📚┃suno-school

[thorn cipher]

Wondering why when i make a rap song, the vocals just start to get raised deep into the verse and eventually turns into shouting, i want it to keep it all on the same tone, any tips?

thanks, will try this out

oh man, no wonder my prompt kept failing

Genre: Rap/Hip-Hop
Tempo: 86 BPM
Key: A minor

Drums: Boom-bap kick, sharp snare layered with rimshot crack, open/closed hi-hats with loose swing, sparse tambourine + shaker accents

Texture: Vinyl crackle + tape hiss base layer, soulful vocal sample chops, dark detuned piano stabs, gritty organ chords with low-pass filter, sustained eerie strings low in mix, jazzy horn stabs

Bass: Upright/double bass loop, sub-bass for weight.

FX: Reverb-heavy echoes for transitions, subtle reverse cymbal swells, DJ scratches on hook transitions, distorted guitar licks for grit.

Vocals: Pure rap delivery, no singing or layered vocals, ad-libs behind main flow for depth, raw, emotional tone, storytelling, reflective, and punchy, male raps with a steady tone/flow, DON'T RAISE THE VOICE IN THE VERSE

no matter how many times i tried it never worked 😭

🤦‍♂️

can i dm you rq?

[quiet solar]

As a general rule, don't tell the AI not to do something as that can backfire. Suno is much better at responding to you telling it what you do want, rather than what you don't.

[frozen dagger]

Reminds me of how police shout "Hold your fire!" all they heard was "Fire!"

[magic forge]

Hello everyone, I need some help with my audio. I hear a faint microphone noise, like a “sheeesh,” in the background. How can I fix it?

[half dragon]

Maybe try to replace/regen that section in Editor?

[magic forge]

I think the problem is in my style prompt. I need to understand it better,what should I include or mention to fix it?

[half dragon]

Is it audible in the instrumental section or during the singing of the lyric?

[magic forge]

instrumental

[quick sonnet]

But style prompts which worked in 3.5 do not the same job in 4.5.

Remaster is also a feature avaible for subscribed users only.

[half dragon]

The simplest way: select the section, then in the new lyric box - just put [Instrumental Interlude] or [Melodic Instrumental] then click Replace
Unless you need something special or different then you added in a style prompt - but to replace a section to remove unwanted sound I'd simply do that.
Same way to fix a wrong pronounciation of a word in singing part.

[stark dove]

Still facing the challenge of making a song in this style below.

https://youtu.be/TklyWxDxo-w?si=oI2jjVB6sxUYR1vT

Even more so when you put lyrics to music...

Like this example of Shiro used in Evangelion.

Aggressive guitar, aggressive background choir.

No main vocalist.

[unborn star]

Doing Instrumental is a lot of work

Since usually if you create instrumental tracks there are like under 1 min, if done poorly

I guess I could make it work, but even for me it takes a lot of time.

I´m specilized in instrumental songs

[stark dove]

The problem is that I want it with lyrics, but only in the background, but every time Suno adds a main vocalist....

[unborn star]

It can work but every text must be in (Lyrics)

• you need to say in the style box, no main singer, just background vocals

something like this

[quiet solar]

In which case, I think your 3.5 version is probably going to be the 'definitive' version of the track.

[unborn star]

• maybe you just say background choir

[stark dove]

Yes, I've been making music on suno for a good few months now, but I can't do just that kind of music, I'm trying to make a boss battle track, Elden Ring style, with the lyrics played in the background in Latin, without a main vocalist,

Even putting "no main singer, just background vocals" doesn't work

[unborn star]

I saw some people making songs like this

[frozen dagger]

[Latin Choir Glossolalia]

[unborn star]

is that really a tag xD?

I´m working on making Xenoblade Chronicles Music right now

So we are on the same page (gaming music)

Maybe you should show us some examples you made

[stark dove]

Same thing, a vocalist singing with a background choir, Suno always has a lead vocalist...

[quiet solar]

You may be able to hand the relevant section of the song entirely to the backing singer if you use ( ) brackets around the relevant lyrics, though your mileage may vary depending on genre. If you define the backing singer(s) in the style prompt, that's even more likely to be effective.

Alternatively, you can definitely get a soft, choral chanting "lead singer" going on, but you might have to play around with prompting to get it right.

[unborn star]

It needs proper prompting to mkae this work, just adding one tag won't fix it

we didn't even saw your prompting

[frozen dagger]

it's an LLM. it reads it and interprets it. you may find better success with what you are looking for, but its something like that it sounds like. [Latin Choir backing].
Def not say "No main vocalist", as abovoe saying that jsut reads "Main vocalist". You coudl try adding main vocals to excludes, but prolly not necessary

[quick sonnet]

Remaster is not avaible for me:

[frozen dagger]

I would try to find promptin to create just that choiir effect, then when find something close, put it into a track, either stitched on or layered. If you have an audio sample of what youre looking for it will be much easier to replicate than text alone.

[stark dove]

I've changed the lyrics many times, with the (), in the prompt, more than 1500 credits, and they all gave the same problem

[quiet solar]

Yes, you said. Hence why I think you're going to be stuck with the 3.5 version, or accept a distinctly transformed version, if you don't have access to the subscriber tools.

[quick sonnet]

3.5 is not avaible any more, at least for free.

[quiet solar]

@zero But songs made using it can still be played. Or was it a work in progress?

[frozen dagger]

Listening to your example track now. I've had succes with similar concept emulating One Winged Angel before v5 stuff

[polar spade]

Correct. If we're covering to hope it sounds similar, then no style prompts would be used.... But, as you said, it's not working as you want.. . So it's a choice of what new prompts will counter the differences in order to bring it closer back to the original we're trying to cover.

I said it something to try. I'm my experience, we had 3.5, even after slider were introduced. Then I was already upgraded, and finding counter prompts were not as important.

There's no list of opposite styles, so it's a guessing game, I would think.

Maybe someone else can think of a more solid solution. Anyone using 3.5 continually would have more experience by now of what to try with 4.5all than I would.

[quick sonnet]

Yes they are still avaible, can even make covers of them, but can't select v3.5 for generating new tracks.

[stark dove]

It's an excellent example, how did you do it?

[unborn star]

no link to your track?

[quick sonnet]

But 3.5 can't be used any more, at least by free users.

[frozen dagger]

focused on instrumentation used , in my case "Heavy Tempani" was a big influence. At the end of the day the models are likely trained on the music you're trying to emulate. It's finding wordd weights to trigger sounds. Phyrigain Mode and Stacatto stuff helped with the structure i was aiming for

[unborn star]

I´m interesting in your result, just want to listen to it

[frozen dagger]

https://suno.com/s/P33VRokstWwdtIN5
I ended up not using this and used a more FF classical sounding thing but yeah

[unborn star]

oh you should know doing a longer song is more challenging than a 10 second

[frozen dagger]

this was a clip there are many extensions that used it. it was for a segment of a song that ended up being like 8mins?

[polar spade]

At some point there's a limit to what magic can be applied. That's the idea of their coming up with newer models.

Basic is just a permanent free trial with potential. That potential may not always come to light without upgrading.

[stark dove]

the same problem when I tried here, when the lyrics really start, the vocalist appears and disrupts everything

[quick sonnet]

I know, but what worked already in older versions just not doing it any more. Already the 1 min samples which were generated for promotion purposes, have shown that old prompts in a new version generate garbage, while the quality of that samples was beter they mised the style prompt 100%.

[frozen dagger]

so you're main issue is trying to create track that has these chanting backing choir vocals, but that's it, correct?
If so, firstly are you wanting non-intelligible chanting, or (I assume) you've written lyrics?

[stark dove]

I'm trying to make a song that only has choirs and instrumentals (guitar) like the example sent from Evangelion.

No main vocalist.

I have the lyrics done.

[frozen dagger]

If you're using Lyrics in the lyrics prompt, its going to try and sing them like a normal main vocalist would. Prompt tags could have it start the way you want but it could vere off on it's own. If that's what it is doing, then you may have to extend from where it was working, before it stops, and reinforce the prompts that are giving you success with backing choir.

i dont even know wwhat triggered woodhose that time

[stark dove]

The last one I did was this https://suno.com/s/3OcBKyomewL2pko9

[quiet solar]

I've just been testing, and it's possible to strongarm Suno into doing it - the problem most likely is that most examples from its wider genre have an explicit lead singer, so you have to keep reminding it to use the full choir.

You're probably not going to get it perfect in a single run, but you should be able to build up a song using Extends where it gets parts right, so it builds off the idea that both the Verse and Chorus are chanted.

For example, I got a chanted verse and chorus here: https://suno.com/s/PXDs8yuw9yqJjYD2 (0:38 - 1:28) on the third cycle of tweaking the prompts. If you're familiar with how to work around Suno's limitations for forcing vocalists to do what you want (the general advice for duets is highly applicable here), it should be possible to use a mixture of extensions, edits to remove/replace vocals and multiple regenerations to create the component pieces for the song, that you can then put together in Studio or an external editor.

[frozen dagger]

gah it has words like faith in the prompt and its gettign flagged haha

https://suno.com/s/dg6P2sF2LvBhst5A
i would add djent guitars, but this achieved with no lyrics in box

[unborn star]

You have a lot of Space in your Style Box, use it

[frozen dagger]

i re-iterate no lyrics used in this, despite it clearly having lyrics. lending more to the "it pulling from sources unseen" with the model training.

[stark dove]

Okay, but how would I put the phrases I want in the music without using the "lyrics"?

[frozen dagger]

most likely through repetitive use of tags that are working for you and more extends like Kantrhic is saying. Not likely a one and done.

[stark dove]

I'll try a little more, it was already more credits than it's worth, I should have given up, it's not even worth it anymore

Now it's my stubbornness...

[frozen dagger]

waiting for this gen to finsihs for tests. using [Latin Choir Chanting] at start of each stanza. see how it does

[quiet solar]

I don't think you're going to be able to get it to drop the lead vocalist. Your best bet is to introduce one you can live with and accept they'll come in from time to time, which can be excised by splitting the vocal stems out. Or limit the vocals only to a chorus and focus all of your prompting efforts on getting that chanted.

If you're trying to do all of this in Suno, without use of Studio or an external editor, it's going to be a painful ride. However, you can see there's nothing particularly special in the prompts or structuring that I used - it's not a prompting issue, but a model issue holding you back.

[frozen dagger]

https://suno.com/s/nHiyhSNAZpx9DrCa

seems fine to me... again tweak as desired but repeat the tags.

[quiet solar]

And OBLMOUS just demonstrated perfectly why focusing specifically on a chorus only, which Suno is very good at repeating, is such a viable option.

[frozen dagger]

and you can always separate vocals and fill with new music or even cover the vocals for new twists

[quiet solar]

It's when you introduce a verse/chorus structure that it gets a lot more complicated to direct Suno as to what you want it to do with vocalists.

[stark dove]

This one almost worked, I just need to add the guitar and get more aggressive, I'll try to go with that

[frozen dagger]

i added only one "djent duitar" bit in the prompt, but i would def reword the style prompt to get more djent edge for that gamer feel.

[stark dove]

(I decided to test it on elevenlabs, and it worked the first time... sad day for suno)

[torpid junco]

Kinda new to Suno... is it possible to use section tags (e.g. [verse], [chorus], ...) when creating an instrumental? I see all the examples using the tags with lyric-based songs

[frozen dagger]

yes. Some tags can effect song in instances, but they do 2 things effectively. break song into sections for editors, and instruct ai what part of the song structure it is, and therefore how to treat that section.

[frozen dagger]

an example might be having an [Outro] tag but after it there is an [Encore] and it could create a better "false ending".

[torpid junco]

I'm currently wondering about syntax, I don't have any up-to-date docs to go by.... for example, if I put a description of the instrumentation I expect in each section, the resultant song contains a voice that will sing those words (when all I want them to do is control the instrumentation)

[frozen dagger]

place that info inside of brackets, friend. [Piano solo], [Virtuosic Percussion], [Xylophone Interlude]

[frozen dagger]

anything inside of [brackets] treated as instruction/info. use them in the lyrics prompt where you want them to appear in song. anything in (parenthesis) is sang as backing vocals.

sometimes () work but not reliable. brackets reliiable

[jaunty swallow]

Hello everyone! I used the same lyrics on Suno with 490 different prompts to create 980 songs and made them into a website. I need some feedback, but I can't post the URL - I just got warned about it once. Do you know how to share URLs properly?

[polar spade]

We're saying the same thing.

You did ask, if covering a 3. 5 can be done in Pro and it sound the same. I'll test again, but I think it can except the vocal typically needs sliders to get closer.

[unborn star]

You need to keep testing to improve your skills and prompts. Think about it, you got some new informations =)

[finite folio]

I discovered how I can remove the lyrics from my songs 🙂
I'm marking Lyrics:
[Verse 1 Instrumental]
[Chorus Instrumental]
[Interlude Instrumental]
[Verse 2 Instrumental]
[Chorus Instrumental]
[Interlude Instrumental]
[Verse 3 Instrumental]
[Chorus Instrumental]
[Interlude Instrumental]
[Chorus Instrumental]
[Verse 4 Instrumental]
[Interlude Instrumental]
[Chorus Instrumental]
[Outro Instrumental]
And it works for me 🙂

usually "Simple" instrumental and when making a copy, it doesn't always work for some reason.

[brisk igloo]

since i'm not sure if my message went through, i'l ltry this again. stupid web app...
is there away to mimic this from mass effect3? fate of the galaxy: https://www.youtube.com/watch?v=LB0GwKV5s4w

[frozen dagger]

A sweeping orchestral space epic unfolds with swelling strings, bold brass fanfares, and low, pulsing percussion. Ethereal synths give a futuristic edge, while choir textures add drama. Tension rises in the bridge with urgent rhythms and layered electronic elements, peaking in a heroic finale.

Search for what instrumentation was used to emulate closer. Add your flair

[brisk igloo]

huh. so i baiscally throw this prompt at suno and it, theoretically, should beable to pull this off?

let's give it a try

[frozen dagger]

its not going to recreate the track based on some words.... but that is literally Suno's words. Not mine.
"The Fate Of The Galaxy · Sam Hulick" in stlye prompt. then click magic wand. Doesn't always work, but can give you an idea of where to start, and how Suno might interpret; presumabley based on its training.

i hit create it was pretty close as far as I'm concerned using minimal effort. I wassn't trying to recreate it with more prompt tags in lyrics for structure or defining the paramets more. Just giving Suno a general idea.

[brisk igloo]

right

[finite folio]

Maas Effect all seires use usually music styles Sci-Fi Ambient ( Space Ambient) and Electronic Orcestral (Hybrid Orcestral) and Composers: Jack Wall, Sam Hulick, Richard Jacques, David Kates.

[brisk igloo]

well i guess what i was oging for is how the song basically changed, even when in miner key from i think it was a flat to whatever flat key it used. lol i'm only good at the main not flat'd keys

[polar spade]

I always suggest, run it through a genre analyzer such as Audio AI Dynamics. Write down all the genres as prompts. Paste in the styles prompts box and Create. That can get you a vibe. Then it's tweak, tweak, tweak, in all directions within the sound you want and with whatever progression you wish.

You may wish to put all of this in chatgpt or other to get a phrasal description, if you want to go that direction with as much metatags in the lyrics box. Or flood the lyrics with audio tags and just a few prompts in the style box.

[frozen dagger]

"in the style of ...." and click the magic wand in the style prompt, do it multiple times, can help when you're trying to think of how to describe a style used by someone else.

[frozen dagger]

https://audioaidynamics.com/genre-finder
nice. thanks.

[brisk igloo]

btw, ust curious, how did you get suno to pull that off? as in, how did you get it to do those words that basically described it?

[finite folio]

Sorry for I spoke about Mass Effect music, but I'm a fan of that game myself 😄

[brisk igloo]

nawi 'm a fan of the soundtrack

[frozen dagger]

type the artist/style then click the orange magic wand. Suno will turn it into a prompt for you.

doesn't always wrok. sometimes it will just not make any changes. which case use chatbots. to kinda get same thing. instrumentation used goes a long way.

[brisk igloo]

yeah but, like, did you do it in the mobile app? or is there asuno extension?

[frozen dagger]

oh, desktop pc. web browser.

[brisk igloo]

o ok. where do i do that? i'm guessing it's in the create menu on simple mode? or is it in that new studio thing?

also i'm blind, and can't see images

[finite folio]

https://suno.com/song/b8ffe2a1-0c92-486a-9a81-47c49962c35c I got this musik that like I say before in Lyrics all I verses and etc I put instrumental. Title is maybe Wrong because this is not actually Cyber style I think.

