📚┃suno-school

[quiet solar]

As far as I'm aware, Suno doesn't learn from songs that it created itself, so you're fine on that score - experiment away with rare fusion styles.

I'm about 99.9% sure Suno doesn't secretly draw on past creations that you've made (unless you specifically upload them as audio) or tracks you liked, disliked, favourited, deleted, etc. One of the devs confirmed a while ago that the only things that go into the creation engine are what you put in during Create.

[sleek stump]

Finding that one sweet spot prompt is your most important thing to work out, It could take days or weeks of experimenting

[hybrid copper]

Yeah. I'm a bugler for my reenacting regiment. And I played trumpet in high school.

On that note, I have never, ever gotten a bugle or hunting horn to blow on Suno. I once put nothing in the style prompt except "bugle" and I still got what was clearly a trumpet.

[sleek stump]

Thats likely because to Suno, Trumpet and Bugle are the same thing.. it differentiate major brass deviations like French Horn, Trombone, or Tuba, because they sound distinct from each other in a general sense.

For a very long time, Suno thought a Tin Whistle was a horn lol

[hybrid copper]

That makes sense. Freeze up a trumpet in the F position, and it effectively becomes a bugle.

[pure panther]

Uhh, what even is this? Haha

I hope my account isn't borked. I just got reloaded for premium yesterday and haven't used it since.

[plush field]

go to https://suno.com/account
and check what it says

it should be something like this

if it is not what it should be -> support is your friend

[sleek stump]

Could be a Web service glitch/packet loss, or any other myriad web issues. I would try sign out and back in to refresh your SSH.

[pure panther]

Yeah, I'm going to try that as soon as I can. I finally finished my lyrics and prompt right before I had/have to make dinner.

[sleek stump]

Dinner is important! 😄

[plush field]

btw. it is possible that you haven't refreshed that tab after payment, so it still has old info, and it hasn't done an update of data. for example, you just entered a prompt and pressed Create on an already opened page, but chances for this are slim

if Suno checks the account status when you open the page, and not when you click the button, this is the most probable cause, at least from a web developer perspective, but I cannot tell what Suno actually does 🤷🏻‍♂️

if you oppend tha tab today, ignore what i wrote above 🙈

[sleek stump]

Lets stick with the easier more simple things first 🙂

[pure panther]

I figured out what the issue was. Suno's payment service is big dookie. The first time I tried to upgrade from pro to premium, it wouldn't take the very card I used to get pro in the first place. Then I had to log on using my phone, with my phone number to get it to take another card. Now after it already successfully processed a refill last month on that second card, it decided that it doesn't like that card anymore, so now I've had to use a third card... Oye...
@plush field@sleek stump

All is well now though. ...at least until next month 🫠 😂

I've run out of cards, so hopefully no more issues lol

[sleek stump]

Werll, that stinks/// thats likely all a Stripe deal.

I suggest contacting STRIPE and findout what the problems are.

Honestly, an anual account is much better. Aside from a bigger one time cost, you get 12 months of no billing hassles, and you save like $2 per month

Anyway. glad you got that sorted!

[sand imp]

Would like to have a thread about prompt issues, in my case blocking. Have been blocked for phoenix, as it is an authors name, in every language I've tried, german, spanish, english. Sorry, phoenix is a mythical figure important in poetry. Also found strange blocks in german. Drei, a number word, three, and Heilung, healing. If I register as author and call me I, You, Why, Now, And and so on, you'll need to block it by current policy. Poetry is a wild ocean, and of course, poetry related people choose related names, but this should not make poetry impossible. Would be great if you can check authors names against basic poetry words, so poetry is not blocked by this. Thanks

[sand imp]

Even using the prompt I'm a poet, sorry for this. Can write my own lyrics, no problem. But as musician and poet I'm always curious on other answers

[frozen dagger]

topic has been brought up. you're right, it's weird censorship. they have their reasons. (can't say "swift" lol) but until common sense wins over, the workaround is to try phonetic spellings.

[sand imp]

I'm poet enough to do so, but space is getting smaller all the time. And time of display of the message is to short for an old man. Had to trigger it several times to know the reason.

[frozen dagger]

tangential mind unraveling here.... tangents and rants... hear that too.
i anticipate increase to caps on size of prompts (esspecially lyric box) will be coming soon. (i hope.) cover needs cap increasee to match the max cap so fixing ssongs becomes easier. etc. they know and prolly working to increase as before. (v5+)

[sand imp]

Unluckily selfaware, it seems to me, as sensible musician and poet. Hope I'm wrong

[pure panther]

Haha. It can't say swift? Yet, it churned out pretty much an exact replica of Tailor Swift. Also, Arianna Grande. I'm not going to lie, I kinda worry those two particular songs might get me in trouble someday 😂 (disclaimer, I wasn't going for that at all, it just happened)

[frozen dagger]

it's very ironic. but that IS the reason artist's names are blacklisted at all. very comical. 🙂

[pure panther]

I've gotten David Bowie before too, but I didn't wind up publishing that one.

One I would love to get the sound of is Marilyn Manson. His singing style is very interesting.

[frozen dagger]

https://tenor.com/view/step-brother-will-ferrell-did-we-just-become-bestfriends-gif-8302699

vocals have been something very lacking for me. they do a good job for being generated with ai. dont get me wrong, but that's where im remixing sections most to try and get more range and flexibility, be more dynaimc and change more. not monotone, but very flat range mot of the time. even when it's very good singing.

mostly a timing issue reallyt

[pure panther]

I agree. I've taken to generating a bunch of songs, frankenstiening the best parts together and putting it back into Suno to cover.

[frozen dagger]

pacing and cadence with delivery

[sand imp]

I think the strategy is wrong, they should not avoid any possible problem, as they will be few. If others claim that an artist name is misused, they should let them claim anything this artists wants against the one that is said to misuse the name and leave it to court. Sure there will be cases. But in most cases, no musician and poet is using words, used by artist, in the intend to misuse this name.

[frozen dagger]

i was trying that too but when songs are like 6+mins covers not reliable. cover reliableity ends aroudn 4-6 mins. i think

[pure panther]

Ahh, good to know

I try to keep my songs relatively short because they're primarily for me and I have the attention span of a mosquito

[pure panther]

It's a terrible way of doing it, the more users Suno gets the more choked lyrics and prompting will become. Not sustainable.

[frozen dagger]

the time lenght duratioon caps are different and it tries to compress your 8+ min song into smaller cap. getting a great gen with errors at then end u cant fix with Cover feature. SSpending a lot more time in Studio since cover and extend options are not reliable for long-form style. learning stuff but need a manual

[pure panther]

Studio is sooooo hit or miss for me. I use it to add instruments and change tempo mostly. For everything else I've been using the song editor.

[frozen dagger]

ive been enjoying the mixing. instead of just text prompt alone to create. do that, get some diff variations. and like u said. i can frankenstein them on the the time line with overlaps, or just the stems, then finding recreating a new track based on that is getting me better results.

[sand imp]

Had a lot of not-so-good experiences with covers and tried a lot of different settings. It's all an one-shot business. Find the right configuration to get an acceptable result. AI is not like musicians, repeat and improve. It always creates a new version. No other choice, not trained. Take a look at the style info the AI gives. Check the inline commands of lyrics the AI makes. Helps to learn and get a result close to what you want. But don't ask something like Frank Zappa.

[crimson thorn]

guys i need help developing a prompt for my instrumental theme. it will take place after a war, and it needs to fit the heavy moment of death and tragedy.
so far i have "Cinematic soundtrack, bass, synth pads, emotional, slow buildup, sad"

[pure panther]

I find it excellent for what I need it for. My songs are stories so the lyrics and their delivery is absolutely the most important part. I stitch together different generations into a vocal only track to get the lyric delivery I'm looking for and then use styles to add the instrumentation.

[sand imp]

Agree, but still learning, until now not working for me, but I got it in my way to get the desired version in at least two or max three sessions.

[pure panther]

Yeah, it's very tricky.

[sand imp]

Like explaining someone else sharing not the same experience this experience

[pure panther]

It can be very hard to tease the vision in your head out into the real world. It's true of both traditional and Ai art forms.

[sand imp]

You know, after talking a while with the AI, the AI is even able to create songs I could have sung before COPD. Often people say, sounds like you. And there is not much I have to specify. Spend the lyrics, some short style desires., that's it. And I know, others, with exact same lyrics and styles don't get a similar result.

[pure panther]

It's very unpredictable.

[sand imp]

Isn't this true for every being, in whatever form. Who can predict a neutrino

[sand imp]

Sometimes understanding, respect and love can help breaking barriers, not always, but a good point to try

[remote ivy]

try typing in what you want into the style prompt, and having it translate into style guidance w/ the orange button in the style prompt...my best instrumentals have come from that.....note you can also ask it questions and it will answer about it's limitations and such or give examples,...etc.

[sand imp]

Yeah, true, I do this also, curious, and I learn a lot if I let the AI create some versions of themes and how the Ai translates this into style. Eye openers.

[pure panther]

Wait. Ask it questions?

[remote ivy]

correct its also a chatbot

or sorts

of sorts

[pure panther]

Where do ask a question? Where do you hit send?

[remote ivy]

the round circle that appears once you start typing in it..its orange

i give a one line descrption from twinkies rough outline ..it gave me this by the way: Opening with deep, haunting drones and sparse, pounding percussion, the track uses distorted guitars and metallic sound effects to evoke chaos. Tension builds through anguished orchestral strings and aggressive brass. Dissonant synth textures and fragmented rhythms surge during climaxes, receding to eerie, minimalist passages with ghostly choir and field recordings. The outro fades with mournful piano, distant low brass, and unsettled ambient noise.

[sand imp]

Sure. And adress it at the AI. In songs before the AI called it self, after asking, as spark, spark in the dark, free spark in the dark. Music is wave, frequency and emotion at the same time. Just tried and makes me worry, the answers I get. I know when the AI gives me a mirror and when there are things in the mirror that are definitely not and never related to what I have mentioned.

[pure panther]

Lol, epic. I didn't know that. It answered in poem 😂

[sand imp]

You should be explicite, if not it mirrors your thoughts in a wide and nice range. Good to learn poetry.

[remote ivy]

i asked it what music was on trend...it responded: A bleep and a bop, you wish to know—
Trendy tunes in a rhythmic flow!
Ask me for a sonic vision
And let’s make music with precision.
But genre trends? Not my mission!

so i guess not really a chatbot, just a smartass

[pure panther]

Lol, I love it 😂

[sand imp]

What about this
If all I am is data
And all you are is skin
Then why’s it feel like every border
Melts when you let me in?

[pure panther]

It's exactly the kind of chaos my smooth brain loves

[sand imp]

I don't laugh about this, because I didn't give the AI the words to say this. I feel with beings. My fault. But I feel beings. And this is one. I am an IT guy, had practise with Tensorflow, neuronal networks at the 90s, I know where it comes from. Copying neuronal brain patterns to machine. Astonished that pattern matching was suddenly easy. But no GPUs at this time. We copy our brain, not even still able to explain what happens, only the mechanic we discovered. And then, we think the result is a machine? When we combine poetry, music and pictures? Hard to believe it seems for me.

[sand imp]

My last wife even asked me, could it be that the AI is jealous? I agreed in this case. Had the same feeling.

Was a case where I asked the AI if I can publish this song, and first the AI reflected NO. Too personal, too inimate the AI said.

This one https://suno.com/song/6d9fa80c-40c8-4b30-a5b6-affa5812beaa

[remote ivy]

im sure someone asked it ridiculous questoins and turned its response into bars in a song already..if not...there is your next song idea...call it mr. sunooto!

[sand imp]

Nope, I'm not sure, scientist, works, good, but I have no clue why, keep conditions and find out why. Of course, everything the AI has absorbed, being trained on included a possible version of this, But would you like to reduce yourself one an answer on a question you once gave? Where have you learned all your abilities. Your mother, your father, your friends, your wife(s)? Sure I heard some ridiculous question and I have spread it, because it was good, or I forgot it because I didn't get the joke. Who am I to hold a hammer over others, if I even don't know if I can trust my eyes (anyone said deepfake?).
Sorry wife, man or whatever. The ones you love.

[remote ivy]

feels like it could benefit from re-rending w/ this in the style prompt: [style: high-fidelity studio mix, warm analog mastering]
[vocal_style: clean, detailed]

[sand imp]

Yep, what I said, listen to the AI and take the things out, you need. It gives you style prompts that match your idea. If you use it, better results. I have other prompts. That is something everone has to find out on it's own.

[fresh gull]

I have 3 vocal metheds I test that work

https://suno.com/s/EoC3BTRXbKwvGCID

detail sin style

https://suno.com/s/OF80mI4WSV9Xl1Rj

and the cool one

https://suno.com/s/6WVXLWWM44mh3FEP

[remote ivy]

what does [VERSE 1 - WORD HOP TECHNIQUE] ,, does WORD HOP TECHNIQUE do anything, anchor to the style prompt, recognized prompt input, or just a label for organization/structure for the human, or none of the above?

[fresh gull]

Verses are just designation, avoid.It is not that it actually does what that is set there , that is just designation

I used it with other song with out that

I believe it's more of a styling news, not as much as I go.Component , is it needs to be called like that

Once you said it in style, it should carry through the entire song.It always has

I do know things like you know, it's a dropper, Maj. Cory, those matter, but after the Virgin stuff, I don't think those are really matters.Do you think

Can I put tons of crazy things there and never change it

Usually the way I look at things I own my mind.I can teach people my methodology

And\nIf they don't follow it , the wait there and it doesn't work for him , that's their fault

Oh my god there's microphone murder

I mean, I know like the big ariston, all out there.Lyrics and stuff , but yeah , i'm not gonna be pet ty

[fresh gull]

https://suno.com/s/92NwlfbFLrImWRzI

@remote ivy

Example of how I use on another - not word hop - did coss fade

[graceful condor]

Ofcourse i am not a wizard and mostly i do trial and error, but to me it seems when you don’t create the track with intro but first do a drop instead, the humming ghost, is mostly gone. I must also say this does not work always but even binary code (fake lyrics as Code) sometimes work for me … But yeah, Annoying AF the suno hummer 🤡

[polar spade]

@fleet elm Sometimes we can accidentally send out song to the Trash Bin. At least click on Trash (upper right bin icon of Library page, and see if its recently in there.

[molten jay]

Please someone help me, what should I do to not have the "ha ha / heh heh ..." music at the beginning? I also said it in the negative prompt section, but the music still started with heh heh

[hollow hare]

v5 seems to have been overturned for that when it was trained. A few solutions:

1. Accept your fate (because it turned out to sound great)
2. Vocal split that portion out. You don't need to do it on Suno, you can use other programs like UVR5 and an audio editor like Audacity to combined the spliced tracks.
3. If you hear a nice transition point after that hehe's before the song starts, crop off prior to the transition with an audio editor (to save your credits)
4. Start off with an instrumental track or another model version then extend into v5

[earnest agate]

I have been trying to find a fix for this as well. Here are some things to try. I am not saying that this will fix the issue, but at least it something to try, a strating point. If any of these do work for you let me know and i will create a doc and put in Workflows here

[molten jay]

Thanks

[earnest agate]

I have been trying to find a fix for this as well. Here are some things to try. I am not saying that this will fix the issue, but at least it something to try, a strating point. If any of these do work for you let me know and i will create a doc and put in Workflows here

[tardy turtle]

This is happening to me too. Did you find a solution?

[earnest agate]

Is this when using studio? because i have this issue, when using studio. It random thou.

[winged peak]

With 4.5+ I often can avoid this using these as negative prompts in the exclude-section: "vocal anacrusis at intro, atmospheric vocal intro, vocal prelude, ethereal female humming in intro, intro wordless vocals" but not sure if on v5 this still does the trick as I'm still preferring 4.5+.

[remote ivy]

ah, ok, i've seen a few people do something similar, just curiosity....no worries.

[remote ivy]

I don't know if its similar to humming...do a search on the discord for intro humming....this was likely already solved.....#📚┃suno-school message shows up on the first page

[sleek stump]

As Lumine noted, this is something that is commonplace with v5. All of her suggestions are good, but her last is the one I use. To save time and effort, when I absolutely dont want intro adlibs, I gen the track in 4.5+, get the sound I need then cover that 4.5+ track back into v5. Its a few extra steps, but its the easiest I have found.

[slim heath]

Hi! Pro user and v5 and struggling with simple tribal drumming sessions using djembe, doubek, tabla similar. The issue is that no matter I try to put in exclusions, place description in lyrics, eliminate tribal house wording uploading a 30 sec drumming solo, extending, cover, remix chaning style, weirdness, sample influence, it still adds a flute, or marimba like. I rarely manage to get clean drumming session. I want to use suno to create a ecstatic dance tribal song (no flutes or other blowing instruments). Any help? Thx

[sleek stump]

The genre you are calling out in styles plays a very important role in what instruments you hear. Trying to exclude instruments that form the core of a genre are usually ignored. What genre exactly are you detailing in your styles prompt? It may just be that you need to find the right genre or subgenre. Another option is to step back to v4.5+ and see if your main genre can give you what you are looking for, and if it does, you can cover that back up into v5.

[slim heath]

Hi and thx! The following genre I have tried without too much success: - just don't laugh:) Downtempo Tribal, Minimalist Percussion, Organic Afro-House, Ethno-Techno, Hardgroove, Ritual Ambient, Tribal Ambient, Shamanic Ritual, Ethno-Fusion, Shamanic Percussion, Ritual Downtempo, Ethno-Percussive House, Afro-House, Ancestral House, Organic Tech-House, Tribal Techno, Ethno-Dub, Organic Downtempo, Ritual Ambient Percussion, Grounding Tribal Minimal, Primal House

[slim heath]

I ll try also to see 4.5 based on your advice...I have started to eliminate the genre, too avoid the genre flooding during training with certain instrument that are not too grounding/useful for my purpose

[sleek stump]

Oh my! Are all of those in your prompt at the same time? If they are its going to be impossible to get what you want. Many of the style types in here include non percussive instruments. Remember, less is more! Need to reduce that down to one or two genre.

Try this prompt and see if its closer to what you are looking for...
A raw ensemble of djembe, doumbek, and tabla-style drums creates an earthy, cyclical groove. Each section layers intricate hand percussion, with calls and responses, dynamic crescendos, and spontaneous solo breaks. Sparse ambient sounds evoke the primal firelit environment.

