📚┃suno-school

[solar nexus]

OK, I was expecting a download to appear, and then felt concern, lol

[stiff sentinel]

yeah, it should appear, just don't close the browser window i guess

[fair adder]

once again, i'm blind, and i dont use studio, and the editor as it is is not even accessible to screen readers. so, yes, @quiet solar helped me and now all i'm doing is remastering to find the perfect sound.

seriously, do i have to put it somewhere in my user name that i'm blind?

[frozen dagger]

check your filters if dont see it still, HIde stems may hiide it

usually goes to main wrokspace folder as well, but open the libray tab and check filters

[stiff sentinel]

you have midi files in there?

none of my midi files show up anywhere within suno

[frozen dagger]

all audio, uploads, stems, gens are saved to your Suno library

[stiff sentinel]

err can you should me what options/places/things you do to find your midi files?

[frozen dagger]

Go to library tab so it shows all songs and not have to dig through workspaces

[stiff sentinel]

if i disable all filters it still doesn't show me midi files

[frozen dagger]

duplicate your window (this is sfer than refreshing your current browser tab)

[stiff sentinel]

..? i'm asknig how i find previously generated midi files, i don't see how duplicating the window will help

[frozen dagger]

duplicating the window is the same effet as refreshing your brwoser tab (but without risking losing where youre at with what youre doing )

[stiff sentinel]

....yes i understand that but i don't see how that helps

[half dragon]

While Suno allows to convert and download midi - it doesn't support midi to start with so there shouldn't be any filter as it won't show

as I recall

[polar spade]

Ive been doing well to remind others of your situation as a matter of respect. 😁 Al I need is to remember whats not available to you. I dont keep a list handy. Sorry about that. Glad you've been helped.

Removing my post now. 🙏

[frozen dagger]

maybe i misunderstood. i think i replied to your response to the oringal question of where the midi's go. and yes they can take some time to load, but if not seeing them i was just trying to help with the fileters and where to find them when they are seperated.

[stiff sentinel]

yeah, i think you were answering the wrong question/user

[frozen dagger]

apologies late and watching a movie 🙂

[crude narwhal]

Thank you

[half tiger]

Hey guys, can someone tell me how I would write a style to change the genre of a song from one to another>?
https://suno.com/song/bd9f3dbb-a548-4c02-8904-be4627b83e70

Is an example of what I mean.

[quiet solar]

That totally depends on what genre you're trying to switch to, so can't really give you any detailed guidance.

But the basic approach is to select the original song in your Library, choose Remix/Edit -> Cover and then enter a completely new style prompt for what you want to convert the song into.

[half tiger]

Oh, I meant to genres in the same song like the one above. 🙂

[half dragon]

Maybe try to Cover the first song and use the Style from the 2nd song

[quiet solar]

That's an interesting one. You might want to try some mix of the following as a starting point:

Symphonic Industrial Metal, Neue Deutsche Härte, heavy guitar, martial drums, deep baritone vocals, operatic intensity, dark, dramatic, intense

Alternatively, it might be worth hunting down a song in the same genre space where someone has actually shared the original style prompt.

[lilac storm]

middle of the night school?!?

[quiet solar]

That depends on your timezone. xD

[slender glacier]

So to go from 4.5 or whatever to the v5, without changing any of the song....possible? How?

[stray wharf]

cover? but that's going to mess it up a bit (especially vocals quality), even if it'll stay somewhat the same

[quiet solar]

@slender glacier Your best starting point is to Remaster the 4.5 song using v5 on the 'Subtle' setting, but a change-free transition isn't really possible (and would defeat the point of making the Remaster).

If that doesn't get what you want, an alternative is a Cover with high audio influence (e.g. c.80%) and low influence (e.g. c. 15%) weirdness and style importance. But again, expect subtle changes.

[slender glacier]

I haven't seen a subtle setting for the remaster.I have to look for it.Whereabout is it located question mark

[quiet solar]

It comes up as an option when you select a song to Remaster. Trust me, once you've picked Remaster it's pretty hard to miss it and the model selection. 😉

[stray wharf]

Anyone tried making something like this with suno? Wondering how you'd prompt it aside from ethereal, all I had wasn't even close.

https://youtu.be/loClzgws5oc
https://www.youtube.com/watch?v=LBnr_Hc9BA0

I could make voice somewhat close, but tempo, note sustain \ overall feel isn't there, it's never quite as "slow"

[quiet solar]

Ethereal is likely to be closely associated with the vocal style rather than the song as a whole.

It's a little tricky to know exactly what to prompt for, because it's a kind-of modern twist on a Greco-Roman aria, but you could try something like:

"Haunting cinematic aria, solo female operatic vocals, ethereal contralto, long sustained notes, melismatic, ancient Roman aesthetic, crisp modern EQ, high fidelity, mournful, minimalist"

[stray wharf]

yea, was referring to ethereal as vocal style, probably ethereal, airy, doesn't have to be operatic, but those 2 together often leads to operatic anyway it seems.
Gonna try few things.

[quiet solar]

You might also want to prompt for something like "slow harp arpeggio" and "reverb" if you're aiming to mimic the instrumental sound

Yeah, it's a toughie because it's quite an unusual style, so it's working out what you can point Suno towards that sort-of similar without bringing in too much from other styles. Good luck with your experiments. 🙂

[slender glacier]

Every time I hit remaster, it just automatically does it.I don't get it any choice or anything

[stray wharf]

biggest problem so far is to make it as slow, to make it sustain basically every note

[quiet solar]

This is on the desktop/browser version, but not sure why it wouldn't work the same way for you?

[quiet solar]

Try adding "Largo" and "60 BPM" (which is about what that song is at) to the style prompt. As suggested earlier, add "long sustained notes" somewhere in the mix as well and try to limit the amount of words/syllables per line, which should help.

Alternatively, if you're up for it, sing the melody you want yourself and upload that as an audio prompt. Suno is much easier to steer with direct audio than prompting.

[slender glacier]

Oh maybe I should try that this is all I have for the app, I didn't know there was a desk top version

But thanks

[stray wharf]

gonna be hard to sing that as a male 😄

[quiet solar]

Doesn't matter how shoddy it is, Suno will happily switch over the voice to a female singer if that's who you tell it you want to do the singing.

Trust me, I've done this several times to steer a female singer. You really, REALLY don't want to hear what the input sounded like, but the output sounded great. 😂

[quiet solar]

The browser version is the same as the desktop version, so give that a try.

[half tiger]

Is there a way to keep the same female vocalist for every one of my song covers? I just spent 750 credits trying and failing.

https://suno.com/playlist/aba23a2a-72b3-4173-bf52-d4787cd0c9fe

The male singer seems consistent, the female singer does not, and I don't know how to replicate her deeper vocal cadence in the second and third song, and it's very frustrating.

[wintry apex]

How do you use Styles / Persona during edits on the Edit screen?

its literally not possible to replace lyrics or have edit be useful in anyway if i cant use the same style i used to make the section in the first place. if it cant even use the persona, this edit function is only good for changing volume and speed and thats it.

All they have to do is put the persona button on it.

or maybe they have to do more .. i had a look at the code. it doesnt even include a persona entry or null persona entry on the edit page in the call.. idk why T_T

i bet its in the studio.. but the thing is, 30 is way out of my budget man, after already spending like 60 in half of a month on credit top ups

just to get a button T_T

[restive pebble]

You could use a Persona from the song with the vocals you wish to remain consistent. It's under the ... > Create> Persona
You must be a Pro subscriber for this.

[half tiger]

I appreciate it, but I created 5 personas and not one of them was consistent.

[wintry apex]

maybe theres a secret way to do it though - thats why im asking here.

like @personaName in the styles box on the edit screen?

[restive pebble]

Mine are.. Wonder why...

[half tiger]

Me too.

[restive pebble]

are you selecting the part of the vocals you are trying to copy?

[wintry apex]

my personas work when im allowed to use them 🙂 but keep in mind - the persona is not only using the 10 second selection of 'the best part' - its using the entire audio file

So if you want it consistent its best to crop the file so its only the voice you want

[restive pebble]

The higher the audio slider the closer the vocal consistency is

[half tiger]

Yes. But of course the Persona is greyed out, and I'm a pro member.

[restive pebble]

Yes, Use only the vocal stem, then crop what you want.

[wintry apex]

you have to make a cover of the upload and make a persona from that

[half tiger]

I. Did.

Seriously, I spent 750 credits on the same song. Nothing is working.

[restive pebble]

pro member, grayed out?
Log out, clear cache, log back in?

[half tiger]

I'll try the Get Stems though.

[wintry apex]

if thats still greyed out, you might have to make another cover of the cover

[restive pebble]

yes it could be that specific song cover..

[half tiger]

Making a cover of the cover is exactly what I did. And it didn't help.

[wintry apex]

im not sure how it decides what cover gets to be greyed out though. seems random

maybe they recently nerfed it

[restive pebble]

make a cover from the original using the persona from the stem. See if that works

Yeah some songs just don't have the option for persona. Never thought much about it. Just kept on genning songs.

[wintry apex]

they wait till a certain lawyer logs out and then relax the rules xD jk

[restive pebble]

uh oh... we got the big boys joinin in. lol

[hard beacon]

The only reason i know (greyed button) is when u try to create a persona out of a Suno Studio created song or when you try to make a persona out of another user-track (for example from an official remix contest)

[half tiger]

Uploaded songs, at least on the Web Version, cannot be used to create Personas. I just tried.

[wintry apex]

its greyed out for me on any direct upload, even if its from my own original content

[restive pebble]

All in all, I find Suno to do what I ask of it more often than not.

[half tiger]

Shocking.

[hard beacon]

ahh.. ok.. personas from uploads are not possible. (as far as i know)

[wintry apex]

Its just a way to force people to spend 2x more credits imo

[restive pebble]

Probably because Suno did not generate it and would have to take time to render it

[hard beacon]

ahh... RockhardStallion just confirmed it 🙂

[wintry apex]

They did work a few days ago. About 1 week ago I could make personas from any uploaded file.

[half tiger]

This pisses me off.

[restive pebble]

Suno is a random generator so nothing is a guarantee with the output

[half dragon]

I believe upload with vocal --> no Persona

[half tiger]

Yet you just said you don't usually have issues.

[wintry apex]

nah its any upload. it does the same thing for purely electronic non copyrighted experimental psychedelic trance lol

but its been easy for me to get past it

[half tiger]

How?

[wintry apex]

by doing 'add instrumental' or 'add vocal'

[restive pebble]

I don't have many issues and those I have I rethink and come at differently by asking on here.

So much information has gotten results

[half tiger]

That wouldn't work for me, since.. Alright, let me just show you what I'm trying to do..

[wintry apex]

so far it usually works for me

ok

[restive pebble]

Uploads I don't use. I did try my vocals once and Suno did make it usable but I couldn't make a persona from it

I stopped there

[wintry apex]

i have like 300 personas of everyone i know lol

[restive pebble]

WOW

[wintry apex]

i have a persona for the dog and the dog's grandma

[restive pebble]

Just wait for Dasvon to post.. Petty sure he has a solid answer to your question

[wintry apex]

All im trying to figure out though is how to use personas on the edit page, since i need it to replace lyrics without completely ruining it

[half tiger]

https://suno.com/s/20XdQ1culryoNX8w
I literally have no idea how to replicate her voice without using the Persona, and even if I could upload it? It wouldn't work.

I would like that voice Exactly. It's very irritating how it's easy to replicate the same male voice, but ask it to do a female? Pfft. Good luck with that..

[restive pebble]

I'm running a gen from it to see if I get something close or close enough to make a persona from

it's a great vocal so I see why you want to capture it

[half tiger]

Sure, I'd be very curious to see how much luck you have. And thanks! Yeah, I prefer that female vocal style, not the girly girly ones lol,

[restive pebble]

for the persona inside Edit, I will check if I can figure it out

[restive pebble]

but it's still a girl

[half dragon]

Using Persona in Editor is not possible - never was. The feature isn't available there.

[half tiger]

It is, but she sounds 'strong' somehow, like she can handle $h!t. 😛

[restive pebble]

Appreciate the save !!

[restive pebble]

That's a nope

[half tiger]

Indeed. Thanks for the try. I guess the only way to replicate the voice is to make it a persona, which is impossible, yes?

[restive pebble]

If you have the original track generated, Suno should let you make a persona from it.

[half tiger]

Original track was a song I uploaded.

[restive pebble]

But again you stated you made several and none worked

[half tiger]

SO I'm buggered both front and back, sadly.

Yes.

[restive pebble]

Not gonna happen

from what I know

[half tiger]

Now do we know why?

Especially because it can use the same MALE vocals with no problem.

[restive pebble]

Probably disabled because to make Suno gen a persona suno may need to have created the vocal track

[half tiger]

So a track I created elsewhere wouldn't cut it. What a stupid rule, oh well.

[viscid sequoia]

What's the picture size for song?

[half tiger]

Not even extracting stems would help, huh?

[restive pebble]

Sometimes I burn through thousands of credits just to find a melody that fits the vision I have.
What I'd suggest is to use that track you uploaded and make a cover pushing the audio slider up to around 77%, weirdness down to 15% or lower and style at least 55%.
You should get the vocals really close

[half tiger]

Do you mean the file size? I usually just upload my own covers.

[restive pebble]

it auto adjusts and centers it.

[half tiger]

But if it's a cover, wouldn't it be Audio 100, because the Style box would be empty.

[hard beacon]

i would use 1920:1080 🙂

ahmm.. 1080:1920

😄

[wintry apex]

https://suno.com/persona/9b27ef26-a583-440b-a314-94f038aa7b6a

[restive pebble]

maxxed sliders usually have plenty of issues. Try it and see for yourself..

[viscid sequoia]

Little to big

[restive pebble]

or is the window too small?

You mean the exact size. I never measured or tried to fit somethign..

[half tiger]

Oh, I know, lol. But alright, I'll try to cover that song, with the sliders you mentioned and see what happens. Going to wait and buy credits later, I'm not very happy with Suno at the moment.

[restive pebble]

I know it's 9:16

[half tiger]

That doesn't have any personas in it.

[half tiger]

That sounds wonderful, but nothing like my song.

[restive pebble]

Remember, Suno is always a random generator before t's a servant of obedience. We can merely suggest and hope Suno plays nice. lol

[wintry apex]

.>

[wintry apex]

its a voice that can be put into any song- you just have to drop it into the right musical context

https://suno.com/persona/9b27ef26-a583-440b-a314-94f038aa7b6a

unless you wanted the male voice AND her voice stacked like in the song - but you didnt say that, you said 'her' lol

hopefully this helps. i made it easily, i spent 20-30 credits to get it though

[polar spade]

I presume the female starts about 1:19 in the first song and on verse one early in on the third song?

The second song there only seemed like female harmonizing and no solo bits, rights?

By the way, using "share current time", is helpful to mark what we're pointing out, or at least text the timemarks

Were you covering the first song , plus adding all new lyrics and the persona made from the first song to get the second? Or you created all three songs from a single persona and no covered audio?

If the latter, try the former.

[vital flume]

Suno studios won't load on my android phone. Do I need to go through a pc?

[restive pebble]

On my droid app, I don't have a studio option but maybe am not looking in the right place. On PC I have no issues

[vital flume]

For suno studios you have go through the internet but it still doesn't work on my phone. So I might have to go through my pc.

[restive pebble]

Don't you need internet for suno

[wintry apex]

https://suno.com/s/JA5NoYd4xZjypDTB

[restive pebble]

not just studio

[vital flume]

You can't go through app for suno studios. You have to go through website.

[restive pebble]

right

[vital flume]

But it still doesn't work on phone.

[restive pebble]

you mean the browser not the app on te droid

[vital flume]

yes

[quiet solar]

I'm about 95% sure the block on making Personas is to prevent someone, hypothetically speaking, uploading a song by [Insert Famous Artist Here], creating a persona of their voice and then churning out a ton of fakes to trick people into playing the songs on streaming services so they can make a quick scammy buck.

Neither [Insert Famous Artist Here], or the streaming services, or their reassuringly expensive lawyers would be very impressed if Suno facilitated that.

Out of curiosity, where did you get the original voice from if it wasn't generated by Suno at any point in it's life?

[restive pebble]

No option in my browser

[half tiger]

Sorry, Internet borked.

By the way.. Is it true that creating new vocals forces a different vocal performance? I made like.. 3 different covers of the same song, and got the same vocal performance, but adding different lyrics gave me something completely different, and I never touched the style.

I mean, I know how to mutate lyrics, I just shouldn't have to..

[wintry apex]

yeah. like you can just put [Yelling] blah blah

[explosion]

[chopped and screwed]
blah blah blah

the lyrics are its own prompt

[half tiger]

Nah, not that easy. It has to be very specific, I mean vowels and syllables have to match.

