📚┃suno-school

[rotund phoenix]

Fun fact, I had still 3 days of PRO running well, upon renewing, these 3 days switched to free as well...and no daily credits. I just stopped using the program waiting for tech reply, but it is 3 days already...

[balmy slate]

Guten Morgen. Ich habe gestern ein Abo abgeschlossen und es wurde über Google bestätigt. Aber in meinem Profil bis dato kein Abo. Klicke ich aber auf Abo abschließen, steht da, ich hätte eins. Support hilft 0! Billing E-Mail Adresse existiert nicht!

[steady crane]

My next music engineering experiment will be plugging in noises that isn't easy to get. Like gears

[tender eagle]

Sprechen Sie Englisch? Ich spreche Deutsche Kleine, aber nicht genug um bei Probleme zu helfen.

[tender eagle]

in what sense? into the audio overpaint? '

cover,

[steady crane]

Suno isn't a LLM so I'm not using super complex builds but my engineering mindset remains the same. Essentially I use a bit more structure. Such as labeling my Keys, etc. Going to attempt to make it make robo sounds using that mindset

May or may not fire. 😅

[tender eagle]

Oh, thought you meant using non musical audio sequences in the overpaint/cover audio upload

which works in interesting ways if you do :p

[steady crane]

Basically like trying to..trick it by using some instrument cues to make gear sounds

This may not work 😅 I will let you know 💀

[azure crane]

You were totally right 😅. I have Suno more detailed instructions and it followed it perfectly 😍. So so much better… Thanks for the recommendation 🙏.

[steady crane]

AI loves more clear structure

[tender eagle]

drop the weirdness, and the prompt style weight, on the sliders, also , bump the audio influence for the vocals / instrumentals you added . If you're running vocals, your own, use the add instrumental, not the COVER

using cover is good for music + vocals and you have to be sure to correct the automatically generated vocals, including adding the [headers] like [verse] or [chorus] etc. else it'll do weird stuff somewhat

[azure crane]

Gonna try the instrumental thing also. thanks for the tip 🙏

[quiet solar]

Very much this.

[tender eagle]

correct the vocal's lyrics when you use add instrumental, and don't use anything with vocals for add vocals, gets weird.

[quiet solar]

Ahh, but that's half the fun.

When you pull off a song with multiple singers combining and switching lines coherently and play it to someone familiar with trying to make duets in Suno, you can hear their jaws bounce off the floor, shortly followed by the existential scream of "What prompts did you use?! Tell me! I have to know!"

And you can just smile and shrug and not let on that you spent 4-5 hours regenerating multiple sections, splitting the audio stems and recombining and rebalancing it all in an external audio editor. 😂

[nimble palm]

Pleasantly surprised Suno was able to nail an odd time signature from a cover. Style prompt: [Time-Signature = 5/4] Afro Cuban Jazz

[dim spear]

hello guys, is it possible to create a mechanical cover with suno? something like "create a mechanical cover of 'Artist - Songname'" so it keeps the same melody and you wont get a copyright infringement with it (with a mechanical cover license)

[polar spade]

Make sure you report in the bug channel as well. Support is more about accounts, which could be your (and others) issue, but if a bug thats affecting everyone they would forward to developers, who monitor the bug channel already. https://discord.com/channels/1069381916492562582/1082775596767117312

[polar spade]

No. You are tasked with tweaking and prompting to get the best sound-alike on your own. Suno and the better AI music sites try to avoid initially uploading anything already available to the public and definitely avoid any means of readily emulating sounds just from the prompt " sounds similar to . . . " or equal.

I have gotten some nice stuff "similar too" the vibes of other artists with meticulous tweaking. Granted its just for my own listening pleasure and would be useless to ever play publicly. So in that respect, also not worth for Suno to allow literal Covers of anything - no profit in it and crossing a line in publicity rights infringement for an artist.

As you mentioned "nechanically" similar. And in real life, to be mechanically similar means paying for that license to cover their song. Youre asking if Suno can go around that process for free.

[dim spear]

No, a mechanical cover means you are keeping the same lyrics and melody, you can still re-record the vocals, add different instruments, change the pace, structure etc

Im paying for a mechanical cover license through distrokid but if for example suno doesnt keep the same melody you are violating copyrights even with a license

I just need it to keep the same melody (and lyrics) as the original, everything else will be changed

[polar spade]

Straight covers means the arrangement can not be changed as well.

Nevertheless, AI generative music doesnt guarantee straight covers anyway, by its nature. So, that would still go against the mechanical license. You would technically always be making a derivative, which is NOT covered by the mechanical license from Distrokid.

Its a good idea in the sense that Id like Suno to provide the best prompts for an uploaded song, even if a copyrighted song, and then we run the prompt to get a viby song or instrumental, but without being allowed to use the same lyrics of the original song. Ive even put that in sunos suggestion box.

[lethal summit]

hi ,

Can I controlle the back vocals cuse Ilost all my cradits becues of that!

https://suno.com/s/sZGGCgZhWK7cuNqi

Was worth it haha

[shadow plaza]

I create persona i really liked the voice but everytime when i make new song dont use the same persona voice damn even sometimes changes the gender ..... The style reference is on 100% also tried to separate with stems only the vocals but still cant replicate the voice ....

What im doing wrong ?

[dim spear]

In a mechanical cover the arrangement is allowed to be changed

[wary plover]

Hello guys! Can you help me with a small prompt?
this is the problem. A lot of time ago, during v2 period, I got this stuff: https://suno.com/song/4b4972d3-bcdd-43be-b430-c5d6dc34c34f I want cover it on the new v4.5-all version, but I can't replicate that effect of a duet voice singing at the same time in a lovely harmony; like on the "sound of silence".
This is my best effort, but just got the music atmosphere, but the main voice isn't a harmony duet.
https://suno.com/song/c87d006d-acf3-4456-ad0d-85d268f3b896

[dim spear]

Persona is just the style box being reused it doesnt reflect the weirdness / style percentage from back when you saved it

but any dev keen on the internals feel free to correct me xD

[wary plover]

¿Do you know which prompt on the styles, or lyrics may help to get that same voice effect (but with better quality) that i got on the v2 version?
Am... the v2 version doesn't have a good prompt; was created with the simple suno form.

[dim spear]

yes, just tell suno things like "no back vocals" or only "male backvocals" etc

[dim spear]

You can add it into the lyrics section directly

[Chorus: male/female duet unison]

should work

[polar spade]

OK. Go for it. 👍 I just dont think Suno auto-censor has yet allowed for this. But theres always the suggestion box. https://discord.com/channels/1069381916492562582/1154587937611190354

[wary plover]

Hello! what do you mean that suno don't allow literal covers of anithing? for example with videogame music it works fine to create covers of songs 🙂

For example a nintendo cover:

https://suno.com/song/f5919e0a-9587-419a-871f-c6a64f864f5f

Or you're talking more in a legal way.

[dim spear]

Its perfectly fine afaik. suno either blocks the tags at the frontend "Taylor Swift" tag found etc or interpretes it on the backend differently "cover Taylor Swift 1:1" => "make a cover in the style of Taylor Swift". So ToS wise you should be fine, as long as you dont bypass any of their detections. But releasing a derivate or mechanical cover publically without a license is not allowed.

[polar spade]

Comment removed -- it had no effect

[polar spade]

Just that. Suno makes its best effort to block or remove anything that is not in the exclusive control of the user. Any publicly available sourced music or anything thats available or recognized as public, best attempts are made to keep it off Sunos site. Suno doesn't need to train on publicly available content other than use it to help censor future content from slipping through. But also, in part, its has no benefit to allow covers uploads of things Suno's algorhythms have already deconstructed.

[dim spear]

"But releasing a derivate or mechanical cover publically without a license is not allowed."

unless you are posting it on streaming services like spotify. they will automatically manage the proper licensing and pay the rightholders directly.

[polar spade]

There is a trick to stemming out the vocal, then adding vocals to vocals, indicating the original lyrics and maybe the accompanying lyrics in parentheses beside it. Then add back in the instrumental stem in Studio. Thats all experimental and not exact, but may give a new vibe yet add some harmony mixed with a duet type feel.

I did this with a little pirate poem so it sounds awful but still interesting to get a mix going that otherwise could not.

[quiet solar]

As far as I know, what you're trying to do is legally okay in the round, but going to be quite tricky to do within Suno's TOS.

In order for the song to fundamentally keep the original melody, you're going to need that melody uploaded as an audio input. You may recall this box when you first clicked upload audio:

So your immediate problem is that uploading the original, copyrighted song, breaks Suno's TOS. Even with a mechanical licence, you do not own or exclusively control the rights to the song. And I would expect Suno would block the upload when it detects copyrighted audio in the first place.

I believe this would apply even if you recorded a basic cover of the song yourself and uploaded that audio to Suno, but this is starting to enter into the territory of needing someone with deeper knowledge of the nuances of music copyright.

[quiet solar]

"The copyright detection system didn't detect it" and "I'm allowed to do this under Suno's terms and conditions" are not necessarily the same thing. 😉

[polar spade]

" . . . I see nothing....." 😊

[wary plover]

I've tried it on the verse section, but doesn't work... Ill gonna try it again.

[lapis snow]

Has anyone figured out how to make noise free song? Almost 90% of musics I make has the hissing noise.

[wary plover]

Sad; I doesn't have access to studio right now, but thanks for the tip.

[wary plover]

I got that with a greenday's song cover . I must change the lirics, and suno do that it wants with the melody, but hey, the final work sounds like a spanish variation of boulevard of brocken dreams.

[quiet solar]

That's going to be really tricky to do through prompts alone. You're already doing pretty much all of the fundamental things right in your v4.5 prompting.

Essentially, you need to have a cover version that 'gets' that you have two singers in harmony right from the start, rather than going with a lead singer and a supporting singer who only comes in on occasion. That may take a lot of attempts if you're trying to do it in a single, one-click-and-done cover.

An alternative is to cover the song with a single primary singer, and then extend from near the end and redo the song to create a new version in which you deliberately try to introduce a very different singer (e.g. switching your tenor lead for a bass-baritone lead). You can then split the vocal stems from your extension and overlay them onto the original using Suno Studio (if you have a Premier membership) or an external audio editor.

It probably won't be quite as tight on the harmonies between the two singers, but again, the only way to do that is to make sure the cover replicates both singers from the outset. You can then tactically use Extend to rewrite it section by section as you go along if the song starts out getting it right, but switches back to a single lead a verse or two in.

Really lovely vocals on that v2 original, by the way.

[karmic lion]

Hello I having technical problems with accessing the v5 version of subscription does anyone suggestions other than contacting support which done several times

[wary plover]

Thanks for the answer; ill gonna try wsomething. But yea, that result on the v2 was really awesome. Sad that is hard to replicate. ¿Maybe v5 got more tools to get voices in harmony? but well. simply try more thinks.

[placid spindle]

Does anyone have a secret on how to get rid of the SUNO shimmer / ringing noise that creeps up constantly in downtempo songs? Instrumentals.

[lapis snow]

This is so frustrating, the only way that works for me is to make over 100 songs and hope one or two are noise free!! Suno is the only AI that gets worst by the time!!!

[placid spindle]

100%!!! the songs start clean and then get very noisy in the middle to end

Remix works sometimes

[polar spade]

And keep in mind that may be considered more of a "medley" , "unison" or "accompany" or "accompaniment" than a "duet". Those can be tried, as well as add genres popular in those such as musical, Broadway, variety show, ensemble

[placid spindle]

but does SUNO know about this? I mean its talked about all the time here

[lapis snow]

Im sure they know, but struggling to fix it. Ive been using suno for work and now its been 3 weeks I can’t use it because of the noise

[lapis snow]

I always report the song as low audio quality, Im hoping they see it

[placid spindle]

thats what is happening for me. It seems to be getting worse. Is there a direct email contact for them?

[quiet solar]

v5 is a bit better at taking prompts from the lyrics box, which helps, but anything involving multiple singers is still very unreliable and takes a lot of patience and/or workarounds.

[lapis snow]

[e-mail]

[placid spindle]

OK thanks!

[lapis snow]

Ill also email them now

[polar spade]

Ytou can focus more on extending every 2 minutes or less. Than try to prompt for the whole song. Thats been a help for others.

[quiet solar]

Shimmer was largely fixed in the update to v4.5 all, but it might be there are still some situations where it's regularly causing problems. What genre are you guys creating songs in?

[lapis snow]

Deep house, downtempo and electronic

[placid spindle]

Thats what I am seeing!! My styles are smooth jazz, downtempo, nostalgic

its more hidden in downtempo, but smooth jazz and nostalgic take a huge hit

[lapis snow]

I have tried everything, making clean persona, separating stems, remaster and nothing works

[placid spindle]

I even put [clean master mix] [final clean mix] in my prompts with minimal luck

[quiet solar]

Hmm...and it's consistently coming in from completely new song generations (ie, no cover, no persona, no audio upload)?

[lapis snow]

Even with cover and persona

[placid spindle]

NO 100% new song

[lapis snow]

What genre are you doing?

[placid spindle]

mainly instrumentals

[quiet solar]

Yeah, that won't help sadly. The shimmer isn't something Suno is intentionally bringing in thinking it's music.

[placid spindle]

its getting to the point where I can't use Suno. But I don't know of another AI music platform as good as SUNO.

[lapis snow]

Yes way less problems with instrumental

[quiet solar]

A range - rock, pop punk, orchestral, broadway musical, comedic, metal....no problems with any of them since v4.5+

[placid spindle]

I actually have more problems with instrumentals. smooth jazz and its 60% bad

[quiet solar]

Massive, massive headaches before that though

[lapis snow]

The only way would be just wait for couple of weeks

[placid spindle]

Glad to heat they fixed some of it.

[placid spindle]

why?

[lapis snow]

Same thing happened in version 4.5 and it got resolved in 3-4 weeks

[quiet solar]

It's an issue which required a technical fix last time it raised it's head. If you've seen multiple posts about it, it's likely Suno's support team have had multiple e-mails about it.

[placid spindle]

So, any word on the next release - bug fix?

and thank you !! Its good to see that its not just me.

Also, the extend thing - I can never figure that out! LOL Doesn't work for me

[quiet solar]

That'd be a question for the Suno team, rather than me. AFAIK, they release updates when they're ready.

[fluid copper]

I've found that if I download the stems and put them in Audacity, there's a stem called "Other". That's where most of the noise is. Mute that stem and listen to see if it helped. If it did, delete the stem and output the song. (Sometimes the "Synth" may have to have the volume dropped in some places).

[lapis snow]

Thank you, Ill try that now

[placid spindle]

Just thought of this: if they fixed it in 4.5, should I be using 4.5 instead of 5.0?

[quiet solar]

Trying 4.5 can't hurt.

[placid spindle]

Can I work with the stems IN Suno? Just mute it there.?

[placid spindle]

Gonna give that a try

[fluid copper]

I don't know. You'd have to ask the smart people here. I think you could only work with stems in the new Studio.

[placid spindle]

ok

[quiet solar]

Only if you have Suno Studio. Otherwise, Audacity is a free audio editor with a fairly similar interface.

What you can do is something like:

• Extract all stems
• Download to your computer
• Import all the stems into Audacity
• (hopefully) Delete the offending one
• (if it's not that simple) Use Graphic EQ under 'Effect' and play around with the volume of sounds around the 8k - 12k mark, from memory. Shimmer tends to appear there more often than not.
• Export the corrected file as an mp3
• Upload the mp3 to Suno and continue as you were

[placid spindle]

thanks!!

[fluid copper]

What is the difference (in general) in the two sets of stems they provide?

[lethal summit]

the question is can I make the sing specific line and lyrics?

[thorn lava]

i dont know where this belongs but why am i not able to write lyrics on the website? i went in and generated a song and i wanted to add some lyrics to one of the verses but no matter what it wouldnt let me type, even when i clicked "edit lyrics" i then copied it to try and start over to paste it in... cant do that either? whats up w that?

[vestal folio]

Hello, is it possible for thesuno to generate an instrumental and a topline, and then tell it to write the lyrics for me, without distorting the lyrics or anything else, please?

[lapis snow]

Just tried something and it’s working better. Try to have super short and simple style prompt.

[quiet solar]

The simple split attempts to separate the vocal track from the combined instrumental track. The 'all stems' split attempts to further split the vocals into lead and backing, and isolate specific instruments within the instrumental track.

Please note that they are not actual stems - Suno doesn't create the individual pieces separately and then blend them together, it creates everything all at once and then tries to split them apart afterwards, so individual stems may not sound perfectly clean in isolation.