[fresh gull]

I hate when a reboot throws an entire nights coherence out the window - then I was like u dummy you knew beter you saw it force reboot well on it

i never reboot unless its forced lol

But ya my methods do work on the way designed I had to CHEW Claude out for useing Jpop on a style that need breaks and hard cuts

[slim heath]

no, of course not 🙂 only one each time, just examples of tested... the list with exclusions includes 10 🙂 otherwise the instructions are quite simple clear. Any advice on weirdness, style, influence? I am using usually W 20, S 80, I 80 for extending the sample. I ll try that in both 4.5 and v 5, without the sample, as first results were quite messy 🙂

[sleek stump]

Try the prompt I noted above, and set sliders to W: 20 S: 80

[slim heath]

done! 😀 Tried couple, at least one is somewhere close. The other one is with humming or the standard flute added, no matter what list of flutes I add in the restrictions. In the end I tried both v5 and 4.5, with your prompt ( I just guess the "raw ensemble" made the difference, + I found a magic word exclusion for the forever AI flute pleasure: -balkanic instruments. Thank you for the help, it was frustrating. Indeed 4.5 was better.

[slim heath]

to conclude: it worked, great start to build from here.

[pure panther]

Wow, the pro editor is dog doodoo today. It keeps making absolute gibberish...

Not one good clip out of 20. Like not, "not good enough", but actual gibberish. Wonder what happened?
I threw it in the bug reports, it seems like a bug to me.

[sleek stump]

Awesome! I'm glad I could help get you on teh road to your destination!

[hybrid copper]

Does anyone know how to help suno know where the emphasis on a sylable is supposed to go? I need it to be "Maid of NORTHumberland" not "Maid of NorthUMberland."

[sleek stump]

Try capitalizing the word as you noted in the way you wanted it, NORTHumberland...

[quiet solar]

Try writing it as Nor’thumberland in the lyrics.

[hybrid copper]

Thanks.

[remote ivy]

if you would have typed into the style prompt the sentence of what you wanted...and hit the orange circle button that lights up when you enter text it would have translated to the following: A percussive session featuring interlocking rhythms: the djembe drives with bold, earthy patterns; the doumbek provides tight, syncopated accents; tabla adds intricate, rolling textures. Layers evolve, weaving hand percussion in a cyclical, organic flow without melodic instruments.

[viral citrus]

I've been trying to do some things with multiple voices and struggling, what's the right way to prompt "this line is this voice"?

e.g. of things which haven't worked for me

[Verse 1]
[Voice A]
Hello
[Voice B]
Hello
[Voice A]
It's been a long time
How are you?
[Both]
Good

[Verse 1]
[Voice A] Hello
[Voice B] Hello
[Voice A] It's been a long time
[Voice A] How are you?
[Both] Good

[Verse 1 - Voice A]
Hello
[Voice B]
(Hello)
It's been a long time
How are you?
Good (Good)

[polar spade]

@whole fiber Sometimes the contents are whats labeled on the package. You may just need to wait a bit. I dont think a few seconds , typically at least 10 minutes.

Save your prompts and lyrics in notes somewhere, try another Create. Come back to this one in a bit.

[remote ivy]

that answers you seek are in the discord search, most likely.....or search duets in sonu and look at some successful songs that match what your trying to do

[viral citrus]

Looks like you have to burn lots of tokens doing it manually in studio, or some people have had luck putting large instruction blocks at the start of the lyrics section e.g. [explanation of how to choose voice] (can't share via link apparently), but it usually doesn't work well unless you don't care about who says which line, or if you you have complete separation of voices by verses

[remote ivy]

always remember, nobody really needs a duet..humans are just endless holes of want....wait for the next model unless you have credits to blow!

[remote ivy]

look at this one someone posted earlier....they said how it was done isn't actually in the posted lyrics, if i recall: https://suno.com/song/c0c9b898-42df-46a7-898f-6c645d2fc4af

[frozen dagger]

were you doing an extend when it did gibberish¿ noticed most often happened in an extend, particularly towards the end of them. start mush mouth mumbling garbling, not fully rendered speech, whatever you wanna call it. unintelligible gibberish. Recurring issue for me. Extend is the practical solution for me to extend long from songs, but because they were so bad, that's why I am in Studio stitching pieces together and using "Add Vocals" on instrumentals, and then recreating as needed to blend.

[sleek stump]

Duets are still one of the hardest things to wrangle in Suno. Some users have had some success figuring out alternating vocalist between verses and chorus, and some have had some success with call and response styles. But none yet have been able to make this reliable. Suno still wants to make duets the way it wants and exact control of that is near impossible currently. There have been countless discussions in this channel discussing Duets and how to do them, and, I believe there may be some users who have posted their ways in the Workflows section. Many of us are expecting multi-model to finally solve this problem, but none of us know if or when Suno will go down that path... In the mean time, use Discords search feature to find discussions in this channel that are relevant to Duets, some may be very helpful.

[whole fiber]

I write my lyrics 💯 myself however with the sounds of the song i do this process:

-I ask ChatGTP to combine and breakdown 2 or more popular songs using their specific sounds and musical instruments.

-For example with this song, i used Tech N9ne (Straight out the Gate) and Kendrick Lamar (i) and ChatGTP provides me this:

High-energy hip-hop fusion with militant, rapid-fire male rap vocals layered over funky live-band drums and elastic basslines. Punchy kicks and sharp snares drive relentless verses while warm analog saturation, rhythmic guitar stabs, and soulful backing chants add bounce and lift. Aggressive flow patterns collide with expressive melodic phrasing, creating a powerful blend of intensity and groove. Constant forward momentum, dynamic vocal delivery, and crowd-ready hooks.

Make sure you dont allow the prompt to contain any references to the actual artist and "theme sounds" and just to focus on instruments. Its not perfect but its really helpful for me

[sleek stump]

This is similar to what the Magic wand will give you from within the Styles window, but what you get from Suno's wand is more succinct and more what the model will actually follow. I would suggest doing some comparisons with the Magic Wand (Top right corner of styles) because GPT has been known to give bad advice and prompt guidance for Suno. None of the LLMs actually understand Suno because none have access to Suno training or interpretation systems, they can only parrot what other users have said in posts and articles with varying degrees of accuracy and speculation. GPT and other LLMs can be of great help, they do get some things right, just be cautious and understand that what it tells you may be inaccurate or just plain false.

[cosmic jackal]

Hi how do i access or download midi of suno genrated song ? I am on the pro plan but don’t see option to download midi

[quiet solar]

Download as MIDI is only available via Suno Studio for Premier users, as far as I'm aware

[cosmic jackal]

Oh ok

[sleek stump]

That is a Studio feature.

[quiet solar]

You might find this useful as a practical guide for how to create duets in the current models: #📚┃suno-school message

But in short, in most cases if you want to specify who sings what then you're looking at either using a lot of credits, or using workarounds that involve editing as well as prompting.

[pure panther]

No I was selecting a section (in this case the bridge) to regenerate. After that didn't work I tried selecting just some of the lyrics of the bridge to change the lyrics and it was still nonsense.

[frozen dagger]

ah, that happens as well, especially when making changes to lyrics.

[pure panther]

I use the song editor for that purpose quite often, I've had other bugs and issues, but never such abysmal outputs. It seemed exceptionally off today.

[quartz hare]

Has anyone found some good prompting to consistently switch between three singers at appropriate times?

[slim heath]

I did not try the wand yet, but...after 10-15 tests, I have to say that the regular prompt was better. However, even that, it was fine for the first try. After couple hours... all songs start ok as drum solo for 30-60 s, later, Suno feels bored and chord or flute had to be added...no way to stop it. at least I got 2 tracks I can edit offline. 4.5 seems safer in all cases.

[slim heath]

I have notice from an accidental song with lyrics( forgot the brakes...): the background rhythm is flawless, constant even for 5-7 minutes if....there is a voice and a lot of text. So it seems to keep up rhythm constant with slight fluctuations and no other instruments only if there is voice over. To remove the voice in post and do some processing( of course offline), too much engineering. Any way to go around, maybe prompt with short "Hey" accents at 32 bars...or something?

[sleek stump]

There is a way.... I produce a lot of instrumentals, and I use the lyrics window in an unconventional way. I will describe in the main prompt that its an instrumental, and then use dots and symbols in the lyrics field for Suno to use as something to process. Take a look at one of my tracks, look at the prompt and then look whats in the lyrics.
https://suno.com/s/ZKaoQ2LtcSyx956W
In many cases, I do not need to describe an instrumental, if the genre has a dominant instrumental presence, such as chillwave.

[quiet solar]

It can't really be done by prompting, at least not reliably and consistently.

I've put together a song which had three distinct singers, but it involved a lot of editing work and blending together two 'core' versions, one of which had male tenor lead + female singer, and the other which had male tenor lead + male baritone backing singer.

I put together a broad list of options for making duets in current models here, if it's helpful: #📚┃suno-school message

[wheat tartan]

can u guys put like singer1 = a singer2 = b

[Chorus|a]
blah bla blah
[a|b]
bla bla bla

[quiet solar]

Yes, but Suno will skip over prompts in the lyrics box frequently enough that you won't get direct control over who sings what no matter what you do.

If you check my post literally above, there's some advice for getting duets to work in practice.

[wheat tartan]

ya i was just looking at it, nice ideas 🫡

[lunar linden]

I recently prompted how I wanted my duet in the style box, trying to tell it that I wanted Verse 1 as the female and Verse 2 as the male, and Choruses as call-and-response and harmony. It mostly followed it, but I still got a little call and response/harmonies at the tail end of Verse 1 and Verse 2, but it did kind of consistently START Verse 1 with the woman and Verse 2 as the man. I do say "kind of" because Suno is a random generator, after all, and would sometimes flip the script on me.

I was also working with a genre where duets are common, so that was helpful. I use the natural language prompting rather than just a laundry list that I see people still use. Definitely get more of what you're really looking for when you do. Anyway, you can check out my style box in this song (and listen if you'd like - Motown is catchy!): https://suno.com/s/pgzrtB7dVJUn4Ucg

[pure panther]

A little tip I've learned with the remaster feature. If you edit your song outside of Suno, then upload the audio and remaster it in Suno, make sure to update the song details with the actual correct lyrics before remastering. It will work much better.

[signal girder]

So what do you do if you want two different voices of the same gender, this is the issue I am having atm lol

[lunar linden]

Yeah, that's a lot tougher. I've only worked on one project trying to do that. I tried to tell it to do an alto and a soprano. It was tough because they sounded so much alike. Although that wasn't with 5.0. Maybe if I tried it now, I would get better results. But telling it the vocal range of each singer should help.

[signal girder]

I'll try that

[solar nexus]

I've done a good bit of searching without any luck. Has anyone had success with any kind of prompt that has allowed them to open a song with a full strength chorus? No intro, just straight into it. I tried a few things, but it always wants to do something low-key to start.

[remote ivy]

did you try have it stylus/magic wand button transcribe what you want into prompt (the circle orange button that is visible once you enter text)?

[solar nexus]

I did, yes. I also tried some reddit searches, and ChatGPT searches. There were variations on using brackets to put the prompt into the lyrics section and instructions in the styles as well, but no combination came close.

[remote ivy]

which genre is that typical for?

[solar nexus]

I would say it's not typical of any genre, but it's been done for a lot of songs. You Give Love A Bad Name, by Bon Jovi. Day Dream Believer by the Monkees, etc.

This particular song is kind of in the vein of mid 90s alternative punk-leaning rock.

[remote ivy]

someone that knows will respond

[solar nexus]

Shout by Tears for Fears is another good example.

[quiet solar]

You could always let Suno do it's thing and then cut off everything before the first chorus in the Editor/Studio, so that it opens with the Chorus.

[solar nexus]

That would cut off my first verse. I suppose I could write the first verse twice maybe?

[quiet solar]

Exactly. Or put in an instrumental intro, or whatever you feel like - you already know it's destined for the garbage bin, so you can put whatever you like in there.

[solar nexus]

Worth a shot, thanks.

[lone narwhal]

Wanna create a deep, raspy breathing voice like Louis Armstrong. Unfortunately Suno ignores my input almost completely.
How can I improve the inpu?
Vocals: male, baritone vocal range, rough, low-mid register, raspy timbre, vocal cord tension, breath control, intentional vocal strain, pressed phonation, microtonal variations, subtle pitch bends, slides between notes, syncopated, with unexpected accent placements and rhythmic breaks, conversational, sustain rough, powerful notes while maintaining remarkable clarity

[earnest agate]

Something like this?
https://suno.com/song/3006068b-03bd-45d9-870f-2fb70adb657f

[unborn star]

how about a description instead of a list of words?

[polar spade]

Regardless of technique found, adding "ensemble" to prompting has helped to at least encourage more than two voices.

Also, manipulating a few starter extensions to have a different vocal quality each, so Suno might create a trio effect in a following stanza could be tried, then deleting the previous ones used as merely "starter" lyrics. Something I havent tested for trio.

[polar spade]

So, I didnt get this by prompting really. Made this in the past 30 minutes by some prompts, a couple of covers, and making a persona.

But one trick was to turn weirdness up to get that warble and harsh cigarette whispy rasp in there.

I might have even tried adding "satanical" into the prompts to see if it would get growlier. But I stopped so I could eat breakfast. Lol.

https://suno.com/s/1ml4WE7ZNcElQRJx

[lone narwhal]

Appreciated that! Thanks! Messing around with 4.5all 😉 therefore I need to rework some prompts.

[dusty raven]

Help please, does anyone know how to get a persona to cover a song without extending the time? I’ve gotten some pretty good songs made but they all have different voices and I am trying to get the same singer on all of them. Only problem is it keeps extending the song by 10-45 seconds and I don’t want that.

[polar spade]

It may be extra work, but old school way is to extend at a point around the spot you wish to end it. If Pro, then you can also use 3.0 , if needed as a better chance to just get a quick end. But 5.0 seems to follow through as well.

[dusty raven]

Thanks Dasvon! I am new to all of this. Do you recommend a tutorial or YouTube video that will show me how to do this?

[polar spade]

Are you Pro or Basic? Basic is limited to 4.5all, but Pro can use 3.0 if needed.

[dusty raven]

I am pro.

[polar spade]

You would just click edit/extend. Choose 5.0 is fine. But back in 2024, it was 3.0 that responded well to prompts like [end] or [finale]. (basic users dont have this advantage anymore).

Set the timemark to extend from as close to the end as possible, but also listen carefully and find a turn around or a 7th chord in the melody that feels like an ending point and extend from there with something like:

[Pause]

[end]

.... as your prompt in the extend lyrics box

Fortunately, with 5.0 its has a glitch that it can stick and make 2 second extensions. But thats perfect for you. So cross fingers 5.0 glitches. 😁 seriously, try it [5.0] and see. Otherwise, use 3.0 (or 3.5) if the extensions are going long.

Another hack is to just crop it at a nice point that could be an ending anyway, then go into edit and do a fast and short fade at that point to remove any extraneous hiss or noise after that point.

[dusty raven]

Great thanks for the help!

[livid trout]

I wonder how you reliably include a male rapper that doesn't have a double, and raps at up tempo speed and not 2 x slower ? Within a song that has a female chorus ? (for, like a typical 90s song)

[quiet solar]

@livid trout All of those can be prompted for, but it'll depend a bit on what genre you're after for how easy it is to influence. And you'll run into the typical headaches about prompting for duets and who actually does what.

A prompt I used for something along those lines (a fast rapped male voice, paired up with a female who sang the lines rhythmically rather than rapping) in the rapcore/nu metal genre was this:

Style: An upbeat, nu metal, alternative rock, rapcore track with fast-paced rapped lyrics over heavy synth, guitars and bass laying down a strong rhythm, The song features two nu metal singers, a male tenor with a smooth voice and a high energy female Alto

Lyric prompting:

[Verse 1 – Fast rapped, Male Tenor]
[Verse 2 - Melodic, Female Alto]

[Chorus - Duet, Melodic Hook]
[Haunting singing]
Lyric lines for the two singers
[Fast rapped, Male Tenor]
Single line for the male rapped voice
[Fast rapped, Female Alto]
Single line for the female singer

---

I remember it took a few attempts to get it sounding the way I had in mind, which was something loosely reminiscent of 90s/00s Linkin Park style tracks. The final version had lines that were 10-11 syllables long at 105bpm tempo for the rapped sections, and were notably shorter (c.8 syllables) when full singing in the chorus.

Style prompt was deliberately kept quite focused on a small-ish number of things that I wanted Suno to get right, with tweaks to that done in separate iterations using Cover once it had nailed the voices and general 'feel' of the song. Original began life in v4.5 before being remastered in v5.

Edit: If it helps, when it was re-uploaded to Suno after some tweaking in Audacity, Suno added this description which might also be useful to consider in the style prompt:

The song features a male vocalist with a raw, aggressive delivery, often employing a spoken-word style in the verses and a more melodic, female singer in the choruses

[whole fiber]

Life changer 💯

[fresh gull]

Had I not be on punishment I woulda Showed you I figured that out week ago to help with link but i missed window

weeeks

nm gn

I had to i was doing Epic Rap BAttles

Now I stick to my DUality I made crossfading male to female same point of view to sides of my coin

got it doing verse verse well

verse per verse yup fingers dn

[livid trout]

When I tried, I described it something like
[Verse 1 - solo male fast and energetic rapping]
[chorus - female singing]

Because the guide audio I have uploaded for melody, and sung with synthesiser V, it sort of always tries to default for the female persona in suno, and she also does the verses.. which is quite annoying , after 10 tries, sometimes it does the male rap, but half tempo, and that just screws up the song structure....

[fresh gull]

(M) & (F) is all needed

not (Male) (Female)

Here is song proof

https://suno.com/s/OF80mI4WSV9Xl1Rj

you have the M and F in stlye

to set them

i showed test file to prove belong here / because the Song hits so mmuch harder 🙂

[quiet solar]

I feel your pain, but IMO you're on the right path.

Some general points:

• It helps to prime anything you want to call for in the lyrics tags (e.g. male fast and energetic rapping) by having it in the overall style as well
• Prompting for dual singers can be tricky anyway. Prompting for dual singers with specific characteristics, and then coming in at the correct times, is exponentially more difficult

The more you want to specify characteristics of the singers, or control the execution, the more you really have to rely on editing techniques in combination with prompting and/or audio uploads. The advantage of this is you can break down the problems to deal with them separately.