[frozen dagger]

Ya it in the style prompt box when usiing the Create tab to make songs using web browser on desktop. i double checked web browser on phone, and it's not there. sorry didn't think of that.

[frozen dagger]

Original had lyrics and your prompt helped remove them? Is that what you're saying? So use that repeating instrumental prompting to push the instrumental and remove lyrics that were coming in on a cover?

[brisk igloo]

huh, interesting

[red whale]

Good evening. I'm not sure what place this question is most appropriate for here. I have a generation where I just love how it came out. However, I screwed up the lyrics and I want to replace all of the lyrics while retaining the music. I know this is a possibility but in Studio it seems like you can only do sections, is there a way to just wholesale replace lyrics while retaining the music?

[finite folio]

It did not add words or other vocals 🙂

[frozen dagger]

right, the original it covered had words tho, correct?

[finite folio]

Yes

[frozen dagger]

good job. way to strong-arm the ai. 🙂

[finite folio]

I added some styling, it would make the copy a little more interesting, but the main story is still the original. 🙂

[brisk igloo]

o and btw, how would i do this on for Mr. Robot, specificallhy, this track here
https://www.youtube.com/watch?v=HwT30uIPFqU i've been trying to do this for a while, but not sure

[frozen dagger]

A dark, pulsing electronic track rooted in industrial and cinematic influences. The intro sets an uneasy tone with sparse synth textures, distorted percussion, and low sub bass. Layers build with modular synth arpeggios and aggressive beats, peaking in intensity before a minimalist outro.

[ruby bane]

That's lit 🔥

[fringe shale]

what happend to the prompt feature where it would refresh the suggestion chips?

[frozen dagger]

u mean this orange magic wand button in the style prompt that rewrites your promtp?

[frozen dagger]

didnt hit reply to, see above

[fringe shale]

no the chips

they used to update as you typed if i remember correctly

and ytou could press a button to refresh

now they are static

[plain pumice]

That certainly depends on what version it's remastering it from and to.
In my experience, as long as you're remastering within the same model version, then you're correct.
However, IF from 3.5 -> 4.0, 3.5 -> 4.5, 3.5 ->5.0, same for 4.0 -> 5.0 and 4.5 to 5.0.
Then, be prepared for it to absolutely BUTCHER your song/track.
That's my experience so far, and of course it's because the model works differently.
I've even gone as far as to try and recreate songs with the aim of making them sound as close to identical, trying with a variety of prompts, styles, slider settings, even covering with no prompt and/or no styles, sliders heavily favoring the original - still experimenting but not there yet.
Hopefully I'll I find a methodology that works which I can share.

[quiet solar]

As a general rule, it saves a lot of headaches to nail down the lyrics as much as humanly possible before generating the song.

There isn't a simple way to wholesale replace the lyrics, but if you have Studio your two best options are:

1. Replace the lyrics using the Studio tool section by section, keeping what you're happy with and building on that for the next edit. It is likely the instrumentation will go through subtle changes during this process, and you may need to cover or remaster the final version to clean it up

2. Use Extend near the end of the song and create the whole song again as if from the start with the new lyrics in the Extension. You can then use Studio to remove the original vocals and bring in the new vocals in their place, with a bit of manual adjustment to make sure the timing all syncs up.

[quiet solar]

Yeah, remasters often break as much as they fix between versions. But a remaster set to subtle is a better bet than a cover on default settings.

Suno will never generate an identical copy, but the best practice that I'm aware of for attempting to replicate the same song is along the lines of:

• Create a persona of the original song
• Use the persona in a cover with high audio influence at 75-95% and low weirdness at 5%-15%
• Use identical style prompts to the original with mid-range influence, unless Suno deviated from them significantly in practice - in which case turn the style influence down to low
• Repeat until you get a version that's close enough that you're happy with.

It's something I'd encourage people with limited credits or patience to steer clear of, much like duets or mid-song genre switching or anything else where Suno is required to do specific things in specific ways at specific points in the song.

[plain pumice]

Keep in mind, not everyone has access to Studio.
@red whale
This can also be done with the editor with the "Replace Section" feature, where highlighting the words in question and replace them with the correct words/lyrics, choose fix alignment and then hit replace. This isn't a 100% an absolute fix, and most likely will also change the music in the selected region, but if you use this with precision and as a surgical tool, you're more likely to get good results.

[plain pumice]

Actually, unless they changed it, you can if you use a persona and don't change the prompt, get an almost if not identical match - this was before sliders was introduced - so I think, this would be the equivalent of keeping the sliders favorable to the original song (0/0/100).

[graceful condor]

Honestly at this point fixing misspelled lyrics in a song with Editor or Studio is stressful/unfulfilling and pointless as it could possible be. You have better chances getting a good version (slightly different) by Covering the song again and fixing those issues that way than using this crap of a studio. I wasted 150 credits only fixing two lines yesterady because the Ai does what it wants instead of following guidance. And I just tried the same again with another song and I am done with this Sh.t! I am so sick of it that 3/4 of Suno is not working as it should be.

[eager siren]

my experience with suno so far: 50% "omg this is amazing" and 50% "wtf this is not what I want please stop doing it, not being able to replace only bad parts, not being able to enforce things and prevent others, losing credits in editor and studio and not being able to get things rights)
my best songs are "happy accidents" and the ones i really try to craft to reproduce those "accident" but with more controls are pure failures

(like if I could simply tag a thumb up and down on some PARTS of the music and launch it again (i suppose the editor was created with taht purpose, except it never really works)

[graceful condor]

If you get 1 good day a week managing getting a good spot in Suno, you are already lucky. Each day is like gambling in Gacha if you are lucky or not and in most cases you are not. This has nothing to do with prompting anymore but with pure luck, what spot you got in SUNO (like ping in online games) and how arbitrary decisions the AI is making because it does what it wants. With each day, each week I feel we have less and less influence and how the song turns out in the end. Literally all from Soundquality, Voice quality, prompt quality is a steady decline and inconsistent as it could be. I do not know what they are doing (because this is not fixing what they do) in the background but they a ruining this platform with each freakin day a bit more. Because there is no logic anymore to anything if one and the same prompt generates great songs on one day and then absolut gibberish, low quality dog sh.t all the other days.

[quick sonnet]

Sorry, my fault then.

[plain pumice]

There is definitely a lot of hit and miss, especially when you envision something and get something very different.
But keep in mind, what you think something means, may means something different for the model behind, it's basically trying to interpret and predict what you want from what you're telling it. I'm curious though, what version(s) of the model(s) are you experiencing this decline with?
Because there has been a shift in how the model works with each major version update, which requires the user to make changes to how they prompt.
Could you share an example of your prompts? Perhaps we/I can help tweak it to get better results.

[graceful condor]

Dude it is V5 and 4.5+ too. Because of the mechancis they forced on the whole system even 4.5+ does not even work like it did before. It all changed and it changes daily. Like a schizophrenic person. There is no logic to the system anymore. And stuff that worked literally yesterday doesn't anymore tomorrow or today. Sliders positioned etc. And it has nothing to do anymore on "What I think it means...etc" because we are talking about simple stuff. Stick to the lyrics, sing what the lyrics say and not some gibberish sh-t. And then messing up sound. Suddenly dropping volume, stutters in voices, hiccups, weird instruments or fx sounds no one asked for especially in songs and genres were it makes no sense. The list is endless.

[polar spade]

. There's no fault. I was suggesting to try alternative prompts to control 4.5all to sound like what 3.5 would sound like. We both agree 4.5all cover are not giving similar results as 3.5 covers of 3.5 songs. Most all models have slight changes in covering older model songs.

If a 3.5 Cover of a 3.5 rock song has some great distortion, but the same covered song in 4.5all now has crisp pop sounds - - both covers having a BLANK style box - - then just adding the prompt "distortion" and no other prompt may help. Maybe this is not in the style box, but more peppered among the lyrics this time to offer a little influence over whatever makes 4.5all cover a different way than the distorted sound of the original.

But again, the time to run through prompts would burn through credits at 50 a day. Skip a couple of lunches, save a months pay, get all your songs prepped, and buy 30 days of Pro account, use as much of the 2500 credits possible and cancel it after one week in, so you're not recharged again. This isn't uncommon for users to do.

[noble sleet]

Hey everyone! I’m new here and could really use your help.
I’ve been running into an issue with Suno always adding a humming intro to my songs, especially when using v5 Pro. I’ve tried every tip and workaround I could find online, but nothing has fixed it so far.

The only thing that worked was switching to an older version, but then the sound quality and style don’t match what I get from v5, which is honestly the best for my needs.

Has anyone found a reliable way to prevent or remove the humming intros in v5?
Would be super grateful for any advice!

[plain pumice]

The "random" things you mention, like volume drops, stutters in voices, hiccups, weird instruments or sounds, etc.
Do you by any chance have a higher Weirdness value? Do you have high slider values across the board? Because this also has tendency to mess up songs. Another thing to keep in mind, it's possible these problems also happen mostly when Suno is under extremely heavy load = Peak Hours.
When does this occur mostly, with new generations or remixes? Because from my own experience, this happens mostly when I was doing a cover with either persona and/or audio source, or with conflicting slider settings.

[polar spade]

I am describing what sliders do. It's just Basic has no sliders. So 3.5 has to do the same thing with adding (even repeating) and subtracting styles, instead of using a styles slider shen we try covering a 3.5 song with a 4.5all model, and it comes out sounding different.

[plain pumice]

Hi and welcome 🙂
I haven't experienced this issue, but I'll try to help anyways.
This is a guess, but it's possible it's related to the genre/style.
I don't know if you've tried this, but you can use the exclude feature to avoid this kind of behavior.
A Persona from earlier models might also be able to alleviate the problem.
If none of that works, perhaps try with an intro that specified as something else.

[graceful condor]

I know all that. And tested all that up and down the sliders range and time wise. There is no logic. As I said. Sure some stuff could impact the song in the extreme case like 100% Weirdness. But all the the not red area and tried all up and down and even not using it at all. Same goes for Persona or not. Like I said. Suno at this point has a schizophrenic personality because it really depends on its mood and daily conditions because no day is the same. And if you are lucky you manage to get a good day but they become more and more rare these days.

[eager siren]

my main concern is that there is obviously a kind of under the hood "good spot" that has an extrem influence on the results independently of the prompts, maybe the peak hours or else, but we have no info on it and in fact this should not have influence, this should just take "longer" but not influence the output
i mean one night i only got pure bangers, and since then the exact same prompts generate garbage :/
i mean we as user should be able to get a "index" on the website telling how much we can expect from a session based on server load

[polar spade]

But... 🤔.... What then happens when you remaster it all, when done, in 5.0? Or Cover first, then remaster?

With the assumption that remaster should blend out the hisses or muffle and bring back up "sound quality". It doesn't try adjusting by digital means like a DAW, it pulls out the best bits of the waveform and rewrites the whole thing into 5.0

[quiet solar]

It's a recognised "quirk" of v5, much like v4 was a bit overenthusiastic towards singers screaming in certain genres.

I don't think anyone has found a 100% perfect way to stop it via prompting, but you might have some luck if you specify an [Instrumental Intro] for the song at the start of the lyrics.

You can also retrospectively remove undesired humming by splitting out the vocal/instrumental stems using Studio or an external audio editor, if you're comfortable with a bit of extra work.

[graceful condor]

Same what I experience. And especially if you are a paying customer this is a really shady tactic to give users different spots in the system and they all vary in quality. Because that is for sure happening.

[raw oasis]

I never thought I would see this level of creativity in my lifetime. I completely understand if AI is still training models and does not always do what we want. I still manage to already have 8-9 songs on youtube (what I perceive to be demos for real musicians/bands to cover at some point). It is what it is ... in 2 more years it will be NEXT LEVEL amazing. Patience.

[graceful condor]

Or the Music industry is shutting them down and using them being their b.tch as they did already with Udio recently and who guarantees us that Suno is not next. Udio removed the download function and is Universals official b.tch now.

[quiet solar]

It's certainly possible that it's affected by server load, but over the last year of using Suno I'd be inclined to say it's just random. The same prompt can generate wildly different songs, that can be of widely different perceived quality and which may or may not contain odd mistakes, audio artefacts, etc.

[polar spade]

It's hard to convince folks, if 4% of the outputs are successful, then that can pay for what would have taken a month just to do one song with a 5 piece band and orchestral backing.. 😁

[raw oasis]

Well I do not see AI music ending, will exist in one form or another. Studios have admitted some of their artists are using it. Also I Just finished Mixing a song that (I hope sounds 100% better than the Suno version) so we are learning huge skills at the same time. I am a guitar player and use synts etc so not sure we should be ALL-IN on just AI but BOY does it sure help with the inspiration!! 😄

[raw oasis]

I also feel your concern. I posted on UDIO in #🗣️┃chill-chat yesterday morning. And would LOVE feedback from Suno where they are heading. All we know now is they are looking for 100million funding on 2billion valuation right now.

[quiet solar]

Having experienced the delights of booking a recording studio, session musicians and a sound engineer to blitz through recording a bunch of songs in half a day, even with a 4% success rate I'd take up Suno as a performer for demo/non-professional work in a heartbeat. That's still going to be thumpingly cheaper and faster in the round, and still produce something that sounds good on the far end.

[graceful condor]

For me having autism I love perfection and it is hard for me to let go if a song is just nearly perfect and Suno managed to ruin a part of it which can't be fixed properyl or at all. And sure you can generate again and again but you never the same voice in a certain moment singing a part so perfectly or a solo doing just that what you were looking for and I hate such things go to waste.

[eager siren]

me: hey suno here are carefully crafted prompts and lyrics can you please respect my prompts and make something sounds good ?
suno: NO
me: hey suno here is pure garbage lyrics to test how much you respect my prompts
suno: Hey, I just made a banger with it 🤘

[quiet solar]

Sod's Law in action, ir0n. 😂

[graceful condor]

That is the issue with Suno as a whole they have no transparency and share sh-t with us.

[raw oasis]

I also would like to expierience some Studio musicians recording my Suno demo's. I hope it would not cost me too too much? Mostly it is the Singer that is so valuable. So trying to network on Facebook with singers lol

[graceful condor]

Is there any info on the different characters to use and how to use them?

[vocal owl]

@oak ocean https://suno.com/song/dc818476-b3c4-427b-bd6a-e1ef02b25a94 check this out 6 min+ Track 😄

[quiet solar]

Have you considered using (or learning to use) an external audio editor like Audacity? The vast majority of things I make through Suno end up going there for tweaking, using Extensions from a 'base' song to correct mistakes and make improvements.

[graceful condor]

I tried that but Either I am not skilled enough but I do not find the outcome really professional. I still can here the transition parts from one part to the other and it bothers me. I am very sensitive to certain sounds and I hear and smell every crap even the subtle ones. It is a blessing and a curse at the same time.

[oak ocean]

ok, wait

[raw oasis]

Yes definitely be developing those DAW skills.
This is a A/B after mastering and killing horrible sounds that I am finalizing to video now.

I have Suno Studio sub now but am not super impressed yet. but still early hours of trying it.

[oak ocean]

it is good, but maybe a bit long. But song is nice, yes

[vocal owl]

😄

[noble sleet]

Thank you all so much for trying to help... I really appreciate it.
After testing around a lot, I realized that as soon as I use words like "dark" or "witchy" in my prompt, Suno automatically adds that humming intro, every single time. It’s frustrating because I love the overall vibe and style that those words give to the track.

I’ve tried all kinds of exclude styles, negative prompts, and over 100 variations of intro descriptions to avoid humming. Nothing works, the humming always comes back. The only way to remove it seems to be by completely changing the style, which totally ruins the sound I’m going for.

I even tried using Studio, but honestly… I don’t really understand how it works yet.

[quiet solar]

It depends on how long you need in the recording studio, plus how much you can do yourself and how much you need to pull in different session musicians for, factoring in their own costs and time to practice.

Expect a ballpark of $200-$400 for a half-day studio hire, plus around $250-$350 per musician and around $130 for a sound engineer. Costs will be higher still if you need to hire specific instruments or setups that the studio doesn't have.

In that time you could probably record 2-5 songs, depending on how complex they are and how many different instruments need to be recorded, plus how well-practiced and briefed the musicians are going in. Either way, a ballpark figure of $1,000 - $2,000 in total is probably a good starting point if you're not filling any of those roles yourself.

In comparison, that's about 5 years' worth of a Suno Premier subscription.

[quick sonnet]

Sliders were also avaible in 3.5 for pro users.

[raw oasis]

Thank you that helps. Finding a Pan Flute player (as an example) could be hard but never know.. lightning can strike and there could be someone in my area. OR that is an example where we can keep/repair those samples as much as possible and do both. One thing seems to be certain to @graceful condor point is the little glitches are really noticible.