[quiet solar]

You'll have to be more specific. Happy to help, but I'm not sure what you're trying to do.

[half dragon]

I'm not sure I followed but did you want to replace the lyric on the already generated song? or just want to edit the written lyric to match what's being sung.

[quiet solar]

Very easily if you're creating the song from nothing.

Not quite so easily if you've already got an instrumental with no vocals and want to add in vocals over the top afterwards. You have some options to do that in Suno Studio; if you don't have access to that, your best bet is probably to try doing a Cover of the song, or uploading it as audio inspiration for a new song, and then creating/adding your lyrics from there.

Bear in mind that they'll need to fit the structure and flow of the existing audio to sound natural, so it's not an easy task - and that it's likely that the Cover / new generation process will make changes to the instruments as well.

[polar spade]

You mean you dont already? 🤔 joking -- but sure, even with 3.5 quality not always best, I use a bit of all of them as they have their strengths. Then we might cover or remaster in 4.5 or 5.0 for better fidelity. Or even take offline and clean them up.

[thorn lava]

no. its pre generating

[half dragon]

Oh ok - I was trying to figure out what could be the cause but unless it wasn't in "auto" mode, one can type in the lyric box freely. What browser was it? have you tried clearing cache?

[trim tree]

Hello!
I can't post on my Suno profile. Can you help me with this? Thank you.

[white python]

Is anyone else not able to access their workshop? I cant create and access my workshop, it always just says "Error fetching feed_v3 clips"

[wise kindle]

it's been mentioned a few times in #🪲┃bugs-web so I'm sure they are looking at it

[white python]

Ohh amazing, thank you so much, was kinda worried already, but this helps so much ❤️

[zinc loom]

Hey there!
Short question: Does anyone have problems with noise in extracted stems?
I use v4.5 an v5. Everytime I extract Instrumentals/Vocals from a track, the files generate some basic noise in the background through the whole track.

[wise kindle]

unfortunately isolating instruments isn't an exact science, so what you're hearing may be artifacts from surrounding layers

[zinc loom]

thank you, makes sense now that u told me

[wise kindle]

No worries! You can still try and sheer off the noise in third-party tools, but yeah, it's not uncommon to hear a bit of noise, especially if there's a lot of complicated layers playing at the same time

[elfin gyro]

Why is Suno suddenly today constantly mixing vocals, and I clearly gave him the command what is singing male and what female vocals?? Please fix that bug, because it can't make a normal song, we're wasting credits for nothing! So far that has not happened!

[wise kindle]

Can you show us an example of your style prompt and lyric tags? If you only want one gender, I wouldn't use both options as there's no point and it might just confuse it.

if you don't have a specific vocal style in mind and just want it to be male, the most sure-fire way is to remove it from the style and just use the vocal gender option

[elfin gyro]

Ok, now when I checked "Male" in this vocal gender, now is ok. Hmm strange, because until now, I gave him commands from styles for that and he is listening... Thanx very much for helping, Suno is amazing for sure, keep the great work guys!

[wise kindle]

Glad to hear it! Yeah that Vocal Gender option works really, really well

[clear robin]

If i take my song and i press on the remix option, will it take the same exact song with sound and everything the same and just let me add new lyrics???????? Need help on this.

[raw oasis]

you would need to move the Audio sliders up in the 80% range. but wont be the same song on a Cover, you would want Premiere Subscription for Studio and then could regen just vocals for example.

Often in my experience you can manage to manipulate your lyrics while covering and get the song you want.

[clear frost]

hi i am having issues with the editor, when trying to edit one of my own songs

when i open the editor, a track starts playing. Not the track I am trying to edit, I think

i cant stop it except by closing the editor

[raw oasis]

Do you have Cover + Add vocals in your sub?

[raw oasis]

If you see the Legacy Editor (link) click that and then back, that is what I do to stop the autoplay. That was a few days ago. still happening huh?

[clear frost]

ah its a known thing? ill try to mess with it

[raw oasis]

upper left corner is Legacy Editor link

then I just back right back to previous page.

[clear frost]

ill give it a go, cheers

yuo are a star, that worked 🙂

[raw oasis]

relief.. silly bug needs to be squashed.

[sharp hull]

Hey y'all! I'm pretty new to the suno game! I'm running into a problem.

I have large chunks of text that aren't really lyrically written that I want to put into songs to help memorize them! About 10% of the time my prompts give me absolute bangers!! With tons of instruments worked in! But the other 90% of the time it's kinda just monotone with a beat in the background! Any suggestions on promoting would be amazing

[tawdry junco]

I don't see Workspaces anymore?

[frozen dagger]

it looks different but it's still there

[main hazel]

i generated a song i liked on at pc suno, sonehow used the app later and the song got messed up, i thought once song is released it can no longer be modified? any idea how i can reverse the song to previous version?

[tawdry junco]

I found them in the Create tab. But we don't have pins anymore?

[frozen dagger]

we do. only viewable within the create tab as far as a i know tho

we do not have pins for the "My Workspace" folder. other created folders have pins

[quiet solar]

Even once it's generated, let alone released, I'm not aware of any way to accidentally modify a song.

Just to rule out possibilities, are you 100% sure the original didn't get deleted and moved to trash, or disliked and hidden by accident?

[main hazel]

its still there i think, tx for makin clear, for a second i was rly worried coz song turned out good

[clear robin]

I have the 1 month subscription

[steady crane]

I successfully made V 4.5 do weird mechanical cog sounds

Not too consistent but made it produce it

[steady crane]

Basically sounds like metallic grinding

[steady crane]

https://suno.com/s/YUt42XvuHibZ9wZX

Experiment share

Made it make gears and industrial drum sounding instruments

[gleaming basin]

Hey anyone know where the editor feature moved to?

Im used to seeing it in the remix/edit but its not there. I thought because im using my phone (house internet down tonight) was the reason I couldn't see.

Actually disregard. It's my phone view. Oops I opened desktop mode on the phone and its there. Sorry but thanks yall 😘

[elfin gyro]

I’m on a Pro plan and today Suno is failing to follow vocal commands across all versions (v5 beta, v4.5 Pro, Pro+). Even with clear prompts for [Male Vocal] or [Female Vocal], the engine ignores instructions and sings something else. Pronunciation errors are frequent, and male vocals often sound distorted. This issue used to be rare, but tonight it’s constant — I spent ~300 credits trying to make 2 duet songs without success, because he sings what he want, event if I check "male" or "female" options in vocal gender. Sometimes he is listening commands, sometimes isn't. Please confirm if this is a known bug and if a fix is in progress? I think this is definitely a bug and it is really annoying.

[half dragon]

It's known that duet is unreliable via prompts, no matter what model versions.
If you searched this channel, there are plenty of discussions regarding that.

[elfin gyro]

Ok, so is fix of that bug in progress, or what? I need to finish my album, but can't finish duet songs because of that.

[half dragon]

There is no talk on any fix regarding that.
You can try different methods to finish your song, one being building the song by section with the use of different Personas, and extensions...etc..

[elfin gyro]

What personas and extensions, please tell me more?

[half dragon]

If you have an existing songs with a vocal that you like, you could create a Persona for it, one for Female, one for Male for example.
Then build your first verse using the first Persona, then extend the first verse to add the 2nd verse with a different Persona..and so on...

Keep in mind, it's still taking sometime to get it right .. like everything else in Suno, it would probably not be perfect on first try.

[queen linden]

man the last six are horrible 
getting a lot of middle eastern sounds .... Nothing close to good for a week, suno can keep its own music its not what i want to hear,,,,

[quiet solar]

Hey there.

First thing to note is that prompts are not commands. They're more like suggestions, which Suno may or may not take into account when developing the song. That's not a bug, it's just how the model is.

Changing vocalists mid-song is something that Suno is capable of doing, but not reliably - especially not mid-verse or chorus, or in complex arrangements with no underlying pattern of how the singers switch. In a nutshell, every time you're looking to change vocalists, there's a chance the singer will switch when they shouldn't or won't switch when they should. Suno models have been getting better at this slowly over time, but it's not something you should expect a "fix" for any time soon.

Creating duets at the moment requires a mix of careful prompting, luck and either a lot of regenerations (300 credits is not a shocking number) or deliberate workarounds such as using multiple extensions to slowly build the song up from pieces that are correct, or to prepare vocal stems for both singers that can be blended together in Suno Studio or an external editor.

If you search for "duet" or "male and female" in this channel, you'll soon find examples of best practice around prompting and workarounds for v4.5 and v5 - this question comes up a lot.

[delicate bloom]

anyone know about suno studio here?

[elfin gyro]

Hmm, I wouldn't say, because 2 weeks ago, even with the free model, I got almost immediately what I was looking for and a great duet, Suno listened to voice commands nicely, without problems, and now he degraded a lot, including the best models 4.5 pro and 5 beta, and he also degraded vocally, I don't know what happened to him, but just a few days ago he was much better, the degradation was noticeable and suddenly he mixed everything up and doesn't follow the words properly. A few days ago I made a lot more songs with much less credits, and today I spent more than almost 400, and I barely made 4. I don't know what happened??

[quiet solar]

Congrats on getting a great duet working first time. But your more recent experience is much more typical.

[polar spade]

Just in case you are using the sliders, dont crank any of them (Audio or Styles) to 100%. This can get odd results similar to sliding the weirdness slider up high .

[steady crane]

Technology isn't able to do real duets. We simply don't have models that can do it yet. It requires next generation tech and multi agents. As an AI engineer I'm telling you duets are simply not reliably gonna happen not on Suno.

[quiet solar]

It can be done, it just can't be done reliably. Hah - literally just saw you edit to make the same point. xD

[steady crane]

No. It cannot.

Forcing the model to fire duet when you command it to is currently not realistic. You will at most roll dice and pray.

[quiet solar]

Yup. And some of us have got quite well-versed in the dark arts of trying to load those dice, or creatively trick the lead and backing singer into doing what we want, and having backup plans and workarounds for the points where it inevitably falls over.

But yeah, it's not a "what prompt do I use" solution. It's a "here's various ways you can creatively use Suno as a tool to build up what you want" approach.

[steady crane]

I work with these things in the lab lane. I'm telling you what you're asking about is NOT actual duet in current hardware. This requires multi agent.

To reliably actually control it

Trying to control it, really control it is not possible yet.

You are a scammer.

Shoo

[quiet solar]

Yeah, it's mimicking the end result rather than having a framework where you can actually direct what multiple singers/instruments do.

But it's still possible to work with that to create a duet, even if it's beyond the model's capabilities to reliably facilitate it. It's just not possible to do it by prompting for a duet and hitting 'create', unless Lady Luck is really, really smiling on you at that moment.

[steady crane]

But back to this. When you know what these models are actually doing. Duet right now is at most flavor but isn't real true duet yet. Because it requires a different technology.

(Right now it's still technically 1 agent overlapping 2 different sounds.)

But you aren't getting control. You cant make it actually sing one voice as male and another as female strictly. This requires a model we don't currently have. Since suno isn't actually doing a duet on technical level.

Also apologies if me explaining is poor. Just woke up. Worst time to tech babble 😅

[quiet solar]

Haha - no, it makes sense.

[steady crane]

Gonna prob try go back to sleep for a lil. Woke up a bit early

https://tenor.com/view/helluva-boss-helluva-loona-helluva-boss-loona-tired-gif-13172053635098397737

[placid aspen]

Hi all. Anyone got any working knowledge regarding stems, and re-importing songs back into Suno after mastering and swapping out stems off-site, ie in Moises? I've been getting some great results by using Suno for composition and creating the basic 'vibe' of tracks, then taking the tracks across to Moises and replacing the messy Suno stems with cleaner replacement stems. I'm wondering what happens if I were to then re-upload the clean versions back into Suno for covers etc. Would Suno retain the cleaner stems, or would it mangle them back together again?

[knotty stag]

My Workspace if the default workspace, until you figure out how to create named workspaces for each project or song you work on. You should move songs out of My Workspace and into a relevant workspace, to avoid clutter and not being able to find earlier songs you created.

[cosmic void]

Is this the right channel to figure out how to get suno to properly create call and response within lyrics?

[quiet solar]

Suno doesn't really have 'stems' in the classic sense. When it interprets or generates audio, it listens to the whole thing at once and creates the output all at once. When we extract stems in Suno, the AI creates them from the final version - hence why they sound messy, because they were never really separate tracks in the first place.

So when you reupload to Suno you'll have a much cleaner sounding audio, which may help for things like covers. But yes, once you actually do something, it'll effectively mangle everything together again.

[quiet solar]

There's no consistently reliable way to ensure this happens, but if you search for "call and response", "duet" or "male and female vocals" you'll find some best practice of prompts and potential workarounds you can try.

[polar spade]

Toil...toil... boil... and bubble.... make these vocals sound like doubles. 🧙‍♂️ 🦄

[polar spade]

You should also try, isacanthymia as a style. Its literally the African call and response style singing. Be sparse with it, such as put near the end of prompts, to help with call and response ( and a few harmonies as well).

But at least run a couple just to hear what it sounds like.

[calm karma]

ok so many different how toos out there i am trying to figure out the proper way to tag this. Which one is right?

1. [Intro: Sharp Industrial Beat, Electronic Glitch]
2. [Intro]
   [Sharp Industrial Beat, Electronic Glitch]
   3.[Intro - Sharp Industrial Beat, Electronic Glitch]

[polar spade]

If you tag to get a clear male on one verse and female on another, then on a stanza you want C&R you can often be successful with just using parentheses per response lines such as:

[Male vocal] "show us the way" (yes please show us)
"Bring the spirits to us" (we are waiting)
It is done (it is done) (it is done)

The female vocal often naturally sings the parentheticals.

@daring lava Did a recent song with an audience call and response too as well. I just cant find it at the moment. But it was a simple tag of [audience call and response] or something close, before the stanza.

And @quiet solar mentions to just search for songs and see the many ways to arrange lyrics or tag for them.

And as always, extensions and splices are our friends.

[onyx sparrow]

hi guys is the persona connected with the beats? because i created one persona and i get always the same beats

[half dragon]

How about extract the stems and create Persona on the vocal stem?

[calm karma]

if your trying to get it for a specific vocal style then split the stems and just use a vocal to create the persona from. and play with the mix slider for it.

[quiet solar]

All three should work absolutely fine. The natural language element in Suno means it's not like coding, where the structure has to be put in a specific "correct" way. If it's in [style prompts], Suno knows to intepret it as such. 😉

[onyx sparrow]

that means yes?xd

are the persona conneced to the style of the beat?

is

[calm karma]

it can be

[quiet solar]

Anything in the persona audio can influence the output of songs created using that persona. So if there's a beat in there - yes, that can be brought into the new song too.

[onyx sparrow]

okay tx

[small bridge]

sorry to ask but where do we put general questions? I tried in general chat but it just got swallowed by the convo there. I am wondering if there will be a suno sale for black friday

[quiet solar]

The place for general questions is General chat 🤷‍♂️

[half dragon]

or try #🗣️┃chill-chat 🙂

[polar spade]

They dont have any inventory for a Black Friday and they just finished a big 50%-25% off sale in October. Partly why the support response went from 2 days to 4 weeks.

So probably not likely.

[hoary plover]

I've tried to create a Persona, and it fails every single time. Tried ona different browers, incognito mode....nothing, continues to fail 🙁

[polar spade]

If it specifically is blocking use of Make Persona due to some standard protocol,it would just be greyed out.

What does failed mean? You click make persona, then click GO and an error occurs? Or it just spins and nothing happens and eventually times out?

[polar spade]

Here it is greyed out on an uploaded audio. It won't make personas on those or outputs containing uploaded audios.

[hoary plover]

You type in the details, add image, bio, etc.. click Save and at the bottom a red box appears "Error: An error occured while creating the persona"

[hoary plover]

As an FYI, mine does not look like that, at the top is v5. On the left panel I click 'Create' and then '+ Persona' at the top

[half dragon]

Yours is Create song page where you could attach an existing Persona to create the song.
@polar spade screen is where you actually initiate the creation of a new Persona

[hoary plover]

Ok...so what am I missing here, as not seeing where to go?

[polar spade]

Its under the songs three-dot menu.

[half dragon]

Click 3 dots menu on any song that you want to create Persona from.

[polar spade]

That one is "Add" Persona.... ygou would be adding a persona which you've already made... heres a video

[hoary plover]

Found it...thanks, but same error in creating

[polar spade]

Are personas the only thing you're having errors on. Youve been creating outputs OK recently the past hour?

[half dragon]

Might be the name? try a different name?