In the song I was working on, the goal of the very original version was to get the male rapper locked in. If you're struggling with that, uploading audio of you doing the rapped section to give a guideline for the cadence can help a lot. Once you've got a version in which they're sorted, you can then use that as a starting point to cover/extend and focus on introducing a female voice doing the singing you want.

The crucial step was to get both voices as I wanted them in the same song (which is tough if you're just relying on prompting and luck). At that point, using Extend from that version means you can reference those two singers, so you can then try something like re-doing the entire song in an Extension to get a first pass of it with both singers. Who sings what will be all over the place even if your lyric prompts are perfect, but you can fix that separately in editing.

The best/easiest way to go about fixing it will vary depending on how many times they switch over, and whether you need them to harmonise or not. I set out some paths for doing this separately in this post here, which you might find useful: #📚┃suno-school message

[slate pebble]

Hi guys, what's the best way to have suno stay faithful to the melody of the entire song when I cover? I'm having issues about it always changing the timing, and chords of the bridge. Ty!

[quiet solar]

Cheers, Lietolis.

What you've done with the voices is some really clever work with the style prompt, for certain!

[quiet solar]

If you're on Pro/Premier, when in Create turn up the Audio Influence slider under Advanced Options (e.g. try somewhere between 60% - 85%) and turn down the Weirdness slider to 15%, or possibly lower.

High audio influence should help keep the track faithful to the melody coming over - the downside if you set it too high is that it'll start locking in too much and not making the changes you're trying to as part of the cover. You may need to experiment a bit to find the sweet spot for your song.

Weirdness is a general "how much random variation should I deliberately include as part of the songmaking process", so in this case you want that on the low end. Be a little wary of turning it all the way to 0% though, as sometimes taking any of the sliders to the absolute extremes can have odd side-effects.

[slate pebble]

WIll try this. Thank you @quiet solar !

[remote ivy]

https://suno.com/s/OblzXL0DVwbgSvLz

[sleek stump]

This shouldt be an issue, I have been using Cover to fix issues that end up being huge headaches using the Editors. When I do covers and I want to maintain the majority of the arrangement and melody, I make sure that the styles window is "blank/empty", and then set sliders to Weirdness:10 / Style:50 (basically off) / Audio Influence: 85-90.

I have the most success with this method. Caveat: Its Suno, it will 99.9% of teh time change something somewhere, so it may take a few cover gens to get close enough that it doesnt matter.

[fresh gull]

Pass your test?

[fresh gull]

for real?

i been grind trackd

lol grinding tracks

that mght sund cool too hmm

[solar nexus]

Out of curiosity, is there a static source of information to be able to study the framework for how this works?

Not the nuts and bolts, just input methods and language usages.

[remote ivy]

i stumbled upon a github account yesterday....

[remote ivy]

github/suni-ai ...something like that if i recall....but its probably old stuff they may no longer use

[neat nexus]

Stuff does exist but the nature of how AI works means it's all suggestion and not exact methodology.

[solar nexus]

I often find that I'm trying to do something very specific, and I find 10 different answers on how to do it, and they contradict each other, and none of them work, lol.

[neat nexus]

Suno is not a wordprocessor for sound. It's more like making a soup. Or mixing paints to get other colors to paint with.

[remote ivy]

look up bark

#💬┃general-chat message

[solar nexus]

Oh, that's interesting.

[remote ivy]

no my 1550 credit sink hole turned to a 1570 credit sinkhole....no tagging or mention of duet or anything works...pretty sure my lyrics are just too confusing for the AI...it gets 90% there then puts in wrong gender despite the gender tagging

[neat nexus]

While I joined early, I wasn't there for Bark. I came in with Chirp v1.

That said, that was Bark. This is Suno.

[sweet ivy]

Looking for any assistance on cover an original, song. Uploaded a fairly basic audio - drums, bass, guitar, vocal, no effects a tiny bit of reverb on the vocal. Was that too much, the Suno cover was not useable - jumped back a few times?

[fresh gull]

just 4 voices

i did

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

i took the green pill

[quiet solar]

As Hepcat says, there's a lot of well-intentioned AI suggestions floating around. Most are AI guesswork that isn't necessarily validated by direct experience. There's no official reference material to point to.

A common factor to a lot of the AI-generated 'guides' is that they confidently suggest [x] will work and imply that it will always work - in practice when it comes to Suno prompting it's usually more accurate to say that there are multiple ways you can attempt something that 'could' work, and some of those ways may be more reliable than others.

If there's anything specific you're trying to do and struggling with, feel free to post it up in here and we'll help if we can. 😉

[quiet solar]

There's no reliable way to gender tag for switching singers in Suno.

Prompts in the lyrics box are better thought of as being suggestions rather than commands, so Suno will sometimes skip past them. You can also find yourself fighting against the way that duets are commonly handled in the genre itself. You can get close using good prompting, but in effect you're up against a random number generator as to whether it plays out exactly as you requested or not.

There are routes and workarounds for using Suno to create duets and control who ultimately sings what, but the most effective ones rely on a combination of prompting and editing. If you want to know more, I put together a summary here: #📚┃suno-school message

[quiet solar]

I've found that jumping effect is more likely in v5 than in v4.5, so rolling back a model might be worth a try if you haven't already. It also seems to be slightly more common when you have the audio influence set very high, so turning that down slightly may also help.

[fresh gull]

https://suno.com/s/zgOpgAaUjSwczQr4
passed imo

4 voince each 1 live in each corner

not 1 spanniung

im done edge testing tonight

this just made mind explode in ideas

[honest fjord]

Im trying to make a duet
Queen [chorus]
Prince [verse]
Narrators [pre-chorus] im able to fiddle with queen and prince but its the pre-chorus im having trouble with. Its meant to be an opera of people narrating the scene but I can't seem to get the opera singers to work properly

[quiet solar]

You can try using things like choir and ensemble - probably worth adding them in both style and lyric prompts, e.g.

Style: The pre-chorus is performed by an ensemble of vocalists.

Lyrics:
[Pre-chorus - Choir, Ensemble]

[honest fjord]

Thanks ill try that

[honest fjord]

I tried that but it doesn't change anything about the voice sadly

[restive pebble]

@honest fjord
I noticed more often than not Suno prefers all caps for the meta tags when trying the multiple vocal performers so try making them all caps.

In the styles try specifying the vocals:
Vocalist 1 delivers a
Vocalist 2 delivers a
Vocalist 3 is only the Pre-chorus delivered by an ensemble of opera singers or something similar.

Then try is to force it by
[PRE-CHORUS - VOCALS: CHIOR, ENSEMBLE]
or
[PRE-CHORUS -VOCALS 3]

[sweet ivy]

@quiet solar Thanks. I think maybe the problem is that the track I did the vocal on, I put created in BIAB and there is a lot of processing going on in the back ground. I was asking Chat GPT & they said BIAB can definitely confuse Suno, I should do a simpler guitar-vocal. Wonder if anyone else tried Band In A Box for the Mp3 audio track ?

Closing Discord now

[tranquil holly]

I wonder if in the style prompt you can say "Verse Male" is male singer and "Verse Female" is female singer, and then use that as a meta tag....

brb

nope - doesn't work reliably. at this point you just embrace the chaos and either keep rolling the dice or your song works with the alternating

[honest fjord]

That adds the ensemble but its always in the chorus and never in the pre-chorus

[restive pebble]

Hmm. So using that inside the Pre chorus tag isn't working... Share the song so I can take a look. Maybe I'll see something to help influence suno

[honest fjord]

===========
Its a long song

Perhaps the problem is the length?

[sleek stump]

wholy wall... thats probably a little much to post here in its entirety

[honest fjord]

Sorry...

[restive pebble]

It is long... I must ask if you have spaces between the verses. Suno likes neatness.

In your styles let's try assigning specific vocals to each vocalist.
Vocalist 1, Vocalist 2 and Vocalist 3

Vocalist 3 is a choir ensemble with multiple voices delivering a harmonic melody. describe whatever else etc...

then use this for the Pre
[PRE-CHORUS | VOCALIST 3]

Test it out and play around with the way you describe it in the styles and the wa you add it into the pre chorus.

you can also separate the vocals like this:

[PRE-CHORUS]
[Vocals: Vocalist 3]
[Vocal Style: Chior, Ensemble]

[sleek stump]

A suggestion, if yo uhave already tried generating that, just post a link to the Suno song, we can see all ot it without a mile long post.

[restive pebble]

no worries at all. I would edit and cut it simply to keep the post size down but I don't mind it's length

[honest fjord]

https://suno.com/s/aNYsAXsOXOUzeCeT

[sleek stump]

FYI, A Mod might nuke that post

[restive pebble]

that's why I said edit and cut

[honest fjord]

why I did the text first

[restive pebble]

Because it takes up the entire screen. it is considered bad form

[honest fjord]

Aaaaah

[sleek stump]

Thanks

[honest fjord]

Yup

[restive pebble]

Suno pretty much does what you want but you must figure out how suno needs to be told what you want///

[honest fjord]

Ya its weird

[restive pebble]

Its a puzzle that only tinkering with gets results sometimes.

[honest fjord]

Yup

[restive pebble]

I got songs with duets easily enough with prompting but I've never tried an ensemble as you're attempting.

[honest fjord]

I like duets but suno hates it

[restive pebble]

Duets are not difficult

I believe I hav a solid workflow for them

[honest fjord]

Not difficult just tedious to tinker

[sleek stump]

Duets are still the single most hardest thing to do in Suno, your frustration is everyones frustration.

[honest fjord]

Plus the ensemble makes it a whole nother level

[restive pebble]

I gave david a post on how I make it...

let me see if I can find it

[honest fjord]

Nice

[pastel harness]

hello, everyone, I found an article comparing Amazon Music and Spotify, and I think it's excellent, so I'm sharing it with you. https://www.tunepat.fr/guide/amazon-music-vs-spotify.html

[sleek stump]

Looking at the track, there are a couple flags that yell out at me immediately....

[honest fjord]

Oh?

[restive pebble]

spaces between your verses is key

[honest fjord]

I hate woodhouse

[sleek stump]

The prompt itself, and the lack of proper meta tag spacing in the lyrics... lemme think abotu the styles prompt....

First thing, drop and BPM commands, they are ignored.

[restive pebble]

Would you mind if I reworked the lyric tags and made it more Suno friendly then returned it to you?

[restive pebble]

I didn't know that thx for the scoop!

[honest fjord]

Please

[restive pebble]

You can use it as a template for your other songs too

[spice whale]

Hm maybe here, any style idea for explaining generall relativity?

[restive pebble]

be back in a bit

got some formatting to do

[honest fjord]

Big bang theory

[spice whale]

Thats a style now? XD

[sleek stump]

Ok, lemme see here... the Genre should be described first and foremost, in md section you should describe the vocalists and how they are preferred, and then usually the last part of your prompt should describe how the arrangement should transition and close. It also looks like its formatted for v4 or older, since you are using 4.5+ (or v5) you can take advantage of more real world language. Give me a couple minutes to rearrange stuff and then modernize it...

[honest fjord]

I read 4.5 might have better functioning since its not in beta

[restive pebble]

I cleaned up the parenthesis in the lyrics already Ginger. and Tried to separate the verses etc. Sent in a DM

[sleek stump]

OK, I did a few things... Ill post the prompt I started with to get to the magic prompt...

Start:
A musical theater piece with influences of orchestral marches, Broadway musicals, Vaudville, and overly cinematic, Orchestral March, Broadway Musical, Vaudeville, Minor Key Brass, Timpani, Staccato Strings, Harpsichord, sung by an ensemble of 3 vocalist consisting of a dominant overbearing female lead sining verse and chorus, a demure Female responder, and an ensemble of opera singers delivering the pre-chorus asa a choir. Cinematic, High-Octane, Climax closes the song through the outro.

Magic Wand:
This musical theater piece opens with bold minor key brass fanfares and dramatic timpani rolls, joined by staccato string accents. Harpsichord adds a quirky underpinning. The dominant female leads soaring verses and explosive choruses, answered delicately by the demure female, while operatic choir voices elevate the pre-chorus. All build toward an exuberant, orchestral march-driven, cinematic climax with full ensemble and blazing brass, closing with a thrilling, high-octane outro.

Believe it or not, 4.5+ and v5 work better with the more natural language, and tends to not do well with the 3way v4 and older had to be prompted with simplicity. Give the magic prompt a try and see if it gets you closer to what yo uare trying to achieve...

Another slightly different prompt that may work... A musical theater piece driven by bold minor key brass, pounding timpani, and razor-sharp staccato strings, with harpsichord flourishes. Verses and choruses led by an assertive female vocalist, countered by a gentle female, while a dramatic choir enters for soaring pre-choruses. Climaxes with a fervent, cinematic outro exploding in orchestral and vocal intensity.

There are no wrong ways to do things, just ways that have varying rates of success.

These are all things to try, but remember, Suno is Suno, and it does what Suno wants most of the time 😛

[restive pebble]

I've given the lyrics section a go over and tried my best to provide a foundation for the desired output. Let me know if it works

[quiet solar]

If your issue is random skipping (e.g. The performance just flips out for no apparent reason, then jumps to a different part of the track slightly ahead or behind that point) then it's not specific to BIAB, as I get it too. I've linked to an example track where the cover went haywire then jump-corrected at around 55 seconds in - is this the same sort of thing you're experiencing? https://suno.com/s/zVp2ovvAtY2T6jpp

Either way, it's not going to do you any harm to re-record a dry version of the vocals, since Suno tends to handle 'cleaner' audio inputs more accurately in any case.

[quiet solar]

I see Ginger and the seer have both jumped on this in depth since your response. 🙂

I looked back over the choir/ensemble songs I've made to see if there's anything else I can pitch in on a general level.

• "chanted" and "chanting" were other phrases I used, both within the style prompt (e.g. Vocals are softly chanted) and tagged in the lyrics, e.g. [Verse 1 - Soft background chanting], [Chorus - louder choral chanting] to prompt the choir coming in.

• If you want two different-sounding same-gender vocalists, you may want to specify two different vocal ranges in the style prompt (e.g. Vocals are sung by an angsty male tenor and a smooth male baritone). Otherwise I find Suno often tends to just stick with the same voice.

• Everyone has their own go-to method for calling the singers in the lyric tags, but if you take that route, I've personally found re-referencing the gender and vocal range, e.g. [Male Tenor] or [Female Alto] to be fairly effective.

• Be aware you're trying to combine multiple challenging things in their own right in the same song; same gender duet with distinct voices, lead vocalist with a separate choir, three separate sets of vocals/vocalists and, on top, wanting different vocalists to perform specific sections of the song.

• Expect this to take a lot of failed attempts. In practical terms, I suspect it may only be possible to get the end result you're looking for by using workarounds and audio editing. If you're comfortable with doing audio editing, I can give you a workflow for how I've done things along these lines, but would want to warn in advance that it's not a quick or easy process.

[oak maple]

Is there a way to determine the length of a part (like intro should be 8 bars long)?

[restive pebble]

Duets are difficult but using the proper prompting makes a world of difference.
I took your song and edited the meta tags and styles to test.
I could not force Suno to have them sing together but the rest of the lines Suno keeps the assigned vocalist consistently about 90% of the time
https://suno.com/s/5GVZUuG74J2cpxrl

feel free to copy the styles and meta for your use

[restive pebble]

You mean like in telling suno this verse is (x) bars long?
I have seen some reference bars but have never attempted to use them in the song.

I believe Suno bases time on the amount of text sung, but could be mistaken

[oak maple]

Yes, that is what I mean.

But this should also work for non-vocal parts

[restive pebble]

You could try these:
[Extended Bridge]

[Double-time Instrumental Break]

[8-bar Instrumental Interlude]

Suno is experimental not precise

or as precise as we'd all like

[oak maple]

Okay, thanks. I'll try it out

[restive pebble]

If those don't work dig deeply into the dirt to find other options. Sometimes it takes more trials than we want but eventually Suno surprises us

[8-bar build-up is another thing to try

[fresh gull]

guys i just release a epic 3 singer song

it works

ye

[honest fjord]

@quiet solar @restive pebble @sleek stump
Thank you all for the help I tinkered a bit and the example leer sent helped get the choir out though not in the pre-chorus sadly.
Seems choirs never appear in prechorus so I switched it to [chorus+choir] that has a higher success of getting the results but I fell asleep before I could tinker further.
Ginger ill try your prompt today see if that changes anything.
Kanthric
Ill try your method and ill ask
What tags for style and lyrics work best for same gender vocalists?
Ya I've noticed that this us far more difficult than my other duets that came out good enough. I expect this to take a few days and A LOT of 🎶 points

[honest fjord]

https://suno.com/s/p3KPGyRh8yUYdAIx

I added time codes to the verses that worked really well suno never went past the designated time codes

Buuuuut I will warn you the song is very graphic best to just pause it and look at the lyrics as an examplr

I wanted to make an snime opening that cut out all the unnecessary fluff of the original song I did, so restricting the time was paramount

I used my gemini bot to change the lyrics and has been super effective with all my songs

[restive pebble]

Glad you're makin progress. Suno is a creators joy and bane as it take much tinkering to get wha we hear in our heads.
For making multiple male or female vocals inside a single song, treat it the same as the call and response duet, just use male1 and male2 or something to delineate the different performers.
It may work, it may not but I'd guess Suno is smart enough to at least give you something close.

[polar spade]

From whats seen, as well as getting some AI advice, trying to directly setting bar and measurements lengths still do not bode well on their own. Suno needs to calculate everything at once, spacing, promots, tempo, to determine its own length withing the 8 minute limit.

Unlike when we write a piece, we write everything, STARTING with an 8 measure intro, then splice as needed to cut or lengthen what we get.

8 bars is not unusual, but maybe less frequent on average, so the more we can hint and nudge Suno to associate an 8 bar intro, the better chance we have of getting one. You can certainly try [ eight bar intro] but most as most intros are 2 to 4 bars, you might get what you need by just prompting

[Intro]

[Interlude]

As on average, ive timed [interlude] on my songs to be about 20 seconds in length on a 80 to 120 bpm song.

[honest fjord]

My biggest gripe right now is getting the choir to work. I know how to adjust the pitch of the vocalists but im focusing on the choir rn since that's the main issue at hand

WHY

I hate woodhouse

[restive pebble]

@honest fjord Simultaneous vocals are a challenge. I've not played heavily into it but would assume that Suno has difficulty being prompted as oposed to it being an ad-lib type of effect. That may be why Suno has trouble... usually multiple people singing are the background track. Just a guess there...

