretty much the whole of Suno really 😂
We show off the one song that came out well, and we all pretend that we didn't fail 2600 times on the way to that one success

[raven turtle]

Have the same issue. Using v4 is a workaround

[pastel axle]

Does anyone have a workaround on the Remaster feature concerning it always dramatically changes the remastered track from the original track?

[radiant breach]

I guess i never thought of using 3.5 to make new songs then remaster them

[tranquil leaf]

I've had pretty reasonable luck with duets, but it almost always requires both style and inline prompting.

[tawny beacon]

same thing happened to me. I had to use the editor to change out the intro

[final dawn]

Im experiencing problems is anyone else ?
I can't seem get anything barely past 4min
Extensions are worse going few seconds or ignores lyrics and does instrumental

[humble verge]

If you buy credits close to the end of the month, keep an eye on your balance. SUNO just took 3000 credits from me when my sub refreshed.

[lament shale]

Is there a way to have the drum beats neverending? Suno likes to go into a quiet state in the middle of a instrumental song

[stoic pond]

yooo me too

[final dawn]

I hope they fix this cause I enjoy doing long songs which im known for and now i can't.

[stoic pond]

Yeah, hopefully, I actually used all of my credits because of this doing trial and error

I noticed this is mostly true when using personas and have successes of going over 4 min with just style prompts

[near flume]

Wait for the WavTools integration, currently you have no control for Remaster.

[near flume]

Credits do not rollover. It's use 'em or lose 'em.

[quick sonnet]

Is there a way to get the AI ​​to respect line breaks? The AI ​​is supposed to sing each line as it is, at least during chorus. If the AI ​​pauses briefly in the middle of a line and then sings the rest of the line along with the next one, it ruins the rhyme, and the music generated for it doesn't match what's intended.

[rare oak]

Hello ! Is there a way to delete all unliked tracks ?

[polar spade]

To answer your first question: not really, at least pre-emptively. I just go back and tweak a dozen times until its right. And thats for real life non-AI writing as well. Arranging it first is VERY helpful , but not perfect. Thats normally called 'rewrites" in the real world. We spend an hour writing a song, and 5 hours rewriting it. If this were free, probably it would not seem so impacting. I just wrote a quick 4-chord song and, regardless of the arrangement of the lines, I tweaked the melody, and a few lines, a dozen times until the metering sounded right.

Add to that, Suno DOES consider taking breaths. Some lines may be just too long to sing a couple of them in one breath. SUNO will break them up to get a breath in.

As for AI and Suno particularly, it does tend to look for rhymes, and so if theres the slightest rhyme somewhere in the middle of a song, it can make pauses and pull or push a line to focus on making a rhyme work where it wasn't supposed to be important. We may not hear it when we write it, but that may be a small part effecting the off-metering.

[polar spade]

And forget about ".....supposed to sing...." . Suno doesnt "supposed to" anything. 😉 We mostly stink at songwriting. If Suno did everything it was supposed to, we'd have terribly boring songs.

[lost holly]

Hi, you can mark several tracks at once with with STRG + left click or, if they are all together, with shift + left click.
Selecting all unliked tracks this way might be possible if you play with filters.
In general you have to go page by page though, all at once is not possible.

[quick sonnet]

Taking pauses in the line can be useful, especially to match the rhythm, but then you can also take an additional pause after the end of the line.

[rare oak]

Thanks ! I have thousands of tracks to delete, it's starting to make the website lag on some workspaces 😂

[polar spade]

Yea. Ive never gotten around it except just tweaking the line with dashes, commas, whatever needed to encourage Suno NOT to do something it wants.

.

[quick sonnet]

The line structure in poems and songs isn't used for fun. You can simply write everything one after the other and let the AI ​​piece something together. 🙂

[polar spade]

OK. So who agrees Sunos favourite key is F? I either have to tune up my guitar or use a capo, constantly. 😆

[lost holly]

STRG + A to mark all works, but you have to unselect the songs you want to keep then. Or move all songs you want to keep into a new workspace first, mark all and delete after 😄

[quick sonnet]

I guess it also depends on the genre.

[polar spade]

Probably true as well. Maybe F is an reasonable key for most singers to hit. Ive been setting my guitars lately for Suno tuning and alternate tunings. But you cant re-tune a piano. 😥

[swift gulch]

does anyone know what i can do i have the following problem:
When i try to cover a song and use a persona it doesn't follow the lyrics anymore, didnt had this problem a few days ago.

[rare wing]

I've been able to create various songs with Suno, but I'm still struggling with the constant repetition of simple scale-based melodies.
Is there a prompt that can prevent this?
Prompt like 'no arpeggio' or 'no sweep melody' don't seem to have any effect.

[polar spade]

Lyrics are controled under Styles slider, but so is the Persona. Try to find that balance.

First you can try the Style slider about 75%. See if that works.

Next question: you didnt say which kyrics you want it to follow. Are they new lyrics, different from what the persona had? Same lyrics as were in the persona? Or the output is mostly just gibberish?

[polar spade]

Add "free interpretation", "improvisation", and/or "free expression" ... those may get you more varied melodies than you want, but then tweak and find a happy middle.

Also, when possible, vary your line structure so Suno doesn't see a pattern and thus might create arpeggio-like sing-songy melodies.

[cyan jewel]

Hey gamers, are any of you well-versed in writing metal? im having difficulties with getting coherent instrumentation during the chorus

[swift gulch]

i'll try that, with the sliders
They are new lyrics, it just gibberish 2 different lyrics.

[polar spade]

If a new song with added persona, keep Styles up around 75% and Audio slider low 2%-20%.

If a Cover of a song but adding persona:
STYLE > 78%
AUDIO> 78%

Gibberish may be caused by having more lyrics originally in the persona song than in the new lyrics. SUNO "FILLING IN" the space with gibberish and mixing the lines together.

Try making another persona of just a 1-minute segment of the best part of the "personas" seed song, before using for making a new song with new lyrics.

[swift gulch]

thanks for this will try it

[polar spade]

As always. Dont shoot the messenger if it does not work. 😁 but come back and someone may discuss it further. There are too many variables between what works for me and whether the same will work for you.

[swift gulch]

1 output worked the other was some weird singing the structure and gibberish. no i idea what I'm doing different when i made a songs a few days ago 😅

[polar spade]

50% success. Thats good.

[polar spade]

If "what Im doing" is that youre trying the same technique but on a new song with new lyrics, it may not be the technique but just how Suno tokenizes (allocates each sound within the full arrangement of the song) one set of lyrics compared to another. Every small change can make the output completely different.

We hope the same technique though, will give us close to 20%- 40% success, most times though. Getting 1 out of 5 to work is good odds.

[rare wing]

For the first minute or so, there's still some repetition of simple melodies, but more varied ones have definitely started to emerge.
I'm now exploring ways to generate even more complex melodies.
THX

[polar spade]

A trick @wispy laurel suggested is to, if possible, repeat a few verses and choruses, either on first run, or just keep adding extensions of the same thing.... back to back, then , when done, edit/leave in the best ones (and delete the rest). Its an alternative way that can be fun to explore.

[wispy kraken]

Has anyone created a themed song and thought this could work in so many different themes/vocals but have issues keeping the same or very similar chorus or melody intact? What techniques would you recommend to stay consistent to the original but changing up words and vocals to match for a new song?

[neat field]

i cannot believe this is a v3.5 output....

https://suno.com/s/ZEfl0woc5XrzbwuJ

[lost holly]

Hi, i would try this using cover with high audio influence + make sure that new lyrics have matching syllables to the original ones, so suno can include them in the music without having to change the music. There is no guarantee but a chance. Good luck with your attempts.

[gloomy shore]

Usually it helps if you prompted like this;
[Verse1: Duet]
[Female vocalist]
Lyrics....
[Male vocalist]
Lyrics...

[lost holly]

That also can work in my experience, true.

[gloomy shore]

Then try putting the prompt [female vocalist] before the lyric again? To make sure the ai follows

[worn flicker]

The way I do it does take multiple tries, but it is the same way that I get the Vocals to speed up on purpose. Almost everything I do with Suno I use the way I start a song to influence the rest. So, if I want more consistent and solid lines then I will start the song with line breaks on purpose and then remove them as the song goes on.

[Intro]
Xxxxxx xxxxxx, yyyyy yyyyy, zzzzzz zzzzzz,
....

[Verse 1]
Xxxxxx xxxxxx yyyyyy yyyyy zzzzz zzzzzz,
.....

It may not be perfect but it does tend to tell Suno that you want less or no breaks in the line by setting the breaks before. As long as the intro is not the part you are wanting to have no pauses. It tends to use those commas in the intro to set the speed of the vocals. When it no longer sees those commas it tends to infer you want more packed into a line. When I do speed rap I remove all the commas and double up lines. If all else fails, I use 3.5 to get the pace and then remaster it to 4.5 then remix it from there bumping audio all the way up and weirdness all the way down.

[quick sonnet]

What is the style slider? Do you posibly mean the lyrics one?

[polar spade]

Are you just basic subscriber. Its part of paid accounts to have the influence sliders.

[quick sonnet]

Sorry my fault.

[safe sentinel]

Today is the renewal date of my subscription, but it still hasn’t been renewed. Does anyone know the reason for this?

[unique stone]

It renews at the same time of day as you originally subscribed

[safe sentinel]

Thank you.

[quick sonnet]

A "." at the end of the line seems to be helpful.

[quaint whale]

Alright, so I tried for 2 days to get the "Add instrumentation" (Beta) to work with a live recording of my own vocals and a clean electric guitar.

I tried all sorts of strategys to get it to work. I took my original recording and remastered in with Suno 4.0, and 4.5+ many times (as it seems 4.5 is not available anymore--except for covers, which I also tried endlessly to get something that sounded like the band hadn't practiced in months.

Now at 95%-100% Audio Influence it really was capable of backing my instrumental sections (Great Success!), but the song has like 3500 characters in it, so as soon as it returned to my own vocals, it would sound like 2 highschool, 3-chord punk bands playing over each other--I mean so bad I couldn't listen to it (and I love deathcore!).

https://suno.com/@pbot64 tried changing so many variables and used a tremendous (over 1500 or so) amount of tokens to try and help Suno with feedback on these new features, and got one song that sounds just passible if you are drunk, only because it has another singer, sing most of it.

I am excited for the "Add Instrumentation" and "Add Vocals" to be working properly. In the meantime, many of the 2500 credits I pay for each month went down the drain trying to use these features as I didn't quite realize how they still have a decent way to go to be useable (most of the time).

Regardless, Thank You Suno Devs For This Awesome Project,

Best,
Paul Barber

[sturdy flax]

Oh yeah \m/

[cyan jewel]

Oh hi. I have some questions I was hoping another metal enjoyer might be able to answer

[cyan jewel]

so firstly im having difficulties with suno continuing to riff while in more vocal heavy portions of the song like choruses. i just get palm muted alt-strummed single chords which inevitably becomes a muddy mess

also in general getting suno to actually riff has been trying.

right now my "workaround" for chorus mud was actually to not prompt for guitars during choruses and use other instruments like keyboards, or electric violin. but that's only a workaround and not a solution

https://suno.com/song/891b0024-eba8-4bd8-8864-f41a873a18f1 <- example of a muddy messy chorus

https://suno.com/song/be7f3710-6e02-487a-a29b-52578825eae9 <- my workaround currently

[hearty moat]

Hey I find persona doesn't keep the voice of a song exact every time. Is there a way to prompt a COVER of a song I have an insist it keep the exact same singing voice.

Or if I'm creating a new song based off a current song, how to make sure it creates the new song with the exact same singing voice?