[hoary plover]

Am new to this app, so not overly familiar...but have been able to produce and publish songs. Editor is a bit quirky, as it seems to want to play a song in the background, though not the song that's in the editor, so while I'm editing, there's another song playing LOL!. Sometimes a refresh fixes it, sometimes not. I have one song where it shows the correct image in the workspace, but for some reason it's not displayed in the library. As I'm new to the app, not sure if these are identified bugs, or just random errors. Different browser or different device, etc.. doesn't change anything.

[hoary plover]

no luck, same error.

[polar spade]

I saw youre talking about Shakespeare's couplets too... are those in your lyrics? Its quite possible of using any publicly available text, that even if it didnt censor at the output level, it might at the persona creation level.

Like @half dragon mentions , text too long, or even proper names (capitol letters) might jam it up or even get censored.

Try one with NO text but a title. If it works , edit in text later.

[hoary plover]

Let me simplify it and see..

[half dragon]

Try without it - then if it works, then edit it to put it back in and it fails to save then, you can can then go through it and clean it up

[polar spade]

Also, Suno does a lot of updating. So diligently cleaning cashe closing browsers and restarting things help you and Suno shake hands properly, and yes, incognito can help in that regard as well.

[polar spade]

A persona is just a "prompt" as it were, to capture both the vocal characteristics, as well as instrumentation and feel of the song, so you then apply that vibe or feel it to other songs. By itself, its nothing. So that bio style description ytou have would not inform anyone of the kind of style that persona would pass across to a new output. That make sense ?

Im lazy mostly and just put down " high pitched country girl with harmonica and dirigidoo" or nothing at all.

[hoary plover]

Ok...went bare bones with it...no bio, no pic, changed the name..and it worked. If you could point me in the direction of where to "Edit" the Persona (or is that even possible).

[half dragon]

Sure. Go to Library page
Find the Persona tab on the top
Click 3 dot menu on the Persona > Edit

[polar spade]

Cool. Glad it worked. Many often get a general error like this, so if you recognize the exact same thing with someone else, pay it forward, please , and mention what you did. Thanks.

. . . And welcome to Suno. There is fun in between this kind of frustration, so weve all been there. Just enjoy the ride.

[hoary plover]

Thanks guys, appreciate you

[golden nimbus]

are hooks not working for anyone else? I know that @snow lark is having problems with it, just like I am. it's not the first time either.

[snow lark]

yes i'm not finding hooks anywhere on my page

[small bridge]

that makes me sad. I didn't hear about the sale in ovctober or i would have grabbed it. But thank you for responding!

[quiet solar]

@golden nimbus @snow lark @obtuse perch Respectfully, these are all questions better suited to raise with Suno's support team ([e-mail]) rather than here, which isn't somewhere that the Suno devs regularly monitor.

[elfin gyro]

But what's interesting, a few days ago the free model generated for me almost from the first 3 different versions of the same song and it listened perfectly to the commands for the vocals, who sings what, and now not a single pro model can do it at once, I tried all possible ways, it's not worth it. The real question is how did the free model 4.5 listen to commands, and these advanced ones don't?? I guess they are better and smarter, more powerful, what's the point?? So how to make a duet in the future, no way?? With Suno, the only thing we can do is make songs with separated vocals? It's strange that Suno does wonders with sound and voices, but doesn't understand simple imputations??

[jaunty swallow]

Just posted the result and the prompt on TikTok if anyone is interested in this style: 🔗 https://www.tiktok.com/@suno__vibes/video/7574789945697504568?is_from_webapp=1&sender_device=pc&web_id=7574771711808538129

Hope it helps your generations! 🎹

[quiet solar]

@elfin gyro Again, your most recent experience is the 'normal' experience in trying to create duets and other similar songs, and that has been the case going back multiple versions. I appreciate why it may seem like the 'normal' situation was where you got lucky and everything seemed to work perfectly, but that isn't the case.

Prompts are not commands - they're effectively suggestions that Suno may or may not pick up on. It will also be heavily influenced by the library of music in the genre that it's trained on. If we're talking in the context of a duet, Suno can and will look up how other songs in the genre have handled using a male and female vocalist, and it will also be influenced by other songs that have an explicit lead singer and a backing singer who sings occasional lines, or just does harmonies, or sometimes takes on a specific section of the song.

I had a similar experience - the first duet style song I made in v5 as a test seemed to work almost perfectly, just a few minor mistakes. I assumed that was Suno taking the prompts into account much more than it had previously, until trying to fix that in Extensions and using other methods demonstrated I couldn't make it correct the problem. It was the same influencable-but-not-reliably situation as always, and I had to use the same workarounds as in previous versions.

[shy badge]

A cool little Gemini Gem: Your AI Sound Designer. Turns any input into production-ready Suno v5 templates. Features auto-mixing, phonetic enhancement, and "Liminal" atmospheric texturing. Perfect output, every time. I'm still slowly developing this Gem. For example, the master prompt needs to be blank. For now if you dont want that liminal psybient vibe. Let it know. Tell it something like: dont add any lyrics, styles etc from the master template. That is the workaround for now.

https://gemini.google.com/gem/1AyygOUCwNR_aFvCV95O1-5-gA8i3wNHT?usp=sharing

[onyx sparrow]

never used gemini is it good?

[torpid thicket]

Found that Vocaloid vocals don't work with v5 - Can we get a fix for that on v5.5?

[last wolf]

Can this channel answer questions about songwriting and effects?

[quiet solar]

Depends on your questions and who's here, but by all means ask away...

[last wolf]

So, I'm trying to make a part of the generated music have a break in the instrumental between the verses to emphasize the vocals

[wise kindle]

can't say I've ever tried that before. I don't know if Suno would understand, but perhaps [Vocal break]lyrics or {A capella]lyrics

I'm more inclined to lean towards a capella

[last wolf]

Okay, got it, I'll run some tests here one at a time to see what I can come up with, and I'll send you what worked best for me in case you want to use it in the future too.

[wise kindle]

sounds good. I like those dramatic instrumental breaks, so I'd be curious to see what worked

[last wolf]

I even have a version of this that I needed, which was generated without asking, and I was like, hey hey hey, this is what I'm talking about

[quiet solar]

I've tried getting an [A capella] prompt to work in the past, and it's definitely a bit hit and miss, so you may need to give it a few tries. Suno can be quite stubborn at the idea of ensuring instruments are playing throughout except for brief dramatic pauses.

If you have Suno Studio, you'd have some additional options from being able to silence the instrumental stems where you want to, or import from the version you just mentioned where the instruments did what you wanted by accident.

In the absence of Studio, you could do something similar with an external audio editor.

If you want to create the entire thing purely within Suno and without Studio...then yeah, you'll have to rely on prompting and regenerations. Let us know if you have any success with that, though, since it'd be useful to know if there are prompts that are reasonably reliable for that kind of effect in v4.5/v5. Good luck. 🙂

[merry sage]

It happens if your image is too big.

[last wolf]

The thing is, I was thinking of something like that, but I still don’t really understand all the features, especially since I’m currently only on the free plan. So I improvised by using the music that came with it, without me asking, as the cover to generate the next ones, but then I got kind of stuck because it didn’t reach the same tone I had in mind.

[quiet solar]

Totally empathise. If there's a feature or element I've referred to that you don't recognise, just say - happy to explain further.

If you're on the free version, you're going to be limited as to what you can do because you don't have access to split out stems (Pro), edit within Suno (Pro or Premier for Studio), Remaster (Pro) or the advanced slider settings on Cover (Pro). Any or all of those would be very helpful for this type of exercise.

I'm not sure what changes you were trying to make to your earlier versions, but your best bet is probably going to be Cover on the version that did mostly what you want. Either way, anything you put in square brackets in the lyrics box will be treated by Suno as a style prompt, so you can try putting [a capella] or [instruments fade to silence] or [vocals only] just before the point in the lyrics where you want Suno to pause.

If the overall vibe isn't working out in your Covers, it may be worth approaching it from the other direction and pay closer attention to what you're putting in the style prompt that affects the song as a whole, and whether there are any additional prompts or stage directions that you want to include against the lyrics for specific sections.

[velvet jungle]

Can anyone answer why when leaving a comment it says " you don't have permission to leave comment blah blah blhq...now what's that all about ?

[half dragon]

Maybe you're not allowed to comment or been blocked.

[velvet jungle]

Good observation, not sure which of the two, but the question is why and how and why and how again... makes no sense..

[half dragon]

It could happen by accident - reach out to the person if you know them. They might have blocked you by accident.
But if they did on purpose - then there's nothing you can do.

[velvet jungle]

Mhhh alright now, happened with others as well.., no way to reach him I just follow him, who knows guess I'll just let it be then... strange, well thanks for reaching out my brother, cheers 🍻

[half dragon]

It could also be he turn off allow comment on the song recently.

So it may not be you being blocked.

[onyx sparrow]

i need more space in the style section:>

i already used 1k words

[velvet jungle]

Yeah I thought of that, but I literally tested a dozen of his songs and can't comment on any.

[velvet jungle]

Yeah well I can't ask him since I can't reach him

[fathom jasper]

Guys do you have also weird electric sound like is broken something in the background in the songs ?

[half dragon]

You might want to post a sample if someone could listen and probably knows what it is if they experienced the same. Without it, it's kinda hard to know what you meant.

[onyx sparrow]

when comes v6?

i need better ai that the can blow up the planet faster

they

xD

joke

[prime prairie]

You need to do some serious training on a cappella! This is pissing me off! I have wasted SO MANY credits trying to get out without instrumentation!

[raven turtle]

I’ve had some success by creating a track with vocal and minimal instrumentation, then stem out the vocal and remaster it.

[quiet solar]

Good suggestion. Unfortunately, CoyoteGuarana is on the free version and so doesn't have access to stems. :/

And on that note, @prime prairie - if you're on one of the paid plans, you can extract the audio stem for vocals. Instant a capella.

[still frost]

does anyone know how to make realistic cinematic videos for my song

[hollow crown]

Hi all - where can i find help with my South African card being flagged causing that i cant subscribe

[quiet solar]

@hollow crown Not in the tech prompt discussions. Contact [e-mail]

@still frost You might be better off putting that question into your search engine or LLM of choice.

But if you're talking about making a "realistic" full-length music video, your options are going to rapidly come down to;

1 - Film it yourself, if you have the equipment and skills to do so
2 - Hire a professional agency to work with. Guys I've worked with before will typically charge around $2,000 - $6,000 for a professional but tight budget affair
3 - Invest in a subscription to one of the better AI video creation suites (you won't be able to do more than about 10-15s of video on a free plan) and skill up on how to do video prompting - for the better end, you're probably looking at investing $200 - $800. Obviously no guarantee that it's going to look "realistic" on the far end though.

[red oar]

Speaking of videos... Anyone knows how to get the alternative lyrics video format shown here? Is it a setting, some property of the input, or a test activated for some users? #🐣┃suno-showcase message

[still frost]

10-15 sec is good cause im just gonna repeat it

[quiet solar]

@still frost In which case, search out a few AI video makers with a free trial and knock yourself out. Good luck.

[still frost]

can u help?

its just some cinematic shots

but i cant find a perfect ai or a prompt

[still frost]

?

[timid swift]

Koe help krega

[still frost]

kaisi help?

[timid swift]

Muje suno ai website ke related chaiye

Music nhi bn rha

Glti se upgrade mein click ho gya

[still frost]

upgrade kr diya kya?

[timid swift]

Ab upgrade krne ko bol rhe ha

Nhi upgrade Kiya nhi ha

[still frost]

mat kro na

[timid swift]

Phele monthly subscription li thi 850₹ ki

Kuch he din hue ha

[still frost]

pata nahi bhai

[timid swift]

Lekin jb se upgrade pr glti se click hua ha tb se music nhi bn rha bs jhi aata ha ki your subscription is past due

[still frost]

mene to kabhi nahi liya

[timid swift]

Ok bhai

Thanks

[quiet solar]

@still frost I don't have a good recommendation for you, sorry. Best of luck with your search. 🙂

[still frost]

aight

[hollow pagoda]

hello, i don't know where to suggest this, i hope it get fixed, but Any hardstyle and it's sub-genres struggles in quality, most of the generations that you get in output are low quality, too muffled. specially the drums and synths, and overall very bad results, and it tends to go offscale, (sounds off) sometimes aswell in these genres in suno. i love suno and been here for a long time, i hope you guys can improve the Hardstyle, rawstyle, euphoric rawstyle, hardcore, genres quality and results,
thank you ❤️

[polar spade]

I think its probably tough to have separated clear instrumentation and vocals of pop amd folks and then to get more distorted and harmonic sounds of hardcore and have it all work as you probably noticed when 4.5 came out. Assuming that tweaking in one direction interferes with tweaks in the other directions, my bet is they go towards tweaking the general all over tools , such as Remaster, towards the more popular trends of the 500,000 members and that usually arent visiting Discord. Including what likely is popular in streaming , radio and among professional studios.

Rare is a studio asked to demo for hardcore in the USA at least in Nashville, New York and LA. Hardstyle bands are good for creating their own music, and individual songwriters and artists dont lean towards those genres. Thats often more Scandinavian/Germanic, Ive seen.

All the other style genres from polka, ska, country, pop, kpop (still #1 so far) and so on have a more separated and cleaner sound. Rock seems obviously in the middle somewhere. Maybe even have two models one day. 5.5 and 5.5H (H for hard)

[hollow pagoda]

i agree that the genre is abit of a niche, but i have to correct you that hardstyle isn't really rock or metal, it's an electronic music with loud synths and bass and deep saw synths, etc, i know the name is kind of confusing because The naming of the genre sounds the same, however the music is electronic. one of the genres/ other name that i didn't mention is deutche krach. it is gainig attraction recently in USA and other places in the world, but i hope they do improve the sounds eventually. because i really like it and i like creating it aswell.

allow me to demonstrate what deutcher krach sounds like. "warning flashing video"
https://www.youtube.com/shorts/uAi0mlvh2LM

[sand hollow]

Best way to convince suno to make a trap remix? Lol

[polar spade]

Thanks for the correction. Even new age music and symphonies can have issues when theres too much orchestration or synths in the mix. Maybe more tools are needed. 😋

[hollow pagoda]

perhaps, and also maybe more training data for this niche genre, i hope my voice gets heard🙏

[sand hollow]

Why is it only in this specific page it takes me, it takes tremendously longer to do It’s thing, haven’t ever seen it been completed before lol

[undone basalt]

“Hey friends, how are you? Is there any way to get the same vocal tone across multiple compositions? I mean, as if it were always the same singer performing different songs, instead of the generic voice the platform generates every time we create something. Thanks in advance for your answers!”

[hoary plover]

Yes...just sorted this with creating a "Persona" scroll up a bit to yesterday where I asked the same question. You won't get the "exact" same voice, but it's better than a completely different one.

[waxen prism]

is there a way to change the lyrics that move on display while the song plays?

[quiet solar]

Yes, if you Edit Displayed Lyrics on the song and then re-download the video to get it to refresh.

[waxen prism]

what do you mean redownload, like regenerate?

[polar spade]

Now there is no generate video... its "download "

[waxen prism]

just for a song not a video

[hoary plover]

Ok...so you've created a song. The lyrics are not what is being sung and you want to correct it? You can go into Edit mode find the section of lyrics to update, in the pane on the left update and it generates to versions, then commit the one you want, and save as a new song.

[hard beacon]

for me it sounds like he want to changed the displayed lyrics in the video.. so the the solution by Kanthric and Dasvon is 100% the right workflow 🙂

[polar spade]

Right. Well, I was just answering the idea that its now " download ". But, yes, we would first edit the lyrics in the song page, assuming were just correcting it to be whatever is being sung.

Save.

Then personally, i pop it in Legacy Edit and sync the lyrics again before downloading video. @waxen prism last year we could sync the lyrics directly from the song menu, so the video karaoke lyrics would line up. Now its done in Edit.

[waxen prism]

so i will have to regenerate? i have 1 tweak & it’s legit killin me

you can’t do it in the free version i think you’re telling me

[polar spade]

You needing to change the actual sung words? Or just edit the text to match the lyrics to the vocals alread sung?

[waxen prism]

edit to match

[waxen prism]

edit to match ai put knifes instead of knives & it’s bugging me

[hard beacon]

i still don't get it... Do you have wrong sung vocals in your track or do you want to edit the shown lyrics in the video?

[waxen prism]

edit to match ai put knifes instead of knives & it’s bugging me

[hard beacon]

in the video or in the audio?