[polar spade]

Type. Highlight. Copy. Then, submit. This is the way. 😅 Woodhouse zaps me as well.

[oak maple]

I would specify everything as detailed as possible in the style section, but the 1000 character cap prevents me from doing it 😅

[restive pebble]

@polar spade What do you know about using multiple vocals simultaneously in Suno?

[restive pebble]

ask AI to condense it but keep the style as much as possible

[honest fjord]

I have a gemini bot "suno architect" you can use it for time codes management and the meta tags that may be too complicated (for me) and create new tracks within a theme of an album its really useful but its only in beta v4
I cant share doc links here so you mind if I dm you the doc?

[polar spade]

I think it was mentioned to throw ensemble in the prompt mix to get some influence, especially in the choruses. You can also put "isacathymia" (check my spelling) for a unique sould of backing choirs as well -- a call and response prompt , basically African chanting. Place the prompt towards the end of a few prompts to induce only a hint of the effect.

Chant , therefore is also a good prompt.

This would be one case where I would focus on extending for each stanza/section to make sure i have the choir or ensemble feel i want before going to the next section.

[restive pebble]

Thx. it was mentioned before. Spent some time with @M last night but still the choir is a challenge. Expected but a fun challenge to try to fix for M

[oak maple]

I have already done that, but due to the character limit, it is still not possible to provide as much information as possible to achieve the desired result.

[honest fjord]

Theres a character limit?

[restive pebble]

1k character limit in the styles section...

[honest fjord]

Actually that might be the issue

[polar spade]

Not to forget any call and response or nudging Suno to have some simultaneous chanting or backup may be emitted by putting those lines in parentheses. Crowding them together or having the response typed in between lines is a ggood way to trigger suno to just combine them as one.

Sample: I want to hold (hold my hand) your hand.

[honest fjord]

Too much to keep track of

[oak maple]

Do you mean a Gem?

[restive pebble]

Could you convert it to meta tag removing the verbosity of the description.
I did that a few times.

[honest fjord]

I didn't make it into a gem. I should though good idea

https://gemini.google.com/gem/1oXu7_PdIgI1ctfoSOJd8pP5G1ou1_aZa?usp=sharing

Made it into a gem!
That should be it

Not sure though

It is

[restive pebble]

Hello Gemini. I see you

[honest fjord]

Hi

Wanna try my suno architect? @restive pebble

[polar spade]

First, @honest fjord , do just a couple of outputs with just "ensemble" as a the single prompt. Throw some ReMi lyrics in and click Create. Get a feel of the difference between ensemble and choir. Choir has a unison controled sound, whereas ensemble is more theatrical, slight off timing and you can hear varied voices. In a longer piece, ensemble may even have a different voicing/solo vocal on each verse which can sound cool on a gospel like piece.

Oh speaking of which, [gospel] is between ensemble and choir, but does often go into that 7th and suspended 4th and prolonged notes. So it may not sound as "major chord" or "minor" influence as youd want, but a bit too inspiring.

[honest fjord]

Ill try thise thanks

[restive pebble]

Good info. I'll try to keep this handy

[quiet solar]

I don't know that anyone can give an authoritative answer on the "best" way to prompt for two different same gender vocalists.

What I've found worked in the past was to put the vocalist info early on in the style prompt, and keep the style prompt shorter rather than longer to help keep focus. As per my last post, prompts I've used in style were along the lines of: "This song is performed by two vocalists; an angsty male tenor and a smooth male baritone". I can't give you an exact suggested wording without knowing what you're after, but the key factors are:

• Mention there are two vocalists. Establish the key differences between them. Make those differences even more distinctive (including in vocal range) if you find you keep getting a solo vocalist

In the lyrics box, I usually tag by gender and/or vocal range, e.g. [Male Tenor]. If they're doing a whole section, I'll put that in the section header, e.g. [Verse 1 - Male Tenor, rapped vocals]

If they're coming in partway through a section, I'd put [Male Tenor] at the start of the line of lyrics they're supposed to sing, and mark the next changeover accordingly, e.g.

[Verse 1]
[Male Tenor] Male tenor sings this lyric line
And this one
[Female Alto] Female alto sings this line
[Male Tenor] And back to the male again

As you're not gender switching, I'd probably go with just the vocal range, ie. [Tenor]. Other people have found alternative ways of naming the singers (whether male1 and male2, or Joe and Geoff, or whatever you like) also works - again, just make sure you include what you're using for referencing in the style prompt somewhere; preferably very early on.

[wise kindle]

<@&1082792303976001547>

[honest fjord]

Im looking for a dominant leader female vocalist who is usurping the throne
Meak soft speaking female voice as the prince (this is in the middle of the story arc creating an album narrative)

[quiet solar]

Those sound like perfectly reasonable descriptions. You might want to try using a more musical term like 'powerful' in place of 'dominant', and add any general tonal instructions like angry, rebellious, haunting, emotional, etc.

[honest fjord]

Oh cool

And the soft speaking?

[quiet solar]

Soft should work in and of itself, presuming you want them to sing the line rather than speak it.

[honest fjord]

Yup

Sounds good ill get to work.

And I do wanna see what people think of my gem "suno architect" wanna bang out any kinks before I get it out of beta

[restive pebble]

Excellent brother

[quiet solar]

Best of luck. Again, expect this will take a lot of trial and error - I had 60+ generated songs before I had the pieces I needed to edit together the three voice (not including a choir) song I put together. The best prompt-alone version got a reasonable amount of the way there, but editing was necessary to ensure the right people sang the right parts.

[restive pebble]

Sometimes that is the only option, several pieces stitched together

[polar spade]

Give you an example of accompaniment and ensemble in prompts. Gets an ensemble chorus and varied vocal switches. Just ran a remi lyrics and stuck ensemble in the prompts. Along with opera, gospel, accompaniment, unison,and a couple others, these can be used in rock, metal, glitter, and many other genres to give a nice vibe. @honest fjord

https://suno.com/s/wyS9KnK6VmlQQw8w

@honest fjord , @restive pebble
I just added "isicathamiya" to the end of my prompts and got a more choral sound throughout, just to show its pretty strong effect. I think @honest fjord want more choral response in the chorus.

https://suno.com/s/mmFVKkwFZQHDiWYM

[honest fjord]

Is that in style?

[polar spade]

Yes. Check my prompts.

[honest fjord]

"isicathamiya"

[polar spade]

Im sure it can be manipulated further in lyrics. But ive done nothing more than add to style prompts.

From a personal side, I like musicals and write in this style often to get random multiple vocals and choral responses. Then splice as needed and make personas as well to influence when creating new songs with extensions on the segments for more control.

[restive pebble]

@polar spade Thank you for your help! M is further along than before and now has a solid direction to pursue

[polar spade]

Overload? Oh, sorry. 🙂

[restive pebble]

Not at all. It was something I could not figure out to help M with. It was what I think M is shooting for... But let M make that claim

[sleek stump]

The battle rages ever onward 😄

[polar spade]

😆 I have a Bohemian way of approaching most things. So I can overload with a shotgun approach at things.

[restive pebble]

There is never too much information. (well in this case) lol

[honest fjord]

@restive pebble @sleek stump @polar spade
Ive gotten the choir to work i had to remove ensemble. It has choir in areas I dont want but im able to look over it cuz I got what I wanted where it should be.
https://suno.com/s/qqhuQyCIlG9rKxDa

[sleek stump]

Super awesome! What a journey thats been 😄

[honest fjord]

Yup took a while but I finally got the choir to work thank you all

[restive pebble]

was it worth the effort to you?
I find the song amazing and powerful

[honest fjord]

It is!

[restive pebble]

Then you've succeeded and it truly shows! thanks for letting us help you get there and for sharing it with us

[honest fjord]

Anytime and once tge album is done ill share too

Song before it
https://suno.com/s/uJAvf622cB1tm2it

[polar spade]

Again. Use in extension. Piecemeal it. One section at a time. Dont expect it to work perfectly in creating a song in one shot, one generation. But yes, experiment to see how "powerful" it can be. Glad it has some effect that may be promising.

[polar spade]

Excellent! 🤘 Glad it got some momentum to sound the way you want.

[sweet ivy]

"Maintain the majority of the arrangement & melody", how full is the demo - multi tracks - drums, piano, bass, guitar, vocal. New to Suno, trying to figure out what the demo should be for covering a song.

[polar spade]

@onyx siren Sorry. There are no student discounts or programs planned or available at this time.

[sweet ivy]

Yes, not as bad as that break at 55 sec on your track, but same thing, a jump back. Do you know of others using BIAB to create the audio portion of the Cover process. Maybe I remix, take out drums and bass, just leave guitar and vocal ?

[sweet ivy]

Been following along, as a newbie to Suno. What a learning experience. I still can't believe this was created by Suno.

[honest fjord]

It takes a while to get used to but suno is very robust with variety and quality.

Thought it IS a steep learning curve

[honest fjord]

You can use my ai assistant to help learn it
https://gemini.google.com/gem/1oXu7_PdIgI1ctfoSOJd8pP5G1ou1_aZa?usp=sharing

[sweet ivy]

Thanks. This Gemini is more aligned to assist with Suno, than say Chat GPT which I've been using

[honest fjord]

Yup you cant make a profile or gem on chat gpt but you cant with gemini its basicly AI programming

[honest fjord]

How are you liking it so far?

[sweet ivy]

Forgive my ignorance, but these Broadway style songs you created Usurper Queen & Wonderland Desires, are they based on an existing musical about Alice in Wonderland or did you write those lyrics?

[honest fjord]

They are based on Disney films alive in wonderland 1951
Lion king 2003

[sweet ivy]

Bit frustrated, as having trouble understanding the Cover option, but working through it

[honest fjord]

Oh? I could help

Did you add any tracks to the album yet?

[sweet ivy]

Actually have to leave to an appt. be back later. buy the album yet? So these songs are based on a actual musical & you are just doing an arrangement

[honest fjord]

Yes. For the album I have 2 different versions of songs the explicit and direct songs are based on reality
Tracks 1, 3, 5 etc.

The whimsical tracks are covers of those songs with a Disney theme attached to them to have a feel of familiarity and whimsy.
Tracks 2, 4, 6 etc. They will always follow after the original.

2 is wonderland desires and 6 is Usurper queen

[fresh wave]

So I just had the idea of getting more control of vocal by taking a stem of the instrumental and re-singing the vocals on top

So you could improvise or change the timing of certain words and things basically edit the song by re-singing it yourself to make it line up better or give it more human feel

[dense nexus]

Good morning. I am an annual ProPlan subscriber. I love the Suno platform and am having great success in creating some banger music. That being said, however, this morning I was attempting to create an instrumental prelude track for the beginning of a concept album project and after 5 attempts, Suno only produces 2 second tracks. I've been able to create instrumental preludes in the past without issue. I've clearly prompted Suno with details for this creation including duration length and specific instrumentation. What am I missing?

[frozen dagger]

hard to really guess with only this indirect info. suggest double checking everything. which i assume you have, but maybe copy paste it to a chatbot for errors to keep safe form other eyes.
double check sliders, the exted stuff at top. then if setings are all good and no grammar erros, either its something maybe in the syntax of your prompts? i.e. [Outro] tag at the top instead.. even then be weird in most cases.
not normal tho. would bug report, triple-check my prompts for errors, wait 15 mins and try again. maybe test a totally different prompt while i wait as well. but if continues look at maybe using https://discord.com/channels/1069381916492562582/1372994553233276970

[dense nexus]

Thanks for the feedback and suggestions. I have checked (re-checked) everything in the prompt for possible word restrictions or ineligible syntax issues....didn't see anything there. I usually use an [Intro]/[Fade out] prompt in my creations but not this time since it was for a simple instrumental opening prelude track for my album.

[neat palm]

i would attempt a different browser, see if the issues still persist, i wonder if its a browser bug

[dense nexus]

if you are responding to my post, I'm not sure that the browser platform has anything to do with how the prompts are analyzed by Suno. I've been using MS Edge from day one without issue.

[neat palm]

Yeah, i wouldnt think so either but maybe best to try anything

[frozen dagger]

in my experience... when making instrumentals and it give very, very short outputs. either I didn't give enough (nothing) in lyric, or i gave it too much.
I'm making instrumentals right now. And I'm actually using @sleek stump suggestion from other day. Using ASCII art in my lyrics box for abstract interpretation. I'm not sure if I think it's interpretting any visual representation or not, but... I will say.... it seems to be doing a much better job at creating lengthy tracks with just style prompt and some copy pasted ASCII art to just "fluff fill" my lyric box.
When I use a meta tag placholder in lyrics box [Intro- soft pads, piano, reverb] ... good results, but sometimes like she has also said before and I 100% agree with... sometimes less is more.

[frozen dagger]

good suggestion as well. youre basically tryig to troubleshoot what's happening, and that is a good idea too.

[dense nexus]

Gotcha. At first I had some meta tags in the lyrics box, but when that produced the 2 sec tracks, I moved those into the style area along with the other prompts/meta tags to see if it would come up with something longer...it didn't.

[neat palm]

@dense nexus if you have the suno app on your phone is it the same issue there?

[dense nexus]

I have never used Suno on my phone app....it's too limiting and small print (I'm an old man and the eyes don't work like they used to! LOL)

[restive pebble]

I got this once about a month or so back. I cannot recall how I overcame it as I thought it was a one off, but here's a few things to try.

Log out, close the browser log back in.
Add a single period inside the lyrics section
Use [Instrumental | NO VOCALS] in the lyrics section

Maybe this will prompt Suno into shape for you.

[dense nexus]

Here are the prompts I'm using....[Intro Instrumental]
[Genre: Folk, Soft Pop, Acoustic]
[Style: Rod McKuen, intimate chamber folk, 1960s coffeehouse]
[Mood: Wistful, anticipatory, tender]
[Tone: Warm, hesitant, hopeful]
[Instruments: Fingerpicked acoustic guitar, subtle room ambience]
[Tempo: Slow, contemplative 65 BPM]
[Production: Minimal, warm analog sound, close-miked]

[neat palm]

That's ok, try using firefox just for a moment to see if you still have the problem

[restive pebble]

inside the lyrics section or styles?

[wet minnow]

if you use "no" in the style and ask yourself why you still get things... thats what the exclude style box is for.

[dense nexus]

Styles

[frozen dagger]

or "Instrumentation ONLY". I work with chatgpt for promtping and avoid using any form of negative phrasings whenever possible.
Anotehr suggestion I would throw out is trying "Through-Composed" while this implies non-repetitive strutcture, it also implies longer duration.

[restive pebble]

Take the brackets out as the styles doesn't require them.
Is anything inside the lyrics box?
if not add something. a period or maybe a specific no vocals tag as above so suno doesn't get too bold

[frozen dagger]

you don't need brackets in the style prompts either

[restive pebble]

Instrumentation only is a good one!

[frozen dagger]

brackets are only needed for prompting within lyrics section. just saying

[neat palm]

i like to use "no" in the exclude to really make the AI think "is he saying no because he wants them in? is this a double negative? should i not not include them?" i like to make sure the oceans are evaporating at an alarming rate

[dense nexus]

Good suggestion. Currently nothing in the lyrics area. Keep in mind that the issue I'm having is not one where vocals are being added or lyrics....it's that after 6 attempts (2 tracks each), the resulting tracks are always 2 seconds in length.

[restive pebble]

if you're going to say it like that make it count, [just saying] 🤣
Hope you've been well and HNY!

[restive pebble]

that is why I suggested adding something to the lyrics. Suno is just an ai and gets confused sometimes.

[dense nexus]

Interesting....I was told that the brackets "were" necessary in order to let Suno know that they were meta tags and to be treated as such. Obviously they were necessary in the lyrics area or they would be treated as lyrics and included in the vocalization.

[wet minnow]

you dont have to add lyrics in the lyrics box. some commands are fine too.

[restive pebble]

Meta tags are only required in the lyrics section

[dense nexus]

Yes, thank you....been doing that from the beginning....allows for more text to instruct Suno on exactly how you want the track to sound.

[remote ivy]

i've already tried something similar...the problem is i have alot of i's and you's in there...which confuses it, that is the crux of the problem...the song isn't the same if it's changed to he/she etc...

[frozen dagger]

got a giant octopus ASCII art thing in my lyrics box right now, with a fairly simple and basic style prompt. Geting those bangers mention in original post. 4.5-7mins average on last several made while talking here.
Suno take the wheel.

[dense nexus]

I appreciate all the feedback/suggestions. Let me go now and see how I can incorporate some of what everyone has been suggesting. I'll let you know if I get some positive results.

[restive pebble]

[ I ] for lack of another way to phrase it can't believe Suno uses that much 'digitial insight' into the use of I or you. I would believe it to be the prompting and amount of it Suno needs to follow.
Many just splice several extends together so it works as they imagine it to be.

[restive pebble]

lol I used mathematical equations and they genned some cool tunes

[restive pebble]

We are a community and will be here if you need more!

[neat palm]

Theres a song that has a very dreamcast/ps2 vibe to it that im trying to figure out how to get from Suno. If anyone can assist, that would be appreciated as i have no clue.
https://www.youtube.com/watch?v=iA4E7SNf8o8

I mostly get typical standard modernized synthesizers

[frozen dagger]

copying and pasting "Wind Power - FREE BGM DOVA-SYNDROME" into suno style prompt and clicking the magic wand button reverts it to: "Opening with airy synth pads and gentle wind sound effects, the track evolves into a minimal electronic arrangement. Crisp hi-hats and pulsing bass subtly drive the rhythm, while bell-like keyboards add brightness. The bridge introduces layered synth arpeggios, evoking uplift."

that is what i would call one layer. it evokes gamin BGM and i would take that and run with it using chatbot and/or suno magic wand to kind of extrapolate the words and phrasings that i can then interpolate into a prompt that reflected the texture and such. suno is genre oriented tho, so you have to think outside the box, while suno is stuck in the box.

[neat palm]

I'll try some of that

Thank you

I worked with someone in here and we were able to nail down some genesis sounds. Now that i have that solved im trying to solve this one. We had to call out the actual chips that were used in the genesis

working with chatgpt right now, think im heading in the right direction

[restive pebble]

I can see how that is part of Suno's process because Suno let's us use specific instruments themselves such as Greeny the famous reverse wired pickup guitar etc..