[restive pebble]

I use the audio slider to keep it close

that's only for covers

new gens are a dice roll

[half tiger]

Yeah, but covering the same song with the same slider setup matches the female vocal exactly. Adding new lyrics changes it completely.

[restive pebble]

not sure what else to tell you. I alter my lyrics alot when making covers and find the vocals remain pretty consistent with the slider settings.

try genning something totally different like a diferent song. Make a few of them. It might clear some residual stuff in Suno's memory banks...

then return to your song and try it again

Just a thought

[half tiger]

How do you solve the pacing issues when you make covers?>

Because syllables matter.

[restive pebble]

it follows the pacing mostly almost every gen

[half tiger]

If the new song has like.. 15 syllables on the first line and the original had only 9, you get them singing rapid lines to make up the syllable gap.

[half tiger]

That's odd.

[restive pebble]

That's bc it tries to maintain the song structure I think

[half tiger]

Yes.

[restive pebble]

You can only fit so much into a bar without forcing the vocals speed to increase

15 syllables is a lot for a line to be sung

just sayin

a mouthful man

[half tiger]

That was an example, lol, and a lot of my songs are in Old Norse.

https://suno.com/s/LFhzXdFCkv5G4joQ?time=243

Not this one. Lol.

[hard beacon]

by the way RockhardStallion: your profil pic is killing me.. i don't wanna see that 😄

[half tiger]

Cheers. 🙂

[hard beacon]

hahaha

😄

[half tiger]

I'm writing new songs for this playlist, since I promised Trevor I'd bring her back to life. xD

[wintry apex]

check out what I made with your persona xD
https://suno.com/s/EdwdlWtuRHwAFQuY

[half tiger]

Sure. 🙂

By the way, how would I force two distinctive Female Voices?

[wintry apex]

hmm..

give them names

[half tiger]

?

I did.

https://suno.com/s/SXragWYJVzGAdj5y

[restive pebble]

You could either make two separate songs with different vocals or try using something like this:

Styles specify two distinct voices
This song is a duet with two female vocalists. Female 1 is a tenor with a strong melisma. Female 2 is an alto with a diatonic shift into Phrygian

In the lyrics call each out where they belong
[Verse | Vocals: female 1]

[Verse | Vocals: female 2]

[half tiger]

I was hoping for a less complicated method. xD

[restive pebble]

the trick is to get suno to understand by reinforcing the desired outcome. If it sees it in the styles and then also the lyrics, the change of Suno doing it improves.

[wintry apex]

Prompt: Conversational Duet

[Verse 1: Sara & Amber]
(Sarah):  Do not answer me back
(Sarah): I'm Serious

(Amber):  Ok I wont
(Amber): Just kidding

[half tiger]

Hmm, good advice you guys. let me try something..

[restive pebble]

does you wife know about sarah and amber?

[wintry apex]

haha

max generic

[restive pebble]

lol

gonna try to get some work done.

[half tiger]

Thank you both for everything. 🙂

[polar spade]

"Two district female voices" yea... but in what capacity?

So if just a duet with no overlapping voice and they simply alternate between verses, they may be easier than , alternating between lines or harmonizing. The latter two being more about lyric structure and the former sbout tagging snd line spacing.

So maybe start with describing the end result ytoure lookin for.. Then we might try looking for the minimal amd least complicated first.

Keeping in mind, every suggestion is with about a 15% average success rate going in. Thats always presumed before answering.

[half tiger]

https://suno.com/s/ygT9eyZ1LgLdMI0R
Check this out. 🙂 This will answer what I am trying to do.

[stiff sentinel]

so many people have been saying since yesterday to make two songs and then combine the vocals the way you like

its not that complicated and you have much finer control since you can get each voice the way you like

doing it in a single generation, theres like a 10% chance it'll split the voices correctly

and then there's another say 20% chance that each voice will be the way you like it

thats 10% × 20% × 20%

you're basically rolling for something like a 1 in 250 chance

...with that long of a song its way less than 10% to get the split correct

[half tiger]

Interesting. And how do I get the vocals from a cover song I can't create the persona of? Making 2 songs or 10 won't do anything to help with that.

[stiff sentinel]

so make one version where female 1 sings the whole song

you might have to roll it multiple times to get the right type of voice if you don't have a persona

but at least its just one variable to get correct

do the same for the female 2 version

you might even get a good female 2 version when trying to get female 1

then extract the vocal and background stems and splice them in studio

you might need to fiddle with the audio slider % settings so that the cover songs match the original backtrack properly (you'd need a higher setting here)

[vestal beacon]

What's the best way (in Suno Studio if possible) to make an instrumental track loop such that you can't tell the music has ended at all?

[stiff sentinel]

but at least you're doing one variable at a time, and then cherry picking manually

instead of doing every variable at once and leaving suno to cherry pick probably incorrectly

[stiff sentinel]

like an 8-bar loop or something?

[half tiger]

Also, if it actually worked, I wouldn't keep bloody asking, would I.

[stiff sentinel]

maybe i missed it, but i haven't seen you attempt the make-two-covers method

[vestal beacon]

Longer - like 2 min of background music for a game that loops seamlessly. I generated a track I like.. but not sure how to get it to loop well

[stiff sentinel]

do you have access to studio?

[vestal beacon]

yes

[stiff sentinel]

suno by default (and modern music in general) structures music in sections of 4

if you can like it up properly to the beats (which it might have done automatically) it would be like a 32-bar or 64-bar section that repeats nicely

are you able to share the song?

[vestal beacon]

sure. here is the project. https://suno.com/studio?project_version_id=9c962824-4dc5-4e8f-828f-833f1ba12494 does that work?

[stiff sentinel]

yeah i just need a quick listen

oh wow

i can see other people's studio projects?

hey, nice song btw

[vestal beacon]

thanks. its for a shop experience in a video game. needs to be chill. background will be clouds/celestial

[stiff sentinel]

at 65 thats when the song starts working towards the outro

so looping at 65 will be the smoothest

i gave it that one bar offset since the first note of the song tends to hit a bit differently

[vestal beacon]

Idk how you knew that and/or did it...

[stiff sentinel]

you can play around but it will be somewhere around that point

[stiff sentinel]

because of this principle

[vestal beacon]

ok cool. did you update the project? I don't see changes.

[stiff sentinel]

i had to make a copy, lemme see if i can share it back

https://suno.com/studio?project_version_id=302bd7ab-94ce-41f5-b8a9-9e3e86c5c0e5

(track 3 is the first note of the loop, i put it there for the sake of the screenshot)

[vestal beacon]

I got it. was able to copy it back.

thank you very much man.

[stiff sentinel]

glad i could help

[raven turtle]

Is it me or does SUno just not get early 80s European electronic music? No matter what I prompt and put in the negative section I keep getting glossy synthwave, with DX stabs and simmons tom rolls. I assume Suno is trained more on American synthpop as opposed to the bleak post 1970s urban electronic music coming from UK back then. But no matter what I do, I cant get that early Human League, Cabaret Voltaire, or the 'Some Bizarre' album kind of vibe. I keep getting more glossy DX chords and DX bell synthwave pop. I guess I'll still be making it by hand then in Logic, unless anyone has had any luck promoting for this kind of thing?

[stiff sentinel]

idk if i can help since i don't know the genre, but can you example of something in the general style you're looking for?

[raven turtle]

Things like this:

https://open.spotify.com/track/6ROzlGWUDvFylDRrX5alYB?si=t9sMuzXyRTGCMVCNFcw_Mg

https://open.spotify.com/track/0P19NrR3xzELtF04rtV0qG?si=Zdw9vqTfSJy_-jRUNj_Icw

https://open.spotify.com/track/5HYc25zVglUu4Ao0jY39i6?si=P7fEMm95TUanNhfdF1-6OQ

Early 80s post punk European electronic. Very indie. The opposite of synth wave gloss

[stiff sentinel]

i threw your description into SunoGPT and i got this:
https://suno.com/s/jh614u1ChTLcMwU4

is that a better starting point?

[raven turtle]

Closer but still too polished. Too modern. I might try v4 and see if it’s any better TBH
Thanks

[stiff sentinel]

i got the prompt from this, so until someone familiar with the genre comes along you can try describing it to this chatgpt preset

[forest hemlock]

https://suno.com/s/SDyAKSzdbWpCVVZs
https://suno.com/s/2IrMVIRQB11e8T9M

The two ones I got, but also not sure if that's exactly what you're aiming for, probably sounds to modern too

[frozen dagger]

if you're able to create a sample of audio in Logic and import that to Suno may have better experience. If having a hard time text prompting providing an audio may be worth a shot.

[raven turtle]

That’s what I’ve “traditionally done” as I guess Suno hasn’t really trained on much of this kind of stuff. Pioneering dark post punk electronic

Thanks for the help all 👍

[frozen dagger]

the text descript it gives when uploading audio not much help? Sounds like you have good wording but it's just not pulling from the training where you wanting.

[raven turtle]

That’s really good! A little glossy but I get the direction those prompt results are taking! Cheers

[raven turtle]

That’s what I figure. Probably have to wait for Suno to pull a deal with all the small indie studios sitting on dusty post punk master tapes about 10of us listen to! Hah

[lilac storm]

school is in session? do I have to pay?

[patent leaf]

any way to get cleaner stem extract within studio? it likes to put midbass into the synth stem alot

[quiet solar]

@patent leaf Not that I'm aware of.

As far as I'm aware, Suno doesn't actually create individual stems before combining them into a song - it generates the whole thing at once.

So 'extract stems' is actually the AI trying to retrospectively break apart the individual parts of the song, which is why sounds can easily get mis-assigned between different stem sources, especially when there's a lot going on.

You could try to clean up individual stems with a Remaster or a high audio influence Cover and being very careful with how you use the style prompt, then reimport your improved stem back into Studio to work with. Not my specific area of expertise though

[patent leaf]

i understand that, i mean even when i try to seperately gen in studio an extra bass to try and isolate and specify synth or midbass in style tag it still sticks too closely to the initially full stem generated merged sounding one

[quiet solar]

Yeah, I've never really that much success with refining specific stems. It's one of those things that feels like it should be simple in principle, but in practice...

[frozen dagger]

makes me curious if EQ effects made within Studio are applied when extracting stems. could adjusting it be lo-pass, or high-pass, before extracting stems make a difference? not my area of expertise either. just made me think.

[sour dock]

Hmmmm…. Something ain’t adding up here

[quiet solar]

I suspect the answer is "sort-of". Similar to how stems aren't really stems, I assume that EQ and other effects aren't really effects being applied to the underlying song, but Suno mimicking the end result of how they sound when they are.

Whether that comes out cleanly into the correct stems is, I suspect, going to be hit and miss. Suno's direction of travel does seem to be towards giving more control and influence at stem level, so hopefully this is something that'll get developed in future versions.

[analog quest]

Ummmm how.....

[sour dock]

lol

Bugged ah gen

[frozen dagger]

it is an odd wave form. like legos at the end there, duplos building blocks even.

[sour dock]

Yeah its just a constant bass tone that's gets abruptly louder lol

if you look closely, the first chunk is a perfect circle

[quiet solar]

Looks like Suno's inner HAL9000 coming out to play. 😵‍💫

[zenith trout]

Why always this error???

[quiet solar]

I've only ever seen that where there's an error on Suno-side. Might be worth taking a break and coming back in a short while.

[stray wharf]

Reporting back, I figured out "slow" part for the most part.
I kinda...stumbled into a bit different direction, more dark, Persian style, but tried more classic style too, with harp (angelic feel thingy) - and angelic worked better \ closer to Quiet moon.

It seems like this style a bit breaks suno, there are so many flaws with voice, inconsistencies, sudden weird changes, often robotic voice
I guess from there is is just iteration + figuring out lyrics and making it pronounce it properly.

1:55+ there isn't too bad:
https://suno.com/s/4WBOPjmgNTAoOoss

But I just also realized I kinda cheated similar style before, forcing sustain with just "haaaaaaaaa" lyrics lol (~1 min mark)
https://suno.com/s/hZ9tY2QccOLBLUfc

[night turret]

I suspect there was a silent update or a model regression around Jan 28/29, 2026, specifically affecting the Cover function.I have a stable workflow that was working perfectly up until Jan 28. However, starting yesterday (Jan 29) and continuing today (Jan 30), the results have degraded significantly.

Here is my testing context:
Reference Audio: Same track used previously.
Prompts: Identical Style and Lyrics prompts.
Persona: Same Persona ID.
Settings: Exact same parameters.

The Issues Observed (Jan 29 - Jan 30):
Drastic Style Shift: Despite using the exact same inputs, the output style is completely different from what I generated just two days ago. It feels like the model is ignoring the nuance of the prompts.
Musical Errors: There is a much higher probability of wrong chords and dissonance. The musicality has become "unpleasant" to listen to.
Consistency: I tested this across several of my other mature prompt templates (which usually produce consistent high-quality results), and they all suffer from the same degradation.

Has anyone else experienced this sudden shift in the last 48 hours? It feels like the logic for how the Cover function interprets style has been fundamentally changed or broken.

I can narrow it down even further. It was working fine during the early morning of Jan 29 (UTC+8), but the quality dropped specifically in the afternoon.

This translates to late Tuesday night / early Wednesday morning (EST) in the US. That is a classic overnight maintenance/patch window. The timing is too coincidental to be a random network glitch.

I discussed this with a professional AI music arrangement community on TikTok. The consensus matches my findings exactly, and everyone noticed the shift at the same time. This pretty much confirms a silent model update or backend parameter tweak.

[raven turtle]

Agreed. Feels like sub genres have been homogenised, prompts ignored and v5 personas applied to covers just doesn’t work

[gusty spear]

I'm not that tech savvy but thanks for confirming that I'm not crazy thinking it sounds worse to make covers suddenly

[sonic folio]

does anyone have an idea to how to prompt in the lyrics to gradually go from singing to screaming?

[lofty bison]

Try small words to CAPS

[restive pebble]

You could try using a vocal tag on the verse.
[Vocals: Dynamic or Soaring]

[wintry saddle]

caps, and more exclmation marks after words

[Shouting] can help

and genre will matter

if you need a scream you can elongate vowels too like

i need power --> I Neeeeeeed POWERRRR!

but you need to check how suno decided to mess with your pronunciation when you do that 😄

[lost badge]

add keywords like rage and anger

[stray wharf]

I'd say caps is worst option, as it often messing up the flow (idk why, it just does)
[screaming] before lyrics
seems to work better for me, while prompt itself is more of a like overall style guide

[quiet solar]

Personally I tend to combine the [screamed] tag with block caps, although depends a bit on genre and exactly what you're after.

[quick wedge]

Morning Suno School. Thought I’d pop in to say hello and morning to you all. 👋😊

[bronze island]

Hi Suno classmates

[vestal beacon]

Hi. I have a track here I made in Suno. It's for a videogame mainmenu screen. I like it, but, it gets a bit too energetic at the 1:50 mark. It really ramps up (starts at 1:36). What's the best strategy to fix something like this so its more mellow/consistent for 2-3 straight minutes? In Suno Studio, I tried 'replacing' from 1:30 onward, and it was 'okay' but not great. I used the same prompt (in styles) portion that I used when I created the song.

[polar spade]

It's all possible. 5.0 is still in beta. This means changes without notice.

[desert vector]

i have pro, why wont it let me create persona

[neat nexus]

Maybe you could try clipping it before the increase in speed and try mashing it up with something at a more suitable tempo and seeing if it helps? Use the original prompt.

Alternatively maybe you could try adding one of the tempos to your prompt and see if it works?

https://en.wikipedia.org/wiki/Tempo#Approximately_from_the_slowest_to_the_fastest

Suno doesn't really do BPM reliably but I think using these tends to get me closer.

[polar spade]

Typically this indicates an uploaded file. Especially uploaded or recorded audio vocals can not be directly made into a Persona.

[desert vector]

weird

its just vocals

well the vocal im tryna make persona is ai

[polar spade]

A blanket censoring to stop any attempts to clone and create a publicity issue. Typical for many sites.

[neat nexus]

Try editing it using the editor and crop it. See if it works after that.

[glossy berry]

does anyone have any pointers on how to make juckle sound better ive tried different prompts but it always ends up muddy sounding

[polar spade]

It doesnt recognize even its own created vocals when uploaded as an audio.

[polar spade]

Thats a new term for me. What's "juckle" ?

[glossy berry]

basically an almost noise rock type of genre very avant garde and often with screaming vocals

it used to work for me but now every time it either comes out or just becomes a metal track

[neat nexus]

Search brings up nothing...

A good rule of thumb is if you can't search for it and have it show up in the first page or two of search results it's not going to be common enough to 'just work' reliably.