What's the best way to ensure I get multiple new songs all using the same voice

[wispy laurel]

So.
When you create a persona, don't create it from the whole song.
Crop out just the bit of that song that best represents the voice you're trying to 'personify'. If there's a lot going on instrumentally, you might also want to split the vocal and instrumental stems from that cropped segment.
Use that to create the persona.
Then, when you create a song with a persona, set the weirdness slide down to 1% and the audio slider up to 99% (I find Suno gets weird if the sliders are at 0 or 100, that might just be my own experience though)
Also make sure that the style prompt describes the voice that you're trying to capture from the persona, because if Suno thinks there's a conflict between style prompt and persona it does weird annoying things, so you have to make sure that any description of the voice aligns with what the persona voice actually sounds like

If you're covering an existing song and you wanted to keep the music basically the same but change just the voice to a persona's voice, again aet weirdness 1, audio 99, style 0 and don't write anything in the style instruction (you can still use tags in the lyrics to indicate minor specific changes you want it to make, but for the best odds of successfully making it adopt the persona's voice you can't change too much else at the same time)

You might need to do multiple stages of this - the first attempt might give you a voice similar to the persona but not quite the same, in which case you can then cover that song using the same method to hopefully get closer to the exact voice you want

[junior terrace]

hello everyone, I'm really trying to get a song in the Daft Punk style and I know that they use a Talk Box for the vocals but I can't figure out how to make Suno understand it

[honest heath]

Have you tried "vocoder" or "vocaloid" as keywords?

[neat nexus]

Vocoder definitely does good stuff.

Try combining it with theaterical in the style prompt.

[unreal condor]

is there any significant diff between https://github.com/Tencent/Youtu-agent and crewai?

[humble verge]

I'm talking purchased credits, not monthly credits.

[humble verge]

Use a vocal stem to create the persona. Don't just cut it from a song, as SUNO is likely to copy the melody as well. I recommend using a 'clean' vocal stem with no added effects, such as echo. Like Mori said, best to use a snippet of the vocals, one covering the best range. Then prompt according to the type of voice. If it's an alto voice, make sure to include 'female alto' in your prompt and so forth. I use my favorite persona in all the songs I'm currently working on, and I'd say SUNO gets it right about 80-90% of the time when the genre fits my persona. Not so much if I prompt outside the genre - like if my persona was taken from a slow blues song, it won't work as well for a hard rock song.

[polar spade]

This is a Suno channel. Sounds like you may have meant that for "AI Revolution" or other channel. There would be no discussion of that here.

[rare wing]

I also use the technique of preparing several good segments and applying Extend.
However, the repetition of similar phrases tends to hinder that approach.
I find this especially noticeable in fast-tempo tracks.
I'm hoping Suno will improve in that regard going forward.

[quiet solar]

Yes - adapt the lyrics. Especially if you find there's a line that regularly trips up the AI.

With pre-v4 versions in particular, Suno can struggle if you hand it a line that has far too many or far too few syllables for the structure implied by the line breaks. It can also struggle with the timing if you have a chorus or verse that changes the underlying structure, or even if where the word stresses fall on the second version don't fit the timing and rhythm of the first.

Generally Suno will give quite a lot of weight to line breaks in how it approaches the song, but it will split them or combine them if that's an expedient route to fit them into the musical time available - particularly for the chorus. That's also why it will sometimes push a line outside of the chorus and add a few extra bars with no lyrics to get back on track - or lose all track of everything and start pulling in lines into the verse/chorus early and breaking everything.

Suno plays much more nicely with predictable structures, so having lines of a similar syllable length, or that respect the syllable structure set out in the first verse and chorus, are far less probelmatic. Although sometimes it will just make weird decisions about timing and stressing that breaks everything despite your best efforts.

[neat nexus]

In that case please send an email to [e-mail] about this because purchased credits should never be used when a monthly plan credit is available and purchased credits are supposed to roll over. Bear in mind though that it is Sunday currently in the US and a national holiday on Monday. So it will be a bit before business hours resume. But definitely send that email!

[quick sonnet]

So, as I wrote before, "." at the end of the line seems to be helpful.

[quiet solar]

That could help to some extent in that Suno sometimes respects punctuation marks as indicative of where to put in breath pauses rather than running on, but as with all prompts, it's takes them under advice and may do it's own thing anyway.

[quick sonnet]

Sure, but since I placed ".", the error hasn't occurred anymore.

[quiet solar]

Give it time. I've known it make absolutely no difference at all sometimes - but ultimately, every little helps when you're trying to get Suno to do something specific correctly first time.

[hard beacon]

pffff... slow motion mode in the #archived-ticket-help channel??? well...:-D

[polar spade]

Dont forget to experiment will varying the weird slider while extending to get both variations in meter as well as melody.

[viral pond]

I really wish Suno could fix this. Right now, when I publish a song and later discover a flaw, any edit creates a new song instead of just updating the existing one. That means the original version (with likes and comments) stays live unless I manually unpublish it, and the notifications end up looking spammy. It would be much better if edits automatically replaced the existing track while preserving its likes and comments.

And of course. the original can be renamed just in case you needed it again

[ripe galleon]

can suno make normal instruments like trumpets make very unusually deep notes like a contrabass

[hard beacon]

AI can create neraly everything if you use the correct prompts 😉

[ripe galleon]

i can make a deep trumpet?

without using a reference audio or anything?

[hard beacon]

what kind of style do you want to create?

[ripe galleon]

uhhhh

[hard beacon]

jazz ? blues ?

edm?

[ripe galleon]

something unsettling

not sure what genre thatd be though

or what genres that exist for it

dark ambience?

[hard beacon]

sounds like an interesting fusion genre 😉

let me try to create a untypical fusion song... 🙂 Hold on for a minute.. will send it to you via DM

[ripe galleon]

alright

thank you, robin

[viral pond]

@ripe galleon oh yes.. you can make underwater trumpet if you want to

[viral pond]

you want unsettling ?

check some of my songs

[ripe galleon]

oh really?

[ripe galleon]

like...

a warbled trumpet or a trumpet with a filter or

[viral pond]

yep plenty of unsettling and creepy stuff here

what you'd like 🙂

@polar spadeam I missing something here in the editor?, i deleted a small area which had bad lyrics hallucinations that ruined it at that part, i click the plus sign to get a new section and added lyrics, but it just did not want to apply the lyrics i wrote , it jus took sme lyrics from another part of the song

[ripe galleon]

oh sorry for no response

well uhhh something that invokes fear? an ambience of sorts? pads?

not sure

[viral pond]

anything in Phrygian F, maybe with Lydian notes

[ripe galleon]

perhaps, yeah?

[viral pond]

oh yeah 😉

but it all depends how creepy you want it

i can make things from slightly unsettling to pure nightmare fuel 🙂

[ripe galleon]

,,,, i got a strange idea

[viral pond]

https://suno.com/s/9XTCBo5d4dasrCMu

this one is slightly cursed

[ripe galleon]

this is like a mixture of indian music and dark videogame ambience

i can see what you mean by slightly cursed though

[viral pond]

like dark fantasy then

[ripe galleon]

evil dark fantasy lolo

but

i do have a strange request actually

something that i dont know you could even actually make in suno

do you want to hear it

[viral pond]

sure

[ripe galleon]

imagine like a generic summer hit

but as more as the song plays, it slowly like "contorts" and becomes darker

as if its wearing skinsuit that is a normal summer song which tears and rips and cause of it: turns into an evil evil song due to it faltering

does that make sense?

[viral pond]

well for tht you could extend something adjusting the weirdness to more chaos , I like to use 80-85% for non-linear results 🙂

oh you need a prompt

[ripe galleon]

id explore this if i had pro plan
the werid sliders i mean

[ripe galleon]

i definitely would need a prompt LOL

[viral pond]

an i havent used v3.5 before so not really famiiar there 😐

[ripe galleon]

you could throw together a prompt and i could try using it on v3.5

be it bad results or not

[viral pond]

so you want a traditional summer hit blended with a videogame disguised as Tunak Tunak Tun

[ripe galleon]

yeah

at first its a summer hit

and it slowly changes to something dark and evil till the end

[viral pond]

Catchy summer hit with tropical beats + hidden 8-bit chiptune layers, playful retro game energy, quirky viral vibe like indian meme song, upbeat, fun, absurd yet danceable

maybe something like that

to get it darker i think you would need to create a suno song structure with only meta tags

[ripe galleon]

whats a simple structure you would suggest?

[viral pond]

[instrument: bright synths, playful retro game arps, bouncy bass, tropical percussion]  
[mood: upbeat, quirky, viral, lighthearted at start]  
[progression: gradually shift from fun to uncanny, detuned game tones emerging, rhythms destabilizing]  
[effect: introduce subtle dissonance, glitch textures, distorted reverb growing stronger]  
[structure: intro playful → mid groove fun → ending unsettling, surreal, warped summer hit]  

add this to lyrics

no guarantee if this work in 3.5 since i never used it

[ripe galleon]

ill try it

maybe ill be able to buy pro plan someday so i could actually try this out properly

but idk

[viral pond]

it's jus $10 per month

[ripe galleon]

i have no source of income and i have a low allowance

[viral pond]

oh, yeah know that feeling

[ripe galleon]

it sucks booty

id have to make songs everyday to make it worth the money and im not that everyday suno user sometimes yknow

but watever, maybe someday

[viral pond]

ai music is here to stay

[ripe galleon]

you could say itll never go away /ref

[ripe galleon]

yeah v3.5 sucks donkey

[viral pond]

well its better thn nothing

[ripe galleon]

the outputs i got from these were
catchy 8 bit summer hit-ish songs

[ripe galleon]

youre right about that

[viral pond]

3.5 sounds okay if you just make instrumentals

vocals in 3.5 got that "shimmer"

[ripe galleon]

i do not know if youve looked in #📚┃suno-school yesterday

but

you know how like when you upload jsut raw vocals into suno

generate a song out of them

and make a full song out of that

the original recording of the vocals sound just,,,, weird and robotic

well not yesterday, the day before yesterday but you know wat i mean

[ripe galleon]

#📚┃suno-school message

does it do that at all for v4.5? have you tried it

[viral pond]

oh that. I do this all the time intentionally... inducing hallucinations

[ripe galleon]

you mean like

you also just insert raw vocals and suno just makes it also sound weird or

[viral pond]

yeah, the garbled hallucinations adds to the unsettling atmosphere

[ripe galleon]

damn

i wonder if its the noise in the recording or just how it processes vocals

[viral pond]

the beginning there sounds just like hallucinations

[ripe galleon]

i see

[viral pond]

I can show you one i used 100% hallucinations

[ripe galleon]

im curious, show me please

[viral pond]

https://suno.com/s/GS3GD2ZhiO86Ttg0 Fake language, no real lyrics... pure made up.. seek to 1:44

[quick sonnet]

Is this for instrumental track?

[ripe galleon]

yes

i think so

[viral pond]

you can use it for both instrumental and vocal, it just need to be put at the very top

[ripe galleon]

pretty awesome

[viral pond]

yes, it's one of those unintentional features that should never be "fixed"

[ripe galleon]

perhaps a setting made for it? idk lol

[viral pond]

i think there are many people who love this kind of hallucination when it vibes perfectly

[lost holly]

Suno doesnt like that though^^ i wager the log my account just for prompt injections that create hallucination, so fast they are fixing them by now 😄

[quick sonnet]

If this goes into the lyrics prompt, what goes into the other prompt?

[viral pond]

well if they ever fix it. i'll just apply japanese and write random japanese katakana characters

[viral pond]

no problem, you can combine it. the style prompt is the main lead prompt.. any meta tags in the lyrics are secondary

only important thing is to have them [square bracketed] otherwise it will just sing the instructions, which is weird, but sometimes it works some how also

i mean the way suno sings the instructional phrase 😂

[quick sonnet]

On some other AI it has generated some lyrics and it sung the out istruction, but in a place it reale have feated the rhyme. 🙂

[viral pond]

sometimes it understand the text is a instruction and ignore it as lyrics

but not everytime.. dice roll

[quick sonnet]

I know it even understands instruction in other laguages. However I prefer tu use english istructions with [] even for non english lyrics.