[waxen prism]

audio… the sing along lyrics that scrolling during the song…or is there a setting to turn off the lyrics scrolling banner

[hard beacon]

<@&1372996242325508176>: do you understand it??? 😄

ok MoonRockRose... so you want to edit the lyrics inside the audio?????

audio and scrolling doesn't make any sense to me

[polar spade]

This is where you would change the text to match what the vocals sing.

But it honestly sound like you want to change the vocals to sing "knife", that would be called changing the vocal lyrics, i suppose is a good term.

[worthy kraken]

Hello. I am trying to get two types of the voices, but i can never really get what i want. Maybe someone will have some tips?

Voice 1 - really deep male baritone or melodic bass. Examples (i never mention band names in my prompts, it's just to let you know what i want) - Lord of the Lost, 69 eyes, Deathstars.

Voice 2 - the adorable female voice, like those that are popular in japan - very feminine and high-pitched.

I tried hundreds of prompts, even tried to describe vocal type directly in lyrics, nothing helps. Maybe someone made any of those voices, and can offer advice?

[waxen prism]

audio is good- no audio prob

[waxen prism]

no vocals match… it’s a grammar problem in the banner display lyrics

[hard beacon]

MoonRockRose: don't understand us wrong.. sometimes you talk about audio and then about grammar problems in the banner display lyrics.. that are two different things..

so.. are we talking about grammar problems in the audioutput or about grammar problems in the video with lyrics..

?

[waxen prism]

yes! 🙌 detective Clark

[hard beacon]

ok.. let's try it on a different way.. what's your motherlanguage?? english? german? french?

[waxen prism]

english

[hard beacon]

ok... maybe then my english is bad 😄 ...it's a big big difference when we talk about grammar problems in the video or grammar problems with the lyrics in the audio.. so.. what is your specific problem?

what do you want to change?

the vocals IN THE TRACK or the vocals IN THE VIDEO?

[waxen prism]

i guess vocals in the video 🫣

[hard beacon]

so the audio you got is good?

[waxen prism]

yes

[hard beacon]

ok.. so you've downloaded a video and in this video the grammar of the vocals is wrong?

[waxen prism]

yes… i haven’t “downloaded” it’s in my playlists

[keen iris]

Hello, suno team!
So I'm getting the same error on any track every time I click on 'more options' in 'Editor', pls help 🥹

[waxen prism]

yes it says knifes instead of knives

[hard beacon]

Ok.. get it... Next question... mobile app or web browser?

[waxen prism]

mobile app

[hard beacon]

pro or premier plan?

[waxen prism]

the free one… i’m still debating a subscription

[hard beacon]

uhm... afaik video download is not included in the free plan... now i am confused

[waxen prism]

don’t worry about it… thanks for trying to help 😇

[hard beacon]

i really want to understand the problem and want to help you.. but.. well.. i am confused 😄

[keen iris]

Do not debating, just get one bruh, I can't convince myself to stop subscription after v5 is out

[waxen prism]

i’ll send a screen shot

[hard beacon]

@keen iris : give us more infos... what browser are you using?

[keen iris]

Microsoft Edge

[keen iris]

Also my pro credits is all spent, and now I have a everyday ones

[waxen prism]

i want it to show knives

[hard beacon]

for me it looks like a javascript error..did you try another browser?? btw: i also use edge.. are you on the newest version?

[keen iris]

I am, will try on another

[waxen prism]

i think so, is there a way to not show the vocal lyrics… like a setting

[hard beacon]

ok MoonRockRose... i am not a pro with the app coz i am using the browser version... there must be an option for "Edit Song Details" or something like that.. on the browser in this detail-window there is another button "Edit Displayed Lyrics"... that's the menu you need to go to change the shown lyrics.. or even delete it

[keen iris]

You can open on pc and click on 'change displayed lirics' in 'add a caption +', after just click save or whatever and save again, it's all

[waxen prism]

ok i’ll try deleting it cuz it says “knives”

[hard beacon]

@keen iris : also check if your pc is running the newest Java Script Software..

[waxen prism]

i’m on the mobile app

can i turn them off from showing

[hard beacon]

go to the track you wanna edit... click the 3 dots on the right side.. go to Edit Song Details... then edit displayed lyrics... change the error or delete it... done

[waxen prism]

edit doesn’t work

i deleted! it’s not showing yay!

thank you!!!! 🙏

[hard beacon]

you are welcome... my billing over 200 US Dollar will be send via email

😛

have a nice day

ok.. next one... @keen iris : does another browser works for you?

[keen iris]

No, in Firefox I get the same error, browsers updated and system is up to date too

[hard beacon]

curious..

i am really sorry, but i can't help you out with that problem..

btw.. you've told me about missing credits... did you every take a look to the mobile app what your credits are up to?

[keen iris]

The same as on pc

[keen iris]

Ok, I hope this thing will fix itself tomorrow, but if no who can I contact to look at my account if there is something wrong with it?

[hard beacon]

hmmm.. never heard about that problem before.. send an email to [e-mail] with all informations you have... and make sure to NOT send emails when you didn't get a response within two weeks... Suno will work from the buttom to the top with emails... if you write an email after one week you will be back on the top 🙂 so please be patient

[polar spade]

Suno sings "knifes" ? Where the heck did it train for English ....reddit? 🤣

[hard beacon]

temu

😄

[calm karma]

i have run this over 20 times. [Intro] [Pulsing Sub Bass] [Epic Grand piano] [Soft strings]. nd its hs never even tried it once. Do I have the format wrong?

[waxen prism]

crazy… it seems to always input something strange in my songs weather sounds or a weird lyric… it sings “roots” as “reets” in one of my songs… driving me nuts…. can you change the pronouncing in the vocals?

[calm karma]

yeah sometimes you have to spell it out like it sounds.

[hard beacon]

we need more infos.... what model are you into???

[calm karma]

sorry using v5

[pastel plaza]

Hi guys im geting alot of hiss in my songs cant get a clean one, can someone tell me what in the prompet can make this hiss ?

Male vocal with a rich, raw tone, soulful and wounded, with gospel-rooted phrasing and whispered intimacy, Emotional arc begins restrained and reflective, growing into a full-throated, burning confession, Genre is dark synth-pop laced with analog warmth and spiritual grit, Mournful piano pulse beneath layered vintage synths, deep reverb, and brooding percussion, Use vocal stacks and subtle choral textures in the final chorus for a swelling, haunted release, Tone is poetic, aching, and reverent exploring memory, fatherhood, and the ache of time, Minor key, 80–90 BPM, Sound design should feel both earthly and sacred like neon flickering through stained glass,

[hard beacon]

ok... what style of music do you want to create?

[calm karma]

sleep token, bring me the horizon, etc

[hard beacon]

hmnmm... gimme more infos... "bring me the horizon" is nu-metal isn't it?

[calm karma]

yeah so nu-metal, melodic metal, etc

[hard beacon]

ok... got it.. what's your exact problem?

[calm karma]

i totally get its not gonna work everytime. but wanna make sure i am building my tags correct

[calm karma]

wanting to know if i am properly formating my tags.

[hard beacon]

well.. nu-metal is complex like EDM... so many ways to get the result you want... but i guess you wanna talk about the right metatags?

[calm karma]

no. not the correct tags. I am wanting to know if i am properly formatting them

[hard beacon]

ok... can you show me refernce track about the result you wanna create?

[calm karma]

let me start over. I am achieving the basic sound i want. In fact its great. But for my intro section i have done the tags like this: [Intro] [Pulsing Sub Bass] [Epic Grand piano] [Soft strings]. What I want to know is... is this the proper way to do the tags for a section? In this case the intro section

[hard beacon]

you've write stuff like [intro] in the lyric section?

[calm karma]

yes.

[calm karma]

i dont think your understanding me

[hard beacon]

i think i do understand you

but overall you doin it right... write things like [Epic Piano] in the lyric section...

but at the end of the day it's still AI... also the sliders (Style Influence, Weirdness) makes a big difference to your output

that's why i was asking about a reference track 😉

[calm karma]

so here are the sections of a song all in the lyric section:
[intro]
[verse]
[pre-chorus]
[chorus]
now lets say i want to give further instuction for the verse section would i do this:
[verse] [Intro] [Pulsing Sub Bass] [Epic Grand piano] [Soft strings]
or
[verse - Pulsing Sub Bass, Epic Grand piano, Soft strings]

[calm karma]

wait do the sliders effect the style instruction in the lyric section? i thought it only effected the style in the style section?

[half dragon]

While it's not always guaranteed - I would use the latter format but also mention them in the style.

[hard beacon]

Well.. i would prefer the [ .... - ...] one... well.. for me the sliders also effect the lyric meta tags

you should try the same prompt with very different slider settings to figure out..

i use the same prompt over and over again for the same style of music... in this case "EDM".... if i use the same slider settings i get the same sound resulut (of course with other melody) over and over again...

if i set the "weirdness" slider to another setting i'll get a complete other result.. so you have to find the best settings for you.. then use it..

[calm karma]

thank you

[hard beacon]

you are welcome mate--

test it... trial and error is the way 😉

Welcome to the world of AI 😄

[polar spade]

Lyrics or more likely the nuances and expression of the lyrics are also considered as style. So something as simple as Suno reading the lyrics as happy , or reverent, or epic, should be influenced by the styles slider. At least that was my impression last year from talk of the new sliders.

I dont know if having similar tags though add more weight to anything matching in the styles prompt box.

[calm karma]

Ok so let me ask this. rather than putting the style for the chorus in the lyric section which is this: [Chorus - Layered Guitars, Powerful Drums, Sub Bass]. Should I be putting that down in the lyric section to reinforce it? And how should the formatting be? like this: Chorus - Layered Guitars, Powerful Drums, Sub Bass.

[hard beacon]

can you send us your complete lyric prompt?

to make it clear: lyric.. not style...

[autumn falcon]

is there a way to make Hard Techno not sound so muddy?

[polar spade]

". . . .Should I be putting that down in the lyric section to reinforce it? . . . "

I think you meant "down in the style prompt section . . .?"

[calm karma]

yersy thats what i ment

[half dragon]

There's no specific format in the style box, you can either list them or describe them ... or list them, then hit that Magic Wan on the style box to turn them into a very detail description.

[calm karma]

thanks

[hard beacon]

ok.. thanks.. and what is your prompt in the style section?

[sullen jay]

So was the "Open song in chat" a test feature or something like for A/B testing in Suno? I loved the feature but now it doesn't appear for me.

[half dragon]

Sounds like something isn't availble at all to the public. Were you part of the test group? you should discuss within that group.

[sullen jay]

I wasn't part of a test group that I was aware of. Just starting using Suno recently and saw that as an option. It was very cool to be able to chat about lyric details and song styles interactively with the AI.

I was going to ask it today if it was able to analyze and display the chord progression of my song. Then the option was gone. Weird

[half dragon]

try others, like chatGPT for now

[sullen jay]

It isn't that important for this, I just wanted to see what it could do. I really enjoyed the interactive process of disussing song vibe and details with the AI before it generated the song. Instead of just using the magic wand or typing in styles

I did see that Suno now has integration with Copilot, maybe this was some testing related to that

[plain pumice]

I notice when I try to extra midi files, and there are audibly two or several instruments of the same, I notice it only gives you one.
Say for example, there is a rhythm and a lead, it often only gives you lead, while in the stem you can hear both.
Anyone know how to get it to exctract both those? If at all possible currently.

[half dragon]

There's no control over that process unfortunately. You can regen the split until it does but that'd be expensive.

[plain pumice]

That's what I thought.
The only other way I know would be to get the instrumental stem, then seperate that into instruments, then use NeuralNote, but even with no instrument bleed, the notes will still be a big mess when converted to midi. And whether it's with Suno or 3rd party, will often use weird tuning - I wonder if there is a way around this - either through prompting or other tools. Because, of course it's possible to transpose the notes, but that's not that easy to do and very time consuming as well.

[quiet solar]

Things you put in the style prompt will generally apply to the whole song, although you can specify things that apply to specific sections of the song (e.g. chorus, verse). Prompts in the lyrics box will tend to apply to that specific line/section.

As far as I'm aware, there's no functional difference between:
[Chorus - Layered Guitars, Powerful Drums, Sub Bass], or
[Chorus] [Layered Guitars] [Powerful Drums] [Sub Bass]

[hushed mist]

"Is anyone having issues with cutting/trimming the tracks in Song Editor?
It suddenly stopped working for me and the split/cut options are no longer appearing.
Did something change in the latest update?"

[neat nexus]

I've been having it crash on me but the legacy editor has been working when that happens.

[polar spade]

I try once maybe, but lately just download into a third party app. Lately between Fadr and Moises. Fadr more often finds some obscure sound or artifact and separates into other or FX or synth or whatever it thinks.

[tired forge]

Can someone please explain what the heck that Weirdness thing means? I'm not really sure on that thing.

[wise kindle]

it prevents suno from playing things safe. It'll try more complicated melodies and rhythms for example. It can give you more intricate vocals too. It's a really fun slider to experiment with, but I wouldn't go over 67ish personally

over 70 and you're in fever dream territory

[tired forge]

AH! I think I have done 75% once. Good to know. Many many thanks

[wise kindle]

hah, I bet that was an interesting one. no prob!

[onyx sparrow]

with v5 i get always cinematic intros

for hiphop or for trap i prefer the v4

4.5

and btw good morningxD

[sour stream]

anyone else have songs less than a min being generated wasting credits ? cause thats been my issue all day 😭 (v5)

ive created 22k songs so i know its not user error

[onyx sparrow]

just write " around 3 minutes length"

but its not a prompt that 100% work on every beat

maybe u need also longer lyrics

[sour stream]

when it works correctly, its given me 3-3.5 min long versions of the same lyrics/style

so its not my lyrics

its creating one valid song and then one song less than a minute every time

yet still charging me the full credit

[onyx sparrow]

yeah test it "around 3 minutes length"

[sour stream]

i would cept im out of credits 😭 gotta listen and earn now, but ill give that a shot. ty

[onyx sparrow]

then ask before u r out of creditsxD

[onyx sparrow]

i have a good question

if i upload a sample do i can use the "persona" in the same time? after i upload the sample that my persona can use this beat?

[onyx sparrow]

i need more space in the style section and in the negativ prompt^^

[rose jasper]

Don't cut yourself too short, you know you're an amazing asset.

[quiet solar]

Yes, you can apply both an audio upload and a persona to the same song in the create screen.

[onyx sparrow]

okay tx

[tired forge]

REALLY hope I didn't repeat a vocal Persona that I already had…

[onyx sparrow]

if i write a 2 different negativ prompt or positiv prompt for the beat with a singer do i have 2 different mixed singer?

[worn thistle]

hello, can i wanna ask, is it possible to get better vocals? i mean, i have a song from Suno and vocals from Suno, but vocals are in bad quality (sometimes) and i wanna ask you guys, is it possible to get stems and make the vocals sound better? (remaster is not doing a good job and cover is more broken)

and, btw, i wanna try the "new" studio in Suno, but first, i wanna know something more about it, soo, can someone please provide me with some functions or how it works (in call pls)

[hard beacon]

You can get a lot Infos about Studio on the official suno Youtube Chanel

[solemn flax]

Hi! My Pro account is missing all Pro features (no Generate, no Remix, no Pro badge), and the billing portal returns 404. This appears to be a backend permissions issue—my Stripe customer link was never applied to my account, so no Pro flags are enabled.

I’ve emailed support with no reply. Could someone please refresh my permissions and re-link my subscription? Thank you!

[shut marten]

I feel like I spend more credits trying to fix suno's ai "glitches" more then making songs in version 5 its a bit frustrating... but I do like the new model because you can tell the difference as far as quality... my current issue is the ai is struggling to say a specific full word (in this case the word is ask) it is cutting off the k and just saying "Ahhh sss" its driving me bananas cuz I am doing things like adding extra k's and nothing is working so far. I may just revert back to 4.5+ (I dont want to but I also do want to keep throwing credits trying to fix this 🙁 ) I've even put in the style prompt: Please pronounce every word clearly and accurately - idk if this will work but I thought I would pop on here to see if anyone else is having that problem and what was a fix that worked for them?

I've spent like 100 credits trying to fix this one issue alone... the crazy thing is... this isnt the original version of the song. the song was done in 4.5+ and I never had this problem.... so idk why its a problem now in v5 🙁

👀

[solemn flax]

Following up — still no Generate/Remix, still no billing portal, still stuck in old UI. This is a backend issue with my subscription link. I would really appreciate help.

Here is a screenshot

[shy jewel]

Guys, Aside form using Extend, is there a prompt or way to make the music lasts for 5 mins?