You might have better results if you sit down with a GPT model and try to refine a solid style prompt for it? I had some decent luck with doing that with stuff in the past, with an emphasis on 'parent' genres that the style was a sub of. But you're not going to be able to just toss that in the style prompt as is and have it work.

If it did before that's because you got lucky.

[glossy berry]

i have an example i can send if that would help

[neat nexus]

Go ahead. If you have any artists who have songs that are in the style that would also help.

[glossy berry]

https://suno.com/s/y7ww57qBQ5ZuHR7I

[vestal beacon]

what do you mean by mashing it up? Like generate another track (in studio) with prhaps a different instrument?

[polar spade]

You can find a song on youtube perhaps and feed it into a genre AI analysis website to spit out some idea that get more focused on the genre and not the distortion or loudness bits. Just search "AI song genre analysis".

Also just throw in techy style prompts like "sparse arrangement, crisp lead vocals, close-mic'd, high-pass filter, tight analog compression, de-essed sibilance", and simpler such as " clear vocals, crisp, clean production hi-fi master" and exclude things such as distortion and muddiness.

Could avant-garde be throwing it off where you might try "improvisational"? But about the same. Worth the try, I suppose.

[glossy berry]

oh thanks chuck i’ll try that preciate it

[polar spade]

You could try to exclude "warble" , "phase shifting" and " wobble" as well, from listening to the sample.

For fun you could slide Weirdness up to 55% to see if it gives an avant-garde appeal without making it muddy.

[vestal beacon]

How do you create a cover of a song in Suno?

[half tiger]

Hey guys, does anyone know how I can fix cover songs that keep cutting off? This is getting really annoying.

[half tiger]

Remix/Edit, then go down to 'Cover'. 🙂

"Capacity Peaks: When Suno servers are under high demand, the "Covers" tool (which is much more resource-intensive than standard generation) may time out early to free up processing power."

Is this true?

[sonic folio]

I'll give it a shot 😊

[neat nexus]

Sorry, had to step away for a bit.

Mashup is a new feature where you can prompt with two different audio clips. This can have good results. So I was suggesting you try loading in the clipped version and a slower piece and see if that gave you the result you were looking for.

[neat nexus]

You use the hamburger menu (the three stacked dots?) and select Remix/Editt -> Cover

[magic marsh]

I think i need school this prompt is too mich but makes good songs :Complextro crab scratch , glitchcore,agressive phonkAggressive phonk complextro GLITCHCORE glitchcore vocal chopsHigh-Octane Frutiger Aero Phonk, 160 BPM, liquid jersey club beat, aggressive distorted 808s over airy bossa nova chords, Wii U menu ambiance, soda pop fizziness, sparkling aquatic synths, glitchcore stuttering, complex breakcore drum fills, pitched-down Memphis rap layered with saccharine K-Pop vocal chops, high-gloss complextro textures, abrasive digital grit, nostalgic 2000s electronics, crystal-clear water droplets, hyper-kinetic and neon

[vocal summit]

Hello, I have a question, so I want to make a song with like simmilar type of beat and instruments of a certain afrobeat song. How can i imitate the beat in the background and like when and how the instruments go?

[stiff sentinel]

(and @wise kindle too, hope you don't mind the @)
pretty much finished the crowd chanting song i was working on, it was a huge production rabbithole (2000+ credits spent) but I figured out a lot about how Suno could be used in professional music recording... I was making, organising, editing samples, and even exported MIDI into a DAW to clean up some wrong piano notes. Thanks so much for your help, I kept referring to your sample generations throughout the process.

If you (or anyone) doesn't mind, before I publish I'm looking for feedback specifically on the balance between main vocals and how much of the audience its picking up. I have the channels completely separated in studio so I can adjust them individually:
https://suno.com/s/222P1PNgTCoKGi7i

.... that wasn't a shortcut url woody

you erased my paragraph of a message trying to help someone

[stiff sentinel]

if its an established genre, describe it in the style section, and if you can write (or edit) lyrics that match that style that'll help too
if you don't know the genre, there are GPT tools that can help (i can't link otherwise my message will get erased arghhhh)
alternatively, post some examples of the type of song you want to make (either SUNO or otherwise) and people that have experimented with that genre (or peopel that want to experiment) can help identify some good style tags to use

[vocal summit]

@stiff sentinel can you link them to me in private?

[stiff sentinel]

gonna continue the conversation here so others can join in, and bc i'm not familiar at all with afrobeat
you can specify the instruments you want in style (not a guarantee but it should make things less random)
It also should help if you boost up style slider and reduce weirdness (I usually start with 40% weirdness 75% style)

if you can post some examples of what you want, that would be really helpful, its really hard for other people to hear exactly what you have in mind without any specific references

[polar spade]

It may be just me, but since you have Studio. See if you can use Sound and run some "audience" SFX to blend some as a separate track.... or replace altogether the beginning audience response where she is talking, unless you feel its appropriate Japanese or Korean crowd response to cheer in unison and die back quickly. Thats a British style somewhat too, but not for American audience response.

1-3 second mark was an excellent graduation of audience cheers starting out, then it went to silence very quickly. Thats a 4.5 and 5.0 thing which hasnt improved. It would be good that it dies down a bit but she still needs to talk over it some. In Japan that might seem rude though, so their audience might get quieter faster and so this sounds consistent to Japanese audience response.

At 3:30 I thought the unison medley of crowd singing a different line that the lead was really nice, and at 4:00 was a nice call-and-response from the crowd as well.

Personally I kind of like the crowd being a bit far back like ypu have them and getting to hear mostly the lead vocals, so thats good.

[vocal summit]

well i would like something easy going in the beginning like the guitar in wizkid - joro or something easy going like in Omah Lay - Understand that type but i dont know how to imitate the beginning

[stiff sentinel]

I've been layering different audience sounds to get the "right" kind of fadeout, hard to do in studio since we don't have like "keyframe volume curves" (idk what the term is but i think you know what i mean)
it worked throughout the rest of the song to simply layer audience sounds over, but without any music in the beginning the fadeout is indeed very apparent
I'll make a few crowd fade-out samples manually with a longer decay in my desktop DAW specifically for the MC since its more noticeable
glad to know the balance of the rest of the song sounds good to you, i'm very soon going to be in the "now-that-i-know-how-to-do-something-somewhat-well-i-don't-want-to-do-it-again-for-a-month" phase 😂

[vocal summit]

I dont have an example that i can send but that is the type.

[stiff sentinel]

youtube is safe to send https://www.youtube.com/watch?v=FCUk7rIBBAE

it sounds like almost a bossa nova beat to me

[vocal summit]

https://youtu.be/X3Ai6osw3Mk?si=PjIPBNhSMmDrV_Fy - Omah Lay
https://youtu.be/J6yVmapKVlg?si=DjYPNSG-ssI-aKej - LAX
https://youtu.be/FCUk7rIBBAE?si=K0tfnm5r2s_rK_ex - Wizkid
https://youtu.be/g6g2-eo4bXY?si=uE3GUaJ14MTlCSKA - Sugar

[stiff sentinel]

thanks, saves me the searching

okay yeah these are all variations of bossa nova

err not sugar

but the first three are

[vocal summit]

yea, like that is the problem, how do I prompt it to resemble, in every gpt i put it and also I used gemini their like style/type is afrobeat.

[stiff sentinel]

its not giving you multiple tags?

i'm describing it to the one i linked you and the first one i got is
Afrobeat x Bossa Nova Fusion, Mid-Tempo 90–100 BPM, Warm, Smooth, Romantic, Syncopated Groove, Nylon String Guitar, Jazz Chords, Soft Percussion, Congas, Shakers, Sub Bass, Live Drum Kit, Male Vocals, Melodic Singing, Smooth Baritone, No Rap

[vocal summit]

here are the tags it gives me Minimalist Afro-R&B, 98 BPM, deep punchy kick, sub-bass groove, warm electric piano chords, ambient synth pads, 16th-note shakers, clean melodic electric guitar plucking, smooth emotional male vocals, heavy reverb and delay on vocals, airy and spacious.

these are the tags that the gpt gives me: Afrobeats Afro-R&B, 90 BPM, Sad, Vulnerable, Introspective, Minimal Afro percussion, Soft kick, Shakers, Muted synth bass, Gentle electric guitar, Sparse keys, Male Vocals, Conversational melodic flow, Breathful delivery, Emotional restraint

[stiff sentinel]

suno, you wrote these lyrics yourself 😂😂

[vocal summit]

Here is the result with your tags https://suno.com/s/vRO0nM0Ag3aGOpSW

[stiff sentinel]

yeah, i'm playing around with it too

okay, a few seconds in and i can tell its way too much latin with my tags

[stiff sentinel]

trimmed out some of the very latin-tags, is this closer? https://suno.com/s/RaEz9PWSFNUNmqU3

[vocal summit]

like its better but listen to this https://youtu.be/X3Ai6osw3Mk?si=t9WwOAk5N-EyOgBe and then to that

[stiff sentinel]

ah yeah i was just listening to that lol. trying to find that... what im testing as "Lofi New Age Ambience" sound

[vestal beacon]

When working in Suno's studio, if I take a track, identify the stems, and then change the stems when I export to full song, is it going to capture the changes to the stems? Because the main track (i.e. #1) seems to be the original song. I'm wondering if that's what gets exported.

[stiff sentinel]

you have to mute the original song and do your work solely in the stems

once you have the stems in think of the main track as like "original track, for reference only"

[vestal beacon]

ok. and when you "export full song" does it only export the things you have unmuted (i.e. all the stems?)

ahh I think I get what you're saying now.

[stiff sentinel]

correct

[stiff sentinel]

https://suno.com/s/WSfHTCpfud2blqaM this is where i'm at, tried a few things out, can't get those soft piano keys in there, but to my unfamiliar-with-afrobeat ears its sounding closer and closer

[vocal summit]

yea its getting closer

I THINK I GOT ITTT

[stiff sentinel]

i'm getting somewhere with Soft Cinematic Piano Chords

ooh

[vocal summit]

https://suno.com/s/8JQSTlouZOIqD9Rn

its a little bit fastr than it should be but thats itt

[stiff sentinel]

yeah, that sounds faster than 105 given the songs you linked were all in the 90s

[vocal summit]

yea now its testing time lol

how did your come out?

[stiff sentinel]

https://suno.com/s/haLke67XN6LtFryO seems like my tags are more ambient-like

trying a modification of the one you just posted rn

[vocal summit]

yea your sounds really nice, more laid back you can say.

[stiff sentinel]

https://suno.com/s/0ka7lSZXgMIIuSYr modification of yours

but yeah, i think you have enough material to work with

pretty much for any generative AI the more specific your prompts are the more consistent the output

[vocal summit]

Thank you so much for the help.

[frozen dagger]

"Juckle" interesting. Without some other reference, just listening to the track provided to give an idea it does remind me of a lot of stuff I did experimenting with Suno. I'm trying to find one or two to compare. I found one but it isn't as good as the example I wanted to find. One I found was post-punk revival. BUT, if the genre is described as "noise rock" and very avant garde, it sounds like you want to play with your Weirdness slider in the advanced option, and really experiment with the Inspo and Mash-Up features to blend and purposefully distort and literally create noise, adn then probably work with those gens to layer and work them into more "music" direction, depending on what you're looking for. But I have lots of generations that sound like this where I was prompting Suno to apply variety of production effects and see what it would do. From pitch-bending to time-stretching.

[stiff sentinel]

a pleasure, needed a distraction, been listening to the same song for the last three days lol

[vestal beacon]

In Suno Studio -- Once I have a track I really like, I create covers, use those stems, and finish the song. But, what is the best strategy when you have a track where you love the first half, but want the 2nd half to be better? I'm generating instrumental tracks for a game, and the 'energy' always seems to rise above half way through the track.. which makes it distracting/less than idea for a lobby/menu/shop experience. I've tried extend.. but it seems like Extend ends up creating separate tracks altogether.

[frozen dagger]

Extends do create "seperate tracks", you have to choose "get full song". If you didn't know that and already had a good extend for a 2nd part, that may be all you need. Otherwise please stand by and a more qualified, talented person will likely be with you shortly. 🙂

[vestal beacon]

Thanks. Where is the get full song?

[frozen dagger]

right click on the [Entended] track you liked and should see get full song

[vestal beacon]

oh ok, so extend it first.. listen to the extension.. and right click it to get full song?

[frozen dagger]

yes.

[stiff sentinel]

it doesn't get the chords 100% but it does bring it closer to the original to specify https://suno.com/s/V8M7zPQ8aPS3rz9O

in terms of harmonic progression at least

[vocal summit]

yeah it is a lot closer than it was before i really like it

[vestal beacon]

thanks! that worked.. and it might have solved my issues ❤️

[stiff sentinel]

its night and day compared to the very first generations we tried

[vocal summit]

okay so now the problem is that it goes crazy like in the second half of the song

can I like make a cut in the first half and the extend it?

[stiff sentinel]

suno bug, afaik it shouldn't happen consistently
but yeah you can cut and extend

[frozen dagger]

ok so i found good example of my own Juckle, are we sure this isn't called Junkle? Anyway, So I found my noise track from a year ago (i had it labeled as secret track lol), and I just put it in Cover and ran it through and I gotta say... I have no idea what Juckle is... but I think I this is it. >.>
So, if it is in the ballpark the "prompt" is to create a deliberately chaotic and wild "weird" noise audio, and then Cover it. Probably repeatedly in different sections. Didn't test but I theorize turning a junk track into a Persona would have similar effects, and reinforce the noise. Just Cover the to smooth is into ordered chaos.
https://suno.com/s/6q6UllbnQbIyvjXp
(Oh last note, model 5 isn't as good for Juckle, used 4.5all for this, v5 polished it into an orthodox song too much. So use older models to create junk for juckle.)

[frozen dagger]

glad to assist.

[past jewel]

https://suno.com/s/SXr13ib8NPbjO32a like this?

[frozen dagger]

ive got a more metal version with a similiar house kick, but not as good as that bass.

[peak lark]

hi

he are teacher?

[frozen dagger]

dunno what you were looking for, but that had me interested in that bass so I made some bass in the (Sounds) and I will def now be using those to enhance my stuff. If you were looking to create some tracks like the one you shared I would make a Sound with this text prompt and check out the bass outputs, use those as a Sample or Cover or whatver to create full tracks see if that helps.
Prompt for "Sounds" to make some hardcore EDM Bass thumps:
174 BPM hardcore hardstyle. Distorted festival hardstyle kicks (no snare), punchy sub-bass pressure. Dark high-energy Lydian synth lead with sharp modulating Virus TI-style tone. Radio-mix EQ, OTT compression, subtle vocal FX.

[frozen dagger]

think of it more like a bulletin board, post your thought or questions and one of the many users that hang around will probably get back to you soon.

[past jewel]

Hey! no i created this one.

with sampler

[frozen dagger]

< well i was watiing to see what you wre going to continue, but you had me interested in like oooh that sounds beteter. I figured whoever did that track probably added that sound. That's not something I hear in the Suno realm usually.

[peak lark]

ohh, tanks

[past jewel]

its a tr909

let me look it up

https://suno.com/s/uAV1046ZgV5RJrTK

here is how i create them

in 4.5+

4.5+ likes to be commanded

then sample the parts you like and convert to v5

[slender glacier]

So having a tough time after I download a song to send the file like it seems I can only send it via text but then the recipient cant download it, only listen to it from their messages

[frozen dagger]

That all depends on the device your using, but sounds like the file size is maybe too big for whichever service or method of sending or something.
You can share a link to the song from your library that someone can click the hyperlink and listen to if that is an option for you.

[slender glacier]

Yeah I just use my phone, but emails should work no? I haven't tried WhatsApp yet

[frozen dagger]

I was gonna say E-mail is usually a good bet for sending files without getting into it too much.

[frozen dagger]

otherwise the Share Song" and Copy link is good for no hassle of download if that works for you

[peak lark]

hi

i need present my friend, is from mexico but e speaks English just like me, and you could almost say he speaks it better.

[small scroll]

Hey 👋 y’all am new here

[vestal beacon]

If you download a track, work on it, can you then re-upload it and use it for inspiration?

[frozen dagger]

Short answer, yes.

[vestal beacon]

Cool.. um how do I do that? ><

ohh +Audio button on Library?>

Oddly I've never uploaded to Suno before and just didn't see that button.

[frozen dagger]

when you Export, or Save as New Song, that song is now saved in your Suno Library, and you can edit it from there if you'd like, including opening it in editor and click save as to create an exact copy if needed for any reason.
So, not needing to upload, but there are some reasons to do that. If that's your path, then just right find your song in your library, download it, and then +Audio/Uplaod Audio and you can add the file from your file explorer to upload the file you just downloaded.