[viral pond]

https://suno.com/s/aNp8coju3LTGSeNP here suno failed catastrophic understanding the style prompt... it sung it 😄 thats rare. but i think i made it too complicated

[bronze trellis]

why cant I pay for 1 month like I did before? only 1 year option

and I cant download .wav files of projects i've generated on premium

[viral pond]

did you see this ?

if its not there. maybe try another browser

the whole page is made out of java script, it can be buggy sometimes

[ripe galleon]

yo liminal soundscapes

have you ever tried anything with deep pitch shifted vocals?

[viral pond]

hmm nott sure, might have happpend by chance some times, but canot remember, ussually i use female vocals, but deep male baritones, sure awesome

[ripe galleon]

hm

im talking of like

extreme pitch shift

like down a whole octave

i dont think youve created anything like that, right?

[viral pond]

hmm, im not sure i can remember, but im sure it is possible

[ripe galleon]

oh

well maybe

[viral pond]

hehe well think you can use [demonic pitch shifting] meta tag, not exact sure which parameter works in 3 here

this track is pure evil to fix

especially when the editor does not understand its own lyrics

probably wasted 500 credits sofar 😄

i have to replace tiny parts at a time

[viral pond]

finally made it, and save as new song... error .... ☠️

[ripe galleon]

i didnt see this until now

😭

well not sure, the meta tags you gave me for that masquerade summer hit song didnt translate well

so

[tropic island]

So i remastered a song that I did, and I like the new voice but not the new music.. is there anyway INSIDE Suno that I can cut the vocals out of the original and replace it with the new vocal that sounds better? And, if i cant, can i do it outside Suno and upload it as a new song that can be shared?

[ripe galleon]

what genre of music is this actually

[tawny beacon]

I don’t think you can actually do it effectively within suno. I’d recommend extracting the stems for the vocals and instrumental and uploading them in a DAW and creating a new song that way.

[wispy laurel]

The closest you can get inside of Suno would be to split the stems from both versions (you can just do Vocal and Instrumental, you don't need the 12 separate stems).
Then make a playlist that includes ONLY the vocal stem that you liked, and the instrumental stem that you liked.
Generate a new song (NOT a cover), using that playlist as Inspiration.
Lyrics should exactly match what you used previously. Leave the style instruction blank, or just a single [tag] if it won't let you leave it blank - if you have to do this, [female vocals] or [male vocals] is what I would use, as the thing least likely to alter the output. Set Weirdness 1%, Style 0%, Audio 99%, and generate.
This is not totally reliable and it's not the same as just pasting one stem onto the other. It can produce something that's almost identical to the original, combining the best parts of both. But, it also can change other stuff that you wanted to keep. If you're absolutely committed to doing this completely inside of Suno, you can try this. But you will get a better result, much more easily, by using an external DAW to just combine the stems yourself

[tropic island]

Yea I did that taking the better vocal but the original music.. it came out real nice.. Thank you GoldWave! https://suno.com/song/19b59b81-2f6c-410f-9dc5-e6533bfc753f

i had never uploaded audio.. so was not sure how to go about that.. but it came out nice I always appreciate other voices on things on here. Thanks!

[tropic island]

THanks for that Mori.. would have never thought of that at all.. I will always be learning new things on this new music horizon that is meshing with traditional music.

[wispy laurel]

You can use the same method to combine other stuff from different songs.
I managed to use this to insert a guitar solo from one song, into a different song that previously didn't have a guitar solo there. It's hit-and-miss by nature and it does use up a lot of credits if you're trying to make a very precise and specific change though

Generally if it's possible to do a specific thing in a conventional DAW, it'll usually be easier to do it there than trying to do it in Suno. But it's interesting and fun to find new ways to use (and abuse) the fancy new toy 😂

[polar spade]

That would probably come under "Chet Atkins Style"

[humble verge]

I sent an email to their support. It would be nice if there was a 'receipt' kind of thing in your account info where you could see how you spent credits. Just a list, like so:

[date] [Title] [Remix] -10 Credits
[date] [Title] [Extension] -10 Credits
[date] [Title] [Full Stems] -50 Credits
[date] [Title] [Remaster] -10 Credits
[date] [Title] [Remaster] -10 Credits
[date] [Title] [Creation] -10 Credits
[date] [Title] [Replacement] -10 Credits
[date] [Title] [Remix] -10 Credits
[date] [Title] [Remaster] -10 Credits
[date] [Title] [Extension] [Refund] +10 Credits

And so forth. Could help you keep track of your actions and your credit usage. Even better if each listed title was a hotlink directly to the track in question.

Also, my credits were just refunded and they gave me an extra 2000. 👍

[near timber]

Bro my whole pro subscription was cancelled out of nowhere and I didn't even cancel it

Like broooooo

[gritty plaza]

Yeah it’s possible. Worked for me. (Although I’m dealing with the other suno issues for my song - artifacts, mispronounced words, too much reverb) - but it got my duet / voices down to a T

[graceful condor]

Heads up just found out you can't use the word "shank".

[quick sonnet]

Is it posible to upload an acapela song and let suno create an instrumental part around it? Do I need to let the lyrics prompt blank, or just write an istruktion there to use the uploaded audio?

[tired star]

Does the summarise feature break songs right now? Sometimes it just messes up the audio completely compared to style tags😅

[pastel sluice]

Hey all, does anyone know how I can change one verse's voice from female to male in the edit screen?

[polar spade]

Is it a locked from use, or when you log in you have a basic account?

Often many, who use the SSO to log in have inadvertently logged a second account. Then thinking the SSO is auto-logging into their original account, notice a Basic subscription and feel something has happened to their Pro account.

Assuming that might have happened -- look through your account info while logged in. Is it the same email? (Write it down) -- The same user name? Does it show 50 credits or "0"?

Log out and go to Create New Account panel.

Enter the email you think was your original account. Rather than create a new account, it should automatically log you into an already created account. (This was a change made around Feb of this year).

If it starts to creates a new account, then back out before creating a new account. And note that that email is apparently NOT the correct one for the original pro account.

[sweet siren]

Really interesting info and approach. Will definitly try this!
Thank you so much!

[polar spade]

Assuming youre above a beginner level:

You can prompt, in the Replace lyrics, box , the same as if using extend.

Note: Adding the brackets and [female vocal] or whatever prompt you prefer, will take up more space than was there before, so Suno may not sing the verse correctly, even if it sings in female.

If needed you might need to add a small section before and after the verse before highlighting and clicking replacing lyrics to add the prompts.

[polar spade]

• i.e. --> using Inspo as a DAW "sequencer". 😉 like Audacity but with a vibe.

[wispy laurel]

Basically yeah
Even with audio turned all the way up it still improvises a little though so it never comes out exactly the same as before. So if you wanted an exact transfer of one vocal track onto a different instrumental, Suno (currently) can't do that
I guess it's more like taking Band X who wrote a song, the singer from Band Y who covered it, and locking them in a studio together. You (theoretically) get the best bits of both but it's still a new, different recording of the old song

[pastel sluice]

tried but wasn't successful

[polar spade]

Describe the result and whats your benchmark for "failed"?

Have you, before, extended a verse and change prompts and tags to change from male sounding to female sounding vocals? This would be quite similar in scope amd technique.

Ill run a test when I get home.

[unkempt shale]

OK. I just want marching drums. I don't want horns, or guitars, or anything els but Drum cadences. I've tried multiple ways just to get that. Anyone have any tips that will make sure the drums don't start playing to other instruments? Even if I stem it out, it's more about the cadence than anything

[polar spade]

Separating into the stem would be the simplest way. You're reliant on the total instrumental arrangement rather than just a drum drop of a cadence. So, once stemmed as a separate track, you may need to pick the best ones and duplicate the best spots for continuity and/or extend for some variety.

[polar spade]

Thanks for trying.

Just to show it is a viable tool in certain circumstances:

So i took a random all female song. Click on one segment called '"Prologue" -- yours probably say [Verse]. Changed style to "male vocal", and took advantage of the separator [Prologue] and renamed it [male vocal] . Clicked regenerate. First two outputs were changed to male vocals.

[grizzled scroll]

DId you label the male and female vocals a certain way in the lyrics box?

[polar spade]

Shank or skank. Its on the books. Not allowed. 🫡

[polar spade]

The little video and screenshot above your post can show an option of many for a simple process of switching gender frequencies in the edit UI. Changing [alternating] line per line could be more challenging but not impossible-- often by getting the first "gender" then the alternating ones then follows.

[gritty plaza]

The very first thing I did was prompt in the lyric box AND the style box that this is a duet featuring a male and female, and told it to pay attention to the meta tags in the lyric box that would indicate who’s singing.

It took a few generations, but it got it right. Then the one it finally got right, I kept using that one as my “base” and would cover or remaster that to change up the song as I wanted

[polar spade]

. . . I'll add.... you might wish to omit the original style prompts as Suno will typically pick up the style from the rest of the song. Just using the gender prompt only can give it more weight for that one segment. If it sounds too "off" continuity, then you can add 1 , more or all prompts back. So copy the prompts before trying this. Again, see the photo a couple of posts above showing just the prompt [male vocal] in the styles box #📚┃suno-school message and compare the sound in the video, it didnt change the vibe of the song, just things like cut out the reverb and lowered the sound, which though, was appropriate for that part.

[pale mist]

whats the promt for a boy kid voice ?

[lost holly]

Kid/child/young male vocal? Its hard to specify one prompt here, since it depends in my experience on the other styles you want to use to some degree.

[sweet siren]

Would it be beneficial to create a playlist with stems of favourite parts to use in new song creations? For example: you like the vocal from song A, electronics from B etc. and make a library to form a sound if you want to make an album? Hope this makes sense. I am new to Suno, so apologies if I ask things that have been asked.

[lost holly]

I would say try it, might work.

[wispy laurel]

It does work. Set the audio influence higher if you're hoping it'll directly copy things from the playlist, or lower if you want a similar style and 'feel' but not actually the same music
I do this, same inspiration playlist combined with different style prompts, and it's quite effective in creating stuff that believably sounds like different works of the same band/artist
Then when it gives me one that I really like, cover that with extremely high audio influence using the persona that contains my usual "singer", to make sure that the voice is right while keeping the rest pretty close to what was generated from the Inspiration playlist

[digital kraken]

I have a song for which I'd like to combine two styles: cumbia (verses) and raggamuffin (choruses).
Do you have any tips for getting Suno to properly recognize these rhythm changes?
I've tried a lot of things (even specifying the BPM), but it keeps the same style throughout the song.

[neat nexus]

Have you tried using style prompting for the Cumbia verses, then before it gets to the chorus, resume from an earlier time and then radically change the style prompt to match the raggamuffin style? You may need to find the sweet spot with sliders to try to force the change. Once you have that, then try covering that output and seeing if it will make the leap the way you want it to naturally. You may also need to play with the sliders for this. It can be a pain but it should be possible.

[digital kraken]

I used these :

[Verse 1 – style: cumbia, melodic phrasing, tempo: slow]
Crecí en la ciudad, soñando en un rincón,
con riddims de calle, fuego en mi corazón,
pero un viaje me llevó, por senderos de calor,
y encontré en la cumbia un latido de amor.

[Bridge – Transition, syncopated percussion + bass drop, switch style: cumbia → raggamuffin, switch tempo: slow → fast]

[break: silence]

[Chorus – style: raggamuffin, fast toasting, tempo: fast]
¡Cumbia’muffin, ven a bailar!
Fuego en la calle, ritmo sin parar.
¡Cumbia’muffin, fuerza y raíz!
Cumbia y ragga me hacen feliz.

[neat nexus]

Ah. no..I don't know how well that would work in lyric prompting.

[winged tusk]

Hi ... what is the delay between 2 posts on showcase please ?

[winged tusk]

What "¡" is for ?

[digital kraken]

This is the punctuation in Spanish.

[winged tusk]

Oh, I C , but what is the goal of it ?

[neat nexus]

I tried to show you what I meant. Please forgive me if I make a mess of everything I am not familiar with the styles at all. So I'm leaning heavily on GPT and my previous experience with making Suo songs....