[wise kindle]

Not unless you add interludes, extra sections, etc. to pad it out. There's no guaranteed way to tell it "I want at least 5 mins"

[shy jewel]

I will try interludes and extra sections.

[quiet solar]

That's not completely new to v5, Suno mispronouncing or dropping letters from the end of a word has happened occasionally in previous versions.

You can style prompt for things like clear pronunciation, but I've never personally noticed it have much of an effect. It generally boils down to how easy it is for Suno to look up what you put in style prompts against the library of music it's been trained on - I'm guessing "clear pronunciation" probably doesn't come up much in musical terms.

My go-to fix for any pronunciation errors is to rewrite the word or phrase phonetically. That's not easy with the word 'ask', especially if it's at the end of the line in the lyrics, but you could try something like 'ahsk' or 'assck'.

[half dragon]

That looks up-to-date UI to me.

[quiet solar]

@solemn flax Appreciate your frustrations, but there's nothing we as Suno users can do for you here. You've raised an issue with support - you need to wait for them to get back to you.

[polar spade]

1. I see you are using a 5.0 version Cover of a song. Thats Pro tools only. And there is no more Remix banner or button since about two months ago. All are Create. So im not following the issue from your screenshot. Sorry, what am I not seeing, or should I see?

2. clicking on your credits, you do not get a subscription page? But a 404 error?

3. you have 57 credits, that more than the maximum 50 credits for Basic, unless you have 7 extra credits for some reason.

4. where would we see a Pro badge on that screenshot? The subscription page would help show subscription status. So if you're getting 404 when clicking your credits, that could be just a connection issue, sure.

Click credits for subscription page. Click "Update Payment" to see Billing page and history, in case you may not have tried that yet.

[quiet solar]

Actually, that's a really good point, Dasvon. That's definitely the Pro UI on display.

[polar spade]

@solemn flax No more Remix banner or button.

[balmy slate]

Hi. I speak English. However, I don't think anyone can help me outside of support. I had Suno as premium for a month. Then I canceled the subscription so I could renew it (it didn't work!). However, €34 was debited from Google Play and used by Suno, I see the credits BUT I can't create any songs that can then be used "commercially" if necessary! Support doesn't respond at all!

😔😔

[fluid copper]

How can I get just a 4 bar intro, instead of long, long ones I've been getting. I tried a few things including [SHORT INTRO]

[tender eagle]

Yeah unfortunately that is not something I have any suggestions on. Thats definitely back end issue.

[tender eagle]

How are you writing the intro. And what is it doing? Is it leaking into the first verse as well?

[fluid copper]

It's like: [SHORT INTRO] [VERSE 1]blah, blah, blah [CHORUS] etc....

[tender eagle]

Ai in general does not have much respect for negation. Telling it to make a short intro is just going to see "intro" more than it sees short.

But AI does tend to love additive prompts

So don't tell it to end the intro. Tell it to begin the music for the verse. Try that.

[balmy slate]

Yes, unfortunately. Everything worked for a month and now something like this. The funny thing is, I have 2 accounts. On my free account it's not even possible to take out a subscription on the app. They say it's not possible. I already have a subscription. Such nonsense. I'm currently rather annoyed that I apparently didn't spend the money on anything! 🥺

[tender eagle]

Akin to [drop]

[fluid copper]

Later today when I go back on, I'll try that

[tender eagle]

Yeah I have no idea about accounts at all. Your best bet is probably to go to https://discord.com/channels/1069381916492562582/1372994553233276970 probably more likely to find useful help there for that kind of thing.

[onyx sparrow]

whats the best way for hiphop beats gemini oder grok or chat gpt?

on gemini i get always stutter idk why

sound like space beats or like science beats idk how to destribe this

[half dragon]

are you saying you were using gemini/grok/.. to generate suno songs? or I think you meant lyrics or style generated by those? maybe its too early for me ..lol

[polar spade]

I assume you mean using Grok or Gemini to create prompting. And that Chatgpt prompts give a stuttering effect to the music or lyrics when pasted into the lyrics or styles?

That seems like a bit of trial and error, honestly for anyone to get what works best for their genre and styles they want, including lyrics used.

[hoary plover]

A question.. I have a friend who is lead singer for a popular indie band around the DMV. They have some albums, etc.. How would I go about taking a vocal track from her and and using Suno to add music to it so that the music is in synch with the vocals, rather than having Suno add vocals from lyrics?

[half dragon]

1. Does she give you permission?
2. if you have Pro - maybe on her vocal track, try Add Instrumental feature.

[onyx sparrow]

yeah i used yesterday and today gemini and i get a very strange space sound

but my prompt was street rap

and hip hop trap

idk

sounds like trance

xD

psy trance or techno

xD

with the stutter ofc xD

xD

idk what gemini model it is but its 200% pure trash

someone know what this means?

[half dragon]

Looks like a user built prompt generator

[gloomy fiber]

https://open.spotify.com/track/1cerNkFSf5Xwgf3z92aRZM?si=4Gdavg2lSuGQlv0XOM8r_w

[onyx sparrow]

omg sry i m so dumb

this is idk what it was like a preset or a generator

xD

damn

[polar spade]

If the vocals are pitched and on beat, however that is interpreted by Suno, then you can click edit/add instrumental to the uploaded original audio. See @half dragon post above. Granted its up to your prompting to get it a style you light like and the metering amd tempo will be loosely based on the metering of the vocal.

But try it.

Attached is 21 second of my singing do do do. And no its not someone prodding a goat. I like how Suno described it as a vocal exercise or warm up 😅

And then the next one adding a Suno instrumental to it.

Theres no guarantee that it won't add background vocals, harmonies, or augment the vocal enough that it just doesnt sound like the original. Mine , since its pitched corrected by Suno, are about 98% NOT sounding in character as the original. But the ones that work are fun.

https://suno.com/s/ABWJKN25AqE4ucAT

https://suno.com/s/0vHd0KCzyEra1Ihb

[brazen trail]

With Remix Cover, the generated songs have a lot of metallic, robotic resonance. How can I remove it? I tried using DAW tools to fix it, but it’s still not completely gone. I also tried the Suno Remaster option, but it made things worse.

[cold zenith]

I would like to know how someone has copyrighted this sentence : " Let there be Light " ???

[wispy niche]

Anyone have a solution for promoting studio that can make it generate sounds to free time? Every time I try to add strings it’s off time to a free temp tracks

Tempo

[neat nexus]

Take it up with Asimov maybe? (The last question)

[wispy niche]

Is he another user?

[polar spade]

GOD has full eternal copyrights. 🤣

Phrases can get caught in the filter. You can write Support to see if they can whitelist it.

Further, sometimes it can be that the tokenization or splitting up of the phrase matches perfectly to something on file. Add a space , or put 2 spaces between words, drop half the phrase down, rearrange in a way that still works but passes the filter on such an innocuous phrase.

[restive pebble]

Well almighty then 🤣

[neat nexus]

Science Fiction writer: https://users.ece.cmu.edu/~gamvrosi/thelastq.html

[quiet solar]

And it takes the patience of a saint to get direct permission from him to cover one of his tracks... 😂

[shut gyro]

Anybody got any got good tips to make a song sound like it was created in the 80s. I can get the instrumental but the vocals seem to polished to be from the 80s

[wet maple]

Has anyone found a good balance of sliders using inspo that achieves a pretty consistent sound or close to what your end goal is?

[polar spade]

Check out the 80s channel. Look at the prompts used there or see if any have made any public personas to borrow. https://discord.com/channels/1069381916492562582/1262592742945456148

[wet maple]

Have you tried using meta tags like [Baritone lead over '80s groove and organ], [Analog Warmth, tape saturation, parallel compression, wide stereo]], [Overdriven Marshall-style guitars provide a rich, mid-heavy tone]

[polar spade]

Add " blue-eyed soul" and you've got Michael McDonald.

[polar spade]

I still question what the sliders are affecting or pulling out of the maximum four outputs. Since inspo often has a greater sense of the first output, and just niche riffs and vibes from the others, which one or all do the slider affect, or does it do the inspired output all at once in the background then act on the output and prompts before the final output is posted.

So. . .I dont know.

[wet maple]

So you're saying using the default setting is the most optimized way of generating the end goal?

[polar spade]

Not really saying anything but "I dont know what the sliders do" in that case without running 300 various patterns and compare. So, sure, default sounds good until we know. 😁

Dang-it. Now adding inspo - slider tests to my list.

[wet maple]

As someone who works a lot with inspo, I think i've tried just about every variation of sliders (close to +1000 gens), and I have yet to find a pattern that works universally which is what prompted my question in the hope that maybe I was missing something. If your tests do produce something close to a pattern please let us know, that would be immensely helpful!

[tired forge]

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

This sounds good, right?

[restive pebble]

He keeps forgetting ....

[onyx sparrow]

its good but not god tier

but for a non musican its good

there is much background stuff inside and other things

[tired forge]

I don't know how the hell you get God tier on Suno since I've never heard God tier from anyone I've heard share a song. Though this might not have been the correct one to share. I forgot if that's what I wanted to share or not since I thought about sharing a song last night and I had to go to sleep.

[onyx sparrow]

okay then i lie: its perfect! good job man.

i hope u feel better now

^^

[sleek stump]

This really isnt the right channel to just share a track and ask if its good 😛

[polar spade]

Well, then you would appear to have an experienced opinion of Sliders vs. Default settings - over anything Ive done.

Betwix the two, do the sliders seem to give the occassional, optimal or interesting effect (at least 1:15 outputs) or they have all sounded
kind of similar no matter the slider setting..... and does just "default" seem to be consistently surprising every few outputs, or does doing nothing just echo the first output in the inspo list the most without much effect?

[frozen dagger]

omg omg omg. i just saw the animate option... curb my excitement, maybe it's subpar. but that is the right direction there! 100 credits... i better find something better than this static cloud image to experiment the animate feature. So stoked.

[fluid wren]

anyone know if a generated song can be re-run to just clean it up on the back end. Song Instrumental and vocals came out good - but the mix is grainy and trash

[wise kindle]

yeah you can sometimes get a better mix by covering it, but it may be slightly different from the original. AI is random by nature, so you're unlikely to get the exact same composition but with better quality. I sometimes throw in words like "balanced production" or "normalized production". I honestly don't know if they have much bearing on the output or if I just got lucky on that particular spin of the wheel.

[fluid wren]

most of the generations that have come out like that havent been worried about saving or modifying - Im not playing with studio or anything yet - so might be something that can be done there with the stems:

are the effects baked in or on their own tracks?

[wise kindle]

everything is created at the same time afaik (so baked in)

[frozen dagger]

u can try the remaster function. usually low or normal options. but same kind of lottery without manual editing or other tools.

[fluid wren]

where is that?

[fluid wren]

just for shegrins - what does it mean if that's grey'd out - make persona is clickable - but im still on the basic new guy starter set here

[wise kindle]

means it's not a valid action. Like if you look at "Make Persona" on an upload it will be greyed out because that's not something anyone can do

[frozen dagger]

EQ in studio, now too? SSuno like The Jeffersons. Movin' on up. Time too update that "What's New" bit with a couple of things. What else hiding in here?

[fluid wren]

well having it run as a cover - style promptd to keep it the same and normalize levels - and of the 2 runs 1 is pretty damn close but the mix is definetly better

[normal mulch]

hello, i have a question on plan and billing if someone can help ? im on pro plan, paid annually, but i would like to try studio for a month, how to proceed ?

[fluid wren]

likely need to scroll down to the #1294421750397796435 channel

[normal mulch]

can t write in there

[wise kindle]

no, it's purely an informational channel. If you want to write to billing you can email them: [e-mail]

/billing

[flint haven]

SUNO BILLING HELP

For billing/payment issues, email: [e-mail]

Include:

• Your Suno @handle
• Description of the issue
• Transaction ID(s) or payment receipt(s)

Before mailing:

• Check if your card was charged
• Verify funds were available
• If Stripe rejected your card, try using the iOS or Android app to pay via Apple/Google Pay.

Do not send non-financial issues to this address — it may delay your response.

[normal mulch]

already wrote email, got answer after 2 weeks, but not answering my question

asked again, 8 days without return

[wise kindle]

I noticed you wrote in #archived-ticket-help with "hi".. keep in mind there's a 2 hour slowmode, but you can edit your comment to provide more details than a greeting

[frozen dagger]

I dunno how they would scale that or anything, but just personally, if I was already paid up annually and I wanted to try Studio for a month cus I was unsure I would just create an alt account for the $30. IF I was unsure about swapping my plan for whatever reason. I dunno what the math is either lol. Just saying. /shrug.
It's a good question. I don't see a clear answer other than the FAQ that you can upgrade or downgrade your account, but I know I've lost some credits myself not paying attention when changing my own plan a while back.

[glad ibex]

what do i do if there's a clip or distortion in the vocals and i want the same exact thing just without the clipping, how do i do that?

i want the same melody, same singer, same words etc

[merry sage]

I often use the feature to save styles/prompt... but I have no idea where they're stored?

[polar spade]

I dont think there is any protocol for a one month trial while in an annual subscription. Likely not something thought that far ahead before Studio was a thing.

You could write to support to see if they would consider anything like that, or even at minimum, throw that in the suggestion-web channel.

Back in September they did that very things and gave out a few limited trials of Studio to those Pro users answering an invite. So it may be at least possible. Seems users favored the chance and many upgraded thereafter.

But how that would be turned on and off in an annual sub after a month is totally a thing for Support to talk to you directly....emmm... for you to talk to Support about.

The alternative , of course, is to buy a premium subscription for one month. And 1 week later , click cancel to set it to stop in one month. Burn through the credits as much possible to really test out Studio.

[half dragon]

Click where the arrow is to save - and to recall it by that icon on the bottom.

[polar spade]

Im lazy at the moment to check.

Has it been working finally as before?

[half dragon]

Yea they fixed it

[merry sage]

Nice, thanks! I knew how to save, but couldn't find that icon for recovering them!

[finite folio]

Has in the Suno the number of characters in the Prompt (style) been changed from 1000 to 100?

[half dragon]

Nope.

[finite folio]

For quite a few songs, the style box showed 100/100 in the "80s New Wave, Industrial, Dark Ambient, Metal, blues, Arctic Ambient, Techno, Psychedelic" and it was red.

[half dragon]

I'm not sure what you meant - this is in Create page style to create new song.
Did you mean when you Cover another song? and it only populated 100 characters max in the style box?

[finite folio]

There was at one monent 100/100 not 1000/1000 for me. Now seems again all normal.

Wait ok again one song like I sayd

83/100 Why I see that ?

[half dragon]

Ah Edit Song Details. That's not the same.

[finite folio]

Ok but why there have only 100 possibile?

[half dragon]

From what I'm aware, I don't think you can modify the original style descriptions after the song was created.
Maybe that box for display purpose and it has only a certain size.

[finite folio]

You mean like for a short story summary?

[half dragon]

Yeah short style summary.

[red oar]

As other said, the filter Suno made is overly sensitive and intrusive. Clearly they have chosen to be defensive, perhaps as a quick fix from take-down threats not made public. I've seen complaints here on perfectly fine single words. But in your case you have a phrase, which very likely can be found as part of copyrighted lyrics, so it's the problem "what parts of copyright lyrics are ok to use". I have no idea how they tried to balance that. (Incidentally, I've never seen anyone saying "This got blocked, and that was well done".)

The given phrase is used already for Song names, album titles, movie titles, which in one way would indicate that it should be protected (think "Lord of the rings" or Spiderman might be blocked to avoid lawsuits). You and me know that that "Let there be light" is "public domain" but it requires an additional layer of knowledge than "take the list of all published songs" (or whatever). By the way, I asked GPT: It said yes to "light" but "risky" for "rings". So that might be an idea for them to process their filter. But the user will be clever enough to break it down to get pass even "spiderman", so the filter might be void anyway.

I'm guessing the problem isn't actual lyrics as much as the influences. Say you have "Oasis" in the lyrics and the model picks it up to make music similar to the well known band - regardless if that was intended by the user or by the model - then it doesn't matter that the word itself isn't copyrightable. The lawsuit will point at the "multiple violations" and "no effort to prevent misuse".

[normal mulch]

as i already paid one year on Pro, i would like to know what happens if i pay on month Premium... then stop it after , if i o back to my Pro plan for the remaining months 🙂

[dire swan]

Im pulling my hair out. Can anyone give me meta prompts that actually work in v5. Im also having a hellova time with one song. For some reason there is one spot that suno seems to be having a stroke. It garbles these two semtences. If i move them somehwere else, they are fine. Ive tried various phrases and vocab.