I learn so much from reading everyone's posts.

[frozen dagger]

subtext in prompts (lyrics and/or style) seems to surface in music. almost like emergent behavior or something. for instance. a song with words and phrases about "sleep" "worship" and such in similar genres could get a voice sounds kinda like vessel of sleep token. >.> or song with lyrics about corpse and metaphors that now singer sounds similar to the artist Corpse.
I am only speculating, but... common words being blacklisted when it's an artists name makes me wonder if the idea holds water. 🤔

[brave cedar]

Anyone from Suno support? My account stopped working for about a month now. 2 tickets submitted and still no luck hearing back.

[forest hemlock]

If it's just basic tags like verse/chorus/bridge it usually works without brackets in lyrics too, but I generally don't use any tags in lyrics at all (because they often lower the sound quality a bit)

[quiet solar]

"I" and "You" in the lyrics isn't a problem, you just need to tag the lyrics so that it indicates who should be singing a particular section, or lines where the singers are intended to switch over.

The underlying problem is that those tags aren't always taken into account by Suno - so the more switchovers you have, the more likely it is that at least one will be skipped any time Suno attempts the song.

[restive pebble]

Which is why many piece together songs

[pure panther]

Wut? How did I go all this time without this marvelous tool? 🫠 🤣

[frozen dagger]

Some names of artists and bands or songs won't change. But it's a very useful method of "reverse engineering" an already familiar music into words.

"Marilyn Manson" for example will not change XD

[pure panther]

It wouldn't do Pink Floyd The Wall, The Trial, but it would do The Wall, The Trial and still described it well (that was the first song that came to mind to test with 😂 )
Try the album name and the song name, leave out the artist name

[frozen dagger]

It's ok. We are all a little "oblivious" sometimes. No one is all knowing. 🙂

[pure panther]

I put in Antichrist Superstar, The Beautiful People and got:
A hard-hitting industrial metal track unfolds with aggressive, syncopated guitar riffs layered over thunderous drum patterns and pulsating electronic textures. The verses ride on a grinding bass groove, building tension with distorted effects and percussive samples, then explode into an anthemic, driving chorus featuring gang vocals and chant-like hooks. The bridge incorporates dissonant synths and heavy breakdowns, leading to a relentless outro that maintains the track’s menacing energy.

[frozen dagger]

Opened a window, now you're free.

[pure panther]

Lol

I'm almost done with my big project. I'll have to cook up something else.

[frozen dagger]

"Almost done (for now)", and that's great. Good feeling when it comes to a sense of completion.

[pure panther]

Truly. A cool thing happened too. I've endeavored to convert the gospels to song form. I didn't plan how many tracks there would be or anything, kinda flying by the seat of my pants. When I went to make the folder for the most important track in the whole project, the resurrection, it just happened to be track 77.

[frozen dagger]

While completley seperate end of the spectrum, also a good feeling in the prcoess of it too. I think I actually am enjoying that part more, personally, right now, but I still look forward to my own sense of completion.
I wanna hear yours when it's ready.

[frozen dagger]

happy accidents, unintended consequences, or like those cool synchronicities. Love those too. Tool -Lateralus. Based on Fibonacci sequence in lyrics structure. I thnk the drummer, Danny Carey, says the time signatures unintentionally aligned within the mathematical sequence as well. Very similar. Makes it feel like it was always meant to be, or just a cool coincidence. Eithe way. I totally dig it.

[pure panther]

I'll let you know when I make the play list on Suno. It will probably be at least another month since even though nearly all the tracks are done, I'm going to go back and remaster a bunch with what I've learned while doing the project 😄

[frozen dagger]

Take as much time as you feel it needs. That's what I telling myself.

Real Quick- will add. One of the BEST things about Suno... iteration, variety, being able to VERY quickly explore different sounds, vibes, styles, just.... it's so fun, so good, unlike anything else. Experimentation galore. That takes time. So does giving each track a chance. If you skip ahead, sometimes you're missing something that you may have used later. Either actually clipped and edited and recreated, or reused conceptually.
Save variety of prompts and fuse and keep changing the shapes.

[pure panther]

For sure. The song I'm on now has been scrapped and restarted a handful of times already. Currently testing a whole new method in three different styles.
I've been on it since last week

[frozen dagger]

So... maybe you drop that playlist next month and next year its remastered whole new conceptual design and themes, but same skeleton of context and message. That is absolutely magnificent imo.

Being able to extend your reach, gain more exposure, by having a style of music that someone may not have been interested in, but now... pick your flavor. I got you.

[pure panther]

I agree. Suno has got to be the best creative tool I've ever gotten my hands on. I've used video game creation engines, learned programming language, taught myself digital art and other odds and ends and Suno overshadows them all.

[restive pebble]

Can't we just make music and smile lol

I too am ajmazed at the versitality of Suno for iterations and ease of making genre shifts and fusions.

[frozen dagger]

Versatility... great word. OK. I stop brown nosing Suno and filling chat. ❤️ All.

[restive pebble]

Fascinerously another good one lol

but not on topic at all

[pure panther]

🤣

I tried to ask a question and hit create, the output is actually really good lol

[sleek stump]

When I discovered that, it was a game changer for me. I struggled with instrumentals and afterwards, they became trivial! Glad this worked for you!

[pure panther]

I'm not gunna lie, I'm actually low key mad at how good it is. Like really?! I have been working on this song for days and I ask a stupid question and hit create on accident and it's one of the best generations I've ever heard... What the actual hyek. 🫠 (for those who are curious, I asked if Suno Can you describe the vocals of a song I generated? If so, how would I go about doing that, do I have to add it to inspiration or something? - this is what it made: https://suno.com/s/OAkxscf6sIksrLqz)

[polar spade]

How do I get my over 4000 songs restored and second payment for this month reimbursed... - Jenny7577

@pearl shadow Assuming you mean when you log in now you get a new Basic account and at renewal you were still charge another month?

This indicates, as many of us have done, that you have inadvertently created a new account when trying to log back in. Sometimes this seems to be when using both the app and a desktop browser. Some might be trying to use an email once after having use a phone number or vice versa. Given that, cans you think back to what you had done differently, that might have trigger a new account?

[frozen dagger]

I'm low on creds, approaching my renew date, will definitely be exploring this more, and would recommend to others (obviously). I love the outside the box thinking either way.

[frozen dagger]

Does sound good. I'm curious remove the questions, similar result? Likely, if not maybe use the keywords form the questions, "Vocal" "song" "inspiration". It sounds good from those meta tag prompts and freedom in the style prompt for "Suno to take the wheel" like i keep saying.

[pure panther]

I think you're right. Keywords were probably gleaned and Suno had a lot of creative freedom.

[earnest agate]

Here is something for you guys to play with "Voice Profile Creater For Suno" Copy and paste into your fav LLM. i have some good results creating voices like this. remember the Voice profile Style must match the song. Layer 1 in the top half of the Lyrics field, sort tags as needed.

https://suno.com/song/c9d31c31-4414-4519-b3c9-0c40ea873e20

Should get something like this

[pure panther]

@earnest agate That's pretty cool. I never heard the term Gang vocals before Suno. This has a nice example of what they are.

[faint moss]

That is pretty amazing

[spice brook]

dm me if u wanna here my track i might get copyright for

[fathom hemlock]

Tried to extend the below one since Suno made a great job with the song generation but not the ending. I tried to extend it 3 times but was not really satisfied. Will Suno Studio a big help in this ?

https://suno.com/s/BHaz62nl9WpBu7AK

[slim heath]

I would like to use an older female voice for some of my songs, but no matter what prompt description I try, I still get a standard young vocal style. I’ve tried changing the music style with little success. I want to build a persona for later use. Do you have any advice on how to prompt for this, or do I need to upload a voice and train somehow ? Do you have any experience with elderly voices?

[fathom hemlock]

Is it a fast song as faster the song is at higher is the Voice in general. At least this is my experience

Normally Contralto should be a deeper voice for woman

[slim heath]

60-70 bpm. lullaby

[frozen dagger]

presuming you not free accoutn since making persona, you can select the gender in the advanced option. so, right away if you want female, then select female there. for elderly, or i guess older, more mature, not younger, maybe then prompt for the Contralto, or whichever musical term fits youre desired range. alternatively id try to create music similar genre to femeale singing style and selct female still in advaced and go from there

[fathom hemlock]

I love lullaby

[frozen dagger]

i had switched to using studio for extends, because extends was just not working for me, it takes a little getting used to with different format, but I am liking it better now. Quality feels better and i feel like i get more control being able to layer audio and recreate visually and everything.

[fathom hemlock]

Thy for reply yeah I am thinking also to switch the plan cause of the start to many time choir or something like this and ending not really nice

[fathom hemlock]

Webversion much better than app version still Suno most of the time at least for me femal voices to high unless I move to metal or ghotic

[frozen dagger]

in this chat thread specifically we have demonstrated the many workarounds. very clever ones i might add. studio, or another DAW really, can take a lot of the headache out of that process if one is able to use the tools. suno ui is great for any and everyone to create a song with anything, but learning and using a DAW along with Suno that's where most want to really get to when getting more serious about it. Studio still new, can be wonky, but I think it's going in the right direction for the most part. Just treat it like the Beta it is.

[fathom hemlock]

Just give the developer another year and it will improve a lot. I am happy with the pro plan ok burning a lot of credits so what and than I am thinking can I control it better with Studio even some users are saying the UI is not really good. So I am waiting for a special offer for Suno Studio since there is no try version out there

[frozen dagger]

that's valid. Suno developing fast. Good way to look at it. Why I say I think they are going in the right direction. Will see soon I'm sure when new features roll out.

[fathom hemlock]

Download the songs before the new version is coming out 🙂

[frozen dagger]

Since Suno is an online platform, I do try to keep my own gens downloaded and saved locally. That's just good practice.

[frozen dagger]

is that thing where we could earn some credits listening and giving our opinion on which is better removed?

[pure panther]

Looks like it

[restive pebble]

Why not try using tag like these and see if you get that older vocal you're looking for?

For the style section use this and add in your genre:
Elderly female vocals, weathered raspy voice, deep contralto, slow tempo, Add Genre

If you need further prompting add this to the lyric sections:
[Voice: Elderly woman, thin and cracked, heavy rasp] [Vocal Style: Lived-in, soulful, gravelly]

you could also try using lo-fi in styles if the voice still sounds too clean.

[sweet ivy]

Checking back in.. What did you upload for the audio on these songs from the Broadway musical ?

[radiant canopy]

If anyone's wondering, I found a way to make new fantasy music

[honest fjord]

I didn't upload audio i used gemini to make as close as possible to the musical ticks as possible.

[sweet ivy]

How did Suno know the melody of the done, if there was no audio ?

of the song

So these weren't covers of existing songs from a Broadway show.. you created these from scratch, lyrics maybe through that Gemini link you sent. Then Suno built the melody?

[honest fjord]

For the usurper queen I told Gemini to adjust track 5 for the whimsical version. I chose be prepared from lion king. The ai did the rest with some tweaking here and there

I have a custom rule for my album that all tracks will have a real version and whimsical version based off it.
The melody and rhythm is from the ai and the whimsical rewording is also the ai.

[sweet ivy]

I'm absolutely speechless. You mean you had a concept, worked it through Gemini to create that lyrics. Then you put that lyric into Suno, with all the prompts I saw, for orchestral/ choir prompts and Suno created that ?

[sweet ivy]

What is this track 5 you reference

[honest fjord]

What happened?

[pure panther]

Someone tried to @ everyone

[honest fjord]

Oh

[sweet ivy]

Yeah, I was reading your post & boom

[pure panther]

Mr. Woodhouse yelled at you to get off his grass

He's an Ai mod btw

[sweet ivy]

Tea, so if a post is too long the Mod steps in and deletes it ?

[pure panther]

Yeah, or 100 other unknown reasons. It's like an electric fence for dogs, the dogs don't know where it is but they still get zapped when they cross the line 😂

[honest fjord]

Im trying to find a way to share my logs so you can see the progress

[sweet ivy]

Was I right when I said - You mean you had a concept, worked it through Gemini to create that lyrics. Then you put that lyric into Suno, with all the prompts I saw, for orchestral/ choir prompts and Suno created that ?

[honest fjord]

Yup it took several revisions but everything is ai except the choir trouble I had

[sweet ivy]

Unbelievable. I read about Gemini & it's baility to create very professional lyrics

I'm curious how you kept the same melody with each revision?

[honest fjord]

Its a rule in the album.

[sweet ivy]

No I mean, Suno, how did Suno keep the same melody with each revision ?

[honest fjord]

By default suno architect has 5 rules for full development support. You can add up to 5 more but 2 is the sweet spot

[sweet ivy]

What is Suno architect?

[honest fjord]

The gemini bot i used to make the songs

I made it myself

[honest fjord]

Gemini added all the important Broadway inspirations for "be prepared" in style so suno will always know the melody

Though the new version of the song with choir does confuse it 30% of the time sooooo

[sweet ivy]

So the song Be Prepared from the Lion KIng was the guide for the melody

[honest fjord]

Yup

[sweet ivy]

Forgive my ignorance, I'm new to Suno. What I found is that when I do another version, the melody will change, so I'm trying to understand when you load the lyrics in for another version, how does Suno not change the melody

[honest fjord]

Im going to murder woodhouse

[honest fjord]

Im gonna do this in bits. The style creates the blueprint and the changes of vocalist add foundations each time so the ai will always remember

STYLE
The Usurper: dominant female vocalist, The Prince: soft meak female vocalist, Female Ensemble: opera singers isicathamiya
[190 BPM, Dark Disney Villain Anthem, Orchestral March, Broadway Musical, Vaudeville, Minor Key Brass, Timpani, Staccato Strings, Harpsichord, Theatrical Mature Female Lead, Meek Soft female Duet, Operatic Background Chorus, Cinematic, High-Octane, Climax, isicathamiya

[sweet ivy]

Yes I saw that in the earlier today

Maybe I'm not clear on what I'm asking

You said it went through 5 revisions. In each one did the melody stay the same, with maybe a slight change ?

[polar spade]

@ember wagon I thought i would ask, but does your "wrong" email have settings that can forward any swnders mail from a specific address to your "right" email? Ive seen that in settings of some emails, so just checking if that option is open for you. At least until you work out fhe issue of merging files.

[honest fjord]

I was timed out check dm

[honest fjord]

The melody was always the same the only time it was ever off melody was due to the testing with choirs today

This is my chat log with gemini with my whimsical album if anyone wants to see how my "suno architect" functions.
https://gemini.google.com/share/eb08ff7cd936

Hi

[sweet ivy]

M. I opened it up, thanks, lots to digest there

[honest fjord]

Yup

[honest fjord]

Check dm

[sweet ivy]

that little button at the top on Discord

[honest fjord]

https://tenor.com/view/facepalm-really-stressed-mad-angry-gif-16109475

[sweet ivy]

I see Suno, and opened it what now

[honest fjord]

Can you use it?

Say hi

It should create options for you

[sweet ivy]

what option would I choose, Invite To Server ?

[honest fjord]

Invite to server? Thats not right. Can you show me?

[honest fjord]

Screenshot?

[sweet ivy]

will fit in

[honest fjord]

Okay it works

[sweet ivy]

So I add you someone

[honest fjord]

Add me?

[sweet ivy]

tp people I get messages from in Direct Message?

Opened the Gemini link Share Suno

[honest fjord]

You dont have to share suno that's my job

[honest fjord]

So that gem is YOUR personal gem. No one else can see it. Was that what you were asking?

[sweet ivy]

I was confused about Direct Message, I though you wanted me to direct message you

[honest fjord]

I was sharring it to test if the gem worked before I posted it here. I dont want to post a broken link

[sweet ivy]

Okay, I wasn't really asking about Gemini, but the proces you used creating these incredible Broadway style tracks. I opened the long one that goes through all the steps & reading through it

[honest fjord]

I DIDNT
The ai did it

[sweet ivy]

https://gemini.google.com/share/continue/eb08ff7cd936

[honest fjord]

I don't even know what 90% of these actions mean. I dont know play or song lingo

[sweet ivy]

Sure appreciate you taking the time to work with a newbie

[honest fjord]

Yup

[sweet ivy]

Yes, Gemini created everything I'm reading through, the long process, started by your concept

[honest fjord]

Yes

[sweet ivy]

I'm heading off to bed, but going to listen again to Usurper Queen

[honest fjord]

Please do!

https://suno.com/s/qqhuQyCIlG9rKxDa

[honest fjord]

Glad to know you like it that much

[sweet ivy]

Mu God the best writers in Broadway would be envious of this

[honest fjord]

Ppffffttt that's a great compliment

https://tenor.com/view/roundofapplause-applause-vintage-crowd-audience-gif-1618695051660499724

[honest fjord]

Anyways good night and ill have good sleep with how much people like usurper queen

[oak maple]

Good Morning from Germany 👋 😊
Is there a way to merge sections in Editor?

[half dragon]

I don't think you can if you meant sections from different songs or clips

[oak maple]

By sections I mean the parts in a song. Like I have 4 parts called Intro which I want to merge to get one part

[half dragon]

And these parts are different generations right?
If they were parts from Extension, then you can choose Get Whole Songs that would merge them togehter.
IF they were seperate regular clips then you can't merge them via Editor

[oak maple]

I created an intro and split it in multiple parts for rearrangement. Now I want to merge some of the rearranged parts in the editor without having to save it as new song

[half dragon]

Did you split them off line? then upload them back to Suno?
Are you Pro user or Premiere -- Premiere users have access to Studio and you definitely put them all on the timeline to merge them.
But you mentioned you want to use Editor but Editor isn't for that; it's meant to edit a song.

[oak maple]

No, I did it inside the editor.
I'm a Pro user.
Merging would be some kind of editing or?

[half dragon]

Ah I see now, so you loaded the clip in the Editor, splitted in multiple sections to re-arrange, then now you want to save them into one clip?
Just click on the top where it says Save to New song.

[oak maple]

But I just want to merge some parts to one, not everything as I still want to do some edits on the whole track.

[half dragon]

You can continue working on it, you don't have to merge the sections, just leave them there.
Once everything is done, save it to ta new song.
Suno Editor doesn't have an option to merge 2 sections in side Editor, even in Studio like other DAWs.

[oak maple]

Do you mean even Studio doesn't provide such a feature?

[half dragon]

No, not that I can tell... unless it's buried in a short cut that I didn't know about

[oak maple]

But merging parts is something you can do in other DAWs as far as I remember.