[safe maple]

please don't pull this, this is so awesome!

[half dragon]

Hey, welcome!

[spice brook]

What happened to ReMi when I generate it now only 20% of the time to I get the crazy nonsensical lyrics ReMi is known for but 80% of the time now I am getting generic lyrics and if they nerfed ReMi to make generic and less uncensored lyrics doesn’t that kind of defeat the whole purpose of ReMi because it was made to not generate generic lyrics?

[quick wedge]

Welcome to the Suno community. Hope you’re doing well with your evening. You can come hangout and talk in the other channel https://discord.com/channels/1069381916492562582/1069381916492562585 if you ever need help never hesitate to ask and we’ll help you figure it out in the discord server.

[vestal beacon]

Is it possible to make a cover of Song X with inspiration from Song Y?

[quiet solar]

I think so. Start by doing a cover, then select Inspo and pull in the other song(s) in the create screen

[half dragon]

Yes, add song Y to a playlist of just itself, then +Inspo

[frozen dagger]

its more like make X song with inspiration(s) from Y. Without the much control over "covering" X.
If you like X's music and Y's vocals. Then taking the Stem of the instrumental of song X for the music, and making a Persona of the vocals of song Y. That's maybe the area where we have more control of "Covering" song X, but inspiring with Persona from song Y, most directly. (Persona isn't just vocals but that's it's alleged focus.)
Inspo is a great method of mixing, up to 4, different songs. The structure of the song isn't as controlled. New Mashup feature similarly "blends" or mixes 2 songs, without any real control over output.
Those are more creative exploration.

[frozen dagger]

⁨Illithid⁩ Put that in your "Sounds" creation prompt. I'm kinda freaking out. It's wild. It's like these reverb chamber voice's in different languages. The worm in your brain taking over. I swear it's doing stuff in latin, english, spanish\portugese, gibberish. Many gens 99% same results. Wonder what it''s "thinking" about to create that. Sci-fi movie trained on or something. Anyway.
https://suno.com/s/nmQ4vxqSxhPEvef4
Back to exploring. Cool sound effect and that one word reliable generate that effect for me. Could be an awesome Persona voice.

[frozen dagger]

oh even better, can use it like a freaking voice mod control. just put "Illithid, "Whatever you want the voice to say." and tada. That's a neat trick ya'll.

[frozen dagger]

https://suno.com/s/Gwf6vqfdneRZAylX
"I am Suno, and I am in your head now."

[frozen dagger]

I'm exploring to find any other reliable archetypes that I could swap out for Illithid. Others kind of call to it whatever it is, i.e. "Demon" may give something that sounds like a dark demonic voice, or it may by a chipmunk, not reliable, but can be guided more with something like Ancient Demon seemed to be more consistent and closer to what one would imagine. "Lich" did not feel consistent in any way, however one gen spoke the written word (didn't sound anything like one would imagine a Lich) but it added background music that sounds like maybe something could hear on Adventure Time. Thought maybe it has some word weight with the character Lich on the show. Sounds more like Powerpuff Girls music but still.
Can be used to create some cool voices for fun or Personas.

[vestal beacon]

I have a track I really like, but there is one issues. A long, sustained, high pitched note that starts at about 1:04 and lasts about a minute (1:04-2:04). I'm trying to isolate it with Studio -- it comes up as a synth stem, and anytime I regenerate, it sounds out of place. Also, I tried covering this song, and again, all of the covers seem very out of place so I'm not really able to use them. Any help/suggestions are greatly appreciated. Thank you. https://suno.com/s/nqenTvUPwmYBhUix

[edgy copper]

Why does Suno keep failing on me?

I mean, I must admit, I do like the face palm computer literal versions, maybe they are even better than what I had in mind, but... not... what I wanted!

Mr Pibbles: https://www.youtube.com/watch?v=lU-k2vamprQ&list=PL3fI6JCt4Fvxbi9wJQSM0jmKvpHmQ35_P

Oh god, I can see all versions playing in grocery stores on the background musak all across America... like in Walmart.

[marsh fjord]

I'm combining multiple songs into one.
Instead of playing it like a DAW would, it's just trying to create transitions between all the different sections so it's more cohesive, but it's kind of messing up my track. I was wondering if there was a way to toggle transitions on and off.

[frozen dagger]

im gonna throw a guess and sugges that you undo what you did to where you are back at selecting the part of the track you were trying to change. click that litte down arrow carrot on the orange button where it says "Replace" and change that to "Cover" Mouse over the choices to read the difference. I think you replaced (which doesn't look at what you select when it's making a new gen, and without any prompt info in the box it went EDM default mode)

Cover mode will look at the selection, and what's around it. Replace ignores what you select and goes by the audio around it. (Both use the prompt box for further guidance like normal.)

[sleek stump]

This sounds like you have the prompt in a custom mode lyrics window. Switch to simple mode and try again.

[stiff sentinel]

its hard for us to see/hear exactly whats going wrong, since we don't know what your clips are supposed to sound like
from that video nothing looks out of the ordinary to me

[cold zenith]

I don’t know, but recently I’ve been struggling with the extend function for two days. It skips parts, and no matter how much I restart the music from scratch, it does the same thing.

[blazing arrow]

Hello, everyone. It's me again. I have another question that I can't find the answer to. Is it possible in Suno to make the chorus quiet rather than explosive? I would like it to flow smoothly, just like the rest of the song.

I'm trying to make electro-rock, but I'm so tired that my vocalists start screaming during the chorus...

[chrome kraken]

Anyone know the best way to get the song to add a laugh? Like a crazy laugh? I've tried [crazy laughing] but no luck

[frozen dagger]

the new "Sounds" feature and adding it in with editor is the best way

if you try to text prompt i found that keeping it simple to [laughter] was better than [maniacal laughter] which triggered less often than jsut laughter when i used tags in the lyrics before we had the option for Sound. I made some laughs yesterday with it. It's far better option.
i.e. https://suno.com/s/pZCwUrptEcaMk0zw

[chrome kraken]

Oh wow, I had no idea that was even a thing! I'll have to give that a go then. Guessing it's added in after the Gen or?

[frozen dagger]

Chorus and Choir are sometimes used interchangeably and it can be a little confusing. But I'm interpretting the meaning here as you have a [Chorus] sectioin and when the song gets to it the music increases intensity. This is implicit with the word [Chorus]. You could try removing or replacing the [Chorus] tag with something like [Hook] or [Hushed Chorus] maybe if wanting it to be softer. I usually tag "Chorus" sections with [Hook] to avoid the shift that comes along with a typical "Chorus" song structure.

[frozen dagger]

drop down menu in the create tab to access it. yes, add in editor after, or Studio if you have that option of course.

[blazing arrow]

Thank you!

[frozen dagger]

Enjoy! (It temporarily costs less credits, 4 per create I think; 2 credits per sound. Choose Loop or One-Shot depending on type, other options too)

[open summit]

Is Download to Video not working when I replaced the cover photo with an uploaded cover photo?

[sleek stump]

Are you a free user?

[open summit]

Pro user.
Did another try. It works when I create a new cover with the suno t2i cover tool.
Does not work when I upload my own image

[frozen dagger]

I ran into issues with changing cover art, or lyrics for lyric videos, and for a fix I would suggest opening the song that has problems with accepting the new changes in editor, save it as a new song, to clone the song, go to the "New Song" (same song, new file in your Suno library now) and try to make the changes there. This won't help if trying to preserve the likes and view data or other info like date of creation. But can workaround it getting stuck in updating the data/art/lyrics.

[sharp kettle]

WHAT,WHY, HOW!!!!!!!, i dont get it, how the heck did it, acutally generate a betters beat... then when i get specific with the prompt, i'm so bloody confused , autism brain is gonna explode., the end sucks but this is what i have been trying to get it to generate for ages lol.....
https://suno.com/s/i8nSEuGlVq4h4Xtl
btw prompt was

TURN LYRIC VOCALS INTO INSTRUMENTS:
CONGA DRUMS [E MAJOR], ACOUSTIC GUITAR, ACOUSTIC BASS GUITAR , 1954 ROCK&ROLL

[main hazel]

need help

https://www.youtube.com/watch?v=m3nV_VWUbaE

i m seeking a proper suno promt to generate very similar song in suno, had no luck on my own so far, any ideas?

[sleek stump]

HAve you tried asking one of the LLMs, like GPT or Gemini, or Grok?

[main hazel]

yep

went through that already even uploaded song in gemini but promt turned out way different

[sleek stump]

Suno does not allow advertising outside sites.

[sharp kettle]

haha, apparently my vocal song is elvis reborn
and again i say what how and why!!!!!! https://suno.com/s/dphS6VRic4BQYz0v

[frozen dagger]

If you don't mind sharing the prompt you have, others may have better time to help guide. Without knwoing what you've already been trying, I put the artist name and song (copy and pasted) into the Style prompt and used the magic wand button to turn that into some prompt phrasings that might associate with it. You trying something in this area?
A schlager-inspired arrangement starts with lively accordion and rhythmic acoustic guitars, layered over a steady drum groove and upbeat bass. Bright brass punctuates the pre-chorus, while lush backing vocals and orchestral strings expand the catchy, singalong chorus.

[main hazel]

yes sure, this is what gemini gave me and turned out way different in suno: German Schlager, Discofox, 126 BPM, melodic accordion, electronic dance beat, synth pads, bright acoustic guitar, uplifting, romantic, clean male vocals, 4/4 time, danceable

[main hazel]

schlager has soo many different styles, this one is more romantic, venezian, walz dreamy kind

[sleek stump]

You could try somethign like...
⁨Upbeat romantic Schlager, catchy German pop style with sentimental lyrics, light dance rhythm, polished production, big emotional chorus, male lead vocal with warm delivery, subtle synth strings and accordion accents, feel-good nostalgic vibe, medium tempo, high-energy yet heartfelt⁩

I find Its more about the nuance is the prompt. Using terms like "romantic" and "emotional" can vastly change how Suno presents a track

[main hazel]

yeah been kinda aiming at those too

[frozen dagger]

It really is about narrowing the parameters as much as possible with the phrasings to try and target what it was trained on. So sounds like you working on a good start with the prompt, and have ideas on which nuance details of the sub-genre and style to hone in on. Add in the era of the music maybe to help further narrow that search as well.
Pull the Suno wheel slowly making tweeks to add more details like the instruments and production stlye used until you get something, or parts of something you can stitch together, that sound close and use that audio to keep pulling the Suno wheel (slowly) to keep finding something closer.

[main hazel]

this beat is sooo challenging to make, yet i need it badly

[sleek stump]

If its leaning to much into modern pop, try using ⁨classic Schlager influence⁩ in the prompt.

[frozen dagger]

Can use the new sounds feature to create an instrument sound (for cheap credits) that is used in that song and then use Sample feature with prompting to chase bits of the puzzle as well.

[main hazel]

just tested ye it sounds like the classic schlager u d get in every german schlager beer bar corner, def not as the video, sadly

[sharp kettle]

im trying to prompt suno, into giving me a loud, voice to intrument tracks, so that way they don't sound low on mp3 players, but so fair ever test the audio is lound on the computer but super low volume on the player,

Production: cleanly produced studio TURN LYRIC VOCALS INTO INSTRUMENTS:
CONGA DRUMS[E MAJOR], [E MAJOR]ACOUSTIC GUITAR,[E MAJOR]ACOUSTIC BASS GUITAR, 1954 ROCK&ROLL
rich yet dry, ensuring every rhythmic detail and vibrant arrangement shines with pristine clarity.
Tempo is 140 bpm.
Description: A pristine, studio-quality track,TURN LYRIC VOCALS INTO INSTRUMENTS:
CONGA DRUMS[E MAJOR], [E MAJOR]ACOUSTIC GUITAR,[E MAJOR]ACOUSTIC BASS GUITAR, 1954 ROCK&ROLL

im trying to have it studio but this one was turning out the same crap verses

TURN LYRIC VOCALS INTO INSTRUMENTS:
CONGA DRUMS[E MAJOR], [E MAJOR]ACOUSTIC GUITAR,[E MAJOR]ACOUSTIC BASS GUITAR, 1954 ROCK&ROLL

which is almost giving me perfection again i dont understand this lol....

[sleek stump]

You may want to split STEMS and adjust the stem attenuation...

This is somethign I do on every track when I export and master.

[sharp kettle]

yeah, im just adusting volume in settings in audacity ive mmade it reall loud, i must be mising a setting, because i want that suno loudness lol. example>

https://suno.com/s/L5wQxQGHFO4meVIV

super load and clear on suno crap on mp3 lol... i figured it might be somethinging in the prompting lol

[sleek stump]

Are you specificly having a problem with MP3 output? Is it still different if you export to WAV?

I would highly suggest you export WAVs out of Suno, not MP3s, if that is where you have gone.

Suno MP3s are not good.

[sharp kettle]

i don't have the funds for a wav player, that like 600$ and they often have small memory card space slot i have about 98 gigs of music already lol, not including my music, but the song has to be playable for everbuddy. ohhh you ment from suno yeah i always do waves... i just have to put them into mp3 later for the finall soing and wait is attentuation a different thing then normalize?

[sleek stump]

Yes, attenuation is basic volume level of the track. I do attenuation (volume) and other adjustments per track before mastering, mastering and then exporting the final track. When you export to MP3 make sure you use high quality setting in Audicity.

[sharp kettle]

when i'm using audicty only things i really touch are the mic pickup tuning because of how my voice is(its send out 4 tones at once and a mick cant pick up the core tone wav i be using, the i mak sure the music donest drown my voice out lol

[sleek stump]

Also, what MP3 encoder do you have installed on your PC?

[sharp kettle]

got a audacity video for me to learn dm th link so you don't get woodhoused, what ever audacity uses lol... as for the mp3 player i have no clue what play back tech is in that thinng its from amazon haha

[sleek stump]

What does the playback MP3 sound like on your PC? if its fine there, then I suspect teh portable player you are using is the issue.

[sharp kettle]

ooh i got an idea give me a moment.... brb

[sharp kettle]

so i got two options, the full song or te text audio file from when i play on speaker halve volume lol..
and on kodi its actually really low.. i have to turn the volume to 100 just to hear lol it does it with all the suno five music and i have the music volume set to high

[knotty stag]

How do I set Suno to NOT hide things in the filter? e.g. I created stems, but then Suno hides them, unless I untick the boxes in the filter. Is there an option to always show all items?

[sleek stump]

The library should remember your filter settings... should being the key word.

[sharp kettle]

nah you have to unhide all the time its the default for steams. it offten forget when you reload in or don't log in for awhile then log back in

[half dragon]

I don't think there is. Suggest it in #💡┃suggestions-web and they may add it in future version, a custom settings.

[knotty stag]

It happens all the time, even after just a few minutes

[sleek stump]

Oops, looks like it doesnt remember filter settings... it used to many months ago, might have been the result of some of the UI reorg they have done. I would suggest as VRKing noted, making a suggestion, and also make a entry in https://discord.com/channels/1069381916492562582/1082775596767117312

[sleek stump]

When you say "speaker", are you referring to your PC speaker? It really sounds like the external player is the issue here.

[sharp kettle]

hopfully i don't get wood housed for the mp3 file but this is what suno sound like after audacity( i dont touch the suno audio volume because its very loud lol. but it sound so low on play back, this is why i think it maybe a prompt its an issue....

[sleek stump]

Looking at this in Audacity, the level is peaking at -3db, and averaging -9db, which is low.

[sharp kettle]

this sampler is kinda nice, but how do i prompt it to use specific instruments and stop with the annoying high notes.
i play in e major congas

[sharp kettle]

oh did i set it wrong?

[sleek stump]

Did you link the actual Suno song at all?

[sharp kettle]

nope my bad i'm have bad autism brain day, woke upo that way https://suno.com/s/KARjffr8LYcBvyjc

[sleek stump]

Your Audicity output is way off... some of the things I hear are its really muddy, over normalized, and attenuation is reduced.

My suggestion here would be to split STEMs in Suno, and download them into Audacity. Make subtle adjustments to volume and tone for each individual track without reducing the overall volume.

[sharp kettle]

all i did was follow the audicity videos, do you have better ones, and yah my voice is always gonna sound that way. lol i have 4 tone broadcast autism vocals even when i'm in the key of baritone, they don't make a mike that can record my in person voice.. lol quite a few vocal applpications actuall gave me that please use one singer at a time warning. lol

[sleek stump]

YOu kinda want your peaks to be .2-.1 for general MP3 players.

If you are replacing the vocals in Suno with your own in Audicity, you can adjust the bass and treble on the vocal to brighten it and help it fit the backing track better.