Style prompt: Opens with slowed, melodic Digital Cumbia—warm accordion riffs, güiro textures, deep sub-bass, airy synth stabs, melodic phrasing, slow tempo, slowcore ballad, drawled female vocals with rervb, After Gradually morphs into a brisk 90s Raggamuffin Dancehall pulse: tight driving riddim, punchy synths, the accordion weaving in, overlaid with dynamic Jamaican toasting, catchy hook, melodic brass

Lyrics: `[slow dramatic Cumbia intro]
[melodic phrasing]
[slowed tempo]

(...)

[Verse 1]
Crecí en la ciudad, soñando en un rincón,
con riddims de calle, fuego en mi corazón,
pero un viaje me llevó, por senderos de calor,
y encontré en la cumbia un latido de amor.

(...)

[break: silence]

[Bridge Transition]
[syncopated percussion and bass drop]

[Rraggamuffin Intro]
[Fast tempo]

(...)

[Chorus]
¡Cumbia’muffin, ven a bailar!
Fuego en la calle, ritmo sin parar.
¡Cumbia’muffin, fuerza y raíz!
Cumbia y ragga me hacen feliz.

(...)`

Sliders 80-70-51

Then I tried it until I liked it. That got me this: [Removed links]

I then extended it from an earlier point in time. I extended from 00:52

Then I edited the style to remove what I didn't want any more.:

style prompt: brisk 90s Raggamuffin Dancehall pulse: tight driving riddim, punchy synths, the accordion weaving in, overlaid with dynamic Jamaican toasting, catchy hook, melodic brass

I also removed the lyrics I didn't need any more:

Lyrics: `[break]
[syncopated percussion picks up and bass drop]
[tempo speeds up]

[Rraggamuffin Intro]
[Fast tempo]

(...)

[Group Chorus]
[male and female vocals]
¡Cumbia’muffin, ven a bailar!
Fuego en la calle, ritmo sin parar.
¡Cumbia’muffin, fuerza y raíz!
Cumbia y ragga me hacen feliz.

(...)`

[Removed links]

Again I know nothing about the two styles so I am guessing largely. I hope this will at least give you some ideas going forward.

[digital kraken]

i don't know, i'm french 🤣

[neat nexus]

🤣

[winged tusk]

Haha, c'te blague ! 😄

[neat nexus]

Are we both the blind leading the blind here?

[winged tusk]

that's a very detailled answer with sounds to illustrate, it should need some time to explore

[digital kraken]

Thanks ! I try this.

[neat nexus]

In hindsight I think that I would completely focus the style prompt on slower stuff in the first part then make it fast in the second part. I think it may still be a little bit too fast in the opening part?

[viral pond]

it's really hard to cover old suno 4 songs and enhance them with 4.5 🙂

it always some glitch, and slurry vocals here and there

[agile jay]

has anyone else had an issue with lyrics no longer syncing up in the editor after changing the bpm from "Follow Track" to "Manual BPM"?

[dense marten]

All my old 3.5 songs are leagues above my 4 and 4.5 ones. I also wanna point out that I use MINIMAL AI in doing lyrics, as I write my own (I just use AI to perfect the grammar), because AI sucks so much and tends to over-word and over-express while also using such simple rhyme schemes and song structures. I notice that when I let the AI do 100% of the writing, it's just garbage and sounds like generic AI songs.

When you use Suno and other music generators so much, you develop some sort of good ear for AI music. Even more amplified if you're a musician and really knows your sht. I just hope that the v5 model REALLY tries to create Raw-sounding instruments/vocals first, then the user can then play around with them. If v4.5 speeds up my workflow so well via Suno's majestic Stems, then I'm guessing v5 will give us more control over those individual stems right from a dedicated App or webapp?

[wet dune]

Hey! I am new to Pro. Got my first song going, and right now it ends on a verse and then fades out.

I would like to have that exact version of the song play another chorus at the end, with the final lines of that chorus going into acapella or a fade.

I am trying everything I can think of in the editor to make this happen (and burning a lot of credits), but extending the song just adds more to whats already there, and trying to change the lyrics keeps trying to "insert" lyrics, even though I am highlighting the whole end of the song. Any thoughts on how to do this?

[fossil nymph]

I specifically had this issue using the new studio update. However clicking on the induvial clip and adjusting it manually use BPM fixed it.

[agile jay]

huh? I don't see that menu anywhere

[agile jay]

This is what I see when I click the three dots under the clip

[quiet solar]

If you want to keep the existing song totally unchanged and just add one chorus, your options are relatively limited.

I'd suggest doing this by Extend first, rather than directly via the Editor. Make the Extend point pretty much exactly where you want it to kick in for the final version, and lyric prompt a few Choruses to pad out the generation and give you a chance of trying a few different attempts to get an a capella version, with the final one being an attempt at a fade - e.g:

[Chorus]
[a capella]
Your chorus here

[Chorus]
[Soft, vocal only]
Your chorus here again

[Chorus]
[Instruments fade off with the final chorus line being just vocals]
Your chorus again

[Chorus]
Your chorus
[Soft finish, Fade out]
Repeat your final chorus line(s) here

That gives you four separate attempts at generating something usable - depending on the music genre, you may find Suno has an inbuilt reluctance to break what it feels are the typical genre rules and have multiple quiet/a capella sections one after another. Also, in my experience so far, prompting for an a capella line is a bit of a dice roll, so you might as well be credit-efficient about it.

Once you have something you're happy with in the mix, combine the songs into a full version and use the editor to remove the choruses you don't want to keep.

[polar spade]

It sounds logical that it would be out of sync. Was doing that before automatically syncing without doing anything and now that has changed?

Are you saying after a few tries, you cant get the lyrics to re-sync , usually by manually moving lyrics a bit? Re-syncing lyrics is a constant and ritual for me after most any "edit" , including saving a new output with a fixed BPM.

I'll test a couple of tries tonight, to see the issue. Just let me know if re-syncing attempts have failed where correcting sync has been possible before.

It sounds logical that it would be out of sync.

[polar spade]

Are you comfortable with the edit UI, to duplicate a chorus and add one to the end.

Click a chorus segment. Click on the segment menu below the segment, choose Duplicate. Click on the duplicate banner to hold and slide to the end.

The rest is experimenting. Remove part of the fade and insert a new 6 second section between the last originsl verse and the new final chorus, perhaps. Fade the end when done.

If you need a video, I can try, but won't be available for about 12 hours.

[echo moat]

hey guys, i had a persona made but i wanted to redo it because i did it with the instruments as well, but really just want one of the persons voice. I deleted my current persona refreshed the page, but it won't let me create another persona based on my songg

[echo moat]

what is the proper way to create a persona of a voice

[lost holly]

Try stem separation and then use the vocal track to create a persona.

[lyric kestrel]

Hi can someone help me with a prompt to get it like this or somewhat close? Thanks :) https://www.youtube.com/watch?v=OlgH-sCFhUU

[echo moat]

i'm trying but everytime i try to upload my acapella clip it says uploading clip forever -

[stark dagger]

Just get a clear stem of only the voice , nothing else, then upload that and you should be able to create a persona via that

[echo moat]

that's what i'm trying to do, it gets stuck on uploading clip, and if i hit continue, than it spins showing untitled in the library list and asks me to create a new track with it

[stark dagger]

Use an online stem splitter to separate your audio stems

[stark dagger]

Weird, worked for me

[lost holly]

My guess would be rap, piano, ethereal. Or describe it in more detail. For example: "The track opens with lush, reverb-heavy piano chords and airy synth pads, creating an ethereal, relaxed mood. Minimal percussion weaves in as rap verses begin, building expressive rhythms. As the track progresses, rap elements and subtle textures become prominent. A stripped-back piano breakdown spotlights hypnotic flows, then atmospheric effects gently close out the piece." But i have no clue if this would result in something similar.

[echo moat]

i'm uploading the vocals in flac format, any issues with that?

[echo moat]

ok it didn't like flac, ogg worked tho

[polar spade]

Did you say you had a previous upload of your song?

If yes.... then either click get stems and do a [vocals/instrument] track separation so youll have a vocal track as @lost holly mentioned or at the least, just make a copy of your file and try making a persona from the new outputed copy.

Oh and you've uploaded now with ogg format. 👍

[vapid elbow]

Every time I make a duet, the woman sings too early, or when a man is on the song, the woman sings, or vice versa. How can I edit this correctly? I'm trying in Edit, but I can't find it.

[quiet solar]

Unless someone here knows some dark arts I don't, there's not really any way to fix that via Suno's current editor. You can re-generate the song until it gets it right, or Extend from a bit that it got right to 'bank' that and then attempt the next section.

As a general rule, Suno struggles a bit when it comes to gender switching the vocals or performing duets.

[hollow hare]

You are honestly way better off using a vocal swapper/voice cloning service to change the vocals to the genders you want how you want it.

Typically, those services generates an output with all the voices swapped to the one you customized. For consistent duet, you should swap the song twice, for each genders, then use a DAW or audio editor like Audacity to trim/overlay which gender sings what part.

Likely a lot more money and time saved doing that since Suno's AI is not specifically trained for duets yet afaik

[wispy laurel]

This would be way easier than trying to do it entirely within Suno
I guess a middle ground, if you're reasonably confident using a DAW, would be to have Suno generate an all-male version and an all-female version with matching style and tempo, then extract stems from both and stitch the vocal together yourself. But that relies on Suno giving you two "matching" versions and you still have to manually edit via DAW either way, so it really boils down to whether or not you want to use an external voice-changer

[manic saffron]

How would one make it so they whisper for a certian verse and etc?

[hearty laurel]

OK, guys, I have one for the techies out there... I am trying to get Suno to do a cover song for me. I uploaded the audio file, complete with lyrics, and inserted the lyrics into my lyrics space. The song I have has some spoken words in the 4-bar intro, but for some reason. Suno won't speak the words. It just plays the guitar riff like it is supposed to, and then goes right into the verse.
Here is the intro I am trying to get Suno to say.
[Intro]
(Hey, all you young songwriters and musicians out there,)
(I want you to gather around.)
(I got a little story for ya.)
(Spoken over clean guitar riff and light drums building)
I tried about 10 different variations, each changing the lyric sheet, but still no dice. Any ideas? Thanx

[hard beacon]

try it with ( ) and put it over [intro]

*without

i am working on a track at this moment where i want a spoken text in the beginning too 🙂

[hearty laurel]

You mean, take out the parentheses in the lyrics and put the lyrics before the word [Intro}?

[hard beacon]

hold on..

DM

[polar spade]

Assuming youre familiar with most formats for metataging:

[whisper] metatag before the verse helps for one...

you may try [spoken] [whisper] or [spoken] (whisper)

....and even put the verse in quotes.

Those are ways I would initially try.

[hearty laurel]

Dasvon, thank you for the reply. Unfortunately, I am not that well-versed in the art of meta-tagging

I just didn't think that it would be that difficult to have someone speaking while the intro was playing.

[manic saffron]

Thanks so much!

[hard beacon]

it's weird... if you try a "Ladies and Gentlemen" in the beginning of vocals the word "Gentlemen" has an pronunciation error in 75% of the outputs 😄

[polar spade]

You may have meant @hard beacon who had responded to your post. I read that they were DMing you in that regard.

[hard beacon]

yepp.. we've talked about it and i have send him some examples 😉

i have? i did ? i've send ? whatever 😄

[polar spade]

leading off of what @hard beacon suggested --> Try using the tag

[spoken: whisper] "followed by the sentence in quotations."

And its baby steps before running then, Basic is a great start to keep from venturing out too far with every paid subscription tool available. Learning some "basic" prompting and tagging first can be less frustrating. Which in and of itself, most things we want, the seasoned users know can be done about 6 different ways with varying results - theres no magic bullet, so dont worry if things dont work.

Success seems to be about 25% of the time , we get something close.

[steady sierra]

I've had extreme issues with the free version just doing it own thing and ignoring my lyric structure and [ ] instructions. I have 3 verses but each chorus is different wording. So I'm trying a suggestion to extend after the first verse and chorus, however, Suno just thinks it knows better and repeats part of the first verse before singing the chorus. I've wasted days on this exercise in frustration. Can someone help me with the prompt to get it to follow just the structure I've outlined and then stop?