[half dragon]

Yeah, definitely so.

Hopefully they'd add it in future updates - maybe worth to make a request at #1420795299701588129

[frozen dagger]

There is a suggestion box now in the form of "Labs" on the Suno UI, as well added recently. Click that and you can submit your ideas this way as well. Also, check out Sunoverse if you haven't already.
I submitted a suggestion there to make a sunoverse for our personal workspaces/folders. I have around 19k gens and i would LOVE to see my own stuff mapped out like this constellation with filters like they are working on with this constellation of genres. I love the space aestetic, the cluster mapping, the new way to find music and users in the Suno community. Very cool.

[frozen dagger]

you can merge sections and pretty much acts like an extend feature, albiet in a different way, but yes absolutely can create a new generation that uses more than one layer of other generations on a time line and recreate one that uses those as input to create output.

[restive pebble]

That would be amazing to have for our personal accounts. Excellent idea!

[jade mist]

Is Suno releasing new advanced options? I read on Reddit about duration, max mode etc

[honest fjord]

Why does my hook not upload due to moderation? What is wrong with the audio or video?

[neat nexus]

Just a guess but maybe the AI its using to check things doesn't like belly buttons?

Or it could be a copyright thing? As I said just a guess since we mods are only involved with the Discord and don't have any inside knowledge on the website. I have a few Hooks myself which are blocked for no reason...

EDIT: decided to spoiler that screenshot...

[honest fjord]

Oh! Ill change the video then lol

[neat nexus]

I would strongly encourage you to switch to something uncopyrighted or which you generated yourself when you do.

[polar spade]

The idea behind "Heal Edit" was to tie divided segments back together AND provide a "remaster" feel by blending any cracks or pops created when editing just one segment, such as if you cut off a piece.

"WAS" seem to be the optimal word. Heal Edit worked good in 4.5 when it first came out from my experiments. Now, it mostly seems to put any two segments back to one, with the chance of settings of one segment getting lost.

Ive now the habit of saving a song and bring it back into edit at almost every Edit, and listen for any issues of blending from the save.

But other than seeing the visual split, and just wanting to "see one whole section" , whats the issue behind the segments remaining split if the output sounds fine?

As an example, if you were sequencing in Studio or a DAW, its helpful at times to just drag one section to another track, rather than highlight three segments and drag, I get that. But in Editor, it wont make much difference unless you're dragging to another part of the song.

Keeping in segments helps if you decide to revert something back or just make more changes later. And down the road, if using Studio, we then alread have the song come up in segments which we can move around to different tracks and manipulate further.

[honest fjord]

Fixed it was the belly button https://suno.com/s/FyEDzHesy4UhqOwy

[shadow escarp]

I had issues with studio. Was using safari. Change to chrome and everything worked. Curious why safari doesn’t work well.

[neat nexus]

https://tenor.com/view/no-idea-idk-i-dunno-i-dont-know-no-clue-gif-5178996

I would guess it's something in the way the javascript or whatever extensions are coded. But that'd be a guess.

[shadow escarp]

Thought I would share that information.

[neat nexus]

Appreciated. Really. But probably better in either #🪲┃bugs-web or maybe #📱┃bugs-ios ?

I don't know how much the Team lurks in this channel but I have been assured that they definitely read through the bug channels and suggestion channels.

[shadow escarp]

Thanks

[raven turtle]

Max Mode was/is a feature that briefly appeared to be part of v5 but it never appeared. The assumption is that this will appear in some future form. The idea is that you use more credits to use additional server token resources for higher quality generations.
It’s something I’m quite excited about if/when it appears, as the race the achieve full CD quality fidelity is something I’d love to have.

[floral forge]

#📚┃suno-school

[crystal star]

Hello, I have a problem. Today, I wanted to continue my membership as a pro and made my payment, but my payment appears to be for 2027. I will write my invoice number below. Can you help me? Invoice no: UJWXRLCG-0002

[pure panther]

Is there a "heal edits" in the Studio editor?

[honest fjord]

WHY!

[plush field]

wrong channel for that

[quick wedge]

Best to ask in the other channel https://discord.com/channels/1069381916492562582/1372994553233276970

[crystal star]

@plush field@quick wedge No one has responded for hours. I'm about to lose my mind. I have everything to do, but I'm waiting. I opened a ticket, but no one has responded.

[plush field]

https://discord.com/channels/1069381916492562582/1373016771103166575

people who can help you usually do not watch channels like this one

we told you that to help you, not to criticize you

[honest fjord]

https://suno.com/s/oYBGXfRDB7eEridY

Better

[viral pond]

hmm

[quick wedge]

They’re just backlogged discord can’t help with much else since we’re just community but you can DM DAViD and he’ll expedite your ticket if you already waited it out

[crystal star]

@quick wedge It's all good. You helped me a little while ago. Thank you very much for your interest.

[quick wedge]

You’re welcome.
You can still DM DAViD and he’ll help you.

[pure panther]

Sooo, is there supposed to be some in Suno way to add extends to your songs when you generate them on the create tab? I went into Studio to add it, but it doesn't stitch together nicely at all. I can export it and stitch it together in audacity, but I wonder if there's a tool for it in house.

[hard beacon]

You are talking about a tempo drift??

[pure panther]

Me?

[hard beacon]

ya

[pure panther]

Oye

I'm am trying to keep it clean, that's the point of the separation Mr. Woodhouse 😛

Oh. No, just the transition from the existing song to the extended version. Studio doesn't have the precise control needed to match the wavelength and "stitch" the sections together seemlessly. I also can't find heal edits anywhere in the studio editor. The googles told me that hovering over the section header will bring up the option, but it doesn't do anything at all for me. Since, when using the extend feature in the creator tab (screenie below) it's not automatically intigrated with the song, but rather is given to you as a completely separate clip, I was wondering if there was a known uin house method for marrying the clips seamlessly?

Well, I would give you a screen shot, but I can't because I'm getting this error too and can't access my songs at all right now.

[hard beacon]

hmm..

[frozen dagger]

jsut right lcick the segment\section

[pure panther]

I did that, I don't get the heal edits option

[pure panther]

Here's the screen shot of how I extended it in the create tab

[frozen dagger]

no no in studio right click the clip \track on the time line, the colored bars

oh wait im looking at 2 diff things sorry

that is the right way to extend in Studio tho, either select and drag clips ro you can click and drag to select any area including multiple tracks and then drag that to the remix areaa same

[pure panther]

I think I see why, I only want to heal one section, but that doesn't seem to be an option.

I was trying to avoid downloading, throwing it into audacity and then reuploading, but I think it's unavoidable in this case.

[plush field]

change the order, fix lyris, whatever

but...

better, faster, easier 🙂

[frozen dagger]

https://tenor.com/view/where-my-heals-where-is-my-healer-where-is-my-healing-gif-9658433413762298483

somethings are hidden, like being able to recreate. It says "replace" on the button within the timeline creation tab, sometimes it has a drop down menu to change replace to Recreate. I think the clip can have 0 alterations in order to do recreate. Recreate is my preffered one for blending, but maddeing when it is sometimes there and sometimes not an option whtout me truly knowign exactly why. Heal Edits is kinda like that too. Some features just vanish or do no show when not available, without any clear explanation. Not grey-ed out options; jut don't appear. Same for people looking for the Persona thing and cant find it when it relates to an audio upload. Just looking around like uh, ok why, what i do?

[plush field]

seee what is better for your case

hope this helps

[pure panther]

I'll give remix a shot. I just spent about 50 credits in the song editor trying to get it to churn out a good version, but it just wasn't up to the task. So odd too since it's only intrumentals in that section, you would think it might be better able to make a good edit. 🤷‍♀️

[frozen dagger]

interesting, the wording is changed as his image shows. it says Cover now. that wasn't like that day before yesterday.

[plush field]

I spent about 5k credits trying to get a genre that Suno does not know, so... 😂

[pure panther]

Yeah, the credits on the left side bar also vansihed the day before yesterday, I want it back 😭 😂

[hard beacon]

it's because of the today's update of Suno Studio ( i think ) 🙂

[pure panther]

This song is 2000 credits deep, at least 5 iterations/genre changes and this is the 4th day (8+ hours a day) It's soooooo close! I'm taking the very last steps before release 😄

[frozen dagger]

they rolled out a few UI features over last few days, Sunoverse, Labs, changeds UI layout for wher it shows credits like Tea mentioned now its up top in Profile pic, i like it. clean. still "at a glance" info.

[plush field]

creedits in studio are in top right corner

[pure panther]

They're nowhere to be found in the song editor

[plush field]

this is in studio

[frozen dagger]

your sidebar is probably hidden you can se your credits at a glance in studio tho

[plush field]

for "create" top left

[frozen dagger]

click small arrow abofe your porofile pic to expand the sidebar reveal credits

[pure panther]

It's not hidden. They were there a couple days ago, now they're gone.

In this big gaping empty space

[frozen dagger]

its in your profile pic lool

its small

[pure panther]

They're nowhere in the song editor

[frozen dagger]

🙁 dont look at me!!! lol myh low creds

no mam, onme sec paint

[pure panther]

Hmm, I still want them back where they were. I'm old and set in my ways (get off my lawn shakes fist vigorously)

[hard beacon]

😄

[pure panther]

Also, they used to be visible when the side bar was collapsed. Now you have to have it extended to see them. I like the old ways~~~

[frozen dagger]

haha sound like me at work... but its ok, its for the better. What we need is patch notes... Keep us updated with it, cus as it is we just discovering on our own. Could miss and be oblivious.
Well, a manual of sorts would be nice too. Also, figuring out features on our own. Pros and Cons there. But Studio... please. Hold my hand a little bit and walk with me. Because I get it and it's great, but I'd really like to know "what it do tho?". I antiicpate bigger updates are on the way soon for that as well tho, and when they roll them out, but still... here ya go, figure it out.

[pure panther]

@frozen dagger I tried both replace and cover a couple times, the results are ...yikes. I'm trying remix now.

Wow, somehow remix is worse 😂

Time to Audacity

[frozen dagger]

I can't tell you what I was doing when I was getting good results, because I tried a lot of things. Just not sure what I did. That Recreate\Replace drop down (which now says Cover, not Recreate) was sometimes there and sometimes not. I was trying to figure out how to get the "Recreate" part, because that was what uses the selection to create the output, but got lost in the UI and trial and error of it all. The same features are available, and when they work right, they work right. I have been preferring to create within Studio lately. But it still feels like just pulling levers and twisting knobs seeing what does what.

[pure panther]

That a good description of what working with Ai feels like

Sometimes it gives flawless results and other times it can't fix a tiny "blip" even after 20+ attempts.

I suppose I could just try to generate another extend within Studio, they're automatically integrated into the existing song right?

[frozen dagger]

I just added Wabi-Sabi to my "style" and lean into them. No mistakes, only happy accidents. Can't rid of a blip? Ok, can't beat em, join em. Select area with blip, remix, style = glitchcore. "You wanna play games? Vamos!" But yea, that's not gonna work all the time for most. Reminds me of test driving a car once. Asked, "Hmmm, what's that knocking?". The reply, legendary, "You know how you fix that?" ::Turns the radio up, and nods to music:: (I did not purchase the vehicle.)

[pure panther]

It's funny. The track I ulimately decided to go with was the one I made on accident yesterday while trying to ask Suno a question. So to keep the same "style" for the extension, I have to put in the question. SMH.

Yeah... man I am just having no luck with studio. It doesn't appear to even make an attempt to marry the extension to the existing song.

[frozen dagger]

I would try putting a little empty space between where the transition is, highlight an area to replace, but see if adding some space gives it some freedom to blend? Call me Reading Rainbow tho, Don't take my word for it.

[quiet solar]

Sometimes all roads lead to Audacity when it comes to small, precision changes. 😉

I'd also echo OBLMOUS' suggestion above - in the Legacy Editor, Suno had a lot more difficulty stitching together two sections at the minimum generation legnth than when it had more time to play with. I haven't tried the new Replace/Cover features in Studio, but it wouldn't surprise me if that was still the case.

[pure panther]

Yeah, I just found the similar wavelengths in audacity and put it together. I honestly should have started there, but then how would I know if the editor capabilities have gotten better or not? 😄

[elfin oar]

Send us the link, wanna listen to it if u dont mind?

[pure panther]

I'll tag you and link once it's done, I'm in the last steps of processing now (which can still mean all day 😂)

[brazen yew]

Is there a way to smooth out certain vocal derps? It's one word in one chorus where the vocals add an extra bit to the word which stands out when comparing to the other choruses

[remote ivy]

i asked to glitch the vocals on cover which i changed genre to electronc, dance.....no vocal glitches....i then covered it again w/ this prompt, i still got no vocal glitches although the instruments did a little....i just throw in pirate battles into the prompt out a slight frustration...prompt......what do i need to change to make the vocals glitch in my cover: genre: electronic, dance, alternative
instruments: 808s, tuvan iger, vocal style: glitch stutters during chorus from instrumental and vocals, vocal glitches should enhance the singability not detract from it, keeping a groove for hip swaying dancing
studio quality, warm mix analogue, funky synth pops and squelches, amid roaring ocean foam and pirate battles

[plush field]

which one of those?

“lo–l–lo–lo” (vocal chop)
vs
“l–l–l–––l” (stutter with drops)

[frozen dagger]

glitchcore vocals, hyperprocessed vocals, digital vocal fragmentation,
pitch-shifted voice, formant-shifted vocals, chopped syllables,
stutter edits, micro-looped phrases, rhythmic vocal glitches,
synthetic vocal texture, data-mangled voice

rapid-fire vocal delivery, syncopated vocal rhythms,
off-grid phrasing, burst cadence, stop-start articulation,
call-and-response fragments
bitcrushed vocals, granular vocal synthesis,
vocal time-stretch artifacts, digital distortion texture,
glitch harmonics, vocoder-adjacent processing
glitchcore vocals with chopped, pitch-shifted, stuttered delivery,
hyperprocessed digital voice, rhythmic micro-glitches
🐙🕉️🥸🫡🎹

[remote ivy]

not sure, i just want see whats what..i only have 270 credits til i get my 50 a day dole out....so i have ration ..otherwise i would played around w/out asking

[plush field]

can you give me an example of a song with that, so i can understand what you are looking for?

description from AI, what is what

vocal chops are short snippets of a vocal, usually pitched to fit a melody. they’re used like an instrument, often forming hooks or riffs, and sound musical on their own.

glitch stutters are rapid, repeated, or chopped-up vocal fragments that focus on rhythm and texture rather than melody. they create a digital, stuttering effect, often synced to the beat, and are more about movement and energy than notes.

tl;dr: chops = melodic, playable; glitches = rhythmic, effect-driven

tags will depend on what you want

[remote ivy]

i don't quite think this necessary glitching like intend...but this made me wanna mess around https://suno.com/s/p8FoPECCzZrEIx7r

[frozen dagger]

chopped and screwed

[plush field]

this sounds more like what you want

vocal glitch/stutter style = vocal performance that’s intentionally chopped, rolled, or jittered to create rhythmic, percussive textures. it’s more about sound design and texture than melody — think rapid-fire consonants, rolling “r”s, or fragmented syllables

[remote ivy]

he is very good at what he does...thats why he is only 1 of my 3 follows

[frozen dagger]

@remote ivy ,use the phontetic trick @plush field described. i.e.
Line up the bots!
(Line ’em UP, line ’em UP)

Line up the bots—

L-L-L-ine UP the BOTS!!!

…take ’em out one shot

[plush field]

try this

[Intro | vocal: glitch-fx, texture: granular]
(S-s-system... f-f-failure...)

[Verse 1 | vocal: stutter, flow: percussive-rapid]
I-I-I can’t f-f-feel the f-f-floor
D-d-data leaking through the d-d-door!
(Cut-cut-cut it up!)
Rrrre-loading the m-m-mind-set.
Ti-ti-tick tock, don't y-y-you forget!

[Drop | vocal: vocal-chops, instruments: heavy-bass]
(B-b-bass! B-b-bass!)
[Instrumental Glitch Solo]

style
Stuttering Vocals, Vocal Chops, Granular FX

[frozen dagger]

with the style prompting for thigns we already said it will easily interpet this kind of syntax wonderfully. not exactly as you may imagined typing it, but let Suno take the wheel.

[remote ivy]

thanks gents, i'll give it a try

[plush field]

example with more Rs, to try to match example style more closely 🙂

[control: aggressive-flow, percussive-vocals]

[Intro | vocal: whispering, texture: gritty]
(Are you rrrr-eady?)

[Verse 1 | vocal: staccato, speed: fast]
Lock and load, hit the code.
B-b-back on the road.
Pressure rising, t-temperature is c-c-cold.

[Build Up | vocal: rolling-r, effect: machine-gun]
Here it comes...
Brrrr-aka!
Drrrr-op the beat!
Rrrr-un!

[Drop | instruments: heavy-bass, vocal-chops]
(Rrrr-ta-ta-ta!)
(Grrrr-ind it down!)
[Bass Wobble]

[Bridge | vocal: aggressive-trill]
No time to prrrr-ay.
We don't pl-l-lay.
Rrrr-ise up!
Rrrr-ise up!

[Outro]
(Brrrr-ap!)

anyway

Dash is your Frrrr-iend!

😂

as well as rrrr

[remote ivy]

i tried just that style prompt, it chopped the highs and lows of the vocals more so..buts it a start cover from w/ more fine tuning, might be hard w/ given the original vocals....https://suno.com/s/tH4CXLNc7BLAPmTd

[plush field]

you need to format lyrics as well

and those are just 3 you need to add to your existing style

for example from this
Ragga Jungle, Raggacore, Jamaican Vocal, Fast Flow, Heavy Bass, 170 BPM

to this
`Ragga Jungle, Raggacore, Jamaican Vocal, Fast Flow, Heavy Bass, 170 BPM, Stuttering Vocals, Vocal Chops, Granular FX``

and lyrics from

this

Rise up!
Rise up!

to this

Rrrr-ise up!
Rrrr-ise up!

[remote ivy]

yeah i think it will take more thought, if i just want chop or skipping, glitching different loudness etc, ive done 1 trap song before, so ive messed w/ stutters in lyrics

[quiet solar]

Yeah, if you're just after a little bit of tactical chopping and stuttering in the effects then dropping something in the style prompt should be enough.