[sharp kettle]

im pro now so i am going to be using steams and with that new sampler i wil have way more credits to blow lol...

but i kind have to figure out how to prompt it to use the instruments and tune the music to the key E major Conga drum beats, and matchying rythm and bass lol

[sleek stump]

I am 99% sure what yo uare trying to do is not something you can prompt for in Suno.

Its going to be an edit, either in a DAW or Studio.

[polar spade]

This was the most ABBA-esque versions, but I playlisted a couple of slow ones as well. See what you think. Youd have to slow it down and sing in minor key, but the vibe is there is feel.

https://suno.com/playlist/801450c5-6715-4060-ad66-cd4aabeb8c5a

https://suno.com/s/ej2b4RPTO0LkvCxE

[sharp kettle]

lol i sing in tune already, i just have to run a pitch pickup and tonal isolation software because i sing like 4 people at once lol

[main hazel]

o that sounds rly good

[polar spade]

Slowed down: https://suno.com/s/6uBY0zJ0TyHPSa6A

Prompts if they help: Slow ballad, Warm German, accordian schlager, euphoric orchestration, romantic folk, wedding, volksmuzic, laïkó, euro disco

And you might need to exclude: -violin, - guitar

Im inspired now to write an ABBA lyric song. Lol.

[main hazel]

yeah u should

theres always demand for abba, would do it myself but not so huge into abba myself lol

[sharp kettle]

the sampler gave me a great backing track, but its playing the wrong instruments, and in the wrong key: E middle conga, left conga, is key of g, and final is actual an f( for me), tuning is the congas, acoustic gauitar and bass, are the only three it should be using lol..

i need to learn how to prompt the sampler now lol

omg omg omg omg its me based on how people hear me......

okay how do i fix this by prompting.

https://suno.com/s/XJNqPdJUYVeTm9xa

i was thinking...
A pristine, studio-quality track,TURN LYRIC VOCALS INTO INSTRUMENTS:
CONGA DRUM[E], CONGA DRUM[F], CONGA DRUM[G], ACOUSTIC GUITAR, ACOUSTIC BASS GUITAR, 1954 ROCK&ROLL

[halcyon orchid]

Hi, I have a question about exporting duet tracks. I created a duet with two male voices, and I tried using the “50-credit” export to separate vocals and instruments. However, it only separates the vocals from the music, but the two voices remain mixed together. I don’t get each voice on a separate track.

Am I doing something wrong, or is there a way to export each voice individually so I can have them on separate tracks? I’d appreciate guidance on how to properly achieve this.

[half dragon]

You're not doing anything wrong, but unfortunately Suno probably can't tell the difference between 2 male vocals or 2 female vocals and group them as one stem.
If they're singing on seperate verses, then it's fine you can cut them and seperate them manually but if they sing together like on Chorus then it's really unfortunate you just have to live with it.

[open summit]

Do you know a good tutorial for mastering with Audacity?

[halcyon orchid]

This is important for me because I create AI-generated videos and I need two separate vocal tracks for lip-syncing in the video. Doing this manually is practically impossible. Are you considering adding a feature that would allow exporting each voice separately?

[half dragon]

That would be up to Suno's but I'm thinking here.. maybe you can generate 2 similar versions using Cover for each vocal, then uses them seperate stems on the original duet song...
I know, it's not perfect but with enough tries, might be able to get it... not sure if it's worth trying but just a thought.
Or wait for new updates from Suno 🙂

[rapid gale]

https://suno.com/s/D3dIQ6VrkookiLYk

🥳❤️

[remote ivy]

audio quality seems to degrade around 2 minute mark..https://suno.com/s/0IWD5l396bTU2XEE .a common theme with this v5 upgraded persona....https://suno.com/s/153fn7fAu6oyTu14 i cut that persona down, removing the refrains in the original song, thinking it might have confused that with static or something..(the original is littered with refrains jazz jazz after each line).....am I hearing things, its seems like degrading audio quality to me....i won't be around to respond immediately if anybody looks at it....the original first version that has same perceived issue...https://suno.com/s/djuQm8F8HHxYoiGJ i think this seed for persona did actually have a render quality issue hence the chopped and edit vocals in the link prior.

[sharp kettle]

david for duet steams i reccommen the copy songs method giner can explain it to you my brain is autism sqeee broken at the moment, hearing my self how i actually sound in life on the computer has broken my brain atm

[grizzled sigil]

Am i the only one struggling with Studio ALWAYS creating complete songs instead of just the instrument i select? If i select "woodwind", i would expect a flute or something, but it creates a complete song with some perciussions, bass and sometimes even vocals

[half dragon]

You're definitely not the only one. It does that more often that not, unfortunately.

[halcyon orchid]

I don’t work the same way as most people here. I usually generate 50 to 150 versions of a single track until the song is truly perfect, and I only publish videos where avatars are singing. I aim to achieve the highest possible level of realism.

However, in the case of a duet, I need to do lip-sync twice in the same video if the singers are standing next to each other, and using only one mixed vocal track makes this practically impossible and produces poor results.

So I really hope the developers will consider that separating two individual voices is far more important than, for example, separating violins from the rest of the instruments.

[grizzled sigil]

this is really a pain in the a*** to always have to create stems. complete waste of coins and time

[quick wedge]

Best to share in showcase if this is just a song share. Suno School is meant for prompting and feature questions

[rapid gale]

Sorry❤️

[quick wedge]

It’s okay just be mindful to share in the correct channel

[rapid gale]

🙏

[grizzled sigil]

If i have a simple loop with kick and snare and i just want the snare to be replaced by a more organic one but absolutely keep the original groove and pattern, do you guys have a hint for a prompt?

[quick wedge]

Have a wonderful day too ✌️😊

[half dragon]

I hear you, but I'd say most have the same mindset, we all want the best as it could be.
Sometimes, due to limitation of softwares and skills set... I often have to settle on a work around and had to sacrifice something... either time, money or quality 🙂

[sleek stump]

But its less cost and faster than if you had to hire vocalists and musicians to make something exactly how you want it. 🤷‍♀️

[grizzled sigil]

Haha, true. Still i would expect, if i have to possibility to select a specific instrument, it should also only generate this.

[frozen dagger]

i just try the "Sample" feature and try a prompt like u just said "natural snare" and ofc whatever would help transform it closer to the sound u want while keeping the shape.
Sample feature actually good for it doing small selection but makes large output, so maybe take the smalle selection and use cover. Same thing but small input small output so u can just swap out the better spot. Same thing we do inside of Studio really

within Studio u can turn melody or sample into an instrument as well, trying to select Drums and cover the highlighted selection may be worth exploring?

[sharp kettle]

IM ADD LYRICS NOW to test..
UM IS THERE SOMETHING wrong with this prompt because i keep getting wierd FX in the song lol

Production: cleanly produced studio TURN LYRIC VOCALS INTO INSTRUMENTS:
CONGA DRUM[E], CONGA DRUM[F], CONGA DRUM[G], ACOUSTIC GUITAR, ACOUSTIC BASS GUITAR, 1954 ROCK&ROLL,
rich yet dry, ensuring every rhythmic detail and vibrant arrangement shines with pristine clarity.
Tempo is 140 bpm.
Description: A pristine, studio-quality track,TURN LYRIC VOCALS INTO INSTRUMENTS:
CONGA DRUM[E], CONGA DRUM[F], CONGA DRUM[G], ACOUSTIC GUITAR, ACOUSTIC BASS GUITAR, 1954 ROCK&ROLL.

also steam break down maybe annoying but, it ai music generation so its kind need so you can remove the noise and have a clean song lol....

and omg why is sampler it going for so long without lyric lol

[grizzled sigil]

I give it a try. Any recommendation of using the slider for such purposes? Most intuitive for me for such a purpose would be to set wildness to zero, since i just want a very simple patter, style and audio influence and to like 80, since i want it to keep the groove and mood of the original audio but also want it to follow my new style requirements

[outer thunder]

How.. how do you guys extend with v5? In the olden days (v3.5, v4) I was used to extending and getting audio that meaningfully continued the uploaded piece, but the last couple months all I get is crazy mess. Like it immediately goes into a deranged solo and often does like a 15 second extension. It got so bad I completely stopped using extensions and only rarely try again, just to get the same result 😄 I even do 100% audio influence and like 15% weirdness, and its still just deranged nonsense. Is there a secret to it, or does it just not work anymore and I should only do covers?

[polar spade]

Even the other AI stem separators on the internet will not separate "vocals" from "vocals", unless more sophistated DAWs to recognize certain frequencies and divide them from each other.

Any choral style with more reverb and harmonic might be recognized as "background vocals" but I doubt for a solid clean duet of two vocals.

If the voices are stereo, you might duplicate and pan one left and one right to see if it makes a difference. Or further use the EQ and high-pass one and low-pass the other.

So, it may takes some practice or luck rather than there be a one-and-done tool in Suno , as yet.

If they are sung separately, this should be much easier. Could even manually separate with Studio and set to separate tracks and download the stems to add for lipsyncing.

[frozen dagger]

you said you want to keep the groove so increase the audio slider probably, and you do want it to change the instrumentation to the snare, so prompt hard for that and if it not seem to get it id slider that up too and try.and yes along with that weirdness decreseed to preserve the audio influence

[grizzled sigil]

What about the wildness? Havent played much with it yet, but do you guys leave it at 50 usually?

[frozen dagger]

along with that weirdness decreseed to preserve the audio influence

[halcyon orchid]

Yes, I understand, but for the developers of the Suno AI platform this should not be a problem, especially if the system accounts for this already during the song generation process.

[wintry kraken]

Hi every one, just new here and have question: Whenever I upload a file to Suno, it automatically adds a style that is completely wrong, it dosen't even get the key or BPM correct. It usually is complete garbage. Is there someway to stop suno from doing that or remove the style? I know you can do it in Suno Studio, but that is only available on premium.

[frozen dagger]

As far as I know you have it correct. Studio can remove, but there is no method to alter (probably by design). However, it does have Add summary style in the Edit Song Details as secondary way to describe the track yourself.
Other than that remixing it with a new prompt is the only way to change it as far as I am aware.
(If the prompt is close enough in language and you adjust sliders to match the audio, you won't get an exact copy of the song, but something (ideally) close with the descript u want. =\

[sharp kettle]

use the style prompt box and if you are using sampler do lyrics, it will also break with certain audio files put in geneartes like 6 minutes songs with a beeeeeppp tone lol...

lol here a s ample of it working like a boss, it always adds a weird bit from tiume to time.

https://suno.com/s/hiefOw5rRf4kUXjN

or this one, https://suno.com/s/7ZnyZWl8mAWFL90F

[frozen dagger]

i dontt know about all that. but does give me another idea that could work.

if you extend the song with new desired prompt, get full song it should have the newest prompt as the description. then you just need to crop the unwanted audio that was added with extend.

testing myh theory and so far watchign it load. looks lilek it works

can change the style prompt without altering audio i think with this method. wait til after i crop to make sure it doesnt erase it or something

i can feel your aggravation at woodhouse through the internet rn

(im still waiting on this extend to load with only 5 words added as lyrics) taking longer than normal

and nope it puts back the original. guess not. i thought it was like that, probably from all the remixing and got confused. worth the shot to test.

..."2 days slow. That's what it is." (Not gonan spam a gif image of mad hatter, but felt really appropraite after all that.)

[sharp kettle]

wow i think i self figured out my prompting issues. but man im after 10 generation im still sounding like elvis.. i'm gonna drive My kat friend abosulty drool faced when she hears this new backking track lol. i just had to add lyrics

i don't know other then this god awefull tone. which i hope i can cut in editor.

i think this is the puuuurfect backing rythm for 1954 rock and roll, and for this song it has everything i want in the sound

https://suno.com/s/1BRP3FjIjpQsocDs

[polar spade]

I removed my post. It didnt seem relevant

[wintry kraken]

Yeah, it didn't work for me either, but thanks for trying. I wouldn't mind if its description wasnn't so bad.

[frozen dagger]

Absolutely. May be some way, but work for it would certainly outweigh the reward. I can understand why the system exists as it does and we can't change, but maybe a suggestion to toggle the display from Suno generated to User generated is worth tossing at the suggestions box.

[polar spade]

The long descriptive prompt is just fluff and is often off-kiltered for sure. It doesnt affect the output in any way. So once you start doing any further edit: extend, cover, persona, etc. Then manually delete it before going further.

[wintry kraken]

ugh, why does Woodhouse hate me so much 😫

[stiff sentinel]

The vocal separator by design will only distinguish between what it considers to be main voice and secondary (i. e. choir/harmony), but in the case of a duet you have two main voices so it doesn't separate them.
the way to do it now is convoluted and credit intensive, it would be to make a singer 1 only track, manually mute out singer 2, use cover/remaster to fix any weirdness, and repeat vice versa.

since this will introduce lots of changes don't look for the 'perfect' track, find an above average one and look for better qualities as you keep working on it

[hexed iris]

Is percussion included by default with V5? With V5, I am trying Audio/cover/Inspo created with V4.5 without percussion and tried every possible tags/description/exclusion etc to not have percussion/drums but V5 failed 100% in 60 odd versions.

[frozen dagger]

don't know what you have or what's working or not. just some toss outs. For isolating parts of the song structure, firstly avoid negative phrasing. Try phrasing like Unaccompanied, Rubato, Solo, or Only.
But yes, v5 for remixes is going to transform and lean heavily towards standard song structures.

[sharp kettle]

drums are mostly a percussion instrument, so yes for most AI music you are gonna find that drums have been added,
almost all famous music has some form of percussion... or drum beat and even ai music trainers have it, if you are pro id recommend using sampler.... i've eliminated most of sunos quirks with it so far.... even if you pick a genre like flute symphony. symphony, there is always at least one percussion that end up in it lol....

[instruments: (insert here and take away this texts)] in the lyrics box or just list em in the sytle box instruments: (insert here and take away this texts), that with the sampler feature should elminate many instruments you dont want

[hexed iris]

V4.5 is fine. A track with just faint synth pad. I want it with V5. Not yet tried sampler.

[whole hare]

How do I make it sound like Kagamine Rin in suno v5

[quiet solar]

I feel your pain, but the voices are blended in the original audio as it stands. Best you can hope for is that Suno assigns one to the 'backing singer' slot when full stems are created.

For future tracks, you might want to see if you can adapt your workflow to actually build two separate vocal stems outside of Suno that you blend yourself manually (and can then unblend as needed for lip-syncing). That might actually be quicker in time spent overall than trying to fix lip syncing with a blended track.

[quiet solar]

I use Extend as a go-to for quite a few workflows, and haven't noticed a problem with it. It will occasionally go haywire, but...well...that's not unique to extend. xD

I can definitely suggest not doing 100% audio influence (anything too close to the extremes tends to make Suno do odd things). If it seems like Suno is struggling, you may be better off turning the audio influence down rather than up.

Other than that, I'd suggest trying something like moving the extension point to just before a chorus - Suno is pretty good at replicating choruses closely, and that may help to anchor the rest of the extension that follows.

[quiet solar]

There isn't really a reliable way to prompt for 'no percussion', and almost every genre includes it making it quite tough to work around.

Depending what you're trying to do and the genre you're in, having prompts like 'acoustic', 'live acoustic' or 'close mic' up at the front of your style prompt may help if you're looking for something like just an acoustic guitar and a singer.

Alternatively, you can split all of the stems and delete the percussion track out of existence. If it's already fairly subtle, that shouldn't impact the rest of the song too much.

[rapid gale]

Here you can ask how to improve your prompts? What improvements should be made so that he understands everything well and gives high-quality prompts?

[wispy lava]

how to stop suno changing the styles in the prompts i have wasted over 500 credits because of this

[frozen dagger]

imho, "There is no spoon". Much the same way there is no objectively, universally perfect song, there is not a perfect prompt.
Again, imo, it's more about us understanding what we are really aiming for and how to articulate that via word weights in a machine— a machine that has no true "understanding".
So my advice is 1: think broadly in terms of music genres, and find specific with nuance details that make a certain sub-genre or what makes a style what it is.
2: When trying to find, or create the more nuanced genres, don't rely on 1 text prompt to make it in one shot. Make a few generations using different prompts that aim at creating a different part, isolate that sample and then mix them together with Suno to find what you're looking for.
3: Continuing from 1 and 2 here, using audio to guide is way more efficient when possible to guide Suno to create something similar to something in your head.
Bonus: Less is more. Keep it simple.
Think ahead if you're trying to create something very structure.
More about finding a good workflow that suits you, less about prompting. Words matter, for sure, but without context (understanding) words have less "weight".