[dire swan]

If you go to upgrade, theres a spot to downgrade. With options to prorate or wait.

[normal mulch]

ok great then i can give it a try and go back to my plan

[dire swan]

I beleive so. Yes.

But

BUT. Id email them im billing. It may change you upgrade for full year. Then refund for full year when you downgrade.

[red oar]

100 credits is maybe not the worst if you're on subscription, but the time and frustration is adding to the injury. Yet, the AI game is one of trial and error.

Here are some general things from my experience...
Do you base it on audio input? That influence can be hard to change with prompt.
The voice and pronounciation is very influenced by the music style, in my experience, Reggae or Rap, etc. Prompting around it is hard.
"Please pronounce every word clearly" - any prompt change may alter the input enough to change output, but I haven't seen evidence that it's a reliable prompt.
There is some kind of "phrase context" so if you have "ask" on a rhyming position with "buss" or in a phrase with another word where dropping the end is natural, that style may spill over.
Of course, if there's a lot of text, starting the next phrase might get win over ending clearly.

I'm usually not writing English lyrics, but here are some additional nonobvious ideas you can use to "break the trend". Write word in "quotes" or emphasis. To add stress you can use case or multiples asK, askkk or word breaks ask-k, as-k, as:k, "as k". You might have different effect if you add or remove punctuation ".,!;" after the word. And as others said, write "phonetically" e.g. "ahskh" or how you envision it to sound. Sometimes it doesn't matter how you write, because the main thing is to get the AI to "rethink"; "askq, askl, asx, asg" could give you the sound you want.

To save credits, you can add in several different ways to change the word in the lyrics and then pick the one that worked best.

Since every output is random, there is no guarantee that the prompt that worked in one version will not fail the next (or opposite), but systematically changing the lyrics is usually working. I have access to Suno Studio so I often hack it there.

[red oar]

hard to say something without your prompt, but

meta prompts that actually work in v5.
what did you want to acchieve?
It garbles these two semtences.
The generic hint is to rewrite the lyrics phonetically. But in your case it might be the "context" that's off. Did you have any [effects] indicated near this? Everything is interconnected, so the most unexpected change can influence everything.
If i move them somehwere else, they are fine.
maybe you can use that to you advantage? what if you insert an interlude before that verse? Or add the same verse several time, then use the Editor to cut out the fails? Change order in lyrics and reorder in Editor? Not a prompt-talk for sure, but a useful tool to have.

[dire swan]

Its part of the chorus. And. I can put this exact same chorus elsewhere and it works fine

As for meta prompts, id seem online where previous versions could do alternating voices (like men and women) on lines depending on the female or male prompt.

[red oar]

I can put this exact same chorus elsewhere and it works fine
did you try editor?
As for meta prompts, id seem online where previous versions could do alternating voices (like men and women) on lines depending on the female or male prompt.
That was not reliable, I think, and still works if you're lucky.

[dire swan]

Not sure yet how to use editor. While ive soent 3000 credits on this song, im still new.

[hard beacon]

i like everything you are new into... find the best way for you and take a look to existing tracks you like

[heady sparrow]

Tell me what this error is? : Error generating clip
Clip generation timed out, please try again

[hard beacon]

to what message is this a response? 😉

[heady sparrow]

This is not an answer. And the error that occurs is in my suno.

[hard beacon]

ahhh... ok.. well.. timed out.. try again later.. that's the solution 😉

and make sure, if you use the app, that it's up2date

[heady sparrow]

I've been having this error for 2 hours now, how long should I wait? ))

[hard beacon]

hmm.. app or web?

[heady sparrow]

app

[hard beacon]

what country?

[heady sparrow]

Russia but VPN via America

[hard beacon]

well.. maybe the VPN has some connection problems at the moment.. that would be the logical explanation

[heady sparrow]

The VPN works reliably. There are no problems.

[hard beacon]

maybe the VPN works from you to america.. but maybe from the final destination server the connection to suno is not working well... that's the downside of having to use VPNs

Suno Services are all online.. No problems at other locations

[heady sparrow]

without VPN the same error

[hard beacon]

But that could well be down to Russia 🙂 All connection tests are positive. (BTW: that's not meant negatively)

is the app on the lastet version??

[formal wagon]

Hey guys. I can’t comment or reply. It’s been like it for hours now. I’m beyond frustrated. Tech support won’t reply. Any ideas how to resolve

[heady sparrow]

Yes, the app is the latest version. The browser has the same error.

[hard beacon]

then it must be an error with your connection...

[heady sparrow]

I discovered that Suno works fine on my other PC, even without a VPN. Do you know why?

Why does it not work on one computer but works on another?

[hard beacon]

what are you talking about with comment or reply???? Tech Support is in weekend mode.. we have sunday 😉

[formal wagon]

On the app man. I can’t comment on anyone’s music or reply to commenting on myn

[heady sparrow]

No, I'm saying that I launched Suno on my second computer and it works there, I don't know why it doesn't work on the first computer.

[hard beacon]

@formal wagon : easy going...

[formal wagon]

Huh ? Easy going ?? I don’t understand man

[hard beacon]

Give us more informations... App ?? Web ?? Location ?? Plan ?

@heady sparrow Send you a DM

[sharp kettle]

holy prompts in 4.5 all i put in the entire song composer song style and it didnt ignore most of the prompt no i just gott figure out with parts it did ignore

oh god i think its time for a type i cant type lol

[hard beacon]

😄

[sharp kettle]

if i could figure out what it ignored i could get the fire music to be purfect id rather not have to cut and recompose

in audacity because i'm esc lifer poor to get suno studio lol

yes i'm using my own voice and lyrics just using suno as back round music... gotta tak extra steps due to life long poor ness

[hard beacon]

Audacity has more options in post-production then studio 😉 Try to get the best result with cover function...

[sharp kettle]

can some re type this so it works

Steady use d major G major A major and B minor drumbeat tempo and rhythm,
add exotic dance tempo bass guitar to back the drums up,
use d major chord G major chord A major and B major chords mix
use only acoustic lead guitar tempo to back the drums up,
add back up rhythm music to Chorus1 Chorus2 Bridge verse1 verse 2,
there should be no break in the music for 3 minutes and 15 seconds,
exact song length is 3 minutes and 15 seconds,
the song should contain even mix of dance rock exotic dance and rock and roll from start to end,
only Acoustic Guitar, only drums, only bass guitar, no high e notes,
the instrumental intro is 15 seconds, the instrumental interludes are 30 seconds, the instrumental outro is 15 seconds,in the baritone style of 1950's rock & roll baritone pioneer,

Fun fact the song beat is inspired by hunk of burning love

[hard beacon]

ooff.. that's not the way i work with style prompts 😉

[sharp kettle]

i went hard core on this prompt because it has a 1950s, 1960s 1970s 1980s 2025 beats beversion all the same lyrics

the 1990s is r and b not rock and rollk

[hard beacon]

maybe you should use chatgpt to create a prompt with all these infos

there is a good GPT out there optimized for v5

[sharp kettle]

it wou be nice if suno ai 4.5 reconnized this musical note since its standard music 101 instruction

[15 second Short Instrumental Outro: 140 bpm tempo: Drumbeat: Acoustic Guitar: Bass Guitar]

then i could just cut the times out of the style

[hard beacon]

there are 1000+ ways to get the result you want... you have to figure out the best for yourself

[sharp kettle]

i still love the fact that 4.5 all recognized the full prompt... because that is composer level and people can't say its deritive since its orginal

[hard beacon]

yapp.. 4.5 is easier with prompting.. i usually do my creations in 4.5+ and cover them in v5 🙂

but that also depends on the style you want to create

😛

woodhouse strikes back again..

[sharp kettle]

oh come on i cant isnult the crap man?

[hard beacon]

looks like 😉

[sharp kettle]

fine robin honey darlin thanks to butmans actions i had to use som of my suno save funds. so i be too poort right now. but i did listen to some of the 5.0 music its kinda fire

[hard beacon]

what style are you creatin?

[sharp kettle]

my main singing voice is mainely rock and operatic, but i stick to many of the rock, r and b and oprattic

[hard beacon]

ok.. so the "handmade music style" 🙂 i've figured out that Suno has some issues with the highs on "rock music" in the drums... but keep going on.. i think you will find your perfect prompt 😉

[sharp kettle]

but i do have 3 songs
1 in sea shanty

1. 1950s Ragtime, Baritone, Lounge Singer, 130 BPM, Male Vocals, Dance Jazz
2. balkan,klezmer,gypsy jazz
   and 4 in edm, dance rythm, dance rock

[hard beacon]

if you got one you love... save it.. and use it over and over again with different slider settings

..btw: talking about the prompt 😉

[sharp kettle]

most of my songs have a rock, r and b, musical, duet, and or dance mix. i also make a instrumental version in audacity so karaokeres and streamers/ dancer can use my music. although i prefeer that streamer play my music on their streams lol.. oh i have good prompt cues i use over and over. but i like when i cant set the beath or notes in my prompts
so i can get the sound i like

[hard beacon]

at the end of the day it's still AI.. so you need some luck to create the one you like.. but i like your mindest 😉

[sharp kettle]

i find blending the styles works really well. also the cover feature in 4.5 all work reall well to get the music right.
but its very hard for me to sing just a 1 minute version. most of my songs are around 3 miniutes. i do have few that are faster songs but they are mainly love songs. i also have 10- 30 second singover bits which covers dont work with lol

my favorit sing over bit is ooooh my goood nooo he didnt. but trying to get a good sound bit prompt for that is 100% too hard for my autism sapian brain

[hard beacon]

others complain about "why my song is only 1 minute long" and you want to have it shortend... that's my sense of humor 😄

[sharp kettle]

actuall i would like it longer because i actually know how to do a voice over song routine to have your back up intrumental feature 100 work

its call rythmic vocal singing and its one area of music i happe to be really good at

[nimble matrix]

Is there any BF discount available?
Been searching and upto nothing

[sharp kettle]

although the song i did for suno i did the in melodic singover twoo bad the drive it was on crased and correpted i have to re sing it

yes i wrote a song about suno and its awesomness

[hard beacon]

negative 🙂

coz BF is on the 28th ;-)))

[nimble matrix]

Gosh!!! Been overwhelmed with all the online sale going on totally forgot the actual date…

[hard beacon]

hahaha.. 😉

i know that problem..

[sharp kettle]

companies really want that money.. looks at 4000$ compter price. damn you autism and gaming industry..
inkscape now need a really powerful pc to run since windows keep breaking it.. audictity will some time crash with its new features. so again a powerful pc is neeed. but companies want that monies so bad the start black friday deal super early..
only problem is my wallet has on't 52 suno bucks in it lol

[hard beacon]

52 suno bucks = 52 or 520 credits? 😄

[sharp kettle]

nope that my savings for luxery spending after 11 months of saving

[hard beacon]

ahhh.

😄

[sharp kettle]

most people dont belive me when i tell them es lifers in north americ only make 1080 mx if the are lucky

so it takes them like 13 to 20 years to buy any kind of tech.. that why i orginal sought suno out for making my musical beats

and oh boy were they fire

the 3 on tic toc have over 1000 hits total lol and they are only demos previews

[hard beacon]

I'll be honest... I can't follow the context anymore 🙂

[sharp kettle]

lol i just wrote a song verse out i think lol...

[hard beacon]

ahh

ok

😉

[sharp kettle]

also its like 2 am if you have to hit the bed that fine two..
btw tic toc is a social media platform where music blows up befor artist youtube it or spotific it.

also ESC lifer stands for econmic support check. wheich is the proper term for goverment assistance emplooye and way les degrattory and 1080$ is the pay amount lol

[hard beacon]

well.. it 11:37am here 🙂

lunch time 🙂

[sharp kettle]

zzzzz... writes types in play a lullaby tune about being nocturnal...

puts in own song lyrics while sleeping..

bed time oh sweet bed time

while others are at lunch with the sun shining i'm alseep

[hard beacon]

Huuuuuuh...... did everyone just test the "animate cover" function for 100 credits??? is it worth of it?

[sharp kettle]

their an animate cover function?

[sharp kettle]

i total just saw that.... that crazy. but if i were to make cover art i wouldnt animate but i would love to generate my cover art since most of my songs have a theme... but thats more file space then i have right now

[hard beacon]

Dasvon is writing............ 😄

[wise kindle]

lol he stopped

you jinxed it

[hard beacon]

hahahahahaha

😄

now i feel bad

[sharp kettle]

omg i think i figured out the prompt.!!!!!!!!!!!!!!

Steady use d major G, major A major and B minor drumbeat tempo and rhythm,
Add exotic dance tempo, bass guitar to back the drums up,
Use the D major chord, G major chord, A major and B major chords mix
use only acoustic lead guitar tempo to back the drums up,
add back up rhythm music to Chorus1 Chorus2 Bridge verse1 verse 2,
The song should contain an even mix of dance, rock, exotic dance and rock and roll from start to end,
only Acoustic Guitar, only drums, only bass guitar, no high e notes,
baritone style of 1950's rock & roll baritone pioneer,
there should be no break in the music for 3 minutes and 15 seconds,
the exact song length is 3 minutes and 15 seconds,

[hard beacon]

if it works for you = PERFECT

[sharp kettle]

[15-second Instrumental Intro: Drumbeat 140 bpm tempo: Acoustic Guitar: Bass Guitar ]

doesnt reconize this when it generates and its still sing me-ow! wrong and LIVEWIRe wrong but thats okay because im gonn get rid oof the vocals anyways lol

[wise kindle]

if you've got the patience for it, Suno is decent at reading chord tags in the lyrics

An [D/F#]alarm [G]dies [D/F#]right before [Em]work
So I [D/F#]sprint but [G]still [D/F#]end up so [Em]late

Allows you to better customize the movement of the song too. If you don't need specific arrangements it's probably overkill though

[sharp kettle]

here re right thes for me

[safe oar]

Idk, what kind of things should I add in my prompt to sound studio quality, or add professionalism to it, I'm new here so need some help

[sharp kettle]

[Short 15 second Instrumental Intro: Drumbeat 140 bpm tempo: Acoustic Guitar: Bass Guitar ]

[30second Instrumental Interlude: Dance Rock 140 bpm Tempo: Drumbeat and Bass Guitar]

[30 second Instrumental Interlude: Dance Rock 140 bpm Tempo: Drumbeat: Acoustic Guitar]

[15 second Short Instrumental Outro: 140 bpm tempo: Drumbeat: Acoustic Guitar: Bass Guitar]

so it plays the rythem the prompt has been hiting the song beats any most of the lyrics just not the song length

[wise kindle]

There's some in #1252267444773326848. You could see if they work for you: #1417162963709857913 message

[polar spade]

I was saying. I am not sure one can bump DOWN to a one month Premium while paying $8 a month for a Pro annual. If that were true, everyone would do that too.

Here my post again. I tried to write it before that you could read between the lines. But im not saying it in DMs.
#📚┃suno-school message

Email them. Find out. No special form or formula here.

As I understand if you just try to click and DOWNGRADE to Premium monthly then as ALL teir changes go, that means canceling your present subscription. , it could and would forfeit your annual. Thats the point of annual subscriptions, to commit you to a year by giving ou a discount.

What's stopping you from opening a premium account for a month? Cancel after a week so that it drops to Basic in 30 days. Or continue and use two accounts at $38 a month until you dont wish to anymore. Its not so hard, and takes having /creating another email if its not uncomfortable to do so.

I doubt any peer user has experienced that to give you experienced advice of dropping their annual sub just to use Studio for a month.

No need for me to response further. Someone else may. Just throwing out options. Good luck.

[hard beacon]

holy.. a response to a message that was sended 4 hours ago 😄

[sharp kettle]

the prompt for studio audio in any ai is usally insert music genre here, "Studio audio",, insert tonal type here( baritone), inserying sing type. then with suno you just set the voice to male or female in the new 4.5 all settings

[safe oar]

Thank you🤗

[normal mulch]

i want to use it on already generated track, so can't open a new account to "try" it. and my emails are not answered 😄

[polar spade]

Wake up. Respond to posts. Go to work. 🤣

[wise kindle]

can you download the wav and upload it to the premier account?

<@&1082792303976001547>

[hard beacon]

😄

[sharp kettle]

the fact one of my songs keeps generating cats for the cover is hilerious the song is not about a cat but about a anthro humn cat hybrid lol

thats spelled with a k its kat

[wise kindle]

at least it got it half right. heh

[dire swan]

Are there any prompts that seem somewhat consistent (i know… im optimistic) that can tell it when to use a male voice and a female voice? Or two different male voices?