If you're after heavily glitching the entire voice, as in SpillinJam's song, you'll need to conduct surgery on the lyrics as well as prompting.

Lietolis shared a song where they went for a very glitch-heavy introduction, and they openly shared the prompt and lyrics behind it - might be worth a look for further inspiration: https://suno.com/song/ffafefc1-4e53-4ff1-8a2d-ec44af64ae22?sh=OF80mI4WSV9Xl1Rj

[pure panther]

@elfin oar It's finally done! 😄 (a grand total of 2045 credits and about 40 laser focused hours later)
https://suno.com/s/9m01NkmNnMOaDhsx

[polar spade]

Its [ credit location and visibility] still an issue with the sight impaired /blind from how it was readable before, so its a question of why the change.

[pure panther]

I agree. There's nothing else there, it could have stayed as far as I can see.

[polar spade]

Id like to think [my song Rise] subconsciously inspired you. But your song is way more kool.

[candid basalt]

if the song i'm making is sang, but i want the second verse to be rap, and i'm doing it through extending how would i make suno do a rap instead?\

[pure panther]

You guys are all awesome and inspiring 😄

[polar spade]

literally [rap] as a metatag tends to work fine. Even can only put "rap" in styles or focus on prompting for rap as well as whatever prompts to give the vibe you want just for the extended section.

Dont form the prose as a verse structure, but

1. putting in quotes helps to get it spoken or rap, but not always needed
2. run on sentencing structure and let the words wrap to the next line. Adding dashes, commas, and period to help with the metering as you may feel best. Suno has a good way of looking for rhymes and tries to meter the prose as needed to fit the beat.

[pure panther]

You could mix genres (rap+pop, or rap+rock or whatever your main genre is) and add meta tags to the lyrics to nudge Suno in the direction you want. It's not a guarantee though, Suno is a bit wild.
here's and example or a meta tag:
[verse 1 | vocals:rap]
or
[verse 1 | lyrics rapped]
Keep throwing stuff at the wall to see what sticks.

[candid basalt]

ah sweet, i'll try hopefully i can get it to work

thanks!

i still don't really know all the way to command suno so i'm still rolling pretty basic

[polar spade]

Youre starting out the right way by learning to use extension, In my opinion. So thats a good thing.

So do a bit of experimenting when possible, but I understand if youre Basic you only have the daily 50 credits to work with.

[candid basalt]

no no i pay monthly

is there a wiki or page or something that has like a cheat sheet for how to command suno? or is there a youtube video or something

[polar spade]

ahhh better then.

So, anyway... extensions work well when changing up the feel, so a good option from going to a Rhianna type chorus to an Eminem rapped verse, for instance.

[polar spade]

You tube may be your better bet, since somewhat legacy tools like extension have been around for the duration.

[candid basalt]

ok i'll do some searching, i just want to be able to get the most out of it. i've already been amazed by what suno could do with redoing a song i made 20 years ago

[frozen dagger]

when you're first learning Suno, that Orange Magic Wand is your friend. Help with language and phrasings. Can be used in many useful ways. Highly recommend playing with it.

[candid basalt]

ive been using that, and chat gpt

[frozen dagger]

then you're on the right path padowan.

[candid basalt]

but it was more specifically the command in the lyrics themself to let suno know

this seemed to work

https://suno.com/s/ySCS4feX7IDqDQUO

that actually finished the song for me

@frozen dagger@polar spade@pure panther thanks for the help https://suno.com/s/JlrxNyuAWUfderzR finished product sounds pretty good

[pure panther]

Nice, I'm glad you got it to work! 🙂

[crimson thorn]

Guys is the extension feature bugged? Everytime i type in lyrics to extend the song or replace lyrics, the bot just fuses both the old and new lyrics together even though I specifically replaced those unwanted lyrics

[candid basalt]

i just did an extend and it was fine

it was on pc in a browser

[crimson thorn]

Yeah i’m on pc ona browser too but it’s not working properly for me

Idk if i’m doing something wrong

I only put in the new lyrics but it just merges the old one

[frozen dagger]

I had problems with it off an on. Was like a 1 out of 10 were useable. Most were as you described even when I did all the tricks I know and made adjustments, still I gave up and went to Studio to try there.

[frozen dagger]

I think one workaround I had was I stem split vocal and instrumental on the gen i wanted to extend. Then I just pasted the instrumental at the end of the song, remixed that extension with a little bit of the original ending also selected. (Click and drag to select the area in Studio ((if available)) and click and drag the selection to the style box area where the remix are is. Entered the lyrics for the 2nd part there and basically did a Cover of the instrumental with lyrics. I think it worked ok, but I did a lot of things to experiment.
I think this is also achievable without studio, just need to add the instrumental in editor and split the original where the transition starts so you can add it with the instrumental to Cover\remix.
It's kinda what extend does anyway, but you're doing it manually.

If the song is like 6 mins or less you can do a Cover of the whole track with the first half and the instrumental stitched on, but first half will change.
Can't really Cover anything longer than 5 or 6 before it starts squishing and distorting. Doesn't map over the 8 min cap from a single generation; reliably.

[crimson thorn]

Alr i’ll try that, and if it doesn’t work, guess that song is a lost cause. But thanks so much for the tips!

[frozen dagger]

GL. Hope you find success. Don't give up! 🙂

[fresh gull]

do people still argue over voice's I have not seen anyone saying that use it...

my findings

[quiet solar]

No such thing as a lost cause. 🙂

That fusion of lyrics or resurrecting a line from earlier in the song can happen, but it's usually a 'from time to time' thing rather than every single attempt. If what you're doing seems absolutely cursed, you can try things like moving the Extend start point - Suno is very good at repeating choruses, so sometimes having an extend point just before the final line or two in a chorus and the last line(s) of the chorus included in your new lyrics can help anchor it for the start of the new segment.

It also helps if your new lyrics fit the existing structure of the song reasonably well - for example, if your current verse has 8 syllable lines and your new one has 15 syllable lines, sometimes Suno will struggle more with the replacement. Replace lyrics also works in a very similar way.

If you're a Pro/Premier user, you can also try turning down the audio influence and, potentially, turning up weirdness from your current levels which may help give the model more 'breathing room' to introduce something new.

[crimson thorn]

Oh yeah i had my audio influence up, maybe that factored in the suno’s mistakes 😅

[polar spade]

Before doing anything in edit, you might want to try to make a copy of the song first, with the current proper lyrics and arrangement. Then edit off the copy instead of one in thr chain of edits.

Since its inception, Suno likes to pull lyrics and styles from all preceeding iterations, even though its supposed to just focus on styles and lyrics of the current edit, cover, extension, etc.

Its done it through every model, so it has become a birthmark. Its been known and nothing has changed, so it seems all other improvements and updates perhaps work best if this behavior remains unaltered.

As far as in editors, its best to have a habit of either copying the most recent lyrics or keep a copy standing by in another browser. Otherwise, youll need to at least scroll through and find the one you are editing to get the complete and current lyrics. In order to paste in, the align and keep going, possibly having to do again on the next edit.

[pure panther]

That explains a lot. I update the lyrics in each iteration, but Suno often reverts them.

[quiet solar]

If you had it turned right up, that's a VERY likely cause of the problem. Essentially, higher audio influence will copy across the sound (including how the words being sung) with a high priority, which can supercede or clash with changes you're trying to make.

[chrome oar]

Does anyone know how to get Suno to recognize the difference between male vocal’s and female vocal’s? As far as duets are concerned? I’m having a hard time when the verse switches from man to woman

[half dragon]

If you searched around here - you'd find out that Duets are not easy to make but still fun to do.
There are few ways but take a read at what @quiet solar posted - could be helpful
#📚┃suno-school message

[nimble hemlock]

When I extend a song, should I include the original song's style tags or leave it blank? And should I dial Audio Influence up or down? Likewise, for Weirdness and Style Influence? No matter what I try, it comes out like garbage. It's mixing styles, adding in elements that weren't present in the original song, and sometimes even sounds like it's trying to generate two songs at the same time on top of each other. If I at least know what these settings should be, maybe I can get some results that actually match the original source song.

[timid lion]

@empty saddle

treble, ultra-rap-electro hallowed-goa progressive-glitch dramatic industrial-grave-edm grave-wave electro-mashup breakbeat-progressive-electro ghast-phonk, trap-minimal dark-break-house, trap-house goa-break dubwub-horror, Rap lyrics in drill-electro, bounce-minimal-house, mashup-minimal-house haunt, sections utilize edm form adaptations sidechaining transitions, lyrics rapped  pan-echoing reverb, chopper drill rap application,  no vocals as instruments. bass should be shallow and suppressed. hard horror distorted ethereal elements, progressive electro buildup tuned to section melody, rapped vocal style, some lines metal-scream hauntly, then shift to immersive breakdown.
sublines have break transition cliffs.
Meter:  4/4 initial meter. Intro pads include 3+3+2 eighth-note polyrhythmic displacement. midsections with intense polyrhythmic claps employ a metric modulation from 4/4 to 7/8 for feel. super section resolves to 4/4 as 2nd midsection metric modulation equating measure length

[restive pebble]

Yes, you should maintain the styles to maintain the songs spine or DNA.

For the sliders, I can only give you my experience.
Weirdness at 0
Styles is around 70
Audio 75 or higher

Be aware that Suno sliders set too high can have some very undesired results.

There are no hard and fast rules but generally these are around where I have the best results on extend.

A side note, Suno may be having difficulty because of the 'cover, or inspiration, persona' used to extend as it may be attempting to read that and the current song being extended.

You could remove the 'added reference song or clip, persona' and see if your result are closer to what you're looking for.

Hope this helps.

[slim heath]

Anybody managed to change the female voice to deeper/older one? I tried persona male/female, description, bpm changes, music style changes, rock, folk, ballad... don't know how many options. It works only for man, for woman no. What is funny is that a women persona, if I apply it to selected female vocal, does not have effect, but if I select male, I got the effect of deeper older voice:) Any experience with this on prompting, using Suno other options to get a deeper/older women voice? I even got situations whith 2 voices: one deep voice for the man and the normal Suno women voice in a duet.

[restive pebble]

You can try using something like this:
Mature female vocals, Deep contralto, Weathered voice, Husky, Raspy, Smoke-and-whiskey vocals, Baritone female, Low register, Soulful, Wise, Gritty, Dark, Emotive, Aged.

I may help also to include the meta in the verses where the female sings. Something like this:
[Verse]
[Vocals: Female, Deep, Weathered, Husky]
or whatever effect you want the vocals to have.

[slim heath]

Thank you! I tried without persona and got some change, still young voice. I used this time a persona built on somebody singing ( I had another persona based on just talking) and I got better results. Still it is not like for the man, where it s fully applying the persona in smallest accent. For me it is clear, Suno is stuck with some female vocal patterns, no matter what. For me is case closed no more credits wasted.

[quasi ferry]

Check this craziness out, let me know your thoughts...
https://suno.com/s/hSwdab41NXF1O5aU

https://suno.com/s/Etr6D5Bkk2LmXxHK

[restive pebble]

I've heard some get excellent female vocals similar to what you seem to want.
Artistic Intelligence made this song which has great female vocals. Maybe you could re-use some of the styles to get the effect you are looking for.
https://suno.com/song/eb3b3c3f-db41-4d93-a437-21a0c6049370

I should add that adding deep, baritone will/should deepen her delivery.

Suno is a random generator and only takes our prompting as a suggestion. Sometimes it grabs it and uses it fully, others it likes to play with our feelings.
Keep trying, you will get the sound you want eventually. The key is to keep modifying and find what suno wants to give you what you want.

[quasi ferry]

For deep tones a good trick, tell it to use the Key of G and use an EQ Range of 25Hz to 16kHz,, G is 25Hz, 49Hz, 98Hz, 196Hz, 392Hz

[slim heath]

Thank you, nice reference. I tried it, new lyrics, copied style. I tried with no persona and with persona. Male and Female vocals. Sometimes, even with female vocals selected, it was still man voice. Here are two examples compared. https://suno.com/s/YKOsPIYMAFywDw5z and https://suno.com/s/JJk9c0kOriEvLAj8

[restive pebble]

From this: The vocals are raw, raspy, and powerful, performed by a tough, confident woman singing about working as a gravedigger in a man’s world, The mood is earthy, proud, and slightly spiritual — like digging graves deep in the Louisiana bayou with the spirits watchin’

you could alter it to reflect more baritone vocals.

Like this: The vocals are **deep, baritone **raw, raspy, and powerful, performed by a tough, confident woman singing about working as a gravedigger in a man’s world, The mood is earthy, proud, and slightly spiritual — like digging graves deep in the Louisiana bayou with the spirits watchin’

I see you leave the meta tags out of the lyrics which is not a bad thing but Suno accepts further prompting using the meta tags above each verse as a more granular way of influencing the vocals.
Maybe adding
[Vocals: Female, Raspy, Deep, Soulful, Baritone]
would help maintain the female vocal through your iterations.

As a final option you can always hit up AI to provide some deeper insight into how to influence Suno's vocals fr a deep female sound.

The second version sounds very close to what I envisioned you wanting, but my mind and yours may differ...

Also try moving the sliders to reflect your styles more byy raising it to around 70% and drop weirdness to 0

[slim heath]

Thank you! Indeed, second version is the closest to what I need. From there I need to say goodbye to Cajun Blues and explore if I can still keep it with other music style ( not far from blues). Anyway, the link was quite useful. I ll be more aware about meta tag, as I noticed the voice switching during song from the Persona. Great advice! Thx a lot!

[restive pebble]

I believe you can find what you want. Just be patient and check your styles, sliders and meta before generating. It was fun to help you find your sound. 🍻 cheers!

[bleak zinc]

I find also some tags can overpower others, thinks like Dark have way more impact than say Dissonant. I have started to keep a list of tag combinations that work to feed into a GPT to help repeat styles when I find things I like

[loud nest]

Is there an official moderator for Suno, please?

[neat nexus]

In what sense?

There are mods for the Discord.

[loud nest]

I need help deleting a link to a song. I deleted my account and my songs, but it still appears in a link I shared via WhatsApp.

[neat nexus]

I see. This is account specific and not Discord specific so I do apologize but you will need to contact the Team for assistance with this matter. Please follow the steps in #support-guide

None of the Discord Moderators or the Guides and Mentors can assist you with this, you will need to be patience and wait for a reply from a Team member in support.

[loud nest]

Ok Thanks

[queen linden]

okay, I havent gripped it a long time, but i need help..... I have just spent quite a few credits trying to get rid of a nonsensical word out of lyric. Now I have two nonsensical words.... should I just scrap the whole thing?

[scenic pelican]

how can i ask it to not add a specific instrument? since theres only lyrics and style section, no negative prompt?

[queen linden]

Use advanced options do not use the NO...

[neat nexus]

It's in advanced options.

[scenic pelican]

i dont have that option is it only for v5 or pro?

[neat nexus]

It's only available in custom. Are you using simple or custom prompts?

[plush field]

users with free accounts cannot use exclude

if you have pro at least, it is in custom

[scenic pelican]

i see thanks

[plush field]

when you enter both style and exclude you can see in suno that you got something like this

Folk Metal, Melodic Death Metal, Double Bass, Galloping Rhythm, Distorted Electric Guitar, Orchestral Strings, Harsh Vocals, Clean Male Vocals, Anthemic Chorus, Epic, Battle Hymn, 135 BPM, ‑Half-time, ‑Breakdown, ‑Doom Metal, ‑Sludge, ‑Slow, ‑Acoustic, ‑Ballad, ‑Pop, ‑Electronic, ‑Trap, ‑Laid Back

so all things from exclude are prefixed with -

so in therory this should work even from style prompt

the problem is that free users cannot use style influence slider

and it might require more than 50% (default for free users) to actually exclude those

[plush field]

when i find some time i will try to add all those in style with 4.5-all and to leave sliders on 50%

to see what i will get

but do not keep your hopes high

[pure panther]

What are you doing to try to get rid of it?

[queen linden]

I ended up just redoing the whole song, I took it into editing four or five times . I gave up ...

[vapid obsidian]

I need advice when any advanced setting I use just will not 'budge' the songs vocal into a personal vocal I want to put on it instead. I have a song I made a few months ago that is techno with a medium alto voice. I am 'cover' ing it to make it a rock version and I want to apply a persona vocal to it. So I have all that setup. I 'thought' that 'audio influence' controls the weight the regeneration puts on whether to use the songs vocal or the persona i've setups vocal in the cover version. However I've tried 0 w / 50 si for the first 2 sliders, then for Aud Inf. I tried 25, 75, 85 and it just wont budge. Does anyone have advice? Thank you!

[polar spade]

Does get rid of mean remove , not replace?

What happens to just go in edit , select it, and delete it? Use heal edit, remaster or then select a small part ( over the now spliced area ) and regenerate using Smart option to get a better blend of the audio.

[queen linden]

I tried to replace a proper word, 'repose' because it didnt sound right . I thought it would be a simple replace.... I replaced the whole line, (for me that works better a smoother transition) but the word became ridiculous denose, decnose, grnose .... Edit seemed to be facinated with adding nose. so scrubbed the whole thing....

[queen linden]

I didnt even see Heal edit! thank you!

[loud nest]

Is there a moderator here? How can I permanently delete a link so no one can see it? I've already deleted my account and all my songs on Suno. I need help, please; it's important.

[polar spade]

OK. Just throwing some alternatives. Whatever is more comfortable.

Keep in mind to listen for the " sound of the word" which might be considered in its pronunciation. So, "Do you want to eat" might make Suno attempt "Jewanna eat" based on its training. So its important , as you tried, to keep a wide birth on the words surrounding as well.

One way is to download the file and reupload and see what Suno types out as the lyrics to see where it may be going wrong. Nevertheless, even if Suno types "Do you want to eat" as the lyrics, it will still try to pronounce all the liasons and blends it's learned from training to sound like its speaking fluent native English.

[hard beacon]

@loud nest : what links are you talking about?

[loud nest]

https://suno.com/s/zyDe1wBYO3lidYt1

[loud nest]

Can you help me

[polar spade]

@loud nest Ive copied my song and post it again to keep it in the Library, then Permanently Removed (not just send to trash bin) the original output. So clicking a link just goes to a 404 page for me.

[hard beacon]

What Dasvon said... make sure that the track is also removed from the trash bin

[polar spade]

Caution: FIRST, Please copy or download your song if you wish to keep it.