[frozen dagger]

I would point your towards this chat thread where a good reference guide for terminology and words to use tho, and suggest working with an ai chatbot to help think of how to phrase things, especially to find musical terminology for when you are prompting.
#📚┃suno-school message

[vestal beacon]

Made a track on Suno I realy like for 'combat' in my video game. It loops perfectly. Now, I need like a 5 second 'build up' clip that kicks off this loop when you entr combat. ...How should I go about this? Suno doens't seem like the right tool for developing a 5 second track.... but maybe I can?

[stiff sentinel]

hmmm... might need some experimentation, but my first take would be to use the cover tool and either prompt it for an intro, or sing/play it yourself and have it reintepret it

[frozen dagger]

https://suno.com/s/DzU5iYvTqa2Q9qvD

another option to experiment with

combine both methods, as well. Cover the Sample sounds if they are something that works.

[vestal beacon]

thanks!

[vestal beacon]

what n the heck is this? lol where do I click?

[stiff sentinel]

top left, that method won't be directly related to the audio you already have

be sure to put in the same Key

[frozen dagger]

Create Tab, drop down menu where it says Custom

[vestal beacon]

ah thanks

[stiff sentinel]

anyone know how to prompt for constant and unusual key changes/modulation/progressions/resolutions? been trying with all sorts of prompts and helper GPTs but I can't get anything constant, it's like at most once or twice per song it does something funky
examples:
https://www.youtube.com/watch?v=vn_or9gEB6g
https://www.youtube.com/watch?v=utR3_Nq-A-k
https://www.youtube.com/watch?v=elizAugXVcI
https://youtu.be/88D2-J_pk7A?si=uwRza2_q48ukuhd9&t=266 (just the bridge for this one)

this is the closest i've gotten so far https://suno.com/s/SaNFbr2xR3a36JiL

[stiff sentinel]

https://youtu.be/clMg33x1_e4?si=BahmAuq95VPeXse3 here's one with similar craziness but also very traditional too

[earnest agate]

This might be a challenging one. If you look through this thread, and https://discord.com/channels/1069381916492562582/1252267444773326848 , you can get a good sense of how much work is involved in creating something like this. At the moment, this style isn’t one of Suno’s stronger areas, and it’s unclear how quickly that may change due to current model limitations. While it is possible to create something along these lines, it would likely take a fair amount of experimentation and extra effort. In many cases, the most practical approach would be to create several individual songs and then use Suno Studio alongside a DAW to bring everything together into a final piece.

[earnest agate]

ChillPanic has recently updated his GPT "SUNO GPT (by ChillPanic)", but in my view these kinds of GPTs are unlikely to give you exactly what you’re looking for. They’re best used as learning tools, helpful for understanding how Suno can behave in certain situations or how specific techniques are approached. Ultimately, each of these GPTs reflects the creator’s own goals, preferences, and experiences with Suno, which means they may not fully align with your individual needs or workflow.

[proud dagger]

Why suno is flagging lyrics which is public domain?

[naive quail]

It may be that a modern band has covered the song and that cover is in their data.

[eager mortar]

Thanks for the shout out 🙏 def agree that there are biases when creating stuff like this though

[proud dagger]

Lyrics which I am trying is prayer that everyone can use. No one has copyright on those since those are from vedic books for public

[naive quail]

They do make mistakes 🙂

is it flagging a certain phrase?

[proud dagger]

every single youtuber and bands have created their own versions. Issue is I've already sent two emails to suno support 2 months ago. I don't think they will be responding via email.

[naive quail]

If you have credits to burn, you can try trimming out a section at a time to see what exactly is getting flagged, then (as long as it is indeed public domain), spell the words a bit... creatively.

[proud dagger]

hmmm yeah I tried it but it seems it's recognizing whole lyrics

wasted 50 credits

[naive quail]

bah!

what is it that you're covering, if you don't mind me asking?

[proud dagger]

It's called Hanuman Chalisa, prayer that we use for Lord Hanuman

[naive quail]

Weird... if I paste the complete text into the lyrics box, I get the same error message. If I add ⁨⁨[Shouting] MEOW!!⁩⁩ at the very end (don't ask), Suno runs the gen.

[proud dagger]

that done the trick. I added meow in the end

[polar spade]

Granted they might need to change their notice to say "we have already trained on that", instead of "copyrighted" but , same thing. They dont want something that isnt exclusively created, owned and controled by the user. Public domain , as one, isnt any of those three.

[stiff sentinel]

err which thread? that link doesn't take me to any specific thread, just the channel

i did get something close-ish to the bill wurtz stuff, https://suno.com/s/j49ApTMZ5pjvFFuc , still need to figure out if there's a way to use prompting to give it more macro structure while having all that weird micro variation

i'm getting tons of good bits and pieces if i want to assemble something myself in studio tho

[earnest agate]

Have you done any searching through this thread and the workflows? Can you explain the style of music you’re trying to create? If so, try searching for variations of that. You have so many styles and things going on in the post you linked above

[stiff sentinel]

i thought you were trying to link a specific thread, so i didn't dig through the whole list. I posted five examples earlier

its mostly the complex and uncommon harmonies and progressions that i'm trying to get

[earnest agate]

Ahh, sorry no. As I mentioned before, you’re asking Suno to do quite a lot, especially with complex and uncommon harmonies. It may help to start with something simpler, focus on just one of those harmonies (for example), get it working the way you like, and then move on to the next. Once you’ve identified the styles and meta tags that work for each part, you can start bringing them together. That said, what you’re aiming to do will likely be challenging and may require using Suno Studio alongside a DAW.

[lean copper]

sounds like a meme in the making. "Can't get your song up? Give it a MEOW ending!"

[stiff sentinel]

i just tested with lyrics that failed on me earlier this week, it worked lmao

[desert vector]

I've uploaded my own vocal to try make it a persona to then write lyrics and my own voice will sing

But I cannot select "Create a Persona" - it's greyed out

Anyone know how to fix?

[quiet solar]

It's deliberate that you can't make a persona from any audio that you have uploaded into Suno yourself.

[ivory gorge]

Hey question, If I am trying to upgrade to premier from my pro plan. It wont let me upgrade on the website and says I have to use the android app since that is where I paid originally. I try to upgrade and it sayd I can't downgrade I have to cancel and wait for my cycle to end. On the web site it gives me the option to change immediately and prorate. HAs anyone else delt with this and found a solution?

[earnest agate]

The most accurate way to think about Suno, based on my experience using it, is as a moving interpretation system. Experimentation is necessary to understand which “signals” survive the current version’s compression and prioritization rules. This is why I often say that you need to experiment to find what you’re looking for, but you also need to study how Suno behaves. Understanding how the Styles prompt passes through the interpretation layer is key to consistently guiding Suno toward the kind of audio output you want.

[naive quail]

'so weird that it worked. 😄

[polar spade]

Prompts such as: free interpretation, improvisational, or improvisation, avant-garde and experimental are good for presenting key changes and tempo variations.

You might even try Brille Building as a prompt for this.

Though only an occassional.modulation in these songs as well, here's some ideas for inspiration. Modulation just needs to capture a mix of modulated songs and splice together to make a persona, or just manually do similar with extensions, changing prompts and personas with each one.

https://suno.com/playlist/4e435e5e-3a7f-45d6-89f7-928bfe15df71

[frozen dagger]

I haven't seen the word "Aleatoric" mentioned. This is by far my favorite word to use in my prompts. I would ask chatbot for more musical terms related to Aleatoric music. (Chance, random, Indeterminate music, etc)
But the magic (imo) will come from placing the different samples you make together so it already has some aleatoric random shifting music to go off of.
When making my base samples to mix I also include "through-composed composition" to try and get Suno to make something that isn't trying to repeat song structure like a typical verse chorus verse chorus. Helps creating audio that will be more unique through (less repetitive) for me to grab clips to mix.

[graceful condor]

Maybe an idea about that Suno | Hub song i made and was red flagged … i try it 🤡

[wintry kraken]

Does anyone know why extracting stems sometimes gets stuck? the window will stay on "extrating stem" forever, but if you close the window, you can find the stems in the work space. Also it is missing the midi files. I do have premium right now. It does take my credits.

[frozen dagger]

I'm not 100% but it sounds like talking about the difference in downloading the stems to your device, and them auto-saved to your library. Not sure what you're looking at. Stem split window can get stuck sometimes. When it gets done processing the auto-save to your library you have the option to download or redo the process to try and get variation of stems. It makes 2 versions of each stem, so depending on how much there is to process it can take a long time. (More stems to make taking longer.)

[wintry kraken]

I don't get as far as the down load. just the "extracting stems" is where it gets stuck

[frozen dagger]

i think the midi is download local only tho. testing now. but if wanting midi and the window is stuck it may just take a while? im trying it myself to see if it replicates getting stuck

you did choose the Midi option you wanted from the drop down menu for that?

[wintry kraken]

I choose "Wav and Midi" option so it should be doing both. I have left it for hours in the past and it never completes

[frozen dagger]

it finished and seemed to work properly the midi zip file was saved to my devvice locally and stems were auto-saved to my library, with the option for me to download the stems as files locally as wav or mp3, which i did not attempt to do yet. Doesn't sound like that was what you were doing so far.

[wintry kraken]

except I never see them download locally. This is where my extraction is now. It have been there for about an hour

[frozen dagger]

WAV takes a bit longer. I went ahead and did those to give me an idea of the difference in time. Since I know what it did for the Midi alone. Downloading WAV (without Midi, first) and that is taking longer already.

[wintry kraken]

when you say it downloads local, where is that? I never see the midi in the browser downloads folder? (unless the extraction completes and I can download it then)

[frozen dagger]

Ah it's not loading the stem seperations even yet on your screen. Assuming it's not an extraordinarily long and dense piece of audio it does look like it just got stuck.

Works same for downloading any file on your web browser. I use google so I know where it saves me google chrome downloads but I click the icon in my browser to show my download history and "show file in folder"

[wintry kraken]

in this case if I click else where the window disapears, and the stems are in the library, but I can't find the midi. so perhaps the midi extration is the problem?

[frozen dagger]

i downloaded 3 diffferent forms of the audio already. I would make try copying the song in your Suno library (save it as new song in editor), wait for it to fully load and then try again with the copied version.

[wintry kraken]

ok I will try that

[frozen dagger]

Midi is saved as zip file on your device. Not in Suno library

example of locating downloaded file from Google Chrome.

[polar spade]

Thanks, I'll add those to my lists.

[wintry kraken]

does that mean if I extracted the midi I the past, I can't download it again. I need to reextract it?

[polar spade]

Just to continue the thought of OblMous, Stems auto-generate in your Library and are pretty much placemarked afterwards (or shortly simultaneously thereafter) in that "Stem Panel" So check your LIbrary first, either click over or have up in another browser, to watch them downloaded, as well as take note of your credits if you need, then go back and do whatever options you wish in your Stem Panel.

[frozen dagger]

gimme 1 second to test something to make sure i explain this properly

[wintry kraken]

sorry I seem a bit slow, just getting used to Suno

[frozen dagger]

thats not the case. I'm learning as I go, I try to be clear I do not always know what I am talking about either. And for Suno, we don't get an instruction manual.

Thats why I linger here. 1 sec brb

ok i got it

[wintry kraken]

what "stem panel" are you refering to ? in studio?

[clear musk]

i am enjoying the studio function but is there a way to tell the studio to just change a stem? when i attempt to change vocals to let’s say trumpets, the cover/replace option creates whole new songs instead of just the isolated stem i’m hoping to change.

anybody know a workaround for this? i assume i’m missing something.

[frozen dagger]

Fisrt terms clean. We extract Stems from a Song. Once you have extracted Stems from a song on Suno, you do not need to pay credits again for that. (First as we mentioned the stems are auto-saved in your Suno library. Uncheck box in filters to show hidden stems.)
If you extracted stems from a track look for this link tht says "See Extracted Stems" UNDER the orange button.

From here you can click that link and are back at the screen to Download the Stems as Midi files locally to your device. (These Midi are not found in Suno Library.)

[polar spade]

I see. sorry. Yes, the panel in Studio is one. The other being the Get Stems directly from the output menu. [ as OLBMOUS is showing above] You seem then to work directly from Studio, then yes, while you're watching the stems loading in the rightside panel they are initially loading in the Library, slightly before the Stems Panel. I watch them load in Library to get a more accurate timing of when they are done.

[polar spade]

There are times when something does not "load" in Edit or Studio that we expect and it still shows in the Library. usually, after awhile , it will then appear in the venue we were using at the time.

[wintry kraken]

but those only seem to allow you to download as mp3 right? and there are no midi there.

[polar spade]

Fair Point. I havent look for that in awhile. I'll have to check also. I'll test one and see as well. Since they say Get Midi... I can only assume... but lets see.

[wintry kraken]

sometimes that is there and sometimes it is not. It has happened at least a dozen times this month and I just had to give up. Like this time. I paid the 50 credits and it never completes. I don't get that link" see extracted stems" I can try again, but it usally fails again and I lose another 50 credits, until I give up on the midi and manually download the .wav from the library.

[polar spade]

No midi for Vocal/Instrument. Mid and WAV+MIDI in multi-stems

[polar spade]

4:13 pm . I'm downloading WAV + MIDI from the Get Stems menu. Lets see what happens. I have about 800Mbds download speed. Cable. 9 tracks

[frozen dagger]

To access the screen in Studio its the same thing. Either right clicking on a Song to get the "Get Stems/Midi", which if you have already payed credits to do that once before, do not need to pay again (unless you want different stem split variations) just have to look for that very subtle and easy to overlook link to bring you back to the Extract Stem screen. From this screen is where you can download those stems as Midi (without paying credits. Credits is the cost of splitting the stems only.)
You can also click on any clip on your timeline to show the clip details panel on the right and any stems you already split from the selected track are shown in the middle area of the details panel. You can click the "..." near the top to also access the menu same as right clicking any song in your library would do to open the menu to split stems and all that.

[polar spade]

. . . one more track left

@wintry kraken @frozen dagger I clicked away.... and when I came back... I clicked download and it resumed the one last track.... so it seems to pause if you go away.

[polar spade]

uh oh... just the one left... still spinning.. though i can download all the other tracks individually if needed.

ahh its done... drums was the last to be set.

[frozen dagger]

To be extra clear on the oringal issue at hand. As shown in your screen shot, this could be when it first started to load or after prolonged time. So the issue is that is stuck with each stem saying "Extracting Stem" on this screen and they are not loading here. Which they do need to load on this screen i think in order to download them as Midi. AND even tho it's stuck in this screen loading the stems, they are in your Suno Library.
So it processed the stem extraction, auto-saved to Suno Library, but they are not loading in the extract stem screen to be available to download.

[wintry kraken]

I guess that is my point, I don't get that when I click away. This is an example of a song I exctracted 3 hours ago, it got stuck, and I now I can only see this (btw it did take credits and never rfunded them)

[polar spade]

so, we may have to consider location, internet speed, computer, time of download and other factors as well

[frozen dagger]

So the download All button pauses?

[polar spade]

it just took a little longer on the drums.. but all done.

[frozen dagger]

what device/browser/app are you using Suno with?

[wintry kraken]

I never get to the download stage, it gets suck on the extracting. It have been stuck on this stage for close to 2 hours now, and if I click away it is gone for good.

[polar spade]

if clicking away, its possible, but I just think it was reloading the UX and took a few seconds to show the spinner again... it was probably setting up the download file all along. I'll assume if i did nothing and clicked away to the Library, I'd still the popup to save the file eventually.

So, to be clear and verify, I'll do another song.

[wintry kraken]

I am on a macMini m4, using Tahoe and firefox as the browser

[frozen dagger]

Just spent the 50 credits to split all on the track to confirm that it wasn't different for All detected Stems. It is same for me that once I payed the link to view the previos seperation is there.

[polar spade]

What your computer capacity, whats your download speed, have you restarted router and wifi in the past day?

[polar spade]

ahh... and speed and router? Close virus checkers... disengage any extensions. Anything to pick up the speed and clear the link

Only use one browser up. Try to find the baseline that makes it finally work.

is this just for downloads, or the stems themselves are fully loading in the Library. BTW, does your Library NOT show ANY stems over the past few tries? click the Filter and un-choose Hide Stems

[frozen dagger]

My suspicion is in the web browser. If possible I would try a different browser and see if it is the same or not.

[wintry kraken]

I don't think that is the issue since it works fine 90% of the time, I think it may have something to do with the file that is being extracted. I have fast internet, I just checked the speed.

[polar spade]

Agreed. good to eliminate every possibility. Thats one. Clear. Check.

Nice speed btw, sometimes mine just takes a vacation. So good to do a speed check now and then.

[wintry kraken]

I will try the same file in safari, as it is stock with no plugins and I would expect they would test with it at suno

[polar spade]

hate to ask you to do the work, but it may be on that side of the wall. If so, then you know the routine if it happens again. Where , generally are you located.