[quiet solar]

@steady sierra Without seeing your structure, prompts, and what you put in the Extend structure its hard to comment in detail on whether anything in there might be contributing to the problem, but its worth bearing in mind Suno effectively treats prompts as suggestions rather than instructions. Newer versions are a bit more compliant, but all those issues still raise their heads.

There isn't a magic bullet to fix it, but if you share what you're trying to do I'm happy to help.

[steady sierra]

[Short instrumental intro]

[Verse 1]
She saw more in me than anyone else
she loved me more than I loved myself.
After two years she returned my heart to the shelf.
I never get tired of crying.
She’d battled despair and a failing lung
At her funeral her memorial I sung
Within a year she took Dad along.
I never get tired of crying.

[Chorus]
I never get tired of crying
over broken hearts or loved ones dying.
Life has pain and clouds and rain
and tears wash out the hurt.
I kneel in the mud, as the drops pour down
my crying soaking into the dirt.

[Short instrumental and then stop]

[quiet solar]

That may be a little on the short side - Suno doesn't like doing really short generations, which might be why it's improvising a second verse for you by repeating from the first. It might be worth trying to generate using the whole song, but then Extending from the point you're happy with.

[steady sierra]

Did the extend from the end of Verse 2. Put in only the second chorus, verse 3 and the final chorus. What does Suno do? Reuses the chorus from the original song instead of the lyrics I put into the extend. I'm not going to pay for a product that can't handle the simple concept of "read lyrics supplied, sing lyrics".

[graceful condor]

Does someone have examples of setting what vocalist you want in []

Just like [Verse 1 - Female Vocals] ?

[quiet solar]

It's a GenAI, so the way it does things is a lot more fuzzy than "Receive Input X => Generate output Y". Which is both a strength and a flaw.

Most songs don't change words in the chorus, and that influences Suno. So it will regularly default to using the first chorus it was supplied with and ignoring subsequent changes. It can be definitely be overcome (and it's less problematic in more recent models). I've run with different words in the chorus before - I can say it helps a lot if it follows the original's syllable structure, but it doesn't get it right every time.

[quiet solar]

That should work fine - especially if the whole verse is female vocals. You can also use a [female vocals] tag anyhwere in the lyrics itself to mark a change for the next line.

Assuming that what you're trying to do is switch between male and female vocalists, best to know up front that it can be a frustrating exercise. If you're using v4.5, setting the overall singer to neither male or female may help, as does referencing both singers in the style prompt.

[wispy laurel]

Maybe try changing [chorus] to [refrain] instead
Still not guaranteed but it may help
(Suno loves to increase vocal volume/intensity, and/or add backing vocals, when it sees [chorus] - and it does this very much less with [refrain] so it definitely recognises a difference and handles them differently)

[graceful condor]

I guess I have to come up with my own workflow, but it seems like theres so many tools 😭

I got lyrics I like and a song thats pretty good, do I go into the editor and just "replace" sections a lot?

[bleak furnace]

Has anyone ever made this kind of ambient music on suno? this kind of ambient is hard to describe in prompthttps://www.youtube.com/watch?v=r9k74AGYZoU

[quiet solar]

You've got a whole suite of potential routes, depending on what exactly it is you're trying to do.

Cover and Remaster give you the opportunity to generate the entire song, coherently, with some potential adaptations (e.g. style, genre, instruments, singer, lyrics)

Replace/Add in the Editor lets you keep most things unchanged, but regenerate a specific section or add a new small section.

Extend can be used in creative ways to generate multiple versions of a section you want to replace, which you can pick from using Combine Song and then going into the Editor and switching things around / deleting as necessary. Downside is you're more likely to experience variations from the original tone and style.

All of them have pros and cons.

[graceful condor]

Thanks, in the editor is there any way for it to make a (mostly) consistient section reused later, for example if I redo the "chorus" it only redoes one, and wont redo the next chorus in the same way

Cover and Remaster seem to usually completely change the song, and the 4.5 remaster is not very good lol

Yeah the editor does not seem very good yet 😔

[quiet solar]

The Editor will only regenerate the chorus you specifically mark to be replaced.

[graceful condor]

If I dragged them together would it get both lol

😨

[quiet solar]

If you select both, then yes.

[graceful condor]

Oh so you can select both at the same time

I have a problem with the prechorus which is repeated, but I dont think theres a good way to get both lol

[quiet solar]

Yeah, you're not restricted to the section structure as Suno perceives it for Replace. There's a min/max length, but other than that you could definitely rearrange the song to put the choruses together and try to replace them in the same regeneration.

Could you just delete the repeated part?

[graceful condor]

Im not quite sure what you mean

[quiet solar]

If I understand you correctly, the pre-chorus is repeated and you want to remove one of the repetitions?

[graceful condor]

Oh no, I want to redo the pre-chorus, however this makes the second pre-chorus sound very different

[quiet solar]

Right, so you want to re-do one of the pre-choruses but have it sound as identical as possible to what was already there?

[graceful condor]

reasonably similar

[quiet solar]

And leave the other pre-chorus untouched?

[graceful condor]

Both of them need to be redone, but I want the redone ones to match

[quiet solar]

Ah, gotcha.

[graceful condor]

They dont/shouldnt be exactly identifcal, I think then id have to do tiny sections over and over on the borders

[quiet solar]

Assuming you're happy with everything else, then yeah - drag them together in the Editor and replace them both at the same time. Then once you've got a new song with updated ones, rearrange them back to their original locations - you'll likely have to use Heal Edits on all four of their connections to the other bits of the song.

[graceful condor]

Now I have a totally different issue lol

If I redo the "Intro", it then eats the rest of the song after inserting randomly?

[graceful condor]

Oh there's a heal edit, I see

[quiet solar]

I'll be honest, I've never seen that happen before.

[graceful condor]

I've reset the song maybe 50 times maybe is why

Thanks for your assistance btw

[quiet solar]

Changing things at the beginning of a song in Suno is a royal pain and best avoided where possible, but if you have to do it, I'd suggest using Extend and trying to create a new intro after the song has finished - then use the editor to switch it to the correct place afterwards.

[graceful condor]

gotcha

well the editor also breaks lol

[quiet solar]

No worries. There's a lot I've learned through the pain of trial and error - happy to help people avoid having to go through the same learning curve.

Is the editor just failing to separate out the sections in that new wall of green?

Or is that the new intro?

[graceful condor]

It inserts the new wall of green (the intro) into the middle of two sections that I didnt choose to redo

Its doing this basically, and then the first chunk of Verse 1 is gone except for like 2 seconds lol

Anyway I've listened to the song too much, so now it all sounds normal now

[quiet solar]

Yeah, genuinely don't know what's going on there. I find trying to replace the intro section does make Suno do some wild things because it doesn't have anything to reference from immediately before, but I've never seen it do that.

[graceful condor]

Yeah that makes sense, It's just a couple seconds of piano thats uninteresting, but oh well

[quiet solar]

If Replace isn't working out, try using an Extend from after the song has finished and prompting it to create an intro section/instrumental that goes into Verse 1. If any of those work, you can use the Editor to rearrange them to the correct place afterwards in a new combined song.

[graceful condor]

Ill keep that in mind

[wet dune]

Where do songs go after you click thumbs down?

[graceful condor]

Technically nowhere, the default filter is to not show disliked songs

[wet dune]

Thanks! I didn't think to look at the filters

[graceful condor]

Ohhh, if you find a voice you like you can save it as a persona, and then if you get a song you like, you can have the persona sing it instead

Now I just need to get good at fixing the lyrics in songs

[velvet jungle]

Can anyone answer this question, so you can set what songs you want other to remix by selecting the options slider on or off for remixing. So how can anyone remix songs when the setting for remixing is Off not allowing remixes, yet there able to remix anyways regardless of my track being set to off!? How is that possible, what magic trick are they pulling off ?

[elfin isle]

Hi guys, just wondering if anyone's figured out a reliable way of reducing the little audio drop outs (slight ducking that happens throughout some tracks during play back).

[elfin isle]

They go to Lumbridge (sorry - an old runescape joke 😆 ). But in all seriousness, I think they just get hidden from the immediate display but if I'm not mistaken they do still show up in ones library.

[polar spade]

If you never post your song, or share the link, no one would have access to remix it using remix tools. And remix is off, no one can remix by any standard tools. I suppose I could make a digital audio recording of it and upload it to remix.

[polar spade]

Well, im just saying you cant. That doesnt explain how another could. Sorry. Not a good answer, I guess.

[velvet jungle]

Yeah your definitely right... I agree 👍🏻... didn't realize that ..u gotcha...truly aoo5rhe response...

[velvet jungle]

I think I figured out what happened, I had a personal created with the title of the song and they used it to remix there own song 😂 unbelievable, I forgot I named my personal the same as the song hence why it appeared as if they remixed it .

Your correct 😁 either way , was my mistake for naming the persona after the song hence why it seemed they remixed my song lol

[polar spade]

Members dont publish or dont turn on Public, but dont think to also turn off remix. But if both are off, then tools, including inspire dont work.

Oh... ahhh.... yes a public persona is basically the whole song if we turn Sliders to 100%. Good practice is to make personas from best snippets or pieces of the song, rather than the whole song.

[velvet jungle]

I wouldn't even know how to do that, I didn't even know that was a option

[bleak furnace]

https://www.youtube.com/watch?v=ohTaJeWOnHo Hi, has anyone here ever made an ambient tune similar to this?

[polar spade]

If we are paid members, we can use the editor UI, is one way. Simple steps. Open the song in Edit UI. Decide on a couple of key segments that have the best samples of vocals and instrumentation, or maybe a good rhythm, whatever you want to be prevalent. Typically im.wanting the vocal qualities. Determine about a minutes worth. Note those and click on all other segment and choose delete in the menu below the clicked segment. Delete all but the ones you wish to keep.

Now you have a weird combination of 1 minute of remaining segments. Save that output. Make a persona of that new saved 1-minute output.

[polar spade]

And of course, even simpler, dont make any of your personas public. 😄. But I dont want to be a wet rag. Users enjoy setting personas or songs public so its really a personal choice.

[velvet jungle]

Ya gotta e kidding...wow I had no idea that was a option.. wow...now that's what I need to do... I would have never know unle6you told me this now...glad I posted that question... This feature is perfect in customizing what you want to save... Unbelievable ok good information right here..I appreciate it brother 🙏🏻🤘🏻

[polar spade]

No worries. Glad I can provide insight. Lots of good users with more ideas here. So post any concern or question anytime.

[velvet jungle]

Yeah I know ... When creating the persona damn suno forces the public option to on when I first create it then I forget to change it back to off...

[polar spade]

Same here. Still. I have to periodically check my profile to make sure I turn them off.

[velvet jungle]

Thank you my friend 🙂 you were absolutely amazing and helpful a gentleman and a scholar 👍🏻🙏🏻🤘🏻😊

[sweet siren]

Have been reading along with you guys @polar spade @velvet jungle. Really good info.
I sometimes have my persona 'lash' out when I don't want it. I coukd steer that with brackets/instructions how to sing where but best bits and turning into personas migjt work or putting that into various Inspo playlist. That way you can steer the song a bit more. I use inspirations of vocal stems now but they are from full songs.
Your option @polar spade might be even more suitable.
Anyway, sorry for the long text and thank you for the info. Such a helpful discord!

[polar spade]

I try to gauge suggestions so not to get over complicated. Most things can have 8 good ways to do something, and of those some may take 5 steps to complete but best to find something simplest to suggest that gets them out of the gate, not necessarily across the finish line.

[sweet siren]

Nicely said, and true! Thanks!

[toxic kraken]

Hey everyone! When I was uploading audio to Suno, it prompted "Can't fetch uploaded the audio." I still have enough credits. What is happening?

[brisk flint]

Just a glitch sometimes happening. Wait a bit and try it later. In a few rare cases some files Suno doesn't "like" as I have heard and read, even converting the file-format (.wav <> .mp3) and changing the name of the file then doesn't do the trick. So, for now, I'd suggest, be patient and give it a try later. Good luck! 😗

[quick sonnet]

Is there actually an instruction that can be used to tell the AI ​​to sing a specific word?