[quiet solar]

I also have a perfectionist streak that I can't always turn off. xD

Unless the transition is unrescuable because the two tracks sound distinctly different, you should be able to make transitions so smooth as to be invisible. Crossfade Tracks is your friend there, and you have to get good at syncing the tracks up by eye and ear and making manual volume adjustments as needed.

Hiding the transitions in areas where the songs are particularly similar helps a lot - choruses are particularly good for that. Alternatively, you can deliberately hide the transition by putting something else over the top of both; I've used drum and guitar riffs and vocal harmonies specifically for that purpose.

This is what a Suno-performed song that I released looked like in Audacity by the end. The amount of edits, additions and corrections is probably about typical:

[raw oasis]

I am doing fades, limiters, eqs, compressors, gates, db adjustments, but not the other tricks you mentioned. Will level up on cross-fades next. Good tips on arrangment.

[plain pumice]

That's what I suspected, that's genre/style/tag related.
Perhaps Suno doesn't interpret it as humming, but as something else - which when using those specific tags triggers it - this suggest it.
I'm not sure what result this would give, but using both dark and witchy, but you could try to also put those in the exclude fields to see what happens. Have you tried creating a persona using one of the earlier models and then using it with the prompt for v.5, to see if it still does it?

Studio is also a beast of it's own, and definitely takes some time getting used to and learning the workarounds.

[graceful condor]

Yeah I might should look more into that and try it again.

[noble sleet]

I might try this out later. Thank you for your help

[graceful condor]

anyone having issues with creating a workspace? I noticed it changed. It's not working.

I have tried 3 times now .

I even searched for it

Not working in Edge either

It's not Chrome

yeah adding a new workspace is broke

Woah slow down...

[junior acorn]

Hello! I can't create something longer than 50-70 seconds. Im paying for Suno, I can't generate anything longer than the time period specified. Also when a song is generated the lyrics are something like this - In a room thats cold, dark and heavy (not even close to what I typed in). I don't know what's going it did some good generations 15 minutes ago, but now it completely decided to go delulu

[polar spade]

Still are.

From a Basic user perspective, no more 3.5.
Now: 4.5all only with no sliders. AS Basic we must find workarounds. So prompts and tags are it. No cookie cutter technique with prompts, as different genres are affected differently. It's a struggle.

[unborn star]

without seeing your full prompt we can't help
Wrong tags might be influencing your track length

[hollow adder]

🙄

[green pike]

Hiii @polar spade how are you? Did you applied for Helper role on #🔦┃server-spotlight ???

If you don't I think you should

you like live here

always helpfull

would love to see you there as a helper role tbh

[polar spade]

That's very kind of you to say. Thank you.

Lots of great folks in here who give their time to help, so I'm just a small part and hope I add something worthwhile to the group.

I hope all those new positive members who have been helping tremendously as well will consider being a Helper.

Sounds like something you'd be great at!

[torpid junco]

I'm having a lot of trouble creating what I thought would be something pretty simple - a stereotypical (instrumental) phonk song, but at the "drop" a banjo comes in. Is this possible with V5? A common thing I've noticed is it puts banjo over the top of the entire song, even when I specify directives in the section brackets to only introduce it during the [chorus]

[green pike]

I also applied! Sadly I don't know if I'll be at the par... honestly I've been a little problematic with people on general suno, yestarday I almost got t/o for like one week because a wrong altercade with a guy there, hope mods and team of suno could give me a chance...

anyways, I think you should be there too. You are always helping on #📚┃suno-school and you even helped me when I was in need ^^

[hardy gust]

Hey there, someone can explain me if there is an other way than segmentation for generating voice with suno/bark for long text ? Maybe @untold briar 😅 ? Trying to find the best way to make my chatbot talking , actually bark is what i'm expecting for a voice but i'm facing the long text problem ...

[green pike]

banjo on phonk? that's new.... with V.5 it should be possible, it would be easier if you try it with Suno Studio, idk if with normal editor you could do that

[torpid junco]

haven't tried solving it with Suno Studio yet, I'll look into that. Also, it's for comedic purposes 😆

[green pike]

lol share me the result I'm curious about that result

[placid spindle]

HI EVERYONE!! Sorry for asking again, but does anyone know if SUNO is working on cleaning up the noisey (ringing like noise) that show up in some mixes that have a groove to them?????? 🙁
I might have posted this question in the wrong channel

[plain pumice]

You're welcome 🙂 Hope it works.

[olive radish]

I have a constant crashing error when it comes to make a cover song from audio

[placid spindle]

is there a work around the noise?

[polar spade]

If you're talking about v5. 0:

If it's some but not all, then the remaster is there to help. You have four options to choose from.

Otherwise, you can report as a bug, and as more come in with the same issue, they can tweak more. 5.0 is still beta, so they're happy to review all concerns. Suno.com/feedback will get you to a form to fill out. Mention the "groove" prompt.

[placid spindle]

AH!!! YES The Remaster. I have used that and it does work. I leave it on normal, but it still changes the song - slightly. Any suggetions on the correct setting for Remaster to where it cleans up the mix and keeps most of the song in intact??

[polar spade]

Remaster in theory should not change time, which it doesn't and should not change the the song, but as you say, yes, slightly.That arpeggio you didn't hear before may have barely been in the mix, now remaster pops it up more... Some hiss or distortion get regenerated into a crisp synth or saxophone. Who knows.

Typically, remaster 4.5 should benefit 4.5 songs. 5.0 benefit 5.0 songs but could be wonky or great with 4.5. No real help here but to try them with the specific genre you're using to see the affect for that one. Your mileage my change with each output or changed genre.

[placid spindle]

Got it and YES I have seen where it turns noise into an instrument !! LOL

I was just wondering what the settings subtle, normal, high did

Quick question on [ ] If I want to tell Suno this is the Chorus I would but [chorus] then the notes of what I want?

and thanks for your help!!

[polar spade]

Assumming those changes are spread evenly across the song, I would assume each is the frequency in which it makes a variation.

If you're song is just four chords, then every few beats measures it my now vary a Major-7th to a Maj 7th add 9th/F bass or other variation across the song. The vocals may include more harmonizing, vamping, melisma or a note change here or there. More variance, the bridge gets sung in a minor key. In essence it's like keeping the style slider about 95% but slide the weirdness to 55%, 60%, and 65%. (... or 35%,45% and 60%)

That can also include if it holds back on lots of the backing instruments or brings many of the riffs and fills forward.

Not enough to change the song, but enough to enhance the vibe. In my opinion.

[hot axle]

If I wanted a song from sono to use as background music for my youtube channel how would you download it from sono so its in the correct format for youtube

[placid spindle]

Yeah - I wonder how often and WHEN to use (increase) the weirdness and style influence sliders. I tend to leave them at 50%

[hexed iris]

Hey guys!
What is the way to add a soft vocal pad (like an atmospheric vocal layer) along with the main vocals?

[raw oasis]

Hi Kaye, I would download .wav version, (i will usually master engineer), then I add and edit and deliver a video via DaVinci Resolve.
That should be the highest quality audio route for you though.

[raw oasis]

If your not Pro, it might be limited to MP3

[quiet solar]

Suno can export as either .mp3 or .wav. WAV is the larger, higher quality version (although I'll be honest, I can't hear the difference or see a difference at waveform level between the two...).

It may depend what software you're using to produce the video for YouTube, but pretty much all standard software will be able to handle both those audio formats as inputs.

[modern vine]

Using V5. All of my songs have have an intro of vocal"Oooo ahhhaahhh" even when I specifically instruct an instrumental intro. like every single track.
Is there a way around this? I'm getting sick of hearing it in every generation.

[placid spindle]

Has anyone used EDIT?

[raw oasis]

Yes. I may have answer this the other day. One idea anyway.
You could create instrumental tracks and then later cover them with lyrics. My thinking is you may already have some guitar, piano, etc. melodies in place from the instrumental that it would be less inclined to need to introduce those vocal runs. That is what I do now anyway.

[neat nexus]

Developers have suggested the following:

Try making a song version you like then using a Persona on the output or using Inspo, these can help you remove the vocalizations.

They also suggested that shorter tags can help eliminate the vocalizations.

Finally it was suggested to add to the exclude styles field this text:

"humming intro, hum, vocalise, humming"

Under Advanced Options.

I hope this helps.

[raw oasis]

I try on occasion 😄.. I get discouraged in Edit/Studio when it introduces completely different sounding options.

[modern vine]

Thanks I'll try this

[placid spindle]

AH!! I am afraid to use it for that reason and or does it work well when I want to just edit out a part of a track

[raw oasis]

This will be a 2nd song I am starting where I cut out part of an instrumental track I like (piano riff in this case) and start a new song with it. So I am liking that angle so far.

[placid spindle]

SO it does work well when you remove a part of the song

?

[raw oasis]

Cropping out a piece and "Covering" is kinda like an influence or uploading your own voice, Suno will still do its own thing and not just that exact piece.
I Will try an extend of my 7 second clip/crop.

[placid spindle]

ok

[placid aspen]

Hi. Has anyone else suffered the issue with Suno producing shorter covers of songs? I'm working with a 7:18 long composition, and all its spitting out are 5min or so variations. Also, the covers all lack a few seconds of the intro of the base track. Anyone know how to extend at the beginning of a track, eg add a few seconds to the very start of the cover before it begins?

[quick sonnet]

Yest it was limited to mp3 for free users, but now not even that avaible.

[quick sonnet]

Extend is not the right feature for it, but if you are a Pro user posibly the editor can help.

[potent inlet]

I'm hoping that I am posting in the right channel ... I have tried to take a look at the knowledgebase and some of the channels in Discord and have not found answers ... so maybe someone here can point me in the right direction:

I just became a 'Pro' member because I wanted to 'fine-tune' the music that I am making. I was hoping that I could provide some specificity to Suno ... such as "For these two lines of lyrics, the voice should be female and melodic", followed by "And these two lines of lyrics, the voice is a deep male voice with a marching cadence" ... etc.

In reality, I have found very little creative control at all ... if any. Even if I put in the style prompt something like "Meditative, ambient, drums only, no voices" ... I will get a chanting voice ... yet I specifically stated no voices.

How can I actually provide any control over what is generated and how much control do we have?

One actual example: I have a song that is a heavy metal song with a march cadence and a male bass voice ... but I want the final verse to simply be a female kulning ... and that seems to make Suno's head explode.

[placid aspen]

I am a Pro user, but find the Editor a bit too 'beta' for my use at the moment. I've been initially creating in Suno, then stemming it out and across into Audacity where I replace the vocal track, re-render it, and then import back into Suno. The new vocal track sounds too 'dry' to be of much use, so the idea (hopefully) is to run it through the cover feature and then, hey presto, everything is good. Problem is that Suno seems to have other ideas.

[unborn star]

It's still a blackbox and you need to learn which prompts are working and how you put them together, to make them work.

It's not that you get exactly what you wanted. That's never possible, if you want something reach like this, then you need to learn it.

[quick sonnet]

Since I am a free user have no experience with it, that advance sliders could be posibly helpful to get the cover, you want.

[unborn star]

You also need to prompt correctly. If you don't know which tags working and you just do it, you won't get fitting results. You also need to think about Genre you are using, if they oppose each other, it doesn't really work. There is a lot to learn.

Each genre does react different to specific prompts as well.

[potent inlet]

Yes, I am looking for information on how to prompt Suno correctly. I would like to learn if it can receive prompts directly in the lyrics area, for example (maybe in parens or brackets?) ... or how to correctly prompt it. Is there documentation?

I'm well versed in prompting AIs ... but I need documentation for what works with Suno

[unborn star]

There are some ye

There is no full documentation

Are you having ChatGPT Pro Version?

[potent inlet]

I have the pro version, yes.

Can you point me toward documentation that might exist?

I see only very, very, very basic "here is where you enter prompts" type documentation ... rather than "Prompts can do x, y, z".

For example, many image-generation AIs will apply extra emphasis to a phrase placed inside parentheses ... does Suno? Tips and documentation like these things.

[final hazel]

Just start here https://discord.com/channels/1069381916492562582/1294418290348921007

[potent inlet]

THat's the first place I went ... unless I missed something ... it was beyond basic and gave no detail or real info

[final hazel]

You can also check out our community work flows: https://discord.com/channels/1069381916492562582/1252267444773326848

Worflows made by users..

[potent inlet]

Thank you -- I will try to look at those 🙂

[fluid vault]

Hello everyone!
Question!
Where did the video regeneration go?

I downloaded the video before editing the lyrics.. and now I need to regenerate the video to show the correct lyrics.. 🙁 but its gone

Every time I download it now it just shows the first unedited lyrics. 🙁

[quick sonnet]

Are you having some kind of pro version of Audacity? Have never seen that features in the normal version of Audacity.

[visual plank]

I think it now checks to see if the style or lyrics or song details changed and it regenerates before doing the download. If there were no changes, you just get the download of the last video version.

[fluid vault]

Yeah, I just found out I have to remaster it, then make sure the lyrics are edited in order to download it correctly.

Tyvm

[velvet hornet]

Anyone else having trouble with extends now? It refuses to follow the lyrics and sometimes just play a instrumental or jumps to a later part of the song.

[quiet solar]

So firstly, style prompts in the general style box should apply to the whole song, or to very distinct sections of the song that Suno can easily identify (e.g. Verse, Chorus, Bridge). Trying to steer it to a specific couple of lines within a verse is better handled by including a style prompt in square brackets in the lyrics box itself.

Style prompts in general, but particularly in the lyrics box, are best thought of as suggestions rather than instructions. Suno may factor them in when it generates the song, and v5 is better at this than previous models, but it's not a guarantee. It can be strongly influenced by the genre as a whole (e.g. it can be very difficult to successfully prompt for a chill woodwind interlude in the middle of a thrash metal song).

Secondly, avoid using "no (xyz)" like the plague. Suno struggles a bit with understanding negative prompting in the style and lyrics boxes, and can very easily misinterpret "no voices" as "voices", giving you the opposite of what you want. You can in extremis add a negative style prompt to the whole song, but again, it's a suggestion and not a command.

Finally, you're also engaging with a well known headache - trying to use separate male/female voices singing specific parts of a song. As a general rule, Suno isn't great at specifically doing exactly what you instruct every time you make an instruction, and it's particularly stubborn when it comes to switching vocalists or genres mid-song, let alone mid-verse.

It can be done, and there's a lot of best practice lurking in here about how to do it, but expect up front that it's going to take a lot of patience and multiple attempts.

[plain pumice]

@potent inlet I agree, the documentation is quite basic, personally I think it should be a lot more detailed and thorough, exactly like any technical documentation would and should be.

[quiet solar]

Some general guidance for the style box:

• Suno breaks your free text down into key words and then looks them up in it's library of pre-existing songs to try to match what you've asked it to do, and allow that to influence it's output. So the more you stick to terms associated with music specifically, the more effective your prompting will tend to be

• It's not as developed a conversational model as LLMs like ChatGPT, so think of it like a stubborn session musician with a reading difficulty. Clarity is more effective than precision - be as clear as you can about the general things you want it to do with specific genres, instruments, vocals and so forth. Leave specific line-by-line instructions for the lyrics box.

• Prompts at the start of the box tend to be assigned more weight, and are more likely to take effect, than prompts nearer the end of the box.

Some general guidance for the lyrics box:

• Suno knows to read anything in [ ] as a style prompt. So, for example, it's comfortable with [Chorus, soft and dreamy]

• It's also been trained on lyric notation, so will fairly reliably take anything in ( ) as being performed by a backing singer(s). You can also use capitals for emphasis, or " " to indicate that part of the line is performed with a different tone - or at least that there's something different about it. You can also give it a steer towards the cadence of the song by using punctuation.

• If you want to call on specific instruments or singers in style prompts in the lyrics box, it helps A LOT if you've specified them in the style box to begin with.

[polar spade]

There a bit of concern that older how-to's are now 50% obsolete or inherently wrong at the start. With each version, what works well with one model may not mesh well with the next.

The best route would be someone doing videos of some how-tos, how what I do may not help your genre, and also show the wrong way something can be done to have a grasp of what to expect or recognize.

Just happy that were all in the same boat so we see the advantage of helping each other.

[potent inlet]

Thank you very, very, very much ... I've copied all of your advice and will keep it. This is the exact kind of knowledge I needed. I can work with the AI taking things as suggestions (they all do ... they work on probabilities) ... but this knowledge of being able to put styles into the lyrics box with [ ] or using ( ) for background vocals/etc ... is excellent.

It's too bad that it's not happy to flip genres mid-song, as that can often make for very interesting musical compositions ... but I get it.

Thank you again. You're a lifesaver.

[upper dagger]

guys how to beat i'm at canto II

[quiet solar]

Happy to help.

Genre switches and complex duets can be done, but it's pushing the limits of what the model is capable of. Good prompts help a lot, but in practice a lot of suggestions are part prompting, part workaround - and even with those, it can still require a lot of patience.

It's something I'd definitely suggest newcomers to Suno steer clear of. There's already a lot to learn through trial and error, like getting a feel for what might trip Suno up and where it may need steering through certain sections, and how to fix things that you're not happy with.

As Dasvon said, we're all in the same boat of learning as we go. And inevitably re-learning when the models change.

[eternal crest]

Hi everyone,
I have a paid subscription, but for some reason, I can’t create any songs — nothing seems to work.
Is anyone else experiencing this issue, or does anyone know how to fix it?

Thanks in advance for the help!

[polar spade]

Commonly new subscription with new payment method has shown to take a few days to review the payment for any suspicious activity as well as associate a "payment authorization " with that payment so the renewal payments are not delayed.

Most first time subscribers see their subscription activated between 3 and 5 days from payment date.

[sour stream]

anyone else have "Lyrics Mode" now under advanced?

[frozen dagger]

it has been added. the "sing without lyrics" looks interesting.

[sour stream]

Id rather have "max mode" lol

[radiant flare]

need some help figure out how to get suno to cough and wheeze and catch breath at certain moments. I tried back a few years ago the song is 500 steps,

[hexed iris]

Hey everyone 👋
I'm using the cover for one of the composition and have extended my lyrics beyond the cover version. Suno keeps ending the song before the full lyrics are done.
Any idea how to make it generate the complete extended version?

[paper dawn]

Check for verse 1 verse 2 . I keep them just verse . If you use ChatGPT to build your prompts that can cause a lot of that as well keep your structure tags just basic that seems to be the only time it happens to me but every once in a while I do get a random like 30 second song out of a full song it just happens

[hexed iris]

Simple structure - Just Verse. Intro/Outro in detail as it is required for an epic composition.
Does Audio have same or flexible duration?

[paper dawn]

I honestly don’t do much extra in my lyrics box I mean maybe some extra words like ad Libs or in parentheses but I don’t add any kind of structure genre and if I do it would be in studio,
For example if I want ad Libs in a song I’m gonna make a backing vocals only track in studio rather than try to get it to do it in the actual song if that’s what you were asking

[sand scroll]

Hello, jus wondering if anyone could help me. I subscribed to Suno studio yesterday and for some reason it stopped allowing me to upload songs, any ideas why? Thanks in advance

[paper dawn]

But do remember here I’ve never been featured I’ve never been on the wall
I’ve never been popular and there’s gotta be a reason for that so take whatever I say with a grain of salt and do what you feel is best

[paper dawn]

Everything you’re uploading is yours? I’ve never heard of that I would send a message to support

[sand scroll]

Yeah all mine, full songs I have recorded myself and acoustic demos

[paper dawn]

Yeah I’m not sure why it would do that it could be sometimes it logged you out so I will try logging out logging back in that could help

[glad ibex]

I'm writing songs in Hebrew and trying to get the pronounciation right. I don't know how to fix individual words so I just keep editing the lyrics, trying in Hebrew and transliterated in English. But I keep starting the song over and I want to know how to fix one or two words. I have pro and when I go to the editor and try to change it I lose my original melody. Please help.

[sand scroll]

Tried logging out, just says something went wrong,could not create file

I contacted support, do they normally take long to get back to you ?

[paper dawn]

i’ve only contacted him once and I got responded to right away

So if it’s saying we could not create file it could be too long of a prompt over 5000 words

[sand scroll]

Do I just send a email to support them ?

[paper dawn]

Yeah whatever email they got in the support like mine was an answered rather quickly

[sand scroll]

Ok, thanks for your help mate

[paper dawn]

So if it’s Hebrew to English the translation could come out to something that is not allowed it could be a wrong translation do your translation in a different AI like ChatGPT and make sure it’s under 5000 words trying to think of anything else I’ve seen that for

Oh you’re welcome anytime I can help out I got you

[sand scroll]

I also get client error 423 if that’s any help

[weak kite]

Just wanted to check in before upgrading to get access to studio:
I have created a few songs from my own lyrics. Suno has made an amazing job with music, melody and vocals. I now have recorded my own voice for each of the songs. I want to access the stems and replace the vocals. I can not do that in the editor since it doesnt seperate the stems. If I go to get stems I only have the option to download.
I am not a music producer and would like not having to get a seperate DAWS for this just mering stems
Would studio be able to open an existing suno song, accept a file for the vocals (I do not want to record straight into studio) that I can replace the vocals with?

[paper dawn]

Yes

As long as you got skill on matching the tempo to the track you’re bringing in yes it’s perfect for your own vocals if they’re at the same tempo it should be easy

[weak kite]

Amazing! thanks @paper dawn ❤️

So studio can accept files as well, not just live recordings?

[paper dawn]

Yep

Its a daw , I don’t use mine as much maybe for plugins

And beats

But I’ve been having fun making beats here I mean I still bring in samples but I don’t have to

[paper dawn]

Yw anytime

And if it comes out sounding weird just change it to original tempo and it’ll sound normal instead of it being on beat

[red oar]

Would studio be able to open an existing suno song, accept a file for the vocals

Yes, you create a new track, and can generate or upload or cut'n'paste stuff.

(I do not want to record straight into studio) that I can replace the vocals with?

It can do that too, if you would want.

have recorded my own voice for each of the songs.

How did you do that? You can mix with anything you have, like Audacity, and avoid Studio if that's the only reason.

I want to access the stems and replace the vocals.
You may want to download the stems and have a listen. You should know that Suno does not work like a real band, so it takes the full song, then tries to separate the instruments. I find that the results are far from perfect most of the time.

I can not do that in the editor since it doesnt seperate the stems.
Yes, Editor is only one track.

[paper dawn]

Yeah I tried to create mine in studio that way it is a multi track because I have noticed that sounds bleed

But if you just want your vocals don’t separate all the sounds just do instrumental and vocals and they come out pretty clean

[weak kite]

Actually I have trained kits.ai with my voice, so I have taken the suno vocal file and uploaded it there ,so it is tempo matched already

[paper dawn]

But I do agree with jay

OK yeah then it should work actually very well cause I do that

[neat nexus]

Nice...

[weak kite]

Trying to avoid yet another software that I newed to pay for. And the DAWS I have tried, the sound is weird, not as nice as it plays in suno

[paper dawn]

Truth

[weak kite]

I am aware, the good thing is that I only need to replace the vocals, so I download all instrumentals as one file, I am not planning to mix the instrumentals

[neat nexus]

Oh in case you didn't know (I forgot myself... 😅 ) we do have a section for Studio related stuff...

#1420792719969878028 just wanted to mention it in case you didn't know (or fogot like I did...)

[weak kite]

Thank you @neat nexus ❤️

[paper dawn]

I did not know thank you hep

[weak kite]

I am not allowed to post in that channel for some reason

[neat nexus]

What? Weird. Let me ask about it. (Did I jump the gun on something....)

[paper dawn]

I think it’s a like read only thread

Hey just willing to help anybody that’s going more than most people I appreciate you

[neat nexus]

I'd actually love to know more about this method of yours and am curious if once you "bake" in the vocal if you can cover the output and fix most of the stem related issues?

[red oar]

There's a "Studo chat" subthread in there, that seems to be open. And "Studio bugs" and "workflows" that may be useful. And by the way, welcome to Discord. 🇸🇪

[paper dawn]

Man I was going to bed a hour ago lol . Goodnight and keep shining fam

[neat nexus]

Are you guys able to post in the individual threads or is it just the creation of new threads that is read-only for you guys?

[weak kite]

Of course! I would love to provide more info and play by play ones I have gone through the whole process

[weak kite]

Now I can see the threads, previously I could only see one feedback thread. I can post in them yes

[neat nexus]

Excellent.

playing-guitar-excellent-fan-iron-maiden-bill-and-ted-gif

[weak kite]

Thankls for all the help guys! Great community!

[neat nexus]

It's what we're here for. Thanks for being a part of it.

[hexed iris]

Anyone 🤔💭

[weak kite]

Have you tried extend instead of cover?

Crang up the Audio influence to max to make it sounds like the original

0% weirdness

[neat nexus]

Maybe not zero. More like 10% or 15%?

Extreme things happen at the extreme edges of the sliders.

Can still be very useful but bear in mind.

A great way to think about the slider settings is this:

[wet maple]

Woah, there had to been a very recent model update, these new gens feel different...

cleaner, crisper, although not wild about how it seems to have pushed off further from style influence having a bigger impact. is it safe to say we're out of beta on v5 now

[bronze harbor]

What would negative weirdness mean?

[neat nexus]

In theory you'd get exactly what you're covering.

[last bay]

Max mode spotted

[neat nexus]

waiting-simpsons-homer-whistle-gif

[empty edge]

Hey i need a mod

[neat nexus]

What's going on?

checks history

Have you sent an email to [e-mail] regarding this issue?

[empty edge]

Thank you i did

[magic mist]

Has anyone else experienced issues the last few days with generations fritzing out? I've had some song generations that have skipped like an old record, started over in the middle of songs or sections of a song, audio dropping out, dropping out and then starting over, etc. I've been reporting them as bugs but I'm blowing through credits with this. I'm on the Pro plan so I have plenty to give but I hate to waste my credits and time.

[modern vine]

Yes I've been complaining about extends not functioning for a while, actually I've seen a lot of users talking about this. Many credits being used up with failed generations
You can try to extend from your main track and then use one of the extensions like a kind of music break then extend from that ("part 3") and see if the third extend will actually sing the lyrics? Occasionally this works. But you might end up with a long song

[quiet solar]

Your best bet is to put something which is phonetically correct for the word, and experiment until Suno gets it right. It's quite tricky to fix a single word in the middle of an otherwise fine verse/chorus purely within Suno.

[quiet solar]

Yes, Studio will be able to handle that - but whether you record yourself singing separately or live in the programme, it will be exactly the same process as using a free external audio editor like Audacity.

[short elm]

So...my covers of uploaded songs that are mine are being used to create new generations, but it's stopping about half way through the song at about 1:45. I am using the 4.5all but it's not creating the whole thing.

[neat nexus]

Why not use v5?

[short elm]

Because I don't have a subscription.

[neat nexus]

I see.

[short elm]

So what's going on? Am I tripping or is this a change to the way that 4.5all operates?

[neat nexus]

Not that I am aware of. Which browser are you using here?

[short elm]

I'm using the Mobile app.

[short elm]

The Mobile app is the only way that I can use the sliders with 4.5all with no exceptions.

[neat nexus]

Can you test it out on a desktop browser to see if it is restricted to the app only? Which mobile app are you using? The iOS or the Android?

[short elm]

Android. And I've tried with it. And it does the same thing but worse. It won't allow me to use the sliders and the exceptions meaning that I cannot make it sound similar to the uploaded sample.

[neat nexus]

On the desktop or just using the desktop view of the mobile browser?

[short elm]

Both.

[neat nexus]

Huh.

Yeah, let me mention this to some of the team and see if there's something going on.

[short elm]

Whenever I attempt to change the sliders on desktop it gives me a pop-up to upgrade to the new version of suno with both versions. The desktop and the mobile desktop view.

[neat nexus]

...waaaaaaait a tick...

What are your slider settings?

[short elm]

My sliders settings are Zero weirdness, 100 style influence, 100 audio influence.

On mobile. Lol

[neat nexus]

Yeah....try adjusting those away from the extremes and you should see the length of your outputs increase.

[short elm]

Ooookaaayy.

[neat nexus]

Try to think of it as each time you adjust the audio, you're taking memory away from the model.

Use a cover? That makes it reduce the max length of the output by a few minutes.

Add a Persona? Subtract some more time.

The model needs to reserve the time to be able to "hear" and think about the audio it's working with. So it reserves time in the output.

Now you can either adjust how much memory you're using or you can try taking a good beginning output, and finding a good resume point from before the end of the track and then using the Extend feature and changing the time to continue from. That is one way to work around the issue.

[short elm]

So I'm just not gonna get a full song out of suno using covers. Cool.

[neat nexus]

That's not what I said is it?

I cover stuff all the time.

If this is a slider issue eating at the song length (on top of the cover itself) then that's one thing, but I have already mentioned the issue in the mod's section and I am hoping to get some responses on the issue you're facing.

[mint granite]

anyone got any idead on how to prompt away the high pitch male vocals in pop?

[neat nexus]

I do all my generating at a laptop so I can't really offer much feedback specific to the apps. But this is as I understand it in general.

[neat nexus]

Depending on the genre you're using, try baritone, try describing the vocals, etc.

Maybe add an accent?

[short elm]

You said using a cover reduces the length of the overall song a couple of mins. Ergo, I won't be able to create music using covers due to my type of music being EDM. Edm songs are 3 to 4 min regardless. And I'm only getting half the song. Imagine a song stops playing just as the beat drops. That what ALL of my generations are doing.

[neat nexus]

Are you unable to extend from the end?

[short elm]

No. It just creates an ending like part of a song and I'm not in the mood for deleting my lyrics until it gets to the part that wasn't sung.

[mint granite]

i tried, but I always end up with something that tries to sound as flamboyant as Mika - Grace Kelly. on top of that, I mix genres like neosoul funk pop r&b. the output is so emasculated it's no longer possible to even find it ridiculous -- it's just sad

[neat nexus]

¯_(ツ)_/¯

I've told you what works.

[hard beacon]

😄

[neat nexus]

I can show that it is possible to make EDM\Dance music:

https://suno.com/s/VyE8eueNw7luTvfI

Of course now you will say it's because I'm using a paying account and on v5...

[neat nexus]

Let me look at my outputs for something that might suggest a way out...

[mint granite]

much obliged. I am no stranger to prompt-fu (or kung-prompt if you will), so any ideas are of value

[neat nexus]

Can you take any vocal inspiration from this track? (Ignore the opening vocalizations.)

https://suno.com/s/lYpNGeNMaSWbS6tE

[short elm]

This is what my songs were using 4.5 before 4.5all. without uploading. I know what WAS possible 😂

https://suno.com/s/3HFY01pLmJUBV15T

[neat nexus]

deep baritone

Deep, raspy baritone vocals

That sort of thing is my go to with this sort of thing.

[mint granite]

that works for me. and my compliments on a rather unlikely genre fusion; that was refreshing to say the least

[hard beacon]

and vocalstyles are often related to music style... "pop punk vocals" are different to "emotional emo vocals" or "Fidget House Style Vocals"

[mint granite]

awesome. thnks for the insights @neat nexus and @hard beacon

[hard beacon]

you are welcome

[neat nexus]

Thanks. And if you're willing to work within generations, a thing to try is playing with the initial version using just divorcecore as the style and adding the [metatags] in your lyrics prompt to make the music happen. Then take the output you like the most, cover that, then add the style you're looking to get and adjust sliders carefully to try to retain that vocal.

[mint granite]

nice. I'll try that.

[radiant flare]

need some help figure out how to get suno to cough and wheeze and catch breath at certain moments. I tried back a few years ago the song is 500 steps,

[hard beacon]

woa... that's a very specific wish....

maybe with specific vocal meta tags in the lyric prompt... but.. phew... trial and error 😄

[radiant flare]

Yeah, it is essentially a song of a guy with breathing problems. I can get some small coughs every now and then but was hoping for some more played up coughs and stuff. Just not sure what tags would work.

[polar spade]

You might mention your sub tier. Premium uses Studio which makes it much simpler.

[radiant flare]

I'm only pro plan

[hard beacon]

if i am right Studio is on Premier Plan only

[polar spade]

Well. [gasp] seems to work OK for a gasp. But as many tags, Suno may make a sound but it puts them where it wants.

But gasp, being kind of a vocal is pretty good to work as a vocal descriptor. Such as

[Male vocal] [surprise] (gasp)
"And that's what you saw? "

I'll assume you've tried already, though?
..

[radiant flare]

A few times i get short coughs and gasps but nothing for rough breathing or weezing breaths.

[polar spade]

Yes. Premier only. So now it makes it harder, or near unlikely.

[hard beacon]

😉

[polar spade]

There are likely gasps within some songs, especially musicals, in audios Suno has trained on, so it may be why gasps work where wheezing may not.

[quick sonnet]

100 audio influence should be very similar to the original, but is it so?

[polar spade]

I say that as I suggest prompting in genres or styles that might have wheezing. musical, Halloween, Industrial punk, breathy, raspy, goth rock, vampire rock, occult rock

Then tag for wheezing and coughing.

[short elm]

Yes. Everything works, it's just the length is too short

[quick sonnet]

Wondering why suno has no length slider, like other AIs do?

[radiant flare]

Alright so far i haven't gotten better them my current iteration but i'll work at it.

[polar spade]

Oh yea. Sure. I'm not guaranteeing that it will even be close. Sound FX are just not an easy task for SUNO in its current configuration. I'm Just suggesting anything that might even go e an FX noise at all.

With Suno Studio, we would just record using coughing or wheezing, drag the track in and put the fxs where we need them.

[ashen carbon]

Is it possible to make a persona with our own voice?

[hard beacon]

negativ

[short salmon]

I have been trying to contact Suno regarding a subscription I have. I accidentally went for yearly and not monthly but I haven’t heard anything back. I send an email back around 10/24 and still haven’t heard anything. Any ideas on how to get resolution?

[polar spade]

There's been an influx of emails due to new users and Fall promotional offers. Billing and Support response to emails in queue appear to be about 2 to 3 weeks.

Click to read regarding refunds - - > https://help.suno.com/en/articles/2550209

Just as an instrumental break, such as after the chorus? Yes.

Into the verse? You mean playing as a background instrument to the vocals? Seems doing that in an extension might bring out instrument in just the extension, then any extension after that it may add the same here and there, is typically Suno style.

Nothing I've tried in awhile, though

[neat nexus]

@narrow cloak What is the name of your Persona?

[narrow cloak]

Aman

[neat nexus]

Is the Persona from an upload?

[polar spade]

You said "into the verse" so it sounded as if you wanted it weaved into the verse at the same time. Maybe I read it wrong.

Before a verse, bridge, chorus, I wasn't meaning to be specific, just saying that [interlude] works for most people to put an instrumental between segments.

[narrow cloak]

From the song i created

[neat nexus]

Was the song from an upload? What is the name of the song?

Was the song a cover of another person's song? A Remix?

Could you post a link to the song?

[polar spade]

[interlude] seems to give a 20 second piece. You can put three, one under the other for 1 minute. Try it out.

[narrow cloak]

Ok

[narrow cloak]

https://suno.com/s/ZjP1xLJXpweRFFg1

[polar spade]

Ahhhh... That. Many have posted success by using things like interlude or instrumental, along with other characters, line spaces, dashes, and so on as if a full song, just removed of any lyrics so the page looks spaced and extended down the page, enough to keep Suno making music.

At least that's the idea.

[neat nexus]

This is you?
https://www.youtube.com/watch?v=IKQKf9WyH3E

[narrow cloak]

No

[neat nexus]

So you are trying to make a persona based on a song you don't own? Is that correct?

[narrow cloak]

Bro i made my own song using suno what are you talking about?

[neat nexus]

Did you write the lyrics?

[narrow cloak]

yeah

[polar spade]

I can not narrow it down any closer than that, really. Suno doesn't follow every trick perfectly. So, according to others, this is a handy way to extending the instrumental to least a chain of music go for a little longer.

You know SUNO, if it gets bored, at any time it may grab an earlier lyric and toss a line back into the end of the song.

[short elm]

No. I couldn't care less about that. What I wanna know is why don't they tell people that uploaded covers aren't as long anymore?.

[narrow cloak]

Are you really gonna solve my problem?

[neat nexus]

I'm trying...

[narrow cloak]

I have to go to sleep

It's late night in india

[quick sonnet]

With such length slider you posibly could get the length you want, not exactly, but very close to it. What you mean with uploaded covers aren't as long anymore?

[polar spade]

I've been looking for someone's example. Can't find one. Just saw on a couple of weeks ago I think.

I'll write a song real quick. See what I can do and what it results in.

[neat nexus]

Sorry. One last question if you're still awake: What plan are you on?

[half dragon]

Something like with empty lines after the tag could help dragging the instrumental break, or solo a bit longer

[Instrumental Interlude] or [Melodic Instrumental]
...
...
(some users even use ascci arts here)

I sometimes use multiple [Instrumental] tags too - not guarantee like everything else but it does work more times than not.

Sometimes it does through out the verse or chorus, other times it bleeds in the first part of the verse then dies down or stops altogether.

[ashen carbon]

any plans to make that possible?

[polar spade]

Not the longest song, but it sings the verses and chorus then the rest after the 2:00 mark is instrumental. https://suno.com/s/k1aVHXBqGQjCgrUW

You referenced instrumental only (to the end of the song?) after the lyrics are sung?

Ahh... Then I see I can't follow what is asked. It's better then to see if someone else has a better grasp. I'll read along and maybe it will teach me as well.

Thanks for letting me try.

I got a nice Native American song out of it.

[chilly epoch]

How do I use a beat and groove of a song and change the lyrics? I tried using cover but it didnt work

[polar spade]

Ahh.. The literal melody. OK. Are you a paid sub? I forgot.

[chilly epoch]

I am.

I have pro not premier

I do need to upgrade to premier to get the studio

[polar spade]

It may be a few steps.

°. Write the song.
°. Split STEMS (just inst /vocals)
°. Cover the vocals as an instrumental so you have an instrumental with melody lines.
° extend that instrumental with the sams lyrics to create a song at the end of the melody instrumental. add extension until you have completed song attached to the melodic instrumental.
°. Bring song into Edit UI and cut and move song segments in front of the pairing melodies so that a verse comes before the melodic instrumental.
° blend, remaster or Cover as needed.

Well. That's why they gave us Studio. 😁
But if you want a melodic interlude that emulated the same melody as was in the verse before it, we can only rely on Suno to do that, which it probably wont, unless we force it somehow.

Again. Tossed a suggestion in the ring. Never know until you try.

[chilly epoch]

I dont have the studio but you dive me an idea

[short salmon]

Awww ok that helps thank you

[chilly epoch]

Honestly the song was reused under 4.0v. And I had detailed out the sound and after 5 rounds it was perfect just needed to change vocals and j decided to write my own lyrics than suno's lyrics

Once I stem and I just download as mp3 and bring it back in as audio?

[half dragon]

You can - but Cover should work too, as long as your new lyric flows similarly as the existing lyric and stays within the melody.
Reupload the instrumental stem would need to go through Cover to add new lyric but it should work as well.
Have the audio refernce slider at higher value to maintain the melody.

[chilly epoch]

Ok. Got it. What about the style influence and weirdness? Which levels should that be?

[polar spade]

The suggestion feels incomplete?

Noted. It may take doing a sample video perhaps?

You might be one who likes things visual, as do many, including myself. . Hope we can have some instructional vids on some of these ideas, in the future.

[weak kite]

Not sure if this is the right channel. I have rendered two versions of the same lyrics. I love the chorus in one of them and the verse in the other. Is there a way for me to merge them through prompting?

[half dragon]

I usually keep the Weirdness level as low as before it turns red.
For style - I usually empty the style box in cover unless I want to change the style - so the style slider wouldn't have much affects if you want to keep the same style from the song you Cover.

[half dragon]

Not through prompting, not possible.

[polar spade]

3 strikes. I'm out. 🤕

Well we went over some neat tricks. Maybe someone else got some ideas reading them.

I shouldn't try again.🙏

Again, reach out to someone who might get the angle just right of what you're asking.

Hope something clicks. Thanks again for letting me try.

[frozen dagger]

If you search for "ChillPanic" on Youtube, can find Suno user,, he psoted new videos in last couple days that adress basiica and more that i see asked about often in here.

[tired forge]

Can we do music by decades? Like 80s hard rock

[polar spade]

It would be benifitial if one would post in the workflows as well, but that's totally up to the user.

[frozen dagger]

ya im not promoting person, i dont know them, but i watched and thought a lot of people seeking answers in here could peek at something new and updated like that to get a better feel of Suno and how some of the stuff works.

[polar spade]

Prompts like "80s hard rock" just as you have it, work fine. You might get a generic 80s sound until you add more prompts to pin down a sound you like. Listen to some of the songs in the 80s channel and see their prompts used. https://discord.com/channels/1069381916492562582/1262592742945456148

[soft atlas]

Anyone know what this new "lyrics mode" is that appeared today?

[half dragon]

that's where you want to tell Suno to use your lyric or let it generate one or none

[soft atlas]

So if I didn't have lyrics, I would just use auto and let it do it's thing? Having lyrics i should leave it on manual?

[tired forge]

I wish I had an answer for you...

[midnight sorrel]

Hi has anyone had any luck with generating an instrumental for a vocal?

[half dragon]

instrumental for a vocal? pls explain more of what you're trying to do

[midnight sorrel]

I am wondering how to add an instrumental to an original vocal.

I upload my vocal, select Cover, then select Add Instrumental, but then no matter the settings it always re-sing my vocals with an AI vocalist.

[half dragon]

oh - yeah - it's known that Suno will not clone your voice
It'd try to generate one as close as possible but would not be the same.

[midnight sorrel]

dang I saw it in a TikTok ad and thought maybe it could

[half dragon]

If you want your own voice singing on Suno instrumental track, it'd have to be done externaly

[midnight sorrel]

well I don't really need it to clone it

just build around it

but yeah I don't think it can

Thank you!

[reef sigil]

Hello, how can I contact technical support? I've been sending them emails, but they haven't responded. I have a donation account, but I can't log in. I haven't received a verification code on my phone.

[tired forge]

Well I had to go with a different decade because going with the 80s just cancelled part of my prompt completely.

[half dragon]

If you already contacted [e-mail], best way is to wait it out - they will respond.
Also it's not recomended to send more than one email as it' would not make it any faster.
Check #archived-ticket-help for follow up.

[quiet solar]

The downside to doing that is that it might divorce Suno's understanding of where the lyrics attached up to the music, although if the song is pretty straightforward it should be able to reconnect them back up again; basically, create a new song with lyrics in the same overall structure and add your uploaded instrumental only song as an audio prompt.

Normally I'd suggest an alternative of Extending the original song but rewriting the lyrics completely for the extension, then using Editor, Studio or an external editor like Audacity to fix it up. But if the original was in 4.0 then it's unlikely 4.5 will keep the music unchanged.

[quiet solar]

It can be done, but not quite in the way you're trying to do it - advance warning, it's not a one-step-and-done process and it can't all be done within Suno:

• I'm actually surprised Add Instrumental worked at all for what you're trying to do - a more typical route is to Create and make a song using your uploaded singing as the audio prompt, and fill in a style prompt for the instrumentals and genre you want. It should create a song with a Sunoified AI singer covering your vocals.

• Use Get Stems on the version you like the most, and split by instruments/vocals

• Download the instrumental track (note: stems are hidden by default in your Library, you may need to show them first)

• If the tempo and key match up, attach together the instrumental track with your vocal track in an external audio editor or DAW. If they don't, record yourself singing over the instrumentals again using your program of choice.

[faint geode]

Hello , I use Pro Version, im Asking about How to control key on prompt or lyric ? i mean i have Chord progression, such as C / Am / Dm / G (Loop Chord) , I want the lyrics + song to follow my chord progression. How should I write the prompt? Both the lyrics and styles. Because before this, I tried testing many times. It appears that the song I got is not the chord as I wrote in the previous prompt. Sometimes, there are dynamics mixed in, causing the key to be distorted. I would like to ask for help from those who know. Thank you very much.

[chilly epoch]

I've learned how make a song sound like my favorite producers jam and lewis

[polar spade]

Can you play the song to give the chord progression and key, or sing the song to record in audio to at least give the melody in the key? I mean do you play an instrument?

[faint geode]

Yes, I can use FL Studio to make my instrumental. Yes, I mean I need suno vocal singing to follow me instrumental. Can I do this?

[polar spade]

You can upload the instrumental, then use remix-edit/ add vocal to paste in your lyrics and have Suno sing your lyrics to accompany the instrumental.

[faint geode]

ok , i will try , Thankyou so much 😊😊

[polar spade]

You won't have any control of the melody. It will make it's best effort to sing to your lyrics with whatever melody works. So each generation could give a different melody.

[midnight sorrel]

Hey anyone know how to upgrade to Suno Studio? I keep getting this error "Your upgrade failed. Your saved payment method was not charged."

[quiet solar]

Further to Dasvon's point, there is an alternative way that lets you control the melody and that's to upload a file with both the instrumental and you singing to it.

Use Create and add the uploaded file as an audio input, fix any inevitable transcription errors in the lyrics from the upload and then use the style prompts to tell Suno what you want the end result to sound like. If you keep the audio influence slider high (80%-95%) and weirdness fairly low (10%-25%) then it should remain fundamentally faithful to your melody while fleshing out the instrumental for the genre you've specified and switching you out with an AI singer.

[narrow cloak]

Pro

[gentle glacier]

As of a few hours ago the recreate feature has disappeared.. has anyone else experienced this? And will this be resolved?

[half dragon]

Which feature?

[gentle glacier]

When importing an audio file in suno suno gives the option to recreate

That isn’t there anymore as of today

[half dragon]

I still see "Cover", "Extend" as options - are those not what you meant? I don't ever remember seeing Re-Create from upload.

[frigid basin]

So how does one get more credits within Suno? Upgrading to a membership?

[errant kiln]

No, Recreate was definitely there yesterday. I used it

[narrow cloak]

@neat nexus Error message is

An erro occurred while creating a presona.

[sleek stump]

Periodically you can review songs for a minimal amount of credits, but normally you would either upgrade to a subscription or wait for the daily free credit refresh.

[half dragon]

1. Upgrade
2. Purchase extra credits
3. Earn free credits by inviting friends, or listen amd rank songs

[gentle glacier]

Recreate disappeared

[sleek stump]

Can only purchase credits if yo uhave an active sub

[half dragon]

nah - Recreate only appears in button in maybe Editor or Studio... not on Upload screen

[gentle glacier]

Yes in studio, but it’s not there anymore more only replace is

[half dragon]

They probably changed it I guess.

[gentle glacier]

Why would they do that it was such an amazing feature

[half dragon]

Replace ~= Recreate ~= same concept

Just wording- feature still works the same way.

[untold bison]

I need a hand on a song im trying to get working right its nearly perfect its a duet its playing out as intended until Verse 3 then the male & female vocals are swapping parts is there anything I can do for that

[half dragon]

Maybe recreate that Verse 3 using Extend - keep generating it until it has the gender you want - then merge.

[untold bison]

Dose merging need a subscription

[half dragon]

I don't think so - I haven't been on a free membership for a while but I recall Extend (which also allows you to get whole song by merging) doesn't need sub.

[broken solar]

Hello! New here! I have a question. I’ve been trying to do some Uptempo on Suno but everytime i do that I just get upbeat music. When I type in Uptempo Hardcore, I just get upbeat rock music. I don’t really now what to type in.. can someone help me?

[astral hound]

So I generated a song on Suno before 4.5 was free. I got v5 demos. They are very good but I generated the song when I was incomplete with the lyrics. Is it possible I can extend the song without losing the style and melody?

I can only unlock those 1min v5 demos to full if I subscribe

But the thing is, I only made lyrics upto first chorus when I ran it on Suno

So if I ever unlock it, I know that Suno had prolly repeated first verse and first chorus for the 2nd half

Here's what I have in mind:
Replace the 2nd half with the lyrics and then extend it with bridge to last chorus lyrics

Am I making sense? Is it possible?

[viral crescent]

Does anyone failed to upgrade from pro to premier like me? 😭

[unkempt summit]

Is there anyone from whom SUNO stole money like from me? I paid for an annual subscription, and a month later SUNO canceled it for me and offered to buy it again. Support is not responding at all. The suggestion to wait for a response from support for several weeks is beyond good and evil. What to do?

[astral hound]

That is why I prefer doing monthly subscription so I won't experience things like this 😅

Even though annual offers tend to be cheaper

But then again, a month is personally enough for me to do more than what I needed to in any platform for I am not a heavy user

[broken solar]

Yes

[unkempt summit]

I really regretted that I bought an annual subscription. I just didn't expect that there could be such an outright scam!

[broken solar]

None of them is what I’m looking for, I’m sorry. If you know the big artists like Lil Texas or Deadly Gunz, you will now what I am looking for! Thank you tho

Oh I get it! Thats usually how I do that too! But it wont work anyways 🙁

[quiet solar]

No. You can extend from the original, and if necessary extend from partway the extension (...and so on) as needed to get things the way you want. Just click "Get Full Song" in the Library on the final Extend in the sequence and it'll stitch it all together.

[iron fractal]

im not able to create persona of any music created using suno

[near flume]

I'm on my second year of annual subscriptions, never had a problem. Make sure you're emailing "billing" and not "support' about your account.

[fair adder]

hello, could someone please give me some advice on how to prompt vocal chops in a song? glitch pop doesn't work and rather does some nice thing to the music but not the vocals, and i've never done this before so i'm not sure about how to write some.

[polar spade]

Merging means to take the extension made from the original and connect it to the end of the original, in this case. Connecting part two to part one.

[quick sonnet]

It's not always the end of the original, you can select the exact time, from where the extension should start, so the extended track will be added to exactly that time stamp.

[polar spade]

I my not know what the exact sound of chopped vocals, but here's a sample of a milder form, I think. My prompts might help you.

Oh... The prompts:
chopped glitch vocals, autotune, pitched-down moans, reverb tail, sidechain ducking, trap minimal, Gnash-sampled vibe

https://suno.com/s/ytqeELqexnpQk9Xu

[fair adder]

... i only understand chopped glitch vocals, but absolutely none of the others xD

[echo moat]

in suno how do you pr ovide instructions in the lyrics section without having the singer sing the instructions?

do i just put them in brackets as well?

[polar spade]

Lol. Try them all then. See if you like the results.

[polar spade]

Basically, yes. SQUARE Bracketed prompts [ ] would not be sung.

As we can now use phrases, that would include instructional prompt sentences and phrases.

Not parentheses.
( not these)

Parenthetical remarks mostly get sung as echos, or call and responses.

[late whale]

(Instruction, second instruction, etc) <-- Might not show up because discord like to make the what I'm trying to do show up a italics, beasically though it's the brackets with two astrericks before and after the brackets, no spaces. I have never had issue with using ( ) brackets vs Square... Place the instructions before the lyrics you want them to take effect in or at the theoretical section of the song you would like them to effect. I have never had an issue with using ( these brackets) as long as you put the double asterisk before and after them noting that they are instructions not to be read. I also suggest for more complicated multiple instructions breaking them up into several lines of instructions instead of grouping them all together seperated by commas in one long instruction.

[polar spade]

@late whale Lucidelusion means using asterisks and parentheses like this.

[late whale]

Thank you! lol yes, that's what i mean

[echo moat]

also are there any recommendations to help prevent bleed over when creating stems? is it better to get the whole song and demucs it myself? i have studio

[late whale]

My suggestion is to try and provide detailed instructions were the lyrics would go normally. Create blocks of multiple line instrumentation, melody, etc you want to have in the song, layer them like you would Lyrics and Chorus of the song, prior to each block add an instruction to increase the BPM (accelerating the tempo). This is assuming you are talking EDM which is what the Lil Texas artist I checked out does (festival edm).

[late whale]

My only suggestion is rending your initial track until you get one that's "Cleaner" sounding with good layering and not too muddled sounding so the extraction can tell the difference in the stems and get a cleaner pull. I use LANDR to master stuff sometimes and if there is not enough Dynamic Range in the track it will tell you it wont master correctly. Same thing I believe for STEM extraction, I could be wrong though. Also I know this sucks if you get the one render you are "THIS IS IT!" in which case I would suggest making a persona of that render and using it to try and get a cleaner my dynamic mix if possible.

[polar spade]

To work for every genre? Maybe not. Obviously create original songs with most of the instrumentation separated as much possible. As @late whale says, clean layering. Fewer distortions and synth instruments.

But depending on what we do with the stem, remastering the stem could... and I emphasize could... take away or convert the bleeds into another sound or instrument.

The problem there is if feeding back and blending in a DAW or Suno Studio, we might get some din from some weird overtones of these newly added frequencies. It's all about experimenting, so try Remaster (4.5 or 5.0) on a few.

[echo moat]

what does remaster do exactly

[polar spade]

Cleans, clarifies, brings instruments or instrumentation up front often to be crispier or can warm muddled tones as well as may correct notes and melody here and there. Takes noises, such as hiss, pops, distortion and tries to rid them or turn them into percussion or instruments or more accommodating sounds.

[echo moat]

ah ok cool, is one remaster usually good enough or do you find yourself doing multiple?

[polar spade]

I won't lie. Sometimes it sucks and can garble vocals and cut off bits as it tries to limit noises. Wanted FX like thunder, wind, and applause might get truncated and sound more poppy/percussive.

[echo moat]

do i have to select the sections or entire track when i remaster? or just hit remaster

[polar spade]

Meh. I'm still experimenting. With changes in models and 5.0 still in beta, it's hard to be consistent. Remaster in one, or more. You never know. Sometimes using 5.0 then 4.5 then 5. 0 again on one output. Don't skimp on credits. 😅

[late whale]

lol @polar spade funny I never tried that, I always assumed it was simply for re-doing your tracks when they released a new version of Suno lol 🙂

[polar spade]

Just choose remaster for the whole output. But noone says you can't crop a segment and remaster just a bit and toss into Studio if that's your thing.

[late whale]

I'm actually testing out the advice I gave to see if it works, problem now is though I've leaned into create a new EDM track and the testing is kind of just secondary lol

[polar spade]

The more we mess with stems though, the wonkier the wonky factor goes up until we just have to take offline and work off our home DAWs. So plan ahead. Start with a plan, THEN create the song. Not the other way around. DAWs and Suno are best for creating, not fixing.

Using the analogy from last year, once you chisel away the marble to get your statue, it's hard to put the marble back on. Plan ahead. Compose a song that works well with the tools. It takes less time and people will be more impressed.

[slim summit]

who decided that the final word in each verse should always take 3 to 5 business days to pronounce and how do i turn this off?

[echo moat]

right here i'm trying to insert a change, but it keeps goign into preview mode -

nvmd got it lol

[late whale]

Can you tell us more about the lyric, the style of song, any information you want to provide regarding the Style description. Sometimes it's as simple as putting a . at the end of the word to keep it from being drawn out, othertimes when you want ot drawn out you can simply add ... at the end. I can also be a part of the style of song you are trying to create, EDM vocals tend to lean into the more breahty etheral realm by default in my experience (not saying that's what everyone gets). So you may need to alter you style description or add instructions before the lyrics to be more specific as to the type of vocal tone you are trying for.

I'm not an expert, but happy to try and help if you are able to provide a bit more information 🙂

[brazen yew]

I upgraded to 5.0 and got the 1 min preview as a full song on the browser version. The app version is still 1 min. Any idea how to get the full version of the song on the app?

[echo moat]

what does "Error creating stem, no context track" mean

i just want to replace like 3 notes in studio

[slim summit]

is power metal. i tried different punctuations. period, exclamation points, comma between the words, pasting the entire final 2 lines as IPA. the weird thing is that the meter and melody is perfect until about the middle of my verse, and then the AI decides "the highest law of the heaven is that the last word of this verse must be loooooooooong. i will break out of the perfect meter, switch to singing a comppletely different song, just to make it work"

it smells like overfitting

[late whale]

Sounds like you did everything I would have thought of. I do feel that pain, lots of time I have rendered a song, first have I'm like "This is perfect" and then all a sudden the AI does some totally random thing that makes no since and I'm like NOOO!" Wish I could be more help, maybe some of the more veteran user have some ideas.

This might be a hated suggestion, I don't endorse using ChatGPT for actually writing songs, but it can be helpful in analyzing issues and suggesting things you can try to fix it. Maybe give it the section of the song that's being an issue, tell it what version of Suno you are using and what the AI is doing and see if it can suggest a solution. Sometimes you have to fight AI crazy with AI crazy lol

[slim summit]

my rhyme schema is A-B-A

the best result i've had so far is adding an em dash — to the penultimate line and ending the final line with a question mark

[rocky garnet]

I've had significant trouble getting verses to come out as rapidly or as angrily as I want to in power metal tracks, I have no idea how to prompt it so that you have that kind of variation

Like maybe I want a standard melodic chorus but I want a particular verse to be a faster tempo, it's difficult to achieve

The last verse of one of my power metal tracks was meant to be completely different but I gave up on it because the song is still good. it's a bit frustrating though

[polar spade]

I believe promos shown on browsers often apply on the browser. But it's logical that if you log in using the same IP it should also show on APP. It may simply take a while to recognize your phone. Are you logging in through your phone browser or desktop. If desktop, try logging in through your phone browser.

[brazen yew]

Thanks. I fixed it by uninstalling and reinstalling the app.

[polar spade]

@late whale Playing off that - write a dummy verse and then remove it after generated if the last line sounds right. Extend an ending to that.

Personally I go to edit UI and surgically remove the middle of the long word and remaster if needed. Often if done precisely, no other edit is needed.

[polar spade]

That will reset the DNS for sure. Good job.

[calm karma]

so i have this outro. basically a chorus but different timing. turned out great. i would like to roll the song again but i would like that outro to be the full band siging. how would i prompt that?

[gentle glacier]

can anyone tell me what happen to the recreate feaure on suno stdudio this is beyond frustrating...

as of yesterday it dissapeard

[forest hemlock]

Could try something like [last chorus - all vocals] and something like "
A climactic final chorus where the full band sings in harmony, powerful group vocals, energetic live sound" in the prompt (with an extension right before the last chorus)

[calm karma]

i'll give it a shot! Thanks

[forest hemlock]

Group chorus and full band could also potentially work in the lyric metatag

[shell zodiac]

Guys I have a problem. I create musics with a perfect Portuguese from Portugal. But now I spended credits and credits and many words are Brasilian. What happened? Why the same instructions dont work in the same way?

[gentle glacier]

anyone has the same issue?

[paper dawn]

Crop in editor

[paper dawn]

I’ll try it online but are you highlighting the track you want to re-create?

[indigo gazelle]

anyone knows how to give a prompt for chopped vocals for deephouse?

[paper dawn]

Stuttered vocals

[indigo gazelle]

do you have an example?

[paper dawn]

[verse]
[stuttered vocals]

Say if you don’t want it for the whole track just put that above the verse or chorus that you want stuttered

If you want it for the whole track just put it in your style box

There’s a really good tutorial in the discord that this guy made on GitHub it’s actually really good I would just go look at that it would help you a lot

[indigo gazelle]

i have received this result with the instruction: chopped vocals, but it took many attempts

i will try with your prompt

[paper dawn]

https://github.com/stayen/suno-reference/blob/main/tag-reference.md

[indigo gazelle]

thank you !

[gentle glacier]

yes ive been using it for the past few weeks and all of a sudden its gone all i see is replace

[hoary marsh]

Not sure if this question goes here?

So I have a song, eh 1:06 long, guitar, vocals, etc. My recording is only 1 verse and chorus. I now have more and want to use that to help guide Suno into putting it all together.

I was able to make a persona out of the recording, but i can't figure out how to add that clip to my lyrics, style ans v5 keywords to use it as the guide.

[paper dawn]

Extend it from where you left off and it should follow everything you’re doing

If you’re just adding more lyrics just extend it but the way you worded it you would be adding the lyrics to the clip . Correct me if im misunderstanding

[hoary marsh]

Well, in my case I fixed some of my lyrics, made a few changes. Added a second verse before the chorus (originally v1 chorus, now v1, v2, chorus). I am just hoping Suno can hear the original to help it reproduce. I have found in past version, when I remix, cover, etc. and add new parts or lyrics it just screws it all up.

[gentle glacier]

thanks for responiding, do you have the same issue?

[paper dawn]

I mean but replace is doing the same thing as re-creating right i? Yeah I didn’t find anything with re-create but I do feel like I have seen that let me know if that’s what you mean I mean you can still get what you need done right remaking one whole track

[paper dawn]

Try editor u can rearrange the track how u want my guess and I say guess is I’m not a professional and I don’t work for Suno I would just say finish your track then rearrange it in the editor

[gentle glacier]

well replace isnt doing really what recreate did, recreate made the track again but diffrent replace isnt doing that and it much much longer to generate and it isnt doing what recreate did

[paper dawn]

Yeah man I’m not 100% sure then I don’t think I paid enough attention to know if we replaced and re-create wouldn’t have been the same word in my head you know

[gentle glacier]

if its gone now does that mean it wont come back? i know they keeop chaning things by the minute

[azure crane]

The Stem download have been broken now in like 1 week. The only thing that is working is the WAV+Midi download.

[half dragon]

I just happened that I finished a song and downloaded the stems working for me, literally 2 min ago.
Did you get any error? or it's just not doing anything?

[azure crane]

what download option did you use? I get the drum file in Wav format when I try to download MP3, Wav or the new Tempo Locked Wav.

[half dragon]

Just WAV Audio

[azure crane]

Hmm, which browser do you use? Mac, PC?

[half dragon]

PC windows 11 with Chrome

[placid spindle]

HI!! When I make a Persona, how can I use it?

[neat nexus]

Select it from here?

[placid spindle]

I thought that was to create one?

Let me look

[placid spindle]

DUH!!!! I feel so stupid. I didn't click on that because I thought it was to create a new one LOL

So a persona can be a vocal style or music style correct?

[azure crane]

Ok. Will try Chrome then…

[wind flower]

Hey everyone — just a quick heads-up from a daily heavy Suno user.

I’m from Japan, and I can confirm that Suno has been acting very differently these last 2 days, and it’s not just a small glitch.
If you’re burning credits right now, you might want to wait.

Here’s what’s happening:

🔻 v4.5
• It reads the prompt inside the song (literally speaks it),
• Lyrics + style instructions are ignored or applied incorrectly,
• Results are nowhere near the usual v4.5 quality.

🔻 v5
• Completely ignores prompts,
• Forces a very generic “default/royalty-free” style output,
• Doesn’t follow genre, mood, structure or detailed instructions.

I use Suno every day for hours, so the shift is extremely obvious.
I’ve already reported it to the team, but I don’t think most users realized it yet — so many people are wasting credits thinking it’s “their prompt” or “their browser”.

This is NOT the usual Suno output.
If your results suddenly feel “cheap / generic / not following instructions”, it’s not you — it’s the current system.

Use your credits carefully until it’s fixed.

Hope this helps someone avoid wasting their monthly credits.

[paper dawn]

I do see something going on because I just brought in my instrumental from fruity loops typed in the lyrics and it won’t stop with the noise like it sounds like they have 16 different tracks playing

And I also just got a new button either last night or today but I didn’t see it last night but it’s generating a new lyric for every new generation and I did see that button but I didn’t have it before so they must’ve done an update between yesterday and today cause yesterday I was fine I made some fire

And that would explain the stuff people were going through last night that just didn’t make sense

[crimson harness]

I'm in the same situation. It doesn't follow commands correctly and produces a very basic sound.

[neat nexus]

Yes, but you can select it and change the listed style and still gain an effect from havin it selected.

[placid spindle]

got it - thankss

[eternal flare]

So I've generated a few thousand songs by now, has anyone else casually observed "magic generations", that are far and away better than those that come before or after them? I maybe get them one in ever 100-200 generations. Same prompts, same settings, one will just be leagues better, and ones before or after dont' usually live up to it. They are sonically better, structurally more precise, if they are covers they will more correctly match the source materials in the exact right ways.

Anyone seen that, or am I just nuts 😄

[velvet shadow]

The day they remove 3.5 is the day I unsubscribe

Gg

[hallow rampart]

hi all, i got an issue whit the subscription , i should get my credit on 1st of this mouth but nothing until now, i pay for one year wtf

[azure crane]

It worked 😁

[half dragon]

Sweet! Glad it did - to unblock you.

[azure crane]

Yeah, it was so frustrating when it didn’t worked..

[echo moat]

What's going on with suno studio? i keep getting unknown errors while trying to replace sections of audio

[polar spade]

Throughout today many posts of different access errors or odd quality issues. May be a system change of some sort.

[polar spade]

Check your next bill date on your subscription page and any renew date on a bill history receipt. Do they all show the same renewal date of next year?

Is your subscription level still functioning - - even if 50 credits only?

[tired forge]

How the heck do I make ONLY the vocals from a song a persona? I cannot remember because my sleep apnea melted away my memory

I can do that. I just don't remember how… Which bothers me. I wish I could…

[polar spade]

You stem out the vocals. Then click create/make persona from the vocal stem menu.

[tired forge]

So extract Stems first? Does it matter which of the extract Stems options I click or is it something besides extract Stems?

Crap. I clicked extract Stems and then it signed me out? WTH?

Did I want to click extract Stems and get this far?

REALLY wish I knew where my laptop mouse was…

[livid wind]

What distributor should I use to get UGC credit? LANDR isn’t giving me the ability to monetize Instagram TikTok etc anymore … does anyone know?

[polar spade]

Once you extract the two stems, it's the Vocal stem, found in your Library that you click it's menu and make persona.

[tired forge]

So once I get to that screen do I refresh the page and it shows the stems separate from the songs? I don't want to screw up further. I've probably used 30 flipping credits already.

[quiet solar]

Honestly, this is a really solid piece general advice for anyone creating something in Suno. It's MUCH better in the long-run to know what you're trying to achieve up front than to have to go back and try to fix large sections of it afterwards.

[polar spade]

Make sure to click off the tick mark in your filter for Hide Stems. Then you'll see 2 vocal stems and 2 instrumental stems. Just choose one vocal stem. If you need to refresh to see them, sure, but you may not see them simply because the filter is on to hide them,

[tired forge]

Oh right. The dang filter!

OMFG... WTF? Since when did I extract Stems FOUR TIMES?!? Ugh

Thanks a bunch @polar spade

[polar spade]

Yea. Each extract gets you 2 pair or four stems. Do that a couple of times and you've already filled up a page with stems. 😅

[tired forge]

I really don't remember extracting so many times, but I hear. It's whatever though because I'm making it a persona

[polar spade]

Got it figured out?

[tired forge]

Had to finish with my phone since I can't remember how to zoom out on my laptop and click the Save button to finalize the persona, but it looks like I did. Now I just have to think about deleting the version of it that has the music attached to it also.

That won't delete the song it's from, will it?

[polar spade]

Cool. Glad it's worked out.

Delete song it's from..... No. In fact

In fact once you make a persona, Suno won't let you delete either the original or the persona

In this case... Taking stems of a song doesn't affect the original song in any way either.

[tired forge]

Sweet. Many thanks.

I'll just hide Stems again now that I have my persona from the vocals stem

[quiet solar]

Not really the right forum for this kind of question - you're better off taking it to General.

However, this isn't just a LANDR issue. YouTube themselves have begun demonitizing certain types of AI music (e.g. static image on the video, 100% AI audio, uploaded as a standard user and not as an Official Artist) , and Spotify are also cracking down on 'slop' level autogeneration AI music and anything which is deliberately attempting to impersonate existing artists.

So yes, you'll need a new distributor (or an upgraded LANDR plan), but it might not be that simple. Best of luck to you.

[tired forge]

It's a little vexing that I've added a 2-3 lines to a song of mine that I'm working on and it's gotten shorter each time I've added a line to the lyrics.

[polar spade]

Is it taking away the longer interludes and putting in new lines so that's shortening over all.?

[tired forge]

I think it must be taking away the interludes for some reason. I didn't think adding a few lines to it would do that. Do I want to make a cover now? Would that do anything for me?

[polar spade]

We never know. Covers can do a few short, a few long. I stay stocked in credits so I throw everything at it. Just know that cover usually go shorter on songs longer than 5 minutes because it allocate a bit of time to use the original song as seed for the Cover. It's like sticking a bit of the original song Invisibly at the end of that 8 minute limit, and so you only have 4 minutes to fit the cover in. So the cover comes up a short one for a 5 minute original.

[tired forge]

How do I see my credits? I never know where to look...

[quiet solar]

It's pretty invisible on the mobile interface, but if you go into Create and look in the top left, your credits are next to the note symbol.

If you can't see it on the left on desktop, you might need to book some time with your optician.

[tired forge]

I see it. I'm used to using my phone because I only have a Chromebook and it's been a total pain in my butt with logging in until recently.

I'm going to have to hold off on creation until more credits come in though I think... I've gone through so freaking much. Goodness…

[quiet solar]

Just pitching in to this headache - Suno loves structure, so adding additional lines to a verse/chorus which it already has a structure for can cause it to do odd things to fit the extra lines in. Sometimes that's weirdly extending the whole thing by a bar and just including the extra line in the middle of it, sometimes it's replacing what would have been a musical interlude.

[tired forge]

Well I didn't just click extend and add extra lines is the thing. What I did was rewrite the song basically because what I added was in the beginning and the middle.

I'd like to click Extend and all, but I don't know what I would add anymore…

[quiet solar]

Ah, gotcha. If you've got to the end of the fixing process, then set aside some credits to run Cover (which you can also use for minor lyric fixes in sections you've already generated) and/or Remaster, as both can be effective in smoothing over any rough edges caused by the editing process.

[tired forge]

I think I can be happy with what I added to the end just now, but I have to listen to it and see how I feel about it.

That… Sounds a little wonky... I think I give up for the evening. Thanks for the help everyone. @quiet solar and @polar spade you guys are awesome. Appreciate ya

[finite folio]

Oh wow! I managed to make the song in Suno AI with a mixed choir 😮 Very simple lyrics, but a powerful chorus 🙂 Too sad… But I have it on another account, and that one is the free version 🙁

[hexed iris]

Should I do an audio or cover version, which is better to keep close to the original including the duration? Both require multiple attempts, so I'll focus on one.

[hexed iris]

Great, share your prompts if okay!

[hexed iris]

Amazing 👌...way to go!

[finite folio]

May I delete this song now here?

[hexed iris]

No Please. I've saved in my Playlist to use the choir style in future!

[finite folio]

Why not 🥺 I tought in here in Discord.

[hexed iris]

Okay. You can 👍. Thanks for sharing

[polar spade]

@waxen jackal [response from #archived-ticket-help message]

First, as would be for all songs/outputs we create in Suno, we would need to keep out own records of origination dates and during which period we created the output(s).

As we may ping-pong our subscriptions between Basic and a paid plan, this becomes even more important:

You may need proof of being in a paid subscription when a song was created as well as your TOS, to prove you have ownership , origination dates, and that all dates coincide with an active Pro/Premium subscription at the time.

As for outputs created in Basic free mode. Those are not owned by you (according to TOS) , and will not be , if you then upgrade again to a paid subscription.

As for the promotional pop-up. That only applies to the one song denoted by the popup , seen as you wished to download to an mp3. As your status was "Basic" at the time, that ONE specific output will be deemed owned by you as for having the same commercial rights as a Pro member, if and only if you upgrade ---> through that popup promo.

No notice will be given, no certificate for the one output, so mark it as your "one upgraded Pro-owned song". Maybe in the title, the lyric description, or elsewhere. You only get the one, and only through upgrading through that promotion. It will not apply if you cancel the popup and upgrade later and, if so, will remain a Basic free output.

[keen iris]

Hi everyone, can someone advise how to increase sound volume without spending credits in 'Edit'?
Suno songs is too quiet for some reason

[honest heath]

download the track, use a free program like audacity to boost the gain

[waxen jackal]

I have understood but isn't this all too complicated for a user having like 400 tracks created with free and pro plans? Can Suno put in tag system next to song's title as commericlly allowed and not allowed.

[raw oasis]

Excellent idea. Might be worth adding that to #💡┃suggestions-web

[polar spade]

I would love that! The more its posted in the suggestion channel, the more they might make it so.

I keep two accounts. One paid and the other Basic. I never mix the two. And if i downgrade paid to Basic for awhile, I dont make songs in that account. https://discord.com/channels/1069381916492562582/1260263379243761776

[raw oasis]

I sure hope I did not seed any of my hits on YT with a basic day 😆 jus kiddin

[keen iris]

Yeah it seems the only way, thanx

[waxen jackal]

Posted suggestion

[raw oasis]

Soooo wish I could get Suno to do a voice similar to Life Screams - Lacey Sturm. I will try husky.. it edges on scream but doesn't..

Trying "Aggressive Hard Rock, 6/8 time, Extremely Passionate Female Vocals, On the Edge of Screaming, Dynamic Drums and Bass, Soaring Melodic Belting, Controlled Vocal Power, Edgy Alternative Metal, Layered Guitars"

[tender venture]

Placing the lyric text in caps helps with screaming

[raw oasis]

I will try that too, with my excludes it might edge where I need it. excludes: "screamo, death growl, throat singing, unbridled rage, yelling, vocal distortion, metalcore"

[tender venture]

Nordic ritual also helps with a low Humm scream

[forest hemlock]

Aggressive hard rock in 6/8 with an edgy alternative metal feel, featuring extremely passionate female vocals that ride a husky edge without breaking. Layered guitars surge with tight, saturated, searing harmonies, while dynamic drums and punchy bass lock a thunderous yet controlled groove. The vocal takes are gritty, soaring, focused, and emotive, with precise articulation, expressive rasp, and crystalline sustain across belted peaks. The mix is wide, immersive, polished, and impactful, with shimmering air, warm low-end, lively transients, and cinematic weight

What about something like this

[raw oasis]

taking a look. thank you.

[forest hemlock]

Aggressive hard rock in 6/8 with an edgy alternative metal feel, featuring extremely passionate female vocals that hover at a strained edge without cracking. Layered guitars crawl with loose, brittle, abrasive lines, while dynamic drums and muddy bass drag a boomy yet uncontrolled groove. The vocal takes are nasal, shrill, unfocused, and fatigued, with sloppy articulation, abrasive rasp, and smeared sustain across forced peaks. The mix is narrow, cluttered, dull, and weak, with harsh air, bloated low-end, lazy transients, and boxy weight.

If I was doing it myself I'd probably also put this in negative and put 100% for style influence

And play around with weirdness

[raw oasis]

Okay cool that is a level up for me; This kind of negative feed. Thank you for the tips!

[finite folio]

If you have another program installed, add another audio track there and add the same song to the second line. This way you can also make the sound louder. Had the same problem and I used a different program than Audacity. for example

But in my opinion, Suno has one problem, it makes music with too much bass.🙄 If the bass is too strong, it can also start to drown out vocals.

[sharp perch]

Hello, Anyone got tips on how to make a radio ad ? Spoken no as a song .

[finite folio]

But what happens .... On the one hand, new songs are constantly added to Suno and eventually there are so many of them that it becomes difficult to keep up with them. The only solution would be that in this case there should already be a separate server where the songs are stored, for example in libraries, so that the user can access their songs there. The user places the song himself in the "music library", for example it is separate from the creation list, but the user can remove it from there again. And...They are kept for a certain amount of time and then they are deleted on their own to make place for new tracks.

[frank ermine]

Hey everyone 👋 I’m new here, nice to meet you all

[tired forge]

I'm unfortunately awake again. Hopefully I will get back to sleep soon, but nice to meet you @frank ermine

[frank ermine]

ohh sorry for waking you up , thanks , nice to meet you too , wish you a nice sleep

[last bay]

Most hang here https://discord.com/channels/1069381916492562582/1069381916492562585

[lyric kestrel]

Hi what would the best prompt for this song be? I'm trying to make a song like this one.

https://www.youtube.com/watch?v=F5MZmWTlSTE

[tired forge]

You didn't wake me up. Don't worry about it

[frank ermine]

ahh ok , sorry my bad

[tired forge]

I'm going to take some Unisom and get back to sleep. This is my longest song ever. I really like my lyrics and I didn't expect it to go this long.

https://suno.com/s/9BkfP7P1enEt4IM0

[unreal trench]

I'm just curious what is wrong with trending again... this morning there were only like 4 titles in global NOW, and now, that it's full again, we got like titles that are almost a week old... why is that again?

[hollow hare]

This is more of a #💬┃general-chat question (well, really more for the Support team and they probably won't know either).
We do not know, only those that works on Suno's algorithm would and I doubt those personally involved are even on Discord.

[unreal trench]

Yeah, I understand what you mean. I gave up on looking at it quite some time ago, but I still notice that constant changes are being made - and they often don’t make sense. Personally, as I said, I’ve stopped trying to understand their algorithm, but people keep asking me about it and expressing their... hmm, let’s say... frustration? It would just be nice if they could fix it once and for all - not only the “global” (Now, Weekly, and Monthly) but also the language trending, because for many creators it’s quite important as it gives them at least a bit of visibility.

[radiant thorn]

I miss the days when I would upload audio and cover it—leaving lyrics and style box blank— setting influence to 100% and everything else to 0%, and it would spit out a cover that was pretty much the same length and structure as the original audio. 4.5+ seemed to stop doing this a little after 5 came out, is it just me? It never even gets past the bridge in some of my tunes

[polar spade]

Edits like that should not charge credits.

Tested: It did not for me. After 2 hours, no reduction in credits.

quick steps
° bring up output in new edit UI
° Select segment to adjust volume
[ If no segment, drag across to create segment.
[ Click lower menu under selection.
[ Choose "split" to create a segment
° click lower menu under segment.
° adjust volume
° Click Save as new file [in menu - top of page] No credit charge should occur.

Done.

[raw oasis]

I am thinking of buying scalar to influence my chord progressions (and melodies) in audio uploads more. I feel like pure Suno generated songs cannot be trusted to create meaningful - AND Repeated - chord progressions. I mean sometimes we get bangars yes.. but also a lot of garbage.

Also just wanting more melodic flavor / off-chords, that can be repeated as well (intro, break, outro).. its pretty good with lyrics doing that, but not necessarily the music sound.

[blazing matrix]

Hi all do anyone know what commands in lyrics suno can understand like the obvious ones: [Verse], [Chorus], [Bridge], [Drop] is there something else that can be useful?

[hard beacon]

[pre-chorus] [interlude] [breakdown] [break] and typical music buildup related stuff

[blazing matrix]

So if I put [Guitar Solo] should work I guess?

[hard beacon]

yapp

maybe stuff like [drumsolo] is working too 🙂

[blazing matrix]

Perfect I’m gonna check that tonight

[hard beacon]

have fun 😉

[neat nexus]

In addition to what others have said the best advice anyone can give you on this is to try and see what works. Throw all sorts of things in brackets to see what the model recognizes and can produce reliably for you. It's also how new things are discovered.

[hard beacon]

this moment you have a wonderful nearly perfect melody you like - then there is a rap part you hate (cover and stuff like that doesn't help) and after 2:30 minutes the quality went down... naaaaaaaaaarf 😄

overall i have the feeling the vocal quality on v5 went down.. in the beginning they sounds "full" and "rich"...

[keen iris]

I tried and it charged me for 5 credits 🥲
Maybe some changes really shouldn't be charged, or atlist volume adjustment should cost little bit less 🥹

[hard beacon]

i've spend 1000x 5 Credits to get a good output today 😄

[keen iris]

Nvm it uncharged me in a few minutes, gratitude ❤️

[keen iris]

Same, but I spend them to get grate outputs 😉

[hard beacon]

well.. i am a perfectionist... one wrong note = TRASH 😄

[quiet solar]

It's better not to think of anything in [ ] in the lyrics as being a "command", but as being a style prompt of some description.

You can literally put anything you like in there, and if Suno can reference it in its library of understood music and decides to pick up on it, it's likely to influence the output. So it will work best with more musical terms. But you can definitely give stage directions to the vocalist, or instrumentalist, or change the overall vibe of the song.

Structural tags like [Verse] and [Chorus] are common across all genres and work very consistently, so they might fairly be considered an exception to the rule - as opposed to a starting basic set of "commands", where there surely must be others programmed in.

[hard beacon]

naaaah... that's not 100% right... in the lyric prompt you are also adding the song structure.. and then stuff like [verse] [breakdown] [interlude] [guitar solo] is working perfect... of course you don't have a [drum solo] in EDM music.. but his text sounds like he is creating "human created music" 🙂

btw.. no front.. i hope we don't wanna say the same thing but with different words 😉

[main holly]

Yo I paid the yearly subscription by accident who do i contact? I tried email but there not responding

[solemn lintel]

I continue to have an issue with personas and covers. Most times I do a song based on a persona, it comes out fine, using the voice I managed to get into the persona. But whenever I do a cover using a persona, it completely ignores the sound of the voice and just uses a generic voice. The whole point is that I'm trying to cover one of my own songs to make it sound like a specific "artist," and it just won't do it.

[solemn lintel]

How long has it been? I'm sure they're pretty swamped most of the time.

[main holly]

The purchase was done on the 5th and I sent them a email that same day and I haven’t had an answer

[solemn lintel]

I’d give them another day I guess.

[drowsy owl]

Is it just me or is the "Create" screen completely broken?

It has been loading infinitely for nearly 3 days now

Then when I try to generate something, it tells me "An error has occurred" but still use my credits

[quick sonnet]

Which suno version? I mean mobile, web or some other?

[drowsy owl]

Web

[quick sonnet]

Have not tried today, but yesterday and days befor had no such problems.

[drowsy owl]

That's the error message I get

[solemn lintel]

Have you done a shift-refresh on the page?

I'm not having any issues, might just be your browser.

[thin flame]

not sure if this is the right spot for it but i have a question about personas, anyone know about them?

i'm trying to figure out why i can't make personas off of some of my songs; doesn't seem to be any good logic behind it, did some testing, seems basically random. saw someone on reddit mention it could be Suno thinking i'm a trailer music company or something so they turn off the functionality for certain songs...? idk. dystopian.

[solemn lintel]

What exactly isn't working?

[thin flame]

that's what i'm wondering. it just says "can't make a persona from this clip", no explanation

it SEEMS to be because it was remastered, but it also SEEMS to happen to my original tracks too

[solemn lintel]

What's the clip? Was it an uploaded song?

[thin flame]

nope

it's a full song i generated

my copyright

...the bot clapped me for dropping a suno link in the suno discord i'm dead

[solemn lintel]

Hmmmmm, that is weird. I've never seen that happen. I basically use personas for everything, only had an issue with it when trying to generate directly from an upload.

[solemn lintel]

Wow, WTF

[thin flame]

Yeah I never use personas so i'm trying them

nah nah i get it

Suno is very risk-averse

but the irony was also fatal

[solemn lintel]

Yeah, but it's understandable why.

So the song you made, it's just entirely your song, remastered? Nothing else special about it?

[thin flame]

STILL i giggle at the idea that an AI company forgets to set up filters

yeah i was trying to link it to you

[solemn lintel]

Like, what was the overall process you used? Any single step along the way could be the culprit.

[thin flame]

brother i have been obsessively studying my process here

i could talk to you for days

but basically i think the "culprit" is that i was using different devices for similar operations, thinking it was a universal app

for example-- there is no "EZ remaster" button on PC, but there IS on the app

so rather than remastering it, it's "remixing" it in v5

otherwise, i literally just typed in a prompt

i remixed it a few times at v4.5, but then i went back to the original to remaster it as v5; i may have remastered one of the remixes, but i compulsively clean these up so i doubt it would've made it this far

the weirder part is it's doing it to all the songs on this album

(it's a theme album; i think it may actually be the thing the reddit guy suggested)

if you go to my profile (link in my server bio) you can poke around the playlist (lonely low road), it's all published

[hexed fossil]

suno isn't works 4 me, how to fix this?

[thin flame]

i'm basically now wondering at what point i should escalate to support; i don't wanna look stupid 🤣

[stable oasis]

Hello all. Squeezing in my question for prompts. Is there a way to get suno to change time signature within a song? Like only for a pre-chorus?

[thin flame]

tell it to

(not kidding)

the editor is extremely awkward to use for this purpose, i recommend deferring to AI's judgement and mixing it yourself in ableton or something

[stable oasis]

I've tried telling suno lots of things. It just ignores most everything

[thin flame]

yeah it's...involved

i do a lot of complex time stuff, and to get it as deep in the pocket as you often want it to you really need to hold its hand

i suggest cutting out the section you want to adjust and just keep iterating til it gets it right

[stable oasis]

I can't even get it to take a time signature for a full song, much less parts.

[thin flame]

well

[stable oasis]

Thanks

[thin flame]

clankers don't have rhythm, brother

that is actually a very very good question and point tho

[faint blade]

hi all - anyone have any experience/preference with utilities to analyze music (for key, bpm, etc)? online or download, just not disreputable or scammy

[raw oasis]

moises.ai will give you 5 free uploads to analyze key (but paywall after).

[solemn lintel]

I’m not sure they would tell you much other than “use the Discord and see if anyone else can help you.” But yeah I’m sure something along the line is breaking it. Have you tried stepping back to 4.5 to make the persona?

[thin flame]

yes

i had to

[solemn lintel]

And it worked?

[thin flame]

sort of-- the remaster was def what broke that one specifically

[solemn lintel]

I’m actually working with 4.5 personas in 5 and they work just fine.

[thin flame]

but others, not so much

[solemn lintel]

Hmm.

[thin flame]

no no no

made the OG in 4.5

made the remaster as v5, tried doing persona off 5

because frankly it sounded better

that def broke it but it didn't say why

and the rule seems inconsistent

i just did it with a song i made just to test this

[solemn lintel]

Yeah. But the way I see it the model is what makes the song, the persona is just guidance. So if you make a 5 song using a 4 persona it will still sound as good as 5.

If that makes sense. Now typing on my phone.

[thin flame]

no it makes sense

[solemn lintel]

Ok

[thin flame]

i don't have a ton of info on how the personas work

or why i can't just define and use aliases

[solemn lintel]

I don’t think any of us do!

[thin flame]

i mean...DO they work? are they even necessary?

[solemn lintel]

It’s getting me worried about making personas now though. Since I haven’t made them since 4.5. Just using the same ones.

[thin flame]

cuz i'll be honest i have had remarkable tonal consistency across my music

(without them i mean)

[solemn lintel]

Yeah. Similar prompts give similar sound.

[thin flame]

oh no, brother

if you want, message me and i can show you what i mean

[solemn lintel]

Ok

[thin flame]

i'm very interested in your notes

[solemn lintel]

When I get done driving and running errands. ^_^

[thin flame]

oh, yes, that's smart

🙃

[placid gazelle]

I am not seeing or able to view what plan my song creations were generated with. I started with a basic plan, upgraded to the Pro plan and I cannot either on the App or the web page get it to display generated under a specific plan. Anyone able to assist me with this?

[oblique pine]

does anyone know how to properly get a consistent girl group with different vocals on a free account?

[half dragon]

I'm not sure that info is available on the song itself.

[placid gazelle]

I remember at one point it displayed it, and even the almighty Grok says there is a way. But I have not been able to get it to work.

[half dragon]

I've never seen it tbh. I can see what model it was created with but not what sub plan.

[half dragon]

Grok can be wrong.

[placid gazelle]

ergo the "almighty" snark. 😁

Just seeing if anyone else could see their own or if everyone is on same page of blankness

[polar spade]

Consistent started when they introduced personas. So, i dont think prompting would ever do consistent. Suno is all about random.

[oblique pine]

yeah i don't think persona's are available to free users but do you know how to get different vocalists to appear on one song?

[placid gazelle]

Yeah, personas were the only way I could truly get any semblance of consistency

[polar spade]

Just your receipts or billing history would be the closest means of drawing lines. Each output never has said Pro or Basic within the outputs anywhere.

[placid gazelle]

I believe at the free level you pretty much are beholden to the Gods of Randomness in a way that they believe you ran away with their daughter

Right. I did reach out to support...maybe they could get back to me by 2027. 🤣

[drowsy owl]

Ok I dunno how but I was able to fix my issue somehow

[placid gazelle]

@polar spade I do remember at one point it did display it as "generated with Basic" or pro. That must have been right before 3.5 i think

I concur that at this point all I can visually see is the version #

[polar spade]

You can try your best then to prompt for a different style before each segment, but you want that same style to consistent on the next output too. So thats where it would not be reliable.

Even within the same style changing prompt at each segment would not guarantee a new vocal, suno like to pull from the vocals that come before to sing in the next one.

Doing extension would be the more succinct way but repeating an extension until you get one that changes vocals before going to the next. Its pretty much all a guess. So June of 2023, we all had the same issue and need.

[thin flame]

if you're trying to achieve specific-sounding voice timbres, you kinda need to instruct the AI the old fashioned way

you can achieve a semblance of tonal consistency by referring to the voices by their ranges and qualities in super specific terms

when i was struggling to get the AI to do growls, i asked it to layer the same voice at different pitches so it added a gravelly quality to the singer's voice and it kinda worked, if that helps

https://howtopromptsuno.com/artist-builder @polar spade close but no cigar, i agree

[polar spade]

I did say never. I guess that means i just dont remember. 😂 Youre probably right.

[thin flame]

personas are not really a suno-unique concept or even a required one for consistency though; you can definitely "build" one using good prompting and patience

[rotund warren]

hey guys im trying to upgrade to studio but it wont let me, its giving me an error issue, anybody know how to fix it

[calm marsh]

I need help it says my phone number is linked to an account and I Would like it removed

[neat nexus]

Send an email to [e-mail]

Then post in #archived-ticket-help

[calm marsh]

Thank you

[solemn lintel]

I’ve struggled with similar issues and it often just comes down to regenerating over and over and over until it does what you want.

[quick sonnet]

Not always, even exactly the same prompt gives sometimes complete different sound.

[south anvil]

Violin back with lil bit of kick

[hazy topaz]

same here and they don’t answer

[worthy marlin]

https://suno.com/s/lOzJzwpTi8gl7cKS

This is the same rock style, I put no style and also put 90s. It doesn't sound like that

[steep haven]

I created a virtual band with a persona, based on a previously created song. Now, when I create a new song with that persona, even if the song is often good, the intro is almost always identical to that of the track that originated the persona. Has anyone else encountered this problem? How did you solve it?

[honest heath]

turn down the audio influence a bit, or use meta tags to guide the intro in a different direction

[manic basin]

I just had 6 songs generate in a row that had the same cadence and intro riff, it was so weird.. now I'm wondering if all users using a similar prompt to mine are getting the exact same sound c.c
and I'm not using the persona thing either..
v4.5-all did it with my prompt from a v3.5 song that had a really unique sound & duet that worked right

[pastel plaza]

hi guys can someone help me plz .
does anyone know why all my songs have a sssssss sound in the backround ? how do i get a clear sound ?

[steep haven]

I've even tried setting the audio influence to zero, but that didn't solve the problem. Could you please give me some tips on how to use the meta tags feature? Please keep in mind that I'm new to this and start from the beginning.

[honest heath]

https://discord.com/channels/1069381916492562582/1435818839257645216

[raw oasis]

In cases like those if extreme audio interference - like it sounds like yours may be - I discard the track as it can happen unfortunately. I try to get clearer songs from Suno in future gen's. If it is a only in a small section (and I have my song close to where I want it) I might try Remaster. it's a pain. 😆

[cloud copper]

Dose anyone know if there is a way to contole what gets sung by who like if i want a tribal like chant in the backround

[placid aspen]

Bit of a long shot, I know, but is there a proven way of getting round the extending at the end of a song problem? Suno just seems to churn out 7 second long bits of unusable nonsense. Granted, most are 'credits refunded', but it's such a frustrating waste of time and energy. I'm using v4.5 pro, as I read that 4.5+ was even worse with this issue. If the only option is to rip up the good stuff thats already there by drilling back further into the song, it kinda makes the whole thing a waste of space.

[graceful condor]

“Hi. I created my account using my phone number and paid for the subscription. But after logging out, I can’t log back in — the verification code doesn’t come through. My region is Kazakhstan. Because of this, I downloaded Discord and I’m communicating through a translator. Your services in the U.S. are so complicated, I’m honestly shocked — especially since I already paid. Please, help me!!!”

[polar spade]

One shortcut can be to stop on the first generation of a couple of short ones. If we're likely to not be experimental and change out prompts, we can expect it to continue.

Listen to the two outputs, pick one that does at least have notes or even a noise or an assemblance of continuing off of the original and then extend off of that extension. This often bypasses the urge to have one clean 1:30 extension and clicking a dozen times in hope one works.

It can work better with the 2 second extensions one often complains about. Just think of it as a link. Suno extends off of the 2 seconds and because its shorter, can make a better blend of it all.

As always. Its worked in the past. And works often for me for how i write and prompt. If it doesnt feel effective for you, drop it and ask for alternate ideas. Good luck.

[south anvil]

Nvm 🤭

[raw oasis]

An exacting Aminor chord progression starting on the Dminor chord if anyone wants to experiment with me on what Suno will do with it.
https://suno.com/s/afNqscZT91m8NlD6

Does not have to be piano btw, you can put -piano in your negative options,; it should pick up the chords.

[south anvil]

My prompt is too simple I don't want ppl seeing it 😳

[raw oasis]

we are family here hehe

show us what your cooking up

[placid aspen]

Thank you, I'll give that a go.

[polar spade]

Some of our best songs are one to three chords. Its all about the final product.

[south anvil]

I'm using gpt I just told it to create musical descriptions inside [] in every chorus verse (don't know what it called🤭) sometimes the song have unintelligible vocal when the description describes there would be choir the results is awesome even if it probably gibberish 😂

[placid aspen]

My understanding of the ai which Suno uses is somewhat sketchy, but one thing I was wondering is this: Does Suno understand time instructions/prompts?

I'm using the following as a Style prompt:

"Final outro of an anthemic rock song, lasting at least 30 seconds. A steady 4/4 drumbeat drives forward with kick on 1 and 3, snare on 2 and 4, as sweeping orchestral strings rise in a grand, emotional finish. Soaring guitars and deep bass build to a cinematic climax before fading into echoing silence. Medium tempo around 120 BPM, mood: triumphant, emotional, and cinematic."

Suno delivers 8 seconds of what sounds like Keith Moon going mental.

[raw oasis]

oh yes. I train ChatGPT to call those areas "song structure" music direction. does some amazing stuff.. only Chat GPT forgets 1/2 my rules all the time 🙁 crazy randomness 😆

[raw oasis]

Yes I will often add time signatures. But it is not quite ready for individual bars yet. That might be best uploaded via audio from a DAW/Pattern creator via .wav / .mp3

song with my time signature.
https://suno.com/s/S5f5o7lAHC72u2Gi

[south anvil]

Here's example, I cant play it tho it's buffering for some reason #🐣┃suno-showcase message

[raw oasis]

Most of your instruction is in the lyric section right? Yes I love that approach, but stuff like Key, Time signature, Tempo etc still seem best in the styles prompt.

[south anvil]

Yes I want to create instrumental music only for my Skyrim 😋

[raw oasis]

Yes that really rocks. you can go all out. i.e this one
https://suno.com/s/LorXRcnkxwewjnxL

Also a 6/8 piece @placid aspen ^^^

[south anvil]

Looks technical! I should try that sometime 🤔

[raw oasis]

The musical direction in song structure so accurate sometimes down to the last detail on an [Outro: drum roll fill then stop]

[south anvil]

Thx for sharing this technique 🙋🕺

[raw oasis]

is it similar to what your trying?

[south anvil]

It's not technical it's like this

Outro:
[As the journey comes to an end, the music fades naturally. The harp plays one last tender phrase, accompanied by the faintest hum of strings. The final note lingers briefly before dissolving into silence, leaving behind only the memory of a tragic, unescapable hell.]

[raw oasis]

I think that is perfect. Just maybe try [Outro: As the journey comes to an end, the music fades naturally. The harp plays one last tender phrase, accompanied by the faintest hum of strings. The final note lingers briefly before dissolving into silence, leaving behind only the memory of a tragic, unescapable hell.]

keeping it bracketed.

[south anvil]

I'll try that

I tried combining it with rap once gpt spit fire with lyric the result is okay tho

[tough ember]

Hi everyone. Has anyone covered a hymn succesfully? I want to do one of a Hymn that is in public domain, but the ai rejects it as copyrighted.

How do I challenge this?

[winged peak]

What might work: Instead of the audio file, find a MIDI file of this anthem, download it, and then look for a free platform that converts MIDI files to MP3. It doesn't have to sound good. You may be able to upload this MP3 and create a cover version from it.

[magic nest]

why does every song have some indian singing for the intro? so annoying🤦‍♂️maybe because i start the prompt with dark?

[raw oasis]

can you share your prompt/song?

Has anyone tried to get a cover of their song like 2 steps down (in key), to match your voice better?

[tough ember]

I translated the hymn, sang it, then uploaded my vocals... no issue with it that way, but it sux that the ai is so clever but can't distinguish copy right from public domain. the fact that an artist recorded it, does not mean the lyrics are suddenly copyrighted.

[winged peak]

Well, it's like in real life, some decisions made by others are incomprehensible. But at least you've found a workaround in addition to my method. ;p

[polar spade]

For instance, if three of your prompts reference an electronic beat, Suno might cross reference with bollywood then emulate an Indian vocalist. It can make inferential leaps like that. Alter the first couple of prompts in some way. Even just remove one, and see what happens.

Its hard to tell... its like you put yellow as a prompt and blue as a prompt, and Suno gives neither but instead gives you green.

[lyric kestrel]

Anyone know?

[barren apex]

Something is wrong with my app ; I am already subscribed and not getting access to any subscription benefits

[scenic bane]

Strange

Also try #archived-ticket-help

[half dragon]

Just make sure you logged in using the correct email that you subscribed with.
Verify if the subscription was successfully completed and payment has been made.
Also, some users accidentally create a new account instead of login as the login and create account screen looks so similar.
If all look correct and still not working, email [e-mail] - if payment did not go through, email [e-mail] for billing issue.
Keep in mind, there will be a waiting period; do not send multiple emails - once you received an auto acknowlegement reply - follow up on #archived-ticket-help
Hope it gets resolved sooner than later!

[barren apex]

Thank you

[normal rain]

Quick question, guys. When the Audio Influence says "Turn it up to resemble your original audio or persona", and I put in a track for upload to use, does it reference the instruments or just the melody? Cause let's say for example I have a melody that is synthwave/retro tv style, and I want to make it more cinematic film score style but keep the melody, how high should I