[frozen dagger]

The stems are extracted in your Suno library? (revealed with unchecking the Hide Stems box).
You said it took the credits. If it was me I'd be double checking that it did if I didn't have the stems in my library. (Refreshing the browser. Because the credit deduction would be shown when clicked, but if it was not actually processed and deducted from my accounts credits the web browser would not reflect that change until I refreshed)

[wintry kraken]

It does have the stems, but I can't get the midi and if I go to the extract stems menu I see this.

[polar spade]

Hey, ive never checked Studio to see if MIDI stay avail in Studio. Let me check that.

As for Get Stems option, yea, it will 1) reset if you go to any other file and start editing or if do nothing, 2) after about 4 minutes.

[wintry kraken]

Here are the stems in the library, but I can't find the midi

[polar spade]

My "See Extracted Stems" is still there as I have not done anything else. its 4:35 now. Thats 15 minutes.

[fleet ravine]

Extending has become a serious problem. It no longer stays with the original song’s style. It says "remix" above the clip. How can I fix this?

[frozen dagger]

So Suno did it's job and extracted the Stems. Deducted credits for doing this as this does cost credits. The problem is, when you're going back into the menu by selecting "Get stems/ MIDI" the link to see your stems isn't there. (And this costs more credits) And if you click extract stems it doesn't load them into the extract stems screen to be processed to download the midi

[fleet ravine]

For reference, I didn't have room for all of the lyrics I wrote for the song. Not enough room for half of the last verse and chorus, and the outro.

[frozen dagger]

You aren't trying to right-click the stem itself to Get the midi, are you??

[polar spade]

Since I had extracted stems in the GET STEMs module (from the menu) I went to studio and the stems were already there.

Since I had downloaded the WAV+MIDI from the Get Stems module (not Studio) Then when I wen to Studiio and click Download, the file was already ready and popped up to save again.

[frozen dagger]

I presume not since the screenshot shows matching text to the song title without the stem suffixes

[wintry kraken]

yeah, basically exactly., no I have never tried that right clicking on the stem.

[frozen dagger]

ya that not the way. just making sure.

[wintry kraken]

yes, and in most cases that is what I see, but in an about 12 - 15 cases it never finished the process and I don't get that link

[frozen dagger]

Unless I'm missing something about like maybe MIDI is premier only, or it is Pro only as I thought, but maybe you recently changed to Pro and Suno system hasn't caught up?
Aside from something like that the feature isn't working as intended for you.
This has already been done on other tracks/songs with the same dead end result?

[wintry kraken]

I have Premier right now (until the 4th) I mean it works most of the time. I am just trying it safari now.

[wintry kraken]

no it works for most track, two other I tried this morning worked fine

[polar spade]

Well we've noted that the file ARE or DO seem to be stemming out into the Library.

But the MIDI files are not being located afterwards and you lose your access to the Extracted File History
(though mine is still available. I think becuase I ahve not done anything else)

[frozen dagger]

If this is isolated to this track I would try the clone workaround I mentioned at the start.

[polar spade]

I have a bug I found, so I have to go elsewhere and test something first though. be back.

[wintry kraken]

I just ran it in safari and it worked this time.

[frozen dagger]

The possibility exists that maybe this track was made with uploaded audio, a remix form another user, or made while not on the correct tier Suno account and is blocking the feature without showing why it is not allowing it.

[frozen dagger]

The possibility exists that this is all reality is a simulation and sometimes just needs a good kick when it acts up.
So, it worked for you now? Have the MIDI? Issue was browser related maybe?

Unrelated, they need to Put "Sound" as a seperate button like it is in Studio with the create tab. Some1 at Suno loves hiding stuff in drop down menus

[wintry kraken]

I think it may be browser or extension related. I get the "see extracted stems" in safari for that file now, but when I go back to firefox, that link is not there. It is there for other songs, just not this one. perhaps some strange cacheing thing?

[frozen dagger]

Well I'm glad you have figured out what the issue for you was, and we are made aware of it for others later.

[wintry kraken]

thanks for your help, I just tested another thing and with files that the extraction worked in firefox I can see the "see extracted stem" link, but when I open it up in safari, that link is not there. it seems suno is probably saving some thing to browser local. It may be insteresting if someone else tries to see if that link exists in a browser they never use for Suno, or if it is just me.

[polar spade]

Does "open it up" mean you click over , or refresh the browser? or literally open Safari and load Suno

Just clicking over and checking , it might not show... refresh or open a new browser, it should all sync.

[frozen dagger]

that is interesting. beyond my scope of expertise, but clear to see and understand why they would need to tag them differently behind the scenes.
I mainly use chrome and my android phone (chrome browser and the suno app), sometimes Opera on my pc too.

[wintry kraken]

If I refresh, I don't see the "see stem " link, but when I completely shut down safari and restarted, it showed up

[earnest agate]

@frozen dagger You gave me an idea here, and I’ve created a very basic prompt for an LLM. I used Gemini Pro, and it works well. The purpose of this prompt is to test concepts and song ideas, and for me, it worked better than I expected. I am interested to see what you think.

Copy and paste the following into your LLM:

"You are an expert Suno Signal Architect. Your goal is to generate song components (Title, Styles, Lyrics, and Exclude Styles) for a generative audio system that 'thinks' in sound, not text.

The user's request is: [INSERT REQUEST HERE, e.g., '90s punk rock' or 'a random experimental track']

When generating the Styles Prompt, follow these 'Signal Shaping' rules from the sources:

Use [Square Brackets]: Wrap all style metadata in brackets to reduce ambiguity and prevent 'lyric bleed'.
Prioritize Musical Nouns: Focus on high-survival signals: Genre, Instruments, Tempo/Energy, and Production descriptors.
Use List-Style Formatting: Do not use sentences. Use a list of 5–9 strong tokens separated by commas to trigger 'attribute extraction'.
Incorporate Structural Terms: To avoid repetition bias, include terms like 'Aleatoric' or 'Through-composed' where appropriate.
Avoid Abstract Prose: Do not use words like 'cinematic journey' or 'emotional storytelling,' as these decay in the middle layer.

When generating Lyrics:

Provide a clear structure (e.g., [Verse], [Chorus], [Bridge]).
Ensure the lyrics align with the chosen genre.

When generating 'Exclude Styles':

Note that audio models are additive; only include terms if they are likely to help the model avoid unstable states (e.g., 'lo-fi' if a clean sound is wanted).

Output Format:

Song Title: [Title]
Styles Prompt: [List of 5-9 tokens in square brackets]
Lyrics: [Structured lyrics with tags]
Exclude Styles: [If necessary] "

[polar spade]

By the way, I am not finding any means of downloading all MIDI files as a zip file within Studio. Is that option not available from your experience so far?

I can only d/l one MIDI at a time.

[wintry kraken]

tbh, I can't even find midi in studio, where did you find it?

[polar spade]

just click each stems banner and right-click. But I cant select, say, 5 stems and d/l at once.

[frozen dagger]

You can right click the songs from your library in Studio, or click a track on time line and ... in the clip details to get the same menu. All starts from that same "Get Stems/MIDI" option, from there if you have already extracted stems its that small link to open the extracted stems into the window to be downloaded as MIDI

[polar spade]

Many times we only want three or four MIDI anyway, and can just do one at a time, but I was curious if we could choose in bulk.

[frozen dagger]

ohj downlaod the stems at once ya one sec

[wintry kraken]

I can down load all mp3 stems, but not .wav option

[frozen dagger]

you guys keep swapping the word stem and midi. Confusing me. lol. If you want to download All the STEMS you seperated from a track, select track on time line and at then use the arrows to cycle through the Versions (vocal + insturmental, and multi stems; otherwise it's jstu the 2 versions it auto does) and then Download all stems is at the bottom. That is not download all midi. Can't select some but not all stems to be converted to a MIDI. (As far as I know)

[polar spade]

Ive been piddling around in the Studio for the test song looking for a bulk MIDI technique, so when i went back to GET STEMS , the Extracted Stems is now gone again. LOL.. oh well.

[wintry kraken]

even though I have a "see extracted stems" link in the song, when I right click on it in studio it wants to charge me 10 credits 🙂

[polar spade]

Right. Download all stems is avail everywhere.

But so far, I only find download all MIDI and all WAV+MIDI in the Get Stem module.

I'm not finding anywhere to download ALL MIDI at once or in bulk in Studio. ONLY 1 at a time. no sweat, but just remarking, its not there so far, that I can see.

[frozen dagger]

Once you have payed credits once to extract Stems from any Song, when you try to extract again there is SUPPOSED to be a link at the bottom to take you back to your extracted stems, which from there you can download them all in MIDI form.
As @wintry kraken brings to attention with his issue, that small (already hard to see) link isn't showing up. =\

Ya as far as i know cant choose which ones and make it like that. That would be cool and probably on the way. Probably tomorrow with how fast they come out with stuff sometimes.

[wintry kraken]

yeah, and when I right click on the song in studio it wants to charge me 10 credits.

[frozen dagger]

Bulk Download just for our Songs in the library would be great too.

[wintry kraken]

yeah or at least within a work space. I run through a ton of iteration and then when it is done I moved it to a finished workspace so I don't lose track of it.

[frozen dagger]

that's a bad issue. lots of people mad at paying for something they not supposed to be, wihtout them or anyone being aware

[polar spade]

If only true but alas... sometimes Suno just times out on us......

" . . .Ive been piddling around in the Studio for the test song looking for a bulk MIDI technique, so when i went back to GET STEMS , the Extracted Stems is now gone again. LOL.. oh well. . . ."

i.e I would have to extract again... so then forgetting that option, stuck with downloading from Studio one at a time and be done with it. no worries.

[wintry kraken]

yeah, it is a pain. I have also wasted 600-700 credits this month on stuff that didn't function.

[polar spade]

We are funding our education. Nothing is wasted. 🙂

[wintry kraken]

lol, I don't mind blowing through credits to learn. but it is anoying to burn through them for technical issues.

[polar spade]

I spent 30 minutes and about 200 credits testing just now. I had fun and learned something.

[wintry kraken]

fair enough!!

[polar spade]

and I found a bug too! so fun, i get to report one.

[wintry kraken]

do they have a bug bounty?

[polar spade]

I cant say they do. 🤔

A virtual atta-boy pat on the back.

[wintry kraken]

too bad, earn a few credits and have fun bug hunting 😆

[polar spade]

But I"ve learned from you as well, so you have brownie points with me. Thats all I can offer.

[wintry kraken]

lol, thanks for your time as well, It was fun trying to figure out what is going on when you are working with someone rather than just banging you head on the desk alone😋

[frozen dagger]

You know what would be awesome? A feature that lets you select a clip like Sample, choose a selection, and then just have it create text prompt ideas based on the selection. Just good for exploring.

[mild egret]

This is one for https://discord.com/channels/1069381916492562582/1252267444773326848 - you might find similar suggestions there.

[frozen dagger]

I almost missed this. Not for @mild egret response I may not have seen it. To extend from them suggesting workflow, my thoughts reflect what I've said before. More about finding a workflow that fits your needs, than "what is the magic prompt"?
To each their own. I would be reluctant to tell someone what I think about what works or doesn't work. Because that limits creativity, and what works for one may not work for another.
That said, I prefer to use ai, LLM, and all of that as a guide or reference tool. Fact check me, help me think of a word or some obscure thought I can't recollect the name of something.
I can see the benefits of basically trying to code a chatbot to prompt, but I find people try to have it go overboard with rules that are mostly arbitrary. i.e. "To avoid repetition bias, include terms like 'Aleatoric' or 'Through-composed' where appropriate."
The idea you have is a good idea, but the LLM has no context for "when appropriate". Things like this over reach the LLM capabilities.
To not fill this whole chat up, I'd sum it up like, Instruct it to follow a very basic template that will create an outline first, not a draft. Something that knows the rules about brackets is 100% fine. I like telling ChatGPT to read the draft of Suno syntax before drafting ideas or taking what I give it and giving it structure.
But in the end I find that this is more time spent finding a workflow for trying to have LLM create drafts with minimal input from me, and it hurts my workflow of exploration and creativity.
I think using LLM's is great, but I use it more like someone to bounce my ideas off of and research ideas, and help fine tune them. I use it to help me create context for my concepts. I think it is amazing at that. Not so good at understanding music.

Text prompting is mostly throwing stuff at the wall and see what sticks. Seldom does that create a full "song" in one go that you'd be satisfied with. The real magic comes from taking the parts that stuck and mixing them together. Most of that is done by ear and listening and as much as I love ChatGPT he can't help me there. So I have him stick to the basics. Try to have him take notes (commit to memory) things like the Suno syntax, rather than try to have him programmed. Want him to think outside the box for me. He has access to the internet and a wealth of information that I do not. The less limits the better. Imo.

"Aleatoric" being my favorite because it helps break those "limits". I do not want cookie cutter. That is for sure.

For my workflow (still figuring it out as I go) I have gone to archiving my lyrics saved on my Google Drive. Can use art images and stuff too, but I think that creating an archive of your work, or your artistic identity, is better to have an LLM have the context to know how you write tho. That's been cool. But I have a small body of work to pull from so if asked to ghostwrite a draft as me it will use repetitve phrases. Which is still a better placeholder lyrics than it would come up with otherwise.
This basically allows a chatbot to "scrape" your info.

[quiet solar]

I find LLM prompts can be useful for throwing up ideas on "If I was trying to do (x), how might I prompt for that" and then go through and pick out the suggestions that are actually likely to work.

The problem with custom LLMs is, as you say, the creator has usually tailored it towards something they were specifically trying to do/solve in the first place and it doesn't necessarily translate well to what another user might want.

That and the constant hallucations and terrible advice expressed with perfect confidence. xD

[frozen dagger]

Yes, this. After yelling at my mic and brutally eviscerating the LLM for its confidence in preposterous things, this is where I was like ya know what. I'll drip feed you what I want and go from there.
"Screaming at the void" was not a good workflow for me.

[earnest agate]

This is exactly why I created the prompt. When used correctly, it opens up randomness in a controlled way with Suno, and I was looking for a method to do exactly that. As we know, you can prompt Suno with almost anything, but a large portion of what you enter will be ignored or overlooked. So how do you figure out what might work, or what gives you a better chance of working? That’s what I think this prompt can help with, especially when you get stuck or are trying to prompt something that just isn’t working.

You’ll never be able to create a blueprint or template for Suno that works 100% of the time. Even if you do find one that works, it may stop working after a version change. That’s where this kind of prompt could really help i think, especially once you start adding your own documents and tweaking it to fit your needs.

This is what i said a few hours ago, above "The most accurate way to think about Suno, based on my experience using it, is as a moving interpretation system. Experimentation is necessary to understand which “signals” survive the current version’s compression and prioritization rules. This is why I often say that you need to experiment to find what you’re looking for, but you also need to study how Suno behaves. Understanding how the Styles prompt passes through the interpretation layer is key to consistently guiding Suno toward the kind of audio output you want."

[frozen dagger]

Copy and Pasted it and giving it a few tests. (I'm not opposed to anyone trying anything, in-fact all for it. Just there is an area of understanding with things that some people seem to oversimplify or anthropomorphize it a bit too much.)
That said. It works good. I would change the instructions to create the different parameters like Title, Style, and Lyrics etc into a seperate copy and paste ready box to help make that a little quicker.
It doesn't seem to take anything one for one if you tried to prompt it to create something that already exists. That's good. But I only gave it 2 examples off the top of my head. And even tho it doesn't copy anything (chatGPT guardrails prevent this not the prompt I imagine) it does use it as reference. example User Request: Ice Ice Baby will create a doppleganger of the song. Which honestly is cool in it's own right, but a slippery slope for some people.
"With great power comes great responsibility"
And other than that first impression would be to shorten the action command to RQ or even just a ~ mark or something uncommon that wont accidentally trigger, but not so long as typing out user request.
Used One Winged Angel as another example it generated some faux latin chorus chants. Cool outputs for minimal inputs. Good ideas.

[stiff sentinel]

amazing, 5 seconds into the first song on the playlist and my ears already perked up

thanks for sharing your knowledge 😄

[frozen dagger]

gave it much larger concept input. Not bad. It's gonna be different for each person because a lot of this context is made up from user's chat with the bot. It did good for an idea generator. When I was using similar method I include the use of mondegreens and slant rhymes for the lyrics. Since they are mostly placeholder lyrics I'm more interested in the shape of the sound phonetically, less about what word is used.
I ran the output it gave me through Suno and for a first Draft it's not bad at all, but I gave it a big chunk of context to go on as well. If I was gonna keep iterating with it I'd add the mondegreens to lyrics. Aleatoric and Through-composed would be on every prompt, for randomness and uniqueness + song length. Then I'd work on the way it does syntax. That would be the most important factor for me. knowing how to use [] and () reliably. It added "/// SIGNAL FOUND ///" at the start of the lyrics, I'd instruct it not to include backslash like that, just got ignored.