[honest heath]

you mean like a heteronym?

[lusty bear]

i need a little help please .....

any help would be much appreciated

[polar spade]

Just post your question. Hopefully a user or two will be available.

[lusty bear]

i'm about making a song for kids school ( KG, Prime ) to be in graduation Day Event , i already have lyrics but i need the best prompt for this since i need it to be song by children choir , or young kids chorus ,
i need no male voice or female , but children choir , like this one for example https://youtu.be/-opKU4WxLD8?t=45 but of course with suitable music theme for such event like kindergarten music events

[polar spade]

Not sure what you mean. Suno sings most lyrics.

Is it misproouncing a word and you need to correct it? You would have to edit the word....or...Do a cover with edits to the word....or....Use edit UI to replace the word....or.... A few different ways to do this.

[polar spade]

"Boys choir" works often to give a boys choir sound. It may be that simple. I think "girls choir" may also give a different feel.

When prompting for a [kids song] or [children's song] Suno may also recognize that and make some adjustments to sound bouncy and childlike, even sometimes use a kids vocal, but often more will just be a higher pitched female voice.

[quick sonnet]

For example to sing the word no as transcribed nooou.

[polar spade]

"nooo" does work. But it may also pronounce the "oooo" as an "oo" sound. You could spell it "knowww" to trick it to keep the long "o" sound. You added a "u".... so that would be "knowwwu"(?)

[quick sonnet]

This was just a transcription and also an example.

[lusty bear]

do you suggest a style of music suitable with that ?

[polar spade]

I always use a genre analyzer app like audioaidynamics and write down all the genres it references. So much easier.

Otherwise, its like a cinematic, ambient, boys choir piece. If you have lyrics then the arrangement and vibe will also be influenced by the lyrics.

[lusty bear]

i tried boys and girls choirs for many time now and every time i use this it gives me mature man or woman who song and no children vocals , choir

[polar spade]

Have you removed ALL prompts other than "boys choir" to give you a benchmark of where to start? I did one just now.... https://suno.com/s/wJUEr50ssd7zvgmB

https://suno.com/s/wgPvxlpWt0I8bt1z

[lusty bear]

so far so good but i want it to be joyful , happy , to be suitable for kids to song in that event like this prompt
Upbeat and playful children’s band style, filled with bright melodies and cheerful rhythms. Instruments should include acoustic guitar, ukulele, light drums, tambourine, and xylophone to create a fun and friendly atmosphere. The vocals should be youthful and energetic, sung in a clear and happy child-like tone, often with group harmonies and call-and-response parts that encourage audience participation. The tempo is mid-to-fast, making it easy to clap or dance along. The lyrics should be simple, repetitive, and catchy, Background elements such as handclaps, whistles, or "la la la" choruses add extra charm. Overall vibe: joyful, innocent, and energetic, perfect for kids to sing along and feel like part of the band

[polar spade]

Yep. Sounds like a plan. Experiement around until you get what you want. Others then may be able to help. Good luck.

[lusty bear]

i got this https://suno.com/s/HYgAlgAWUwugnq4n
but i need to be songed by children choir

[polar spade]

Ive got work this morning. So I can't help further. Sorry mate.

[frosty spoke]

I tried making certain song. and it starts perfectly. and its really good and exactly how I wanted it, to the detail. but it slowly starts to become static , and towards the end its impossible to even understand, its just loud static. I tried to redo it, but it never turned out as good

[sweet siren]

Hi
I have 4 stems (mostly vocal) in a playlist and use it as inspiration (Inspo) for a song. I have had a few times where the song just stops. I don't think the lyrics are too long or the descriptions. I tried shorter descriptions also. Any ideas what it could be?

[urban sonnet]

looking for ideas for prompting to ensure lenghtly track so it does not burn through the lyrics too fast but leaves amble room for drums/beats/instrumentals tried [120 sec, no vocals] but does not really seem to do much, extending a track 2-3 times overall gives 3 different sounds so does not really work either. Any tips? 🙂

[polar spade]

It seems you want a more casual group of kids singing ginger, not a boys choir.

Any listing of instruments are then weighted and tend to compete with each other than complement each other. So instruments would be better derived if you can work out the genre youre going for. Maybe a 70s pop, groovy, bouncy style.

Its helpful, if trying to go for bust on one prompt, to surgically metatag some prompts in the lyrics where you'd like to hear some changes, such as when claps should start.

As always, keep it simple to start. See what "simple" can do first. Then build from there. And others can help more easily with simple as well. That means, if they finally come up with something that works, but it is lavished with lots of prompts and tags, 1) its likely it took them a few outputs to get there and 2) what they give, may not produce the same outcome when you try. Start simple.

You want "Sesame Street" sound, it seems . So go to audioaidynamics and throw in a sesame street video from YouTube in. Write down the genres (not the instruments) and put those genres (not instruments) in the styles prompt to help get a foundation of the sound you want.

[polar spade]

Dp you know how to make a persona. Have you used personas?

[exotic pecan]

Negative prompt instruments doesn't work

[wise rapids]

Any ideas on how to prompt 3 different voices in the lyrics?

[graceful condor]

What is the best AI Generator to create really good Covers for Suno Songs with Words?

[sweet siren]

Maybe the best is a paid one. I used canva and it is pretty good.

[lost holly]

Try extending in from around when it starts to turn worse/static. meaning right before that happens.

[lost holly]

Check out https://discord.com/channels/1069381916492562582/1404828447695569037

[lost holly]

Try working with descriptions like slow/medium, low BPM like BPM:65. All no guaranteed ways, but good luck.

[lost holly]

You can try different things:

1. Try describing the voice/vocal in the [Verse 1, deep male vocal] [Verse 2, young male vocal]
2. Try putting the lyrics you want in varying vocals in () and or "".
3. a combination of the 1 and 2.

[fair moth]

I thought at one point that we could upload an instrumental mp3 file to Suno and have it generate vocals for that track, but lately it seems all I can do is use it as a remix source or remaster the mp3 I uploaded; is that functionality gone?

[quiet solar]

This genre is way out of my usual comfort zone, but I'll try to lend you a hand.

I think using 'choir' may be causing headaches, as choral music is typically quite different to child-focused songs. So I took your lyrics and adapted your style prompt as follows:

"Upbeat, innocent and playful children’s music, filled with bright melodies and simple cheerful rhythms, The song should be mid-tempo, based around simple acoustic guitar strumming with straightforward percussion and incidental use of instruments such as clapping, tambourine and xylophone. The vocals should be youthful and cheery, sung by a child vocalist with group harmonies and "call and response" elements in the chorus"

The overall summary tags Suno gave this were: children's music, acoustic pop, novelty. They might be worth playing around with.

This prompt was capable of some truly cursed outputs, but also led to this - which aside from repeating a chorus at the end, is pretty faithful to your original: https://suno.com/song/cc841ba7-5165-4543-bcc2-f9c5f345de1b

Some of the prompts are being ignored, e.g. there's no call and response because the lyrics don't set up and facilitate one. To give you an idea of how to generate something like that, I used the same style prompt with this as the lyrics prompt for a separate song:

"Upbeat, innocent and playful children’s music about graduation day. Lyrics should be simple, repetitive, and catchy with group harmonies and "call and response" elements in the chorus."

The output needed some minor editing to tag each chorus line with [Call], [Response] or [Harmony], but led to this: https://suno.com/song/cc9c17ef-b45a-470f-b98a-b12a6c2fab63

I've published both on Suno so you can raid the prompts and structure or adapt either of these directly if that's useful. Although the style prompt called for simple rhythms, I think with hindsight it might have been an idea to prompt for a simple melody too. The outputs also tended to come out very fast in tempo, but you can always use 'adjust speed' under Remix/Edit if needed.

[polar spade]

I created an instrumental in Suno.
Downloaded it.
Uploaded it.
Had suno create lyrics.
Added vocals through suno. (Had choice of vocal or instrumental)
Seems to work fine.

[fair moth]

It seems the only option I have is Remix which creates a different version of the audio I uploaded - is there a function I'm missing even as a pro subscriber?

[polar spade]

Unknown.

After uploading i had the option to add vocals. I could have written my own lyrics. So the add vocal/instrumental function is still a tool option directly after uploading.

If there is some censoring of your kyrics, that would be hard to determine.

I assumed you were saying the option was not in the menu or greyed out.

[neat nexus]

I haven't been following this discussion but just to make sure (because it happened to me) did you make sure you are actually covering it instead of trying to extend it?

Like I said, I made this mistake recently and was getting really upset how nothing was lining up until I noticed my mistake...

[polar spade]

@fair moth this is my screen..... Frostbite Kisses was the final song adding vocals/lyrics.

[polar spade]

It seems you want a kids chorus (or emsemble), not a kids choir? Something more like the attached? https://suno.com/s/mG7HAhudcR2higdS

[wet dune]

If I go into the editor and change a line in my chorus, is there a simple way to make that change consistent over each occurance of the chorus?

[polar spade]

Global search and replace ( including audio) ? Thats not yet available. So, correcting one, and then manually duplicating the. Chorus and splicing/pasting where needed, is about the quickest for that point.

You may then get click points where they are spliced which you should see if Remaster blends them in.

[wet dune]

Thanks!

[polar spade]

As there are usually no more than 4 choruses, just using replace to replace the line in all choruses isnt too bad either.

[quick sonnet]

In this context using something like:

[verse]
verse 1 lines
....

[chorus]
chorus lines
...

[verse]
verse 2 lines
....

[chorus]

[verse]
verse 3 lines
....

[chorus]

And every chorus using the lines of the first one, would be really cool.

[graceful condor]

If I use the edit function in Suno and save the song does it reduce the quality or is it the same?

[sturdy dove]

<@&1372996242325508176> ! There you are! I tried to share something with you and a few others last week, but I don't think it got noticed. I have something I'd like to share with you - the final of that project you and all those others helped me with!

Sorry - don't mean to hijack. Tell me where to go to share : )

[fathom siren]

It reduces the quality

Unfortuantely, every edit, download, upload, reduces quality (and not just the part you editing)

I often times do the edits in suno, download, and then splice the edit into the original downloaded wav (best you get). No way to get that back into SUNO at that quality so then I just release a "sunoish" version

[graceful condor]

Honestly I do not hear a difference.

Even with headphones

[flint ledge]

Lately all my songs have too much reverb and the background is almost over powering the vocals. I rerolled like a hundred songs with different prompts. I noticed that it's not as bad in verse 4. Is anyone else running into this lately?

[olive carbon]

Serious issues with the Editor these days. From plain silence with alternates, having the new song has glitches, so on and so forth. Suno not only needs to fix things but they SERIOUSLY need to think about free credits for compensation.

[sullen jay]

Actually, yeah. What is up with that? Used the editor just now and it wouldn't stop generating mute segments. Even after finishing the song, it's still mute. I burned lots of credits with that editor already, that issue is really putting the cherry on top

[fathom siren]

You can just compare the wave forms when you download.

Put your edited track and original in audacity or any daw, invert one of the tracks. Anything you can hear when you play is the difference.

Top is the original, middle is the edited track, (1 edit from 1:40ish-1:43ish) bottom is the difference between them.

If say you downloaded your original and edit in a DAW... and then upload it you get about this much change. (This example has no edits, this is just an original downloaded - uploaded - downloaded.

[fathom siren]

Try changing the edit type Classic, to Fixed or Fixed to Classic.

[viral pond]

reported to suno that i found it very annoying to have all my edit clips start to play sequential when i use the app while driving 😄

they listened to me, made my day

[silver escarp]

Is the site loading slowly for everyone else?

[viral pond]

seems fine here

[small bridge]

i'm having a problem right now where suno is ignoring my lyrics when i try edit a part and is making it all instrumental is anyone else having this issue? Make instrumental is not ticked its the first thing I checked.

[olive carbon]

Still silence... More credits spent! FIX THIS NOW! And re-credit us.