[sharp kettle]

it finally got the whole dang song right for the 1950 version lol

[polar spade]

OK. I see your worry. So more response, if it helps.

1. The response time is about or approaching 4 weeks at the moment. About 30 days. Hopefully shorter. If your email was less than 4 weeks ago, then yes, no one should expect a response in that lesser time.

2. a track is a completed output. Nothing stops me from downloading my output from Pro and uploading it to Premium and continue from there. Its just a file.

3. You have now, three options. I feel you're looking for the option to "try" [free trial] it, thats not there. Those you have to ask for, usually privately, and thats through email, which you've done. It never hurts to ask. See #1.

I hope you're fortunate and get what you're wanting. 🙏

[sharp kettle]

on thing that is wrong is the song time but i can live with that for now lol

[wise kindle]

duets/multiple voices are never a sure thing - there's no way to accurately to do that. You can add square brackets where you want the voice to start in the lyrics and use names, ie: [Tom] or {Sally], but it's not guaranteed to work, like at all

[sharp kettle]

in order to do a duet in any song in ai, and it might not happen gonna happen you have to type, song genre, duet, male vocals, femal vocals, male tone type here, female tone type here, duet styling goes here, also if suno installed it [chorus duet] cjange to verse one etc.. in the lyryics but both prompts may or may not work

[dire swan]

Yeah. Ill be honest, i wish there was a way to chat woth the AI after a somg to discuss fixes and then. Retry

Think if suno and chatgpt had a baby. Haha.

[sharp kettle]

so a song would read
[verse 1 female vocals]
blah blah i met this gurl

[verse 2 : female vocals duet}
blah blah i love her , saw her from across the room

or

[verse 1 male vocals]
blah blah i met this gurl

[verse 2 : female vocals]
blah blah i love him , saw him from across the room
style

[chorus duet]
blah blah our love is eternal

style prompt
song genre goes here, duet, male vocals, female vocals, male tone type here, female tone type here, duet styling goes here

[polar spade]

@dire swan

To answer your question from ticket-help:

Suno debits 10 credits usually at the end of the duration of creating an output, so regardless of 1 or 2 generated on an original creation, it will debit 10 credits, but if only one output is finally produced, suno credits back 5 credits within a minute or two.

[wise kindle]

honestly, I don't think the syntax is super important. In my duet (which actually worked pretty well) I used [Verse 1 / Male Singer] [Verse 2 / Female Singer]

[dire swan]

Ahhh. Thanks for all the help

[wise kindle]

I also added "duet" in the style prompt. Not sure if that helped, but it can't let hurt to let suno know you're planning a duet I suppose

[sharp kettle]

nice use of music 102 brackets i didnt know those worked in suno

[narrow minnow]

Mugo Gusti paring dalan,
tresna iki ora mung sak kedhipan.
Tak simpen sliramu ning atiku,
kaya kembang sing ora gelem layu.

[Final Drop + Outro – pesta kecil]
Kembang tresna… mekar maneh,
(bass nggebug, kendang koplo mlebu)
Ayoo~ ayoo~ bareng aku nyanyi,
senenge ati yen kowe nang kene.
Hee… hee…
(fade out, adlib sinden: “seneng… seneng…”)

[restive pebble]

I have always had luck using the granular duet call and response approach.

[Verse]
[Vocals: Male, Call]
lyrics lyrics

[Verse]
[Vocals: Female, Response]
lyrics lyrics

For the styles I use something like:

This track is a Call and Response Duet. Then go on to describe the vocal qualities of each if you need ...
I'd say about 70% of the time I get something very usable.
YMMV

[wise kindle]

Oooo this is clever

[restive pebble]

That may be a much simpler way to get the desired effect! Woot... great post... 🙃

[restive pebble]

what I was thinking about your post

[polar spade]

Keep in mind, if your lyrics are "written" for a duet and with proper spacing, as well as a nudge in the style prompts , along with proper line spacing and quotation marks, its possible to simply get a duet without any tag prompting.

So in many instances, suno creates a duet because it wills it.

It adjust frequencies, just as it learned from training, if the lyrics and style calls for it. I still doubt Suno knows what exactly a male or female is, only a range of frequencies and characteristics which work best when prompting for a "female", "male" , or those odd emmmm predictable k-pop vocals that are somewhere in between.

Experiment: write and prompt some k-pop songs. You'll get lots of duets and multi-vocal levels on more than a few outputs every time.

[sharp kettle]

3rd clever ring the bel ding ding

[restive pebble]

Morning @polar spade

[restive pebble]

.hope you fond a prompt that works for your duet

[polar spade]

G'day, mate. ☕ ☕

[restive pebble]

Aye mate. Tis a sparkling morning

[polar spade]

Have a cuppa. Enjoy the day!

[sharp kettle]

nah no duest for me for now i dont have a life partner to sing her humannoid part.. i use my own vocals in my music. sometime i have to recut the ai songs in audicity all though mye Kat song is hit the 1950s verson i just have to het the 19 60s 1972, 1980s and 2025 version lol

[restive pebble]

on my second. Finishing up the turkey time challenge entry

[sharp kettle]

the fact there is a black dancing cat on the 1950s cover is hilerious

[restive pebble]

@polar spade what are you up to on this fine seventh day of our week?

[restive pebble]

You have the gift of vocals.. I have no choice but to let Suno make them or people would lynch me

[polar spade]

Some reason, I keep thinking the deadline on those is Sat at 9am. So I was late on my Friday night instrumental entry. 😅

[sharp kettle]

live vocals not mic vocals i stiil brodcast four singing tones at once on mics
i have to use a tonal isolation software to amplify the one people here in irl..... on mics i sound like a totaly 90s annoying teeng like omg god

my baritone and tenor base are both very powerfull by the way then tend to hit each other on the mic and that hint of falseto tenneor from kid voice traing always finds a way to shine

i have what called the freedy mercury vocal chord thing but both for voice and singing. just not the teeth thing

no i dont know the medical name i just call it harmoic vocal chords

[polar spade]

Work. Eat. Sleep. That covers it. Every day.

[restive pebble]

Recalling times due is an issue I also face but I made a killer instrumental for last weeks and somehow lost it after entering it. Now a 404 error...
The other gen was good so I ran with that but didn't enter as the rules are 1 entry.
This time I am making sure I have multiple copies. 😂

[restive pebble]

My pitch is far from accurate and tonality is surely lacking to be kind to myself...

[restive pebble]

Elderly neighbor moving today and called me at 530 to ask for help... The lollipop that I am could not refuse...

fun times at my place

I'm a writer not a mover or a shaker..

[polar spade]

Attach the second to the first ones post. Delete the first link. Or David notes to DM your instrumental and he'll switch in the Playlist. But I prefer not going that route. So, those I miss I just post in my personal Playlist of my challenge songs, just to keep up which ive missed.

[sharp kettle]

lol useing the never ending sing song method even if you tone deaf it makes the voice melodic any ways... how do you thing johny cass sang ring of fire lol and johny cash cant actually sing hes just sing songing...

singggg sooong is like that baaaaby just draw out letters that can be drawn oooout

[restive pebble]

He asked about it but I declined in the interest of fair play

[restive pebble]

care to elaborate and school an old man?

[sharp kettle]

the first part of sing song is vocal chord traning simple sing about anything for thirty seconds and increase it to a minute by the end of the week aft a month try to go up to 2. minute and 30 second within a single breath. but never exced 3 minutes lev eveer!!!
ever like 3 mouths youl be able to sing song for a full 6 minute song

the second part is drawing out sound that tonal training
siiiingggg sooong is like that baaaaby do that for a minute at a time never ever go over 2 minutes! if you drawing out sing song you wil strain your vocal chords

i do both btw in my stream intro song btw so you can actually here how it sounds i even throw in the breath part lol on purpose

you could also sing johny cashs ring of fire for sing song practice or almost any of his songs but they have varying dgrees of difficulty

[polar spade]

When I first read this I thought you were referencing I Walk The Line , where he goes through 4 key changes. It sounds never ending.

[restive pebble]

All of this is confusing but here's what I attempt.. note the word Attempt... has stressors on it!

I write all my lyrics and when Suno comes up with the sound I envision for it, I use a mic, headphones and my daw to capture the vocals I sing. The song plays in the headphones and my voice to he mic.
While this occurs I havea pitch meter to show where I am. It is no where close to matching any proper pitch where ever I try to place my vocals.

I know it takes time to learn and train the cords but it's been a couple weeks and unless the pitch meter s lying, progress has not been made.
Also I have found numerous threatening letters in my mailbox since this training began...
I do not quit but am leaning towards acquiescence very soon... 🤣

[open jolt]

Heyy simple but tricky question. How much weirdness % do you guys put in order to obtain the creativity and flexibility that was comparable to creative versions like 3.5 or 4.5?
Cause to be honest, my best music comes from the 4.5 when - I believe - Suno was semi - positively - hallucinating. I'm not sure V5 is capable of semi-hallucination that turns out incredibly amazing. I feel V5 is - great but kinda strict with the prompt? I loved when Suno could be very flexible and you could expect anything absolutely outstanding you were never thinking about at first. I don't know if I'm the only one who think that?
I let all the settings to default so weirdness is 50%, I just wonder if by default you guys think it's the "weirdness level" of previous versions? It's just my question.

[polar spade]

For what its worth v.5 hallucinates pretty good too. Just have to find a prompt that creates gibberish, then repeat outputs to get a variety.

3.5 may be easier to get a hallucination from just by prompting, [male vocal] and 5.0 might need to fish for a good prompt that triggers it. So maybe thats the difference.

I wouldnt think weirdness would make it hallucinate. Weirdness would take the hallucination or non-hallucinated and just scramble some aspects among the whole things.

Even a meduim-high weirdness can, depending on the styles and prompts , just change out an instrument or vocal, or alternate between them more.

Yea. The defaults are just the best settings the devs had set to work best with every genre on average . That just means some genres probably would need tweaking left or right, more or less, from the neutral setting of 50%, to work specifically with that genre.

[polar spade]

Hallucinated sound effects hit harder too. 🙂https://suno.com/s/PISvqFgHqgx7AlRd for some reason, this prompt kept hallucinating things like whoosh, explosions and taiko drums.

[glad ibex]

thanks!
How do I change the pronounciation of one word? when I'm in the editor and I change the lyrics it doesn't change it.

[half dragon]

These steps in Editor:

1. Select the section contains the word
2. The panel on the left where the lyrics are will light up the "Replace Lyrics" button - click that.
3. Put in your New lyric - depends on the selected section, you put in a line or a word or a phrase to correct the pronounciation
4. Click Replace button.

If it doesn't work, consider to spell the word differently or phonetically and try again.

[glad ibex]

ok thanks

[glad ibex]

the truth is that sometimes i do that and it doesn't change

[restive pebble]

You can post that information here and the staff will look into it.

https://discord.com/channels/1069381916492562582/1372994553233276970

[light mica]

спасибо большое 👍

[half dragon]

I would suggest sending an email to [e-mail] first - describe your situation and what you were trying to do - then wait for a response. It'd take sometime but in the meantime, visit #archived-ticket-help as @restive pebble mentioned for follow up.

[hoary plover]

Hello...looking for some assistance. I have cover art that is 9:16, yet it is being cropped. Are there specific size settings I can resize the image to ahead of uploading? Thanks.

[restive pebble]

yes thats probably more effecitve and efficient

[restive pebble]

I know of none exceptthe standard size you mention you have

[light mica]

thanks more, I've already contacted everyone, but there's no response. There is still no answer. The only thing I get is advice on where to write, but there is no answer.

[half dragon]

Yeah, unfortunately it is what it is. It sounds like only Suno support could help resolving your account issue. So be patient - it feels like we need urgent care but they have many many cases before you too. Do not try to make any more payment attempt until it gets resolved and do not try to send multiple emails as it will not help making it faster.

[light mica]

thanks . but I've already written several messages without the source code and paid twice. Thank you so much for responding.

[half dragon]

Hope this helps.

[open jolt]

Nice. I hope V5 can hallucinate as you say, but I'm afraid with the models becoming more and more accurate and stricts with the prompts , lyrics and mood, I suppose it decreases a lot of the ability of creating amazing semi-hallucinating musics.
From what I've tried, I feel the weirdness % just adds "weirdo stuff" for the sake of it, but it's not able to generate once out of 10 the amazing hallucinating out-of-this-world music I got from some of the previous versions like typically the 4.5 or 3.5. Also when I upload the audios, Suno is unable to accurately prompt the style. Like it doesn't recognize itself what it has done.

[hidden yarrow]

it's not able to generate once out of 10 the amazing hallucinating out-of-this-world music I got from some of the previous versions like typically the 4.5 or 3.5
You made me very curious. Got an example? 1 out of 10 does sound like a much higher chance than I would expect. That would be every 5th click on the "create" button. I feel it's more like 1 of 10000, but only IF the prompt is right for the job (and it seems impossible to know beforehand).

[open jolt]

Well, I'm under the impression - don't know if it's just me - that 4.5 was giving quite out-of-the-box musics with a unique prompt, you could get very different results and musical creativity, at least that's my experience. But yes if we talk about pure half-hallucinations, they were much rarer for sure. And for the rest, I'm not exactly sure if we can talk about "half-hallucinations" , but Suno was giving very creative and very interesting generations with 6 lines of lyrics and a very simple prompt. And those, when it happened, were very unique. You couldn't get twice similar.

[hidden yarrow]

"quite out-of-the-box musics" is a different statement than your earlier "out-of-this-world music"

[open jolt]

I mean the same thing. My mothertongue is not English. But I meant the same thing.

And what do you mean by that?

So hallucinations? I don't know. When the result was totally different and positively bizarre from what you were getting usually with a same prompt, I don't know if we can talk about hallucinations. It's for why I call them half-hallucinations. I can't explain otherwise.

[hidden yarrow]

Well, I asked you for a sample, e.g. a song that would fit your category.
By your description, I would expect something that would surprise me in a positive way by it's creativity and yet keeping it's musicality. Not mainstream. but still appeasing to most people, not just some unqiue A.I. noise.

[open jolt]

Ok gimme a minute. I have a good example that is striking. I have a couple of others but that are more musical and probably less obvious but let's see

Ok this one might not be obvious for everyone but for me it was cause it was super unique among plenty of generations of the same audio, lyrics and prompt:
https://suno.com/s/n4LH00qTqQiO3mei

This one is quite striking: https://suno.com/s/2NGgATzIOyXHk1FB
Normal lyrics, normal prompt, got normal stuff and this came

I freaking loved the result. Impossible to reproduce. Until 1:35 - after I had tried to extend it and see what happened but nothing special came again.
I'm not sure V5 could give positively bizarre generations like that once in a while. But maybe I'm wrong.

[hidden yarrow]

Are those original style and lyrics prompts or did you shorten/summarize them?

[open jolt]

They are original lyrics prompts

and original style. No summarize

[hidden yarrow]

sorry, typo, I mean: The style and lyrics prompts I see on your song page, are they the original ones and not changed after publishing your song?

[open jolt]

yeah they are original

So yeah my original question and concern was I was wondering if V5 was capable of such flexibility and weird creativity. But I'm not sure as the model became very strict with the prompting and all.

Maybe I'll come back sometimes to V4.5 to get creative ideas. If the model hasn't been dumbed down since.

[hidden yarrow]

Ok, I see what you mean. It is very possible that you would need different, may be even more specific prompt to get those results with v5. On other hand, if you would have written such complex prompts for v4.0, most of it would possibly ignored.
You also are using genres that are quite off-mainstream themselves (indie/ambient and symphonic/mystic), so I am not that much surprised. What I noticed with v4 at least (not so sure about v5 with all the adlibs), that the intro had much more freedom and creativity. Once the model got into the verses/chorus, the usual "song training" takes over....

[sharp kettle]

lol walk the line is the one of the hardest sing song levels johny cash does... like i said he doesnt actually sing
his entire thing is sing songing which is a form of melodic talking.
is an alternative form of music training.. funny thing is he he hadnt gotten into drugs and kept at it he would have starting sing baritone country tone...

[open jolt]

Yeah. That's true. But before I could just wrIte "edm, female voice, experimental" and it was giving me original good stuff (at least my taste) like this. Now I type "edm" and it gives mainstream stuff. And I think "experimental" is prohibited since the weirdness setting introduction. Well.

[hidden yarrow]

My best example is for v5 is this one, but it had very special prompts and multiple remixing. Still, it was a unique result within its group:
https://suno.com/s/lHiKtVxQJNoXbNt8

[gilded mist]

I hope V6, whenever it will come, will drastically improve audio quality.
Don't get me wrong, its not awful, but especially with rock or more "fast paced" music genres, the noise is still very audible and it's just so frustrating.