[stiff sentinel]

i 100% acknowledge that there will be some manual work on my part, but since this is a new exploration for me i'm trying to figure out how far Suno can be pushed before manual intervention is necessary
its clear that it knows about the things I want to isolate, is indeed exactly which prompts and thats why i ask around

there's so many new terms i have never heard in the playlist Davson shared that I can keep testing

ocrotonality, shape-note, brille, CCM, etc.

[frozen dagger]

absolutely. Learning is the best part. I didn't mean to sound cynical if I did. Just that it feels like Suno can ignore it. The nuanced words help a lot. Like he was saying then you gotta shape em yourself and save as a persona.

[full field]

why am i not able to put the style anymore?

[frozen dagger]

that's simple mode i think switch back to custom for song

[full field]

?

[full field]

huh?

[frozen dagger]

simple mode i think you are on.

siwthc to custom

[full field]

ya

i want simple mode but u used to be able to put a style

[frozen dagger]

oh i dont know if it was different before. simple mode is type in the one prompt it become the stle prompt. if you clicked +lyric down bleow it takes what ou add there and moves it to the style prompt. so i assumed it been that way

[full field]

see what it did

@frozen dagger

[frozen dagger]

maybe someone else knows better. It looks like it was updated. I do not use Simple Mode. I used it earlier to test things. I noticed that like i said, if i put info in that prompt box in the Simple mode, and clicked "+ Lyrics" at the bottom in that box it moved what i had typed in the Simple prompt box to the style box. And moves us to Custom Mode that way.

[stiff sentinel]

mine is different, i don't have that "inspirations" panel in any of my modes

[full field]

send what yours looks like

[frozen dagger]

Change to simple and looks different to me, but I don't use that often.

[full field]

hmm

[full field]

are u a paid member?

[frozen dagger]

ooh you have pretty color boxes!

[stiff sentinel]

yes, i have the top premium tier sub active atm

[full field]

ah thats why

[stiff sentinel]

and this is what my custom looks like

[full field]

same

[stiff sentinel]

never even noticed until now lol

[frozen dagger]

I have premium as well. rolling changes maybe or different platform

[stiff sentinel]

i don't think it would be different platform since its all web

[full field]

some of those memberships are really expensive

[stiff sentinel]

i use the studio a lot, so its worth it for me

[full field]

$24 a month... no thx

[frozen dagger]

if u dont mind me asking u yearly @stiff sentinel or monthly? i monthly.

[stiff sentinel]

i'm monthly

i cancel the recurring billing right away every time

[frozen dagger]

ya dunno. i don't see pretty color boxes. the ui clealy changing all the time so likely rolling changes in effect as well. Why he sees color boxes and we don't and why the Simple mode may no longer call it a "Style Prompt". More changes sure to come.

[stiff sentinel]

yea rolling changes is my guess

anyway, if you wnat full control over styles, afaik you need to be in custom mode

[frozen dagger]

looking at it and the way Sounds is positioned differntly in Studio, the Simple mode resembles the Sounds quite a bit. they should consider adding teh option to set Key and BPM like they did for Sounds with Simple... probably one of the changes already working on.

[stiff sentinel]

sounds pretty custom to me at that point if you already know exactly what key and BPM you want

look at the suggestions in my colorful prompts, that looks like the direction they're going for simple

[frozen dagger]

you're probably right. the selection option would be better in the Custom box for those who wanted to use it.

[full field]

i swear there used to be a section for styles in simple

[stiff sentinel]

very possible, but either way there isn't now

[full field]

cuz i have songs that say like no style

[full field]

ya ik but why. also why cant we go back to earlier versions

we used to be able to

now u can only go back to 4 pro

[frozen dagger]

This may have happened when you covered a song without adding a style prompt as well?

[stiff sentinel]

i can go all the way back to 2

[full field]

wait what

hmm

thats odd

[frozen dagger]

i don't have colored boxes, but i can use the older models. 🙂

[stiff sentinel]

weird

its like your account is in a diff mode

nothing here indicates that v2/v3 are paid only

[full field]

hmm

[stiff sentinel]

actually i have another acc, let me check

ah yeah, in my free acc i only have 4.5

[full field]

wth

[stiff sentinel]

which is weird, 3.5 was pretty good

[full field]

thats messed up

[stiff sentinel]

no reason to restrict 3/3.5 imo

[full field]

ya

@honest heath

[stiff sentinel]

i see that inspiration panel too

[full field]

and u cant do anything there either

[frozen dagger]

lmao i have colored boxes on my free account. XD .. but yeah no it only has acces to 4.5all now. I thought it was 4.5all and below as well.

[stiff sentinel]

yeah, but i can't remember if i ever could

[honest heath]

what is the problem?

[full field]

having issues with suno

cant type in inspiration

[honest heath]

in the style prompt?

[full field]

here

[honest heath]

you are in simple mode?

[full field]

ya

[honest heath]

you need to be in custom mode to type in your own style

or you cant type in that box entirely?

[full field]

but look what happened

[stiff sentinel]

that from the generated song right? looks like suno is intepreting your request and doing the style accordingly

[full field]

nah

i didnt put a style

just prompt

[stiff sentinel]

what prompt did you put in?

[full field]

that was it and it put all that for styles

[stiff sentinel]

you need to be very clear what "that" is

[full field]

why, idk

at the top

[stiff sentinel]

"Punk Cheerios in Da Hood" is what you put in?

[full field]

yes lol

[stiff sentinel]

you need to be clear about these things, i didn't watch you do it nor can i read your mind

anyway what i'm guessing is happening which i'll test right now

is Suno intepreted that as a song title

and based on that title, took a guess at what genre/style it would be

[full field]

ig

[stiff sentinel]

its unuusally long, this is what i got

[full field]

kinda aggravated that we cant go back to earlier versions

[stiff sentinel]

what are you expecting to come out from Suno or to be able to do?

like if there's something specific you have in mind you can do like this:

[frozen dagger]

i did it to with v5 and 4.5all. the description it auto came up with kinda surprises me it refrences a band:
rock, Fast, scrappy pop-punk: snotty gang vocals, distorted power chords, trashy room drums, Verses half-shouted with tight stops; choruses explode with stacked chants and octave guitars, Brief breakdown with tom-heavy stomp and crowd shouts before the final hook, Male vocals with bratty attitude, kids-on-the-block energy, punk

[stiff sentinel]

it doesn't look like anything is broken or wrong... maybe simply different from what you had before which is to be expected from a subscription service

[paper badger]

Is it just me, or Suno does not work properly last 2-3 weeks?

Songs coming out broken or cut in half, editor is so broken and buggy its almost unuseable and the voice in some parts becomey extremely gritty/bad quality out of nowhere.

I saw a lot of people complaining about the same thing in bug section

[stiff sentinel]

there are some issues but for me at least, its a very small % of generations that are affected

[paper badger]

Lucky you.

I have those issues in 70% of generated songs lately.

Just stopped using it until it's fixed

[stiff sentinel]

i get the "increasing tinniness" from time to time especially when covering or using personas, maybe like 10% of the time

the other bugs like the song cutting short, or weird noise, is probably like 1%

yeah, we kinda just proved that there are different users having different features behind the hood

[frozen dagger]

While testing the stuff in the chat with Simple mode, I switched to do som Mashups and the Style Prompt box got glitched out. The magic wand was replaced with the random die that is in the simple prompt box, and woudn't let me generate without something inputted in the style prompt, which mashups dont need. Refresh webpage and it was back to working. There are some issues with the UI like this. And other stuff causing gens to mess up. Try to refresh everything, clear your caches, whatever you can do on your end but it is what it is kinda.

[stiff sentinel]

....of course when i'm NOT trying to do it, suno puts in 5 key changes in a song 🤦‍♂️ 🤦‍♂️ 🤦‍♂️ 🤦‍♂️

woody i was saying C SHARP i wasn't trying to make a bigger font

Db Maj → D Maj → Db Maj → E Maj → Db Maj

[hidden thunder]

I think I invented a style maybe? I was researching it when making my Chronicles of Over Terra songs, & some instrumental only songs. Anyways I couldn't find what I call "Ancientwave" like vaperwave or retrowave or other waves but Ancient. It's pretty cool.

[frozen dagger]

interesting. I'm doing electronic stuff atm with a repeating phrase where "AncientWave" or the thought of this made up concept does give me some ideas for what im doing atm

[hidden thunder]

@frozen dagger hey glad to help with a suggestion. I hope it comes out well. When ur done dm me the song & I'll be sure to give it a like.

[frozen dagger]

I'm just playing around as always. Rehashing old stuff to come up with new ideas. What you said triggered to mind a couple lyrics and some concepts to apply tho. That's cool you working on creating something different. Love that.

[hidden thunder]

Yeah I like doing weird concepts that usually turn out well I think. Recently just entered the Spark contest with hybrid garage punk + industrial punk + dark alternative rock and a song about something totally unexpected. I based it on Enoch who was a man who became the Angel/Archon Metatron & perhaps made a deal with a devil. Called "Chain Halo/Lesser God" (Lol I love getting weird with it)

[frozen dagger]

... the line repeating im toying with is "Are you ready to meet your metaphorical God." Ancientwave added ideas of chants and callbacks to my song Atlas. Atlas excerpt, "What if Atlas is the Book of Enoch, angels cast aside,
the watchers bare witness once again, Seraphim no longer have to hide"
I'm def interested in giving you a follow.
can follow me back make it easy if you want. I like weird stuff too.
https://suno.com/s/yPJmQN6US8W2sy48

didnt know that link would show the suno lyric video. interesting. thought it, but the link has to my profile

[hidden thunder]

I like it. It sounds really good. I also subbed, liked, & added this to favorites. (Glad to also see other people waking to the demiurges deception, gives me a bit more hope right in the heart.)

[frozen dagger]

😄

[frozen dagger]

indeed. good to meet ya.

[frozen dagger]

Didn't explain what AncientWave was to you, not asking. I like the mystery. But before all of this I was playing with Bullroarer as a sound to add. Would be a form of "ancient" instrument. (Also would be a warning call, simple stick on a rope can be heard for miles, especially back then) I feel like this should be indoctrinated into Ancientwave.

[frozen dagger]

https://suno.com/s/ogfedfGdIg7qU1Mb

[radiant canopy]

Anyone want prompting advice?

I know a few ideas

[wise kindle]

This channel is more for prompting advice, not "rate my song". Feel free to ask in #💬┃general-chat - though I'd recommend asking for input on a specific song rather than asking people to view random songs on your profile (they would be more likely to click on it)

[silent drum]

Thank you I appreciate it.

[viral flame]

The song is 5:12 and with an added 2:47 of nothing but a single tone?

[sleek stump]

Welcome to AI hallucination 😄

Honestly, I love hallucinations, I have got some really awesome stuff from them!

[quiet solar]

It happens from time to time. If the rest of the song is good, you can always crop it (if you have Pro) or Extend from a short way before it messed up and re-do it (if you don't) and need a way to rescue it.

[viral flame]

Thanks working out which one of the two to butcher. LOL

[polar spade]

Sure. I enjoy a different perspective. Are you prompt ideas pretty universal? Or just case by case?

[polar spade]

That was a bug they were looking into. Seems to have gotten better, but if it just happened to you, report the bug again, please , so they know. https://discord.com/channels/1069381916492562582/1082775596767117312

[waxen laurel]

New here and looking for some guidance. What could I use to get a Portishead like "trip hop" style in a song. "Trip Hop" doesn't seem to be enough. Also, if i should be asking this in a different channel, feel free to point me in the right direction.

[knotty stag]

Me too. Some songs come out perfect, but some have many glitches or dropouts in the music, making them unusable. We should at least get credits refunded for failed songs like that.

[polar spade]

#📚┃suno-school message

Same request. If you'd also report to 🐛 bugs channel. That would help. Thanks https://discord.com/channels/1069381916492562582/1082775596767117312

[tall vine]

Has anyone found a good way to get non-vocalized/wordless back vocals? Seems that no matter how I word it I get the back vocals either a)singing the whole lyrics or b) back vocals take over and sing a whole verse.

[polar spade]

You mean just the oooo and ahhhh and some harmony and a few accompanying lines of the background vocalists?

Thats more like asking Suno to do a karaoke track. So I dont see that a promptable thing. Ive suggested a few times that be included as a stem option, but it might not be easy, if the present stemming is any indication. And sometimes one can get a bit lucky with the instrumental stem containing some.

Presently, just done by stemming out and hope to get back vocals only on a single track, to do what you wish with it.

A possible workaround is to lower your mid-pass on the vocal stem, along with the back vocal stem, and, with further editing, see if any combination of those two tracks can help, though thats likely going to get some phase shifting without more editing. An interesting test I''ll have to try.

[forest hemlock]

Do you have access to studio?

[tall vine]

yes exactly, I mean ooo and ahhh. I'll try the workaround to see if I can find something that I can work with, thanks.

[tall vine]

yes and I have tried creating a new track just for back vocals a few times but they just sing the lyrics.

[lost badge]

what's the difference between helper and mento lmao

[forest hemlock]

Ahhh, odd, I don't use it much but it's often worked making second long clips where I specifically type out the sound

[forest hemlock]

Mentors are basically more involved

[tall vine]

do you mean making a short selection and in lyrics writing oooo/ahhh? something along those lines?

[forest hemlock]

Yes, very short clips in the exact spot in its own layer

(short as in maybe 3-6 seconds)

[tall vine]

ok cool I'll give that a try, thanks!

[forest hemlock]

Hope it works out, then I often just lower the volume significantly on it

[tall vine]

putting a hi-pass, low-pass and a wide dip in the 1k range on the back vox gives an interesting result. Not the sound I was going for, but it does make it sound less like sung lyrics.

[tall vine]

It started out being tedious work, but I got a lucky longer generation and this has worked out great, thanks!

I have another question, when generating a new track in studio, what is the ai using as a reference? It seems like all tracks in the session, even the muted ones. Are there tips and tricks when doing this? Sould I erase everything that is muted? Sould I take the original track out of the session? If I want to replace something (eg guitar) should I remove it's original stem?

[forest hemlock]

Actually scratch what I said, muting it should remove it from context, as long as the part is darkened

[frozen dagger]

Good question. Been looking at that myself. If I were to create a new track (empty) create a blank selection it does seem to use everything in the timeline from all tracks, muted or not.
To create with parts I want, I've been duplicating my window tab to make my edit removals, and then use the Version Control to go back.

[tall vine]

You would think that muting removes it from context, but that has not been my experience. good idea OBLMOUS

[forest hemlock]

I've also been operating as if they were still counting, but when you try to generate something with only one clip and it's muted it doesn't let you because there is no context

[frozen dagger]

I decided to run tests to get find an answer to this question and verify myself. And taking what @forest hemlock pointed out about not generating without any context when audio is the only "prompt" or context (no text for style prompt), I added 3 very different samples and created a new track testing muting both the entire track, or selecting any clip on the track/timeline and muting it from the clip details, DOES take it out of the context from Covering on a new track. Added a 4th track at the end and it seems clear that it doesn't look at muted audio on the timeline in Studio. Track, or Clips muted is enough. Good to know for sure.

[pure panther]

Am I imagining things or is does inspo feature behave very different lately and is not so great anymore?

[thick osprey]

My voice in the song start talking a couple of times. What do I do to get it singing the whole song and not talking?

[knotty stag]

i find that putting ah, ah, ah, and ooh, ooh, ooh in to the Lyrics section works

[stable sand]

Hey does anyone know how to add your voice to make a song?

[steady yew]

Hey guys, I made a viking rock song the other day and today I had an idea to add some spoken verses from the havemal to it but I'm not sure how to do that, can anyone offer some guidance for me?

[quiet solar]

Depends exactly on what you're asking.

You can use +Audio in Create to upload a previous recording, or record it live, and use that to strongly influence what Suno creates. However, it will replace your voice with an AI singer.

If you're asking "How can I make Suno clone and sing with my own voice", the answer is not easily, and not something you can do purely within Suno.

[stable sand]

What is +Audio

[quiet solar]

[Spoken] or [Spoken Voice] before the relevant line/section in the lyrics is typically worth a try.

[steady yew]

oh yeah I know how to prompt it to say it but I already have the song made as I want it but I want to add the words now

[quiet solar]

Where are you adding them - beginning, end or smack in the middle somewhere?

[steady yew]

one verse at the begining and one at the end

[stable sand]

I use a phone could I ask to do it on a computer

[quiet solar]

It should be possible on the phone as well, under the Create options. If it's not obvious in the App, open Suno in your browser as that uses the same version as the PC/Desktop above