I can actually prove it if someone for tech would reach out.

[polar spade]

What was it again. You were working on getting a duet to work?

[fathom siren]

Try changing replace model, and ensuring lyrics all aligned

[small bridge]

thanks i'll give it a try

[polar spade]

It was interesting that I had a song with reverb ( and a bit over-produced) but when I edited and only changed the vocal from female to male, its dried up that section for a crisper, dry vocal with less production.

Its hard to find what exactly changes things, but often the adding or removal of just one "random" prompt can change it 180.

[olive carbon]

Reset don't work either. They claimed it to be this... but it's actually that!

[jaunty moth]

Hi, I'm a new user and would like to ask how to use the same vocalist as one of my tracks. I have created the persona with genre and his vocal characterisitcs ie: American Male singer-songwriter, gritty, raspy, raw , but dont seem to get a consistent singer...Is it not yet possible to pull a vocalist from a specific track and insert that singer onto a new track ? Any tips or cheats would Rock 🤘

[honest heath]

pull the stems out of the track and then base the persona on the vocal stem. Not perfect, but better

[jaunty moth]

Thanks for the tip, but if I upload the isolated vocal to the audio tab how would I add my original music to the track ?

[honest heath]

base the persona on the isolated stem, then prompt the musical style. When you isolate the stems they show up in your library. You can pull the stems out via the suno editor, it costs 50 credits iirc

[jaunty moth]

Ok thanks I'll give it a go.

[sweet siren]

I have had that too. I sent a mail yesterday. 3 actually as suno want you to submit one mail for each different problem. Still waiting for a reply. Hope to get a refund also since my other issue is that I suddenly cannot ear credits anymore.
Maybe best to aend a mail at support@suno

[jaunty moth]

I have an open ticket-thread too. They invited me to the group to get some more advanced tips, so I popped in to see if discord can make the learning curve a little less steep. I've also reported a few bugs in the editor which hopefully they fix soon.

[shy spear]

Okay, since I keep getting ignored in General chat, I guess I'll have to come here.

Questions. 1. Am I allowed to to post songs with explicit lyrics? 2. Do I have to put anything in the description alluding to it being explicit?
Because I wanted to try my hand at dirty pop and it came out SO damn good lol
It's basically a Verosika Mayday style pop song, if her songs were less 'sfw' lol

[cyan dagger]

Hello ! I would love to try a fnf like song on Suno 4.5+ but don’t know what style to put and to make it work !

[shy spear]

Thin 34+35 and Rocket, with a Hellish twist.

[graceful condor]

Okay this left me no rest. I just listen to the tracks side by side same portion over and over again and honestly I do even think that the new saved edited version sounds even better because it is louder more clear.

[quick sonnet]

Different people have different hearing abilities; not everyone has perfect pitch. Hearing usually deteriorates with age. I didn't mean that in any way personally.

[echo prawn]

I think it's the same old problem. When covering a song, the AI ​​has no idea where to pause.

😅

[quaint perch]

is it possible to upload chorus, and make the rest of the song around the given chorus in Suno?

[sweet siren]

I did this with success. Using a stem of a chorus and putting it into a Inspo playlist. Then use that as a source.

[quick sonnet]

I asked the same question yesterday.

[cyan dagger]

hopefully anyone can help

[unborn star]

Is there a way to somehow address the song you are using as a reference, aka “Use Style & Lyrics,” in the Lyrics Box or Style Box?

[fathom siren]

I mean at the end of the day it just matters what you "want" to hear so if it's closer to what you want to hear awesome. I was just pointing out every edit, upload, download etc.. changes the original (entire track) to some degree.

[quiet solar]

Yes. In addition to Cliodhna's suggestion, you can try Reuse with the audio upload - if you include the chorus lyrics as those you've uploaded, Suno will usually identify it as the basis for the chorus.

[cyan dagger]

Anyone ?

[unborn star]

what is fnf?

[ornate merlin]

how can i make prompt that avoid drums, percussion, and beats? Im so frustated on that. Hope anyone can help this.

[unborn star]

exclude them?

Advanced Option > Exclude Styles there you go with your instruments you dont want to have

[unreal quail]

Anyone else miss the Shimmer?

[glossy forum]

I also have the problem when replacing a part of one of my creation. The generated part is silent no matter I use Classic or Smart. Any solution ?

[wispy laurel]

Exclude doesn't usually work for specific instruments
You can try it but it's really really unreliable

[wispy laurel]

This is difficult but not impossible
Although it depends on the genre, I find if I prompt a specific list of instruments eg
[Piano, violin, bass, acoustic guitar]
in both the style prompt and in the lyrics (right at the start of both), it usually sticks to it
Although it's a bit reluctant if you're trying to leave something out that's usually 'standard' in that genre

Might take a few (or more) attempts, and if it drops a drumbeat partway through you'll need to extend from just before that point until you manage to get a whole song without drums
Although to be honest, it's probably easier to just let it generate with drums, then separate the stems, then recombine them yourself without the drum track

[cyan dagger]

friday night funkin

https://www.youtube.com/watch?v=zYmBeiqPB_s

here's an example of how it sounds

[graceful condor]

With personas should you use the slider at 100% if you really want that voice?

[cyan dagger]

i real ly hope anyone can try to help me acheive this, i don't wanna waste my credits too much on bold moves

[floral python]

yes but exclude anything you dont want, just like prompting a pic

You have a you're talking about the instrument slider, right?

[graceful condor]

Audio slider which seems to be dedicated to the persona.

[floral python]

Style influence?

That's pertaining to instrument style.

If you are making a remix and you want to keep the same flow and style, you put it to 0. It doesn't affect the persona.

There's a list somewhere on the Internet where you can pick out any artist you want. And it's like 4 words per artist that you can use to call him up or to callback

personas are kinda like a sepperate save....it might effevt some...but not as much as you think...and you can always reinforce the persona.

We can say the N word in rap, but I'm cussing too much right now...lol

You guys will sponsor a contest where it has the N word. But if I say **** we have a problem.

Sometimes it's the most simplest message. Like Mcdonald's theme song.
I didn't say Mcdonald's, so I didn't get a theme song. On my MC doobies, but my brother did. Because he just used that prompt Mcdonald's theme song. Came out beautifully.

[hidden yarrow]

Uh, what's your point?

[floral python]

Tell the employees that made the happy song. You don't have to use the GG, you can use QQ.

me or him?

comes out betta, is my point

The emphasis. being on the q and not the R in that word.

I mean, say right if you're gonna say.it.

ever seen a hundred fifty
rubber band up
ever seen a ngga
start with nothing

• That's a contest you guys are sponsoring. And those are the lyrics that it starts out with the very start.

why Is that appropriate? But me sing the S word will just spin some bots head around?

I'm not looking for a hard time, but why are you guys giving me one??

[cyan dagger]

Si uh anyone please

Make my dream come true

[floral python]

I even have my sensor on man. The sensors, you know, saying the word **** but that bot just jumped right on it.

[polar spade]

Usually, anyone wanting to get a "same sound as...." has in their mind the exact sound and feel, within about 90% of the original. It may be possible, but from experience, my trying to please someone and getting so close, but somehow they didnt here a violin or get a transition exactly like the original..... it would take nearly 100 outputs of tweaking to even get close. You know.... as in bold moves.

So, for that. . .

.. I'm out.

[wispy laurel]

Style slider controls how closely it sticks to the written style prompt.
Audio slider affects both the persona and the song that you're covering or extending from (if any)
Which is really annoying, and we keep asking them to separate those two things, but it hasn't happened yet.
You can also increase the effect of the persona by being careful how you create the persona. Most of the time, what you're actually trying to get is the specific voice. Best way to achieve that is, rather than just making a whole song into a persona, first split out the vocal stem from that song. Then, crop out just the section that best shows the particular voice that you're trying to get (if you like the deep, rough and gravelly quality of the voice, isolate the roughest and gravelliest section) and use just that little bit to make the persona. That way, it's much more likely to successfully replicate that vocal sound into other songs, and it won't get confused by trying to include melody or structural stuff from the old song into a new one that might not fit with it

[cyan dagger]

Oh no no I don’t mean to do a soecifi song

Just the overall voices and consistency

Which I struggle so badly with

As u can hear the voices are VERY different

I don’t want to make anyone feel bad or anything I’m ready to try anything with anyone to learn and see what works

[floral python]

OK. So how does the work boards work then? Your workspaces>?

That's influence too, right?

them boxes need to be worked out a little better.

truth

[wispy laurel]

As far as I know, workspaces basically just work like separate folders so you can keep everything related to one song in one place, and not mix it up with anything else that you might also be working on

[floral python]

I thought they worked like a callback. That's why you can like incorporate the style of the workflow

Desire as to how much you want the influence to happen. Of the box.

The box is called Style Influencer.

It's influencing it. It's not creating it.

[wispy laurel]

No, I think that workspaces are just to organise your stuff
Style Influence means, how closely will it stick to whatever you write in the Style Instruction

[floral python]

And you can give it prompts, don't get me wrong and then slide your fire up but.

[cyan dagger]

In the end it’s your choice if you wanna help me in this very specific style and I wish you the best

[floral python]

from nothin?

no style?

even your lyrics influence it

Has anyone ever tried to run just everything all in one page?

if you mention johnny cash,...in your lyrics...what happens?

[wispy laurel]

Probably it'll refuse to generate the song
It won't let you use lyrics that it thinks are likely to be copyrighted
And that includes names of a lot of artists, songs, albums etc

There are exceptions
But it's hard to predict what will or won't be allowed

[floral python]

you dont hear cash?

shrugs

I can't say the name Cash. I can't say he produced it.

Can't write it in the style box is what I'm telling you.

I mean, if that's your point, you're correct.

[wispy laurel]

No. Because like I said, Suno is cautious around anything that might be copyrighted
You can't say "sounds like Johnny Cash", because if the AI happened to be trained on any Johnny Cash records there's a slim chance that that could prompt it to directly copy something from him. So you have to find a way to describe what Johnny Cash sounds like instead

[floral python]

alright, thank you.

I mean, copywright...there is a fan claus you're also aware of?

I agree with you

there is a diff. in there the laws..and actual ones

[wispy laurel]

Yes. Suno's rules are stricter than what the law actually requires. It won't let you use public-domain lyrics either, legally you do have the right to but Suno as a company currently won't help you do it

[floral python]

it was about sunp bro!

lolx

z

woody, please take it out of your ...lol Or at least understand what context and topic is.

It had nothing to do with politics.

Topic was law.

Talking got sued, if you're wondering. Find out why. And did anybody ever find out if you want, if he lost? If he can still play that song that he made.? Because that would be a real life instance. Of the fan clause I was talking about.

token the rapper

[wispy laurel]

I have no idea who that is 😂

[floral python]

ouch

its my religion 🙂

https://youtube.com/shorts/dZXaDqTCzbU?si=_SPDK682CSU58ITC there wondering if hes talking about tech nina

The song Youtube rapper. It features Tech Nine on there with him.

I'm passionate. I didn't mean to like, spam you there.

[wispy laurel]

Woodhouse is gonna come and yell at you again if you keep posting YouTube links in the Help Me Write A Better Prompt channel 😂

[floral python]

thanks!! I would but... its part of my secret... shhh

thanks again 🙂 Woody. pews catholic...darn you woody

gave me an ideal, have a blessed day

[flint haven]

A reminder that this is not the chit chat channel, go to #🗣️┃chill-chat for idle chatter

[cyan dagger]

Please anyone help me achieve the Friday night funkin style

[hard beacon]

holy cow.. i've just listen to your youtube example... for real??!! this sounds like an 8 year old child is messing around with FL Studio and the Stock Plugins with random vocal samples.. 🤷‍♂️

turn the weirdness slider to 100% and you*ve got what you want 😄

[bronze pecan]

was a while inactive with suno, but 4.5+ apparently has treble again at metal music

Thank you 🙏

[cyan dagger]

Heck no lmaooo fnf has a special style of music

I mean they’re still musics but they have robotic vocals weird to recreate

I really don’t know how to recreate them

I know it might be very weird but I still want to reproduce those vocals with the instruments

Like a real game

[hard beacon]

i am in the music business as producer, dj and label owner since 2005.. and i am honest.. i've never heard trash like that 🙂 but i think you are talking about the vocal chop effect right??

well.. "never" is wrong... but.. you know what i mean 😉

test it with: High-energy retro rhythm game soundtrack with funky basslines, bouncy chiptune synths, playful call-and-response melodies, fast-paced electronic drums, arcade battle vibe, looping and catchy chiptune vocal chop effects

then use the "creative boost" button

[cyan dagger]

The vocals only are some kind of do da de that’s what I’m aiming for

[hard beacon]

ok.. so vocal chops

[cyan dagger]

Yeah !