[open jolt]

Very nice. Very clean too

[hidden yarrow]

Obviously you did not work with 3.x or 4.0 at their time ^^

[open jolt]

Yeah could you prompt that kind of style again? It has a little bit of a massive attack vibe

But it makes me think of soul or something. I'm not into that kind of style so I don't know well but..

[hidden yarrow]

No, like I said it's a multi-remix and then some ... But you can fiddle around with Cosmic+Rumba+Jungle+Erotic Duet 🤗

[open jolt]

Yeah I see. At least you found a prompt

[gilded mist]

Oh I did.
I started with V3 even. I know where it comes from xD

[hidden yarrow]

Alright, then. Are you not happy, how far we (Suno) has come?

And yes, there is still room for improvement! But I dare to say that audio quality is not the highest prio anymore.

[gilded mist]

Imho it should be. It is still very far away from great imho.
It's not awful, it is amazing on many genres, but others, especially if they are nieche, it just breaks apart.
The same can be said about lenght. The longer the song, the more the quality degrades.

Also, if I would hate Suno, I wouldn't be paying the Premier plan.

[hidden yarrow]

I did not get the feeling that you hate Suno for that. I am just surprised that audio quality is your top wish. Mine would be "prompt following" and like @open jolt stated, it seems there was some kind of creativity lost in v5 (or needs different prompts to be awaken)

Oh, for stem extracting I would agree: it is a downer for every Suno Studio project if you start with full song.

[gilded mist]

Hm, maybe to me the focus is different because while I completely agree with your points, prompt following, or different way of prompting, is something I as a person can learn. It sucks to have to do that with every new version etc. totally agree! But after some attempts one gets the hang of it. Some will always be a gamble simple because how AIs generate but some is controllable.
But audio quality is not. Or at least if it is, I have not found a way to do it yet.
Especially in Rock/Metal genres, the shimmer/noise gets awfully obvious to me, to the point that I have to put on some other music to "flush" it out of my hears, like a tinitus even.

[hidden yarrow]

Yea, distorted sounds probably contradict its acoustic training. And also other effects like complex reverb are very difficult top create (for the model) without artifacts.

[sharp kettle]

try this with suno for sing song training.
step zero
learn were you best voice is on the tonal scale, bass baritone tenor soprano or even falsetto...
where you are on that scale actually determines which genre of music you can sing for..

(for example is the new lead singer for linkin park is a tenor/ soprano so she cant acting sing the songs they wrote and rock genre the band likes to play her voice is literally designed for pop and r&B but she could sing Rock-pop)

step one

write a song and use a prompt that says...
sing song this lyric set with Insert music genre here, keep steady easy to sing to beats, the vocalist should sounds like they are talking melodically or rhythmically. insert what ever else you need to into the prompt.

step two sing said song:

step 3 use suno and insert prompt
that should work

(if your already pro have it use vocal match if not cover it)

[gilded mist]

I wonder if we'll ever get advanced control over the model too. I'd love to be able to control the reference area, especially in long songs, so the chorus sounds the same, even if it's 2 minutes later, or be able to input seeds directly, or even be able to Pic quality/generation time. I wouldn't mind to have it all in premier or behind an increase in credits cost.

[hidden yarrow]

For quality reasons, I skipped stem extraction for this song and just added some stuff in Suno Studio, to keep the good e-guitar rendering and of course the voice:
https://suno.com/s/vuz4WpCNlkW9lGHU

[sharp kettle]

nice song was responding to anthoer person what was a person trying to create?

[hidden yarrow]

We were chatting about unique results from simple prompts in v4 and possibly limited creativity of v5

[shut marten]

Thanks @red oar I finally found a version I am "ok" with. Not perfect but better than the other versions suno was spitting out. Will try your suggestions. Thanks again

[soft siren]

anyone that has managed to make extremely chill lofi for studying with perfect audio quality? Seems that distortion and annoying crackle in the background always appear in most songs no matter what I try

[gilded mist]

What exactly did you add in Suno Studio?
I watched the tutorials they have on their YouTube Page but still feel either misunderstanding the controls completely or it severely lacking.
The Extracted Stems sometimes sound a bit off from the original and I still struggle with extending song in Studio. It feels like it's just a expanded Cover-Generator.

Like if there is a way to regenerate a specific part with only subtle variation (for example to improve vocal clarity) I havent found it yet. It always delivers completely different results.

[hidden yarrow]

Harmonica riffs mostly and some effects / cross-fading to fix some vocal artifacts, I think.
I replace / extend in Editor before using Suno Studio.

Actually, I extend only from create panel

[sharp kettle]

oh can't help wit v5 or studio... but ai doesnt do very well with simple commands any more
unless the programmers put in very simple commands to recognize.. the music sounds blegh

the prompt has to be like paragraph of instructions if you want it very unique.

edm is one of the hardest to get unique sounds for the other is echo melodies( which convey emotion in a song with out actually singing that way...

sunno does need to woork on its noise levels. audacity can only do so much to clean it up even with its own ai...
(its promps and music making is no where near as good as sunos though...)

forget up indivual instrument stems even audacity can't even seperate that hot mess

[gilded mist]

Do you know of a good way to fix mispronunciations? Whenever I try to fix or replace even just 1 line in a song, Suno sometimes generates 1 Minute of Song when I just want to fix 8-12 Seconds.

[hidden yarrow]

expand selection area, so it has to replace a whole line or even 2, sometimes a whole verse

[gilded mist]

I feel like I really need to learn to let go. Noticed it that I sometimes get way to attached to a flawed gen, trying to fix it but it just doesn't turn out right no matter what.

[hidden yarrow]

Yea, it's all about compromise

[sharp kettle]

yes actaul i do go to gogle and learn prounciation keys.
use that in the song lyrics
for example using the new horizens nation nation anthem here are a few spell out examples...

Hor-rise-zens
en-sures
Advo-cacy
eco-nomic

then its a matter of playing with the spelling as well

but keep in mind the ai read is not human and it maybebe pulling from at least 26 diffferent ways of pronoucing if the song is in english, 15 for spanish, 3 in chinese, and 4 in japanese, Italian has 6, french has 5... and german have 8... dialect reading..

so ai will always be hit or mis... still havent found a key phrase for livewire

genre pronuion puls are ever worse numbers fyi

rock has over 2000

[gilded mist]

I usually try to put words that would sound like the word I'm looking for (led for lead, as in the material, etc.)
But to be fair, sometimes even I am amazed that the ai can perfectly pronounce very hard words, but then for the love of god, cannot get words right like "Industrial".

[neat nexus]

I feel this so much.

Thank God for phonetics.

[sharp kettle]

if you know how to acces the accent letters on a physical key board learn accent language
song writing also help ai. but ive never figured that out. and also most skools dont teach da accente lanagauges

[gilded mist]

Do you perhaps mean Phonetic? (for example industrial being written /ɪnˈdʌstriəl/ ) ?
Does the AI understand those?

[neat nexus]

It understands when you try thse (don't need to do the special characters necessarily) and yes, homonyms are a life saver.

[sharp kettle]

In-dust-real, you can try phonetics i havent done those my self

prounouciation keytones is musical reading only part of sheet music i actual can learn to read

it also combines read sounds you-are-eh(funn music joke)

Me--ow!

teasin'

Raahh!

disastah

disastar

lie-ve-wire (closest to livewire i could get

ive done thing like this

[plain pumice]

It's actually able to pronounciate non-existing language, as long as it knows the base you're working from and it's consistent.
I've tried using Pseudo-Latin, which is basically latin at the core (western language core), and it worked really well actually.

[sharp kettle]

since you in strudio you have acces to longer song uplads for vocals right? you could just sing the song out and have the generator cover it.. it does a suprising good job or pronoucing. just have to maksure you type in improve singer qaulity in the prompt

my covers keep hitting an elvis like ai to for my baritone. its great lol

and i dont have pro or studio

[quiet solar]

“lyve wire” or “li’ve wire” may be worth a try. That’s the approach I’ve used when Suno gets a bit stuck on the context of how to pronounce “live”.

[sharp kettle]

thank you so much i was stuck on the ai prompt vor ever

[quiet solar]

Well, see if it works for you before thanking me too much, but hope it helps! Good luck.

[gilded mist]

I just tried using the editor to regenerate 2 lines to fix that pronounciation.
Now the newly generated part has some rough transition and the "Heal Edit" button just reverts the edit. 🤣

[sharp kettle]

heal-edd-it .... dont know what kind of song you writing but thats kinnda dirty lol

[sharp kettle]

heal-edd-it .... dont know what kind of song you writing but thats kinnda dirty lol

[gilded mist]

Oh no for once that wasnt part of the lyrics.
In the Suno Editor (not Studio) there is a button called "Heal Edit" that supposedly smoothes out rough transitions between inserted or newly regenerated parts, but for me, he just deletes the regenerated part in the editor instead of doing anything.

[sharp kettle]

off ouch

[plain pumice]

Yeah, it's a weird feature, from what I understood, it was intended to blend sections better / more seamless together after regenerating.

[frozen dagger]

Using Suno via web browser on my android phone. Noticed the UI changed, but now it pops up with a message saying "pssst... you may want to turn up the volume", but there is no volume control I see. Is it just trolling me? (Maybe UI changes still coming, but seems odd)

[gilded mist]

I really feel like Studio and the Editor would benefit from some real developer love 😄

[dusk lance]

how to have multiple voices singing?

[gilded mist]

I myself havent found a reliable way to do this, but depending on how you want to do it sometimes this helps:

A Sung line
(A Sung line)

Or

[Multiple Vocalists]
A Sung line

For some genres one works, for others the other and for me none of them were 100% reliable, sadly.

[frozen dagger]

[Call and Response] [Layered Vocals] [Vocal Harmony] [Duet]
(depending on how you want multiple voices represented: back and forth, layerd, etc) some examples of prompts you can add

haha now its popping up while i use the desktop browser, and the volume control is definitely at max in the Suno browser. So, it is trolling me. 🙂

[sharp kettle]

SINCE I many of you in this channel were interested in sing song for suno music i am online right now and free to teach
dont know what up with twitch lately but it seem to have trouble with my old music files and mic system
so many chirps ruins the music i cut... and my voice but lets do this lol.
i didnt see a promo channel but i gace a clue if yah want tech talk sing son wit me right now lol special stream.

[frozen dagger]

same username?

[sharp kettle]

lol hint recieved well i guess 😆

[sharp kettle]

lol i hope yall enjoyed this tech voice lesson lol lession 😂

[hasty cove]

Can someone help me? I work in hip hop, and Ive gotten Dj record scratching random vocals which is great in my prompts, hoeevr I am trying to get certain words scratched up to see how i can get the prompt to trigger. Ive tried everything i know. All i want to trigger is a DJ record scratching a rapper voice saying "Guess who's back" and I cannot trigger it. It does the scratching, but its just random sounds/not actual words... sounds cool tho lol

[lilac nest]

Has anyone ever used clock-ticking for the drums to tune a song to 120bpm?
Neat trick. Also makes for neat sounds.

[frozen dagger]

telling me lyrics coopyrighted. guess they dont want any smoke from Em

[hasty cove]

well it can say whatever... i just need to figure out how to get the engine to scratch whatever i want it to scratch vocally

[frozen dagger]

small sample, but just using DJ Record Scratch vocals as the style and typing in the words does ok. can add more flair with some phonetics it seem like l-l-l-ladies and gentlemen did ok. It is adding its own background beats youd may have to seperate but it working that way so taggining in lyrics should give similar results (did both in some gens)
https://suno.com/s/LnDP7MTAr2Y94zqX

[hasty cove]

ill try this , did you place the words you wants in brackets or anything?

[frozen dagger]

at first but then removed in example shown

https://suno.com/s/tvkMUsrEwZbnyJk7

[hasty cove]

ill try thism you doing this in the lyric section or just song decription?

thanks for this

[frozen dagger]

at first i did but then removed and it still did it pretty good. the phonetic typing to kinda cue the l-l-l-ittle p-p-parts you want to stu-tt-er-er-er should be attempted but prolly be its own interpretation. can just type normal and let it it interpret from that

[hasty cove]

ahhhh I got it ! ok ill try this thank you!

[frozen dagger]

you can put [DJ Record Scratch Vocals] in your lyrics box too and it should use it where it's placed with the lyrics in the section with it. (or until you put new [tags] that maay overwrite that)

[hasty cove]

thanks a bunch!!!

[ornate minnow]

Hope this where i go for help, Im using Suno and have been for awhile however i stopped because of an issue i have. Any songs i make with our without persona. and trust me i found a persona i love. but my song is the same on anything i do. this would be great if i wanted to change lyrics for that said sound. But im not im making new songs and its still making me us the same sound over and over and over again. i haved wasted so many credits trying to get songs and trying to fix it. i have cleared my browser cache i have log out and sign back in and still same issue. i can't make any new songs and have not been able to for months. i even messaged in email for help and i have not heard back from the email for about week or two now.

[neat nexus]

Would you be willing to document this with some links showing this in #🪲┃bugs-web ?

[onyx sparrow]

where is the master chef of suno? i need more prompt space:>

eh plz

xD

[frozen dagger]

have to turn use your lyric prompt box and often plan extends making your song in sectons with blocks of prompt tags in with your lyrics prompt if need extra space.

[onyx sparrow]

:/

🙁

[frozen dagger]

not as bad as it sounds lol. just means you can use you lyric box too. (audio is it's own prompt as well)

[onyx sparrow]

i need also more negativ prompt space

i need more space of everything:D

okay...

yeah i understood but thats not my style i dont like the magic stick

or to optimize the prompt

s

the magic stick deletes words and thatssss a no go

or he change the meaning sometimes

https://tenor.com/view/halo-master-chief-nod-nodding-pablo-schreiber-gif-26177401

[frozen dagger]

the magic stick has it's uses, but mostly i use it to see what the ai thinks something means, or how it would phrase. we have 5k limit for lyrics tho. and you can ptu all your prompt [inside the brackets] and it will use it. (and then again with extends you can stack all these).

[onyx sparrow]

okay

and negativ prompt?

i need more space i try to mix a beat

[frozen dagger]

i'm not sure on that one. not sure if it uses same:
[-country, -trap, -female vocals] tyle of negative promptint that ai image gens can understand or not

[onyx sparrow]

yes this one

for example -white noise

or -muddy mids

[frozen dagger]

i knwo we have the box for exclude, and each extend you should be able to stack like a work around, but i dont know for sure if it will read it like that with bracket prompts. Negative prompting is a mixed bag anyway with most ai generation.

[onyx sparrow]

okay

i dont use bracket

i just use -

do i need brackets?

[frozen dagger]

any prompt tags you put inside the lyrics box (which is same basically as typing in the style box) should be in brackets so the ai knows they are instructions/prompts

[half dragon]

If you use Exclude box, don't use - either

[frozen dagger]

it looks like the exclude character limit is 1,000 as well

[onyx sparrow]

hier?

here?

okay

yeah i need higher limit

😄

[frozen dagger]

yes that is where you type anything you want to exclude (negative prompting) don't put "-coutnry" jsut "country" seperate with commas like normal list country, trap, white noise

[onyx sparrow]

yesss

can someone tell the masterchief i need higher limit

maybe we get one in the future

[frozen dagger]

if i were trying to use lyrics boxs for more excludes i'd either try
[-country, -trap, -white noise]
or simply
[exclude: country, trap, white noise]
see if it did anything, but that me theorizing. i dont work with negative prompts myself much because i find it counter productive in my experience. just sayin'.

[onyx sparrow]

i only edit the skin

surface

the audio surface

i dont use prompts to forbidden something

nvm..

https://tenor.com/view/game-over-80s-galaga-retrogaming-gif-23439591

[frozen dagger]

i understand what you trying to say i think. you weren't excluding any styles to limit creative boundary, only to exclude bad noise and artifacts to get good quality.
at least, that's what i got... after you edit it lol.

[onyx sparrow]

yesssssssssssss

https://tenor.com/view/incident-gif-2747499635347300991

for mix a beat

soo eh no extra space:>?

[frozen dagger]

in my experience, again no wrong way, i prefer to try to say what i do want as much as i can rather than what i don't want. as specific as possible.

[onyx sparrow]

if i write -no true peaks he check this?

he understand this?

[frozen dagger]

extra space for exclude only thing i can think is to create a track with all your prompts in it as much as you can fill it, then extend near the very beginning of the track and then change the exclude box with new info. Every extend should have all the info from both parts of the extension. (in theory)