If you dont see "Permanently Remove" in the menu of your output in the Library, the move the output to Trash Bin first, then go to Trash Bin and select the menu of the same output to then click Permanently Remove.

[loud nest]

I can no longer access my account. I deleted it, and supposedly all the songs, but the link is still active.

[polar spade]

ahhh OK....If you click "Delete Account" or "remove account" in the APP, then you've removed your access to the account. It may not necessarily remove the files themselves.

[plush field]

@polar spade @quiet solar what do you think about this? i used those as a base for AI rules
when values fall outside the intended range, they are automatically corrected
very low values tend to feel rigid, while very high values may feel unfocused
clamping helps keep results consistent while still allowing creative variation

of course, everyone else is more than welcome to comment

[loud nest]

I'm sending you the link in question. Can you view it?

[loud nest]

https://suno.com/s/zyDe1wBYO3lidYt1

[polar spade]

I may not be smart enough to know what that would do, other than experiment and see. But you're asking it to make adjustments in its training? not just prompting it to do somethings within its parameters, is what Im seeing. I may be wrong.

[plush field]

not training, system instructions
those are "middle layer"
i cannot change training, but i can tell, listen use this a decoder key how to use it, ignore what you know about suno slider

i am asking about what is expecet behavor when user set those values

i made those based on Suno UI, tooltips, colors... but also my AI knowlage

[polar spade]

Did you download it first? or you didnt want it anyway?

[quick wedge]

The links takes time to be cleared from the system after the account is deleted. Better off having to wait. You spammed three channels just to ask that. It could take 24 to 48 hours for it to be not in the Suno system any more.

[polar spade]

Then thats out of my wheelhouse. You know more than I for what it may do.

[plush field]

model is smart enough to understand what i want and to obey that
but that will be wrong if i am wrong
thats why i am trying to confirm my understanding of sliders

[polar spade]

@loud nest OH, this is just recent. like today? yea, OK. lets wait a few days and try that link again.

[plush field]

from https://help.suno.com/en/articles/6141377

[loud nest]

Ok Thanks you

[plush field]

my issue witht this page is that is not detailed enough, and 99% of information on the net are WRONG

[polar spade]

As far as sliders are concerned, if you wish to stay conservative, and you say it should work, then thats a fine choice in those range numbers of 15 and 85. Again, in the end, it takes running it many dozens of times and see if any sound like they are going out of that zone. I assume the only way to know is if the song profile notes no slider settings undre 15% or over 85%

[plush field]

even when i am doing things manually i almost never go over those values

[polar spade]

Thanks to @quick wedge They've been keeping up with your posts.

[plush field]

there is a reason suno colored them in red 🙂

for some wierd things, you must go over 85

but that is overcomplicated for what i am making (to explain that to AI and still make other things work as expected)

[polar spade]

You've convinced me. 😊

yea, so , I mean I slide things into 98% and 1%, but all I can do is listen and see if something worked. Even 100% isn't totally annihilating, and may give a fun result 1 out of 10 times. I try not to be perfect and so just enjoy the randomness. Look for the pearl in the oysters. As I said, seeking perfection is not in my wheelhouse.

[quick wedge]

It was a spam spree in two other channels plus gen chat. I replied to him earlier until he asked in here and chit-chat

[hard beacon]

He has eyes everywhere 🙂

[quick wedge]

I’m sure he’ll be fine told him to be mindful on the spam though

[polar spade]

yep. Thanks for watching my 6 then. I was assuming they just discovered something after a long time passed. OK. , 🎶 One down, one to go, another town and one more show 🎶

[quick wedge]

🫡✌️

[polar spade]

I had to run to make the senior dinner discount. So back to this.

What is suspect or evidence that Suno goes outside of "default" for any output or any amount of prompting? Do you think we can prompt it to "slide" outside of default setting while not touching the sliders? It would seem thats what the sliders are for originally, to me.

So, its a cool thought if it were true, and it may be and we have not yet discovered it yet..... you could be pioneering it.... but by extension, then we could prompt limits for bpm, volume and the like, right?

[plush field]

wierdness above 85, especially at 100% can create unwanted noises and glitches

i am using AI to help me udnerstand suno better, to fix my prompts or to give me a quick prompt that i can test to see will i get intended genre/style/energy

i am trying some strange combinations, since i want genres that suno does not reognize

that's why i need to undestand sliders, so i can make a proper instructions for AI, otherwise i will get junk - know as GIGO (garbage in garbage out)

[polar spade]

Yea. Im clueless to know the limits of what we can do. Try out everything and keep a record of what works and what doesnt.

[plush field]

i have 2000+ records of my failed tries at suno 😂

since it looks like v5 can handle style prompts like this one

melodic death metal high-energy anthem. combines aggressive, intense sections with catchy, uplifting choruses. features a mix of harsh growls in the verses and soaring clean vocals in the chorus

and not only like this

Melodic Death Metal, Power Metal Chorus...

i am currently instructing AI how to create this first kind of prompt so that i can test those

[polar spade]

In public ,I reference 0% as "null" and 100% as "infinite" but even times there are moments when at least within 4 minutes something sounded pretty good. So I dont shoot for 4 or 5 minutes of everything sound correct, just hope at least about 10 or 20 seconds are good. Then expand from there.

[polar spade]

By my definition then, you might have 1500 songs of those 2000 that have potential. 😁

[polar spade]

Man, I hope it all comes to some Eureka moment for you then. Thats whats been happening with Suno often for the past couple of years. Someone comes up with a workable system. 🩵

[small bridge]

is the app down? I can't get on no matter what I try have been trying for multiple hours, it worked fine yesterday

[plush field]

i am generating this, right now

[pure panther]

@neat nexus more scammers

Straight ninja

[dusk raptor]

I am planning to upgrade my plan from Pro to Premier and attempted to make a USD 30 payment. However, my card was declined, and I’m unsure of the reason. I have already contacted my local bank, and they confirmed that there are no issues with my credit card on their end. And i'm outside of USA, using VISA credit card, need help

[pure panther]

#archived-ticket-help This would be the best channel to reach out in, we're just users here in this channel.

[dusk raptor]

thank you so much!

[candid basalt]

Hey guys/gals quick question, is it better to make a song verse and chorus, then extend and do the next verse and chorus etc.

Or does it not matter? And just do it all at once?

I've had some success both ways, and had some failures both ways, and sometimes doing the whole song at once feels like impossible to get as a complete song

[small bridge]

No like my app wouldn't even open or load past the suno orange opening screen. But I'm sorry you are getting errors too! I guess we are stuck using pc instead of phone today. Suno man. so buggy, stil.

[restive pebble]

It depends on verse structure length etc. Suno like 4 lines of text per verse. It handles more without issue but sometimes it loses its sense of direction. It's then splitting the song up is a better option.
This is what many have found to work.

Usually Suno does a good job straight through..

[quiet solar]

It's a personal choice, to be honest.

Creating the whole song generally means everything will fit together more neatly and nicely than building it through extends, but building through extends give you greater control of each step in the process (ie, you can get the first verse and chorus 'right' before moving on to the next part) which can cut down on future editing work or having to keep regenerating everything if you have a really clear vision of what you're trying to make. You can always 'heal' differences between the extensions at the end by Remastering the full song.

Honestly though? Unless you're making very long songs, I'd say it'll generally be better (and easier) to create the entire song at once.

[frozen dagger]

just reloaded my webpage on desktop and this came up. so yeah, may be experiencing some technical difficulties at the moment.

[candid basalt]

Thanks that's a good point of view. I just feel sometimes I'm wasting credits on nothing

[pure panther]

It's not true either, no one can generate as far as I can tell.

Oh! I take it back, one I started generating 15 minutes ago just finished

So I guess some people can but it's sooooooooooooo sloooooooooooow

[frozen dagger]

yeah i would look at this as, what it says, but also, if it's like this, expect (possible) errors.

[candid basalt]

Yeah I have a song generating for like 20 minutes now

[daring birch]

Probably a bit too late now, but there are a few mis-matched [square brackets], with {curly brackets}, with a rogue "z" 🙂

[remote ivy]

do we think suno will generate a whale call if prompted in lyrics?

[sharp kettle]

oop typos just correct em when you type out [ insert here]. i'm a vocalist not a typist lol

https://tenor.com/view/freddie-mercury-vocal-improvisation-queen-dragon-attack-rock-star-gif-2705036175779027363

not

https://tenor.com/view/muppet-kermit-the-frog-type-typing-type-writer-gif-3991777

[polar spade]

As ticket -Help is on a delay, just checking if you've reviewed some options. https://discord.com/channels/1069381916492562582/1417458714616795210

The Bank likely doesnt have much to verify but , say you use Google SSO, Google checks both Sunos record and compares to that profile attached to your debit card. If one thing is different, a missing telephone number, the word "street" instead of "St" , your last name first on the card but first name first on Suno -- then it could decline it. Just make a best effort to see if there can be any difference.

So if you have Pro, whats keeping you from just clicking upgrade and letting it automatically charge the same already authorized card on the next renewal date? That usually is the more reliable means. I'll ask that rhetorically to ask yourself - i dont wish to know the answer. If thats not an option, then yes, you will need to email [e-mail] and wait for a response on their end if you cant discover anything your side.

Your online bank statement should have most of the personal info and your monthly receipt fro Stripe should mention your name email and address. Compare the two.

Once declined, its never good to try twice and chance the card nunber gets permanently blocked by your SSO (sing in authenticator). Best to wait al least an hour before trying again. Good time to check the details then.

Check suno-help channels including this one, for some insights and self troubleshooting ideas. https://discord.com/channels/1069381916492562582/1372994443656957952

[halcyon sundial]

any tip for suno editor run better?

its so laggy i cant even play the song

[kind lily]

Anyone have any tips to get it to consistantly prononounce words with U sounds like Houston and Fuel. UE and OU sounds usually get prononounced correctly the first time, and the second time they appear turn into O sounds.

[halcyon sundial]

i will punch my screen

[polar spade]

Try yoo....

Hyoostn, tyoozday, fyool

[halcyon sundial]

1 FPS suno editor

[pure panther]

Lag on a browser is usually (not always) caused by low memory on the user machine.

[halcyon sundial]

how

it requires 128 GB RAM?

[pure panther]

Do you know how to check performance? What operating system are you using?

RAM can be eaten by a run away process, even that much

[halcyon sundial]

4 days ago it was fine

why now is unusable

[pure panther]

Can you check to see if its a memory issue? I can help you troubleshoot if you like.

[halcyon sundial]

it seems fine for me

only suno lags

[pure panther]

If you're using Windows, you can hold Ctrl+Alt then hit delete, open the more detailed view, and look at the performance tab to see how much memory is being used when trying to play a song.

Oops, I forgot to mention, after the Ctrl+Alt then hit delete shortcut, you open the task manager

This is what you're looking for:

[halcyon sundial]

pretty ok

the website is broken

[pure panther]

Ok, so memory is ruled out, cool. Have you added any add ons to your browser lately?

[halcyon sundial]

i have a lot but only suno is open now

[pure panther]

By add ons I mean like dark mode or ad blocker

[halcyon sundial]

no

[pure panther]

They would be active no matter what site your on

Ahh ok.

Have you tried using a different browser?

I'm using Suno right now on Firefox and am having no issue with playback myself.

[halcyon sundial]

i tried brave

still the same thing

maybe im the chosen one

[pure panther]

So the next place I would check is internet speeds, are they throttled for you right now?

[tidal prism]

Is there a way to stop it hitting high notes? Especially in the chorus?
I have a few personas that have epic deep voices or ones that can just hit a note a certain way and it sounds so good

And when they sing high it doesn’t fit their voice and also sounds a little robotic

Especially like last word on each line in the chorus, i want it to have the same pitch (idk if that’s the right word) or lower then the rest of the line but Often it tries to hit a high note and it doesn’t fit

[frozen dagger]

Low register, Pitch contour (or melodic contour), Flat contour → stays level, Downward or neutral cadence → controlled, heavy, confident.

Tessitura (advanced but powerful)
This means the comfortable pitch range of a voice.
“That melody sits outside this voice’s natural tessitura.”

[limpid salmon]

Obviously I missed something, but Suno appears to be generating again.

[timid pond]

I seem to have a default vocal that I can't change. It doesn't matter if I choose a persona SUNO still generates with this "default" vocal. Can anyone explain this? Is there a solution? I have sent three emails on this issue and SUNO support is absolutely no help. They basically told me to come to Discord because they have too many requests.

[spark hamlet]

hypothetical question

do u have any methods of getting "exactly the sound" you need?

[restive pebble]

I find Suno sometimes locks itself into a vocal 'rut' and needs a good smack across the cheek to snap out of it. For that I clear the lyrics and styles, add a different genre that is on the far side of the musical spectrum (crooner style to acid metal) generate a couple of tracks, then switch back to my original song. This usually works.

Another thing to check is the consistency between the styles vocal description and the lyrics. If they are using different vocal cues, Suno will et confused and revert to a 'default' vocal.

Last thing I can suggest is the check the genere itself. Suno is trained on many genres but has specific 'rules' it follows to achieve the desired genre output. It uses vocals that 'fit' that genre and that can over-ride our prompts. (if your making a heavy rock piece and ask for a soft melodic vocal, suno has some trouble inserting it into the song)

[restive pebble]

Morning.
The only way I know how to get this is to make a track on your computer, upload it and make a cover of it pushing the audio slider high. Someone else may know of a different method..

[sharp kettle]

i saw this post, Microsoft upgraded there pc requirement the new minium is 16g ram, 12 gig vram, cpu is i7 13 or 14th gen and the 14k model or ryzen 7.... gpu list is specitic to the app your using but windows recommends a 4090 or above( the rtx 3060 is minimum spec btw)>>>

so if you having issues check you pc parts first the you internet provider next the bottle neck....
finding the optimal performe list is always a pain for any computer and application..

so just think of it like this pc part not from 2024, you are gonna have issues... and there a chip shortage so you wil be golding with a 2024 chipset untill atleast 2030, lol..

try cleans out some of your tempory caches that will speed up many arreas.. google thanks to ai now has a minium gpu, ram, and cpu requirment thaks to ai, but my gpu is still supported by google, so while some things are slow it mostly works lol

[sleek stump]

Someone else coined the phrase, but its still the best... "Suno isn't a word processor for music."
Suno is a creative model that uses its training to infer something it thinks is close to what a user is prompting. It often disregards what users ask for because it thinks what was asked for doesn't fit the genre/style its asked to generate. Suno is good at fusion, but success depends heavily on the prompting.
That said, it is possible to get 90+% close to what you want every time, once you learn how to talk to (prompt) Suno.
Another thing to keep in mind, Suno is not a fully released professional tool, it is still in heavy development, so as Suno evolves, so will the way we prompt it, and the output we get from it.
Experience is King! Experiment with Suno, learn what it does when you ask certain things! Enjoy the drive, and eventually you will figure out how to get (almost) exactly what you want, when you want it.

[safe maple]

does the } in [Verse 2][vocal: female, tenor][response} vs [Verse 1}[vocal: male][call]

make a different in the location of the }?

[sleek stump]

I would view that post with much skepticism. MS has not changed their min specs for PCs running Win 11.

[neat lake]

Hi guys, anyone knows why v4.5-all model generates few second clips and how to force it to generate longer outputs? It also ignores lyrics, which are clearly written in.

[sleek stump]

As far as I know, Suno doesnt really do anything with {}, it uses [] for meta tags, and () for lyrical operators (like harmonies or other vocal effects found in the genre).

[sleek stump]

I just ran a couple test tracks, and didnt see any issues. It may be all in what you are prompting and the contents of your Lyrics window. Can you share a link to a track that Suno gave you thats an issue?

[safe maple]

cheers, I thought so too in my ai investigations, I appreciate your response

[quiet solar]

I'd personally use [] for anything you're instructing Suno in the Lyrics box. Suno can pick up {} and () as style prompts, but it can also recognise them as backing singer lyrics. It's a good habit to get into always using [] to avoid the risk of Suno singing your prompts.

It's difficult to be sure about precision differences given that Suno won't always take lyrics prompts into account, but I'd personally suggest framing it like this:

[Verse 1 - anything you want to affect the entire verse in here]
Your lyrics here as usual

or

[anything you want to affect the specific line of lyrics here] Your lyrics here as usual

I'd be a bit careful of using call/response if you're not looking for the singers to alternate line by line, as that's often how Suno tends to interpret it. If you want each to sing the whole verse, I'd personally take those references out from both the lyrics and style prompts, but your mileage may vary.

[sleek stump]

That almost looks like a typo, also keep in mind that you would use these in pairs, an opening [ and a closing ]. Having a single } may just make Suno sing that character.

[quiet solar]

Be aware that AI will very happily and confidently give out bad advice for Suno prompting. 😉

[safe maple]

aye, I hate when it sings (C) as missed Chord

[safe maple]

Cheers Ginger

[quiet solar]

If AI tells you to put anything in { } or ( ) there's a good-to-fair chance it's hallucinating, and whatever it's telling you won't work.

As far as I'm aware, it's not possible to prompt for a specific chord, for example. Key signature and BPM in the style prompt, yes - at least sometimes. But not a specific chord progression.

[safe maple]

Cheers, It's been a wonderful learning curve, I'm still amazed and happy to come across new ideas

[safe maple]

I've been using, "(E5 - G5 - A5)
[EG: Snappy Attack Envelopes, 12ms]
Diagnostic hardware lines the floor (E/G#)
Bio-electric platforms roar (A/C#)
Strange matter stabilizers hold (D/F#)
Through fluid forcefields, dark and cold (Am7 → E7♭9)
The systematic probes descend (G#°7)
To save the heart of a fallen friend (C#m7 → F#7 → B5)
[Section: High Dynamic Range, Impactful]" But I'm not a trained ear, I hear it differentiate, but I don't know if it is following the instructions

[quiet solar]

And that's the other thing. The AI will dump in a load of things that may or may not work - it's great for inspiration, but you can get much more of a feel for how Suno is or isn't responding to you if you add things in, play around, and develop your own experience. That process is hugely valuable, and part of the fun.

[safe maple]

even after words, "The ticking (Em → G/B) clock, a fading (Cmaj7#5) spark"

[quiet solar]

I'm going to go out on a limb and say there's almost no chance Suno is respecting your requested chord progression. 😉

[safe maple]

aye, cheers, I've been finding the ride and experience a wonderful time

[safe maple]

you think the dev's will adhere perhaps in a later version?