[hard beacon]

i think this will be very tricky with AI

[cyan dagger]

And do I write anything in lyrics

[cyan dagger]

Ahhhh

[hard beacon]

but give it a try with the prompt and your vocal text...

[cyan dagger]

https://suno.com/song/6c487732-ea34-4a89-8b00-2d4e3ea6c084

That’s a good start !

[hard beacon]

now it's up to you to optimize the prompt 😉

[cyan dagger]

Ahhh I’m not good at prompting at all ! But I’ll try and see

[hard beacon]

A high-energy, retro-inspired soundtrack for a fast-paced rhythm arcade game. The track should feel instantly groovy and playful, with funky, driving basslines that keep the momentum alive. Bright, bouncy 8-bit chiptune synths should carry the main melodies, often interacting in a fun call-and-response style that feels dynamic and game-like. The rhythm section is powered by fast, electronic drum patterns with punchy kicks, crisp snares, and driving hi-hats to create a constant sense of speed and excitement. Melodies should be catchy, playful, and slightly chaotic in a fun way, reminiscent of classic video game soundtracks, with short, memorable motifs that stick in the player’s head. Add rhythmic chiptune-style vocal chops that act like playful digital voices, enhancing the arcade vibe and making the track feel more alive. The overall mood is joyful, energetic, and competitive, perfect for intense rhythm battles in an arcade setting. The structure should be loop-friendly, so the track can repeat seamlessly without losing energy or excitement.

🙂

[cyan dagger]

Wow.

[hard beacon]

let chatgtp (with the correct prompts) your best friend 😉

[cyan dagger]

Ooo it was too long it stopped at so the track

Near the end

[hard beacon]

never forget the "AI Roulette" 😄

[cyan dagger]

Oh right well that sounds good

Still work to do but it’s awesome !

Weirdness to 100???

Style and weirdness is at 50

[hard beacon]

no no.. that was just a joke... you can try it for 1 output.. but i think you will hear the experimental output 🙂

[cyan dagger]

The female in the song is very good ! The male tho sounds hyman

[hard beacon]

like a screaming male vocal?

[cyan dagger]

The vocals of the male are more human like

[hard beacon]

ok..

[cyan dagger]

Not robotic enough yk

Idk how to describe it ahhh

[hard beacon]

weill.. i repeat my words... AI Roulette 😄

[cyan dagger]

Im so so sorry

[cyan dagger]

True !

Any tips to get consistent ?

[hard beacon]

well.. that's how you would ask me about the correct lottery numbers 🙂

[cyan dagger]

Ahaha I see thanks again for the big help really

[hard beacon]

i spend round about 3000 credits to get the result i want 🙂

[cyan dagger]

Woahhhhh

It’s it just about regenerating or doing tweaks ?

[hard beacon]

doing tweaks and (of course) the best quality output... and if i got the best balanced output i am doing the rest in my DAW Software 🙂

you can't create gold out of crap

[hard beacon]

11:35pm in germany.. it's time for my bed 🙂 CYA

[barren prism]

why they removed 4.5-lite of my acc?

[cyan dagger]

Cya thanks really you did an amazing job I’ll always be very thankful

[trim glade]

Why doesn't Suno actually cover a song but simply draws inspiration from the given cover song?

[wispy laurel]

Turn the audio influence slider up and weirdness down, it'll be a lot closer to the original version

[cyan dagger]

Wow thanks ! But how can I duet that same persona for example with a male enemy that will rap like in the game against you

[trim glade]

that not help 🙁 wierdness is 0 and audio influencer is 100%

[wispy laurel]

First thing, this might have now been fixed but it seemed like there was a problem with 0 and 100, so maybe shift it to 1 and 99 instead
Second, if you want it to be really close to the original, leave the style instruction blank or just put either [male vocals] or [female vocals] in there and nothing else
If you're trying to make bigger changes (eg switch from piano to guitar, or play a metal song as blues instead), you need to be REALLY clear and specific about what changes you want
Also, you could try just switching to a different AI model (4, 4.5, 4.5+ or whatever else is currently on that list) - in some cases the older models do some things better than the newer ones. 4.5+ especially has amazing sound quality (sometimes, when it's not bugged), but it likes to improvise a lot more so sometimes just switching to a different version can make it work better

[trim glade]

thank you for the long text i try it 🙂

[olive carbon]

How did you email them... Stupid editor is still not working and still wasting my credits!

[unborn star]

So youwant music like that?

[heady prism]

Seriously considering just putting it in my daw and remixing it like that. lol. Likely won't count tho.

[languid sedge]

When I click instrumental in Suno AI, it's supposed to generate music without lyrics, but suddenly vocals/singing come out too. What's the reason for this?

[granite pawn]

Need help! I’m trying to add inflection into the voice in verse 1. Line 3 to the mess I’ve made can anyone help me

Headlights paint the dawn on an empty county line
Rearview full of ghosts I’ve tried to leave behind
Radio hums a hymn “to the mess I’ve made”
Every mile a promise that I might be saved

[fathom siren]

What is the style?

[polar spade]

The simple answer. It happens now and then. It could be that some combination of prompts weigh heavily towards having vocals.

If you like the song, for 10 credits you can stem out the instrumental if a paid subscriber.

[polar spade]

You want the whole phrase said louder? Try adding an exclamation point (!) It will be louder, more emphatic , scream it, or sing the word "made" longer.

Capitalizing one of the words may help as well.

[granite pawn]

Sliceofbytes the style is Country Folk

[granite pawn]

I just want the “to the mess I’ve made”

[granite pawn]

It’s Country Folk

[polar spade]

Sorry. Yes, the whole phrase "to the mess Ive made". No , not the whole line.

Same answer.

.

[granite pawn]

Yes just the phrase

[polar spade]

Or I should ask, what does inflection mean? Or rather , whats it doing that doesnt sound the way you would want.

[granite pawn]

I just want to raise the voice just a bit so it sounds like remorse

[fathom siren]

[Voice raises slightly, held remorse, pulls away delivery]

[Voice rises, remorse bleeding through]

[granite pawn]

It made a woman sing it 🤣

[cyan jewel]

@serene delta @true summit I did a thing. This is far and away the most complex lyric sheet I've made thusfar.

[granite pawn]

Thanks for the link, i will let my Studio handle it, appreciate the thought though

[polar spade]

Haha.... yea... OK.

[humble verge]

I sure hope SUNO will fix its false flagging of public domain material as copyrighted.

[half dragon]

I seriously doubt they would.

[polar spade]

Its just a placemarker. Its a general.term for "already published material" , so public domain has no bearing.

[humble verge]

well how silly.

[polar spade]

Economical. One error message for 20 different reasons to mean, simply, we're not allowing it.

For the optimal success of Suno, the less repetitive they are when generating, the better. And secondly, it forces everyone to be more clever in creating something that doesnt yet already exist.

If its been done before, why train on it now. There's enough "public" stuff already trained on as well as Suno now training on its own unique creates.

So, cost averaging, it is better to peeve some users, to then give others a better experience at having something we hope is unique.

[humble verge]

Yeah well, I get the argument. At the same time, if SUNO wants to be part of the future of music, making songs using lyrics from the Great American Songbook - for example - or any public domain traditional, like so much blues, is a fairly big part of musical tradition.

[novel kindle]

Hey, guys, I'm tryna make a song “Only Woman You Knew“ by my persona Emile Sandra, it's inspired by Aaliyah and Muni Long, Can y'all plase give me a prompt, the version is v4.5+

[polar spade]

@novel kindle Im kind of a broken record to say, but I throw songs in audioaidynamics.com and write down all the styles it pops up. Then find the 10 or so that are similar to both artists, I suppose, to get you started somewhere.

There are, of course, other genre analyzers as well. I just go to that one.

[cyan dagger]

Ye

[polar spade]

It may not be that they dont have it, in fact, its likely they already have trained on whatever you are trying to upload...... or they did allow uploads initially, and now having trained on it, censors it from being uploaded again. We just dont know. Many who had used public domain text still have their original on Suno, but they cant remix from it, else get a censor error.

[humble verge]

It's not because i'm uploading anything. I wanted to do covers of some old traditionals - but if someone else covered the song (and some of those traditionals have been covered by lots of people], SUNO auto-flags them as copyrighted, which they aren't. They are public domain. Take House of the Rising Sun for example. The original lyrics are from the 19th century. Written by Anonymous. Covered in various ways and versions since the 1930s. Blues, gospel, rock what have you. They belong to us all.

So it's just that SUNO's flagging bot uses too wide a net.

And honestly, I'd rather hear new versions of old traditionals than lyrics written by AI. Personally I write my own lyrics, but it's not like my writing is a fountain of plenty. I'm not Bob Dylan.

That said, it's far from all lyrics it flags. Mainly the well known ones. I did a great cover of a song written by David Lynch for Wild at Heart. So there's ways because the flagbot only knows so much.

[icy valve]

when using the Style Influence slider does that make it ignore meta tags in the lyrics more and less?

[untold briar]

I'm curious too. My gut feeling is that it has the opposite effect if anything, and tags in the lyrics are also increased. But I haven't tested this yet

[icy valve]

hrmm.. interesting, it seems to be basically ignoring my tags

and only going by the style description

[untold briar]

If you find the answer, be sure to update.

[icy valve]

will do

[sullen jay]

Apologies if this has already been discussed, but has anyone else found that the "exclude styles" field is really useless? Every time I put something in there, it feels like SUNO is purposefully adding the stuff I wanted to EXCLUDE into the song. Why is that the case?

[wet dune]

If Suno keeps insisting on putting "ohs" into the intros, is there anything I can do to stop that behaviour?

[untold briar]

Exclude is not very reliable. One thing you can try is just repetition, "piano, piano, piano, piano" instead of just once time. Also be sure you don't use grammar like "no piano" instead of simply the word "piano".

[cyan dagger]

Besides that u think there’s anything else ?

That works a bit better !

[viral pond]

ahhh..trying to move and clean my workspace the workspaces i want to move to just vanishes from the list all the time

[graceful condor]

okay enough wasting credits on trying to get spoken word in. It's just not working out half spoken the other half gets sung and not the best. then this error...

[vale obsidian]

Is anybody else getting to the point of bashing their head against the wall from the the excessively long vocal embellishments on the end of each and "every" verse suno generates? Like, sometimes it's entertaining, but most of the time it's just REALLY annoying and a major PitA to fix.

I'm just about out of synonyms for prompts and exclusions to try to circumvent it. At this rate I'm going to have to start making up my own words like Maynard James Keenan by this time next week.

[floral python]

if your doing it on a remix, it will already have the lyrics... there already added, that should be removed ...its getting doubles that way

[unborn star]

What does it do?

[floral python]

listen to yoursongfirst part... riight?

[cyan jewel]

its an html that has assets embedded into it along with the song. it tracks lyrics too.

[floral python]

then when you remix, it will have those in the box

take those out..

[cyan jewel]

its basically an mp3, visual assets, a lrc file, and lyrics toggle. html was just the best way to do that.

[floral python]

its causes a repeat

remix,,,clear lyrics

there added when there already there

[cyan jewel]

a screenshot of what it looks like

[floral python]

Im doing suno 😒

[cyan jewel]

Im replying to Kunai

[floral python]

well and another I dont wanna mentionm

? so your not even talking about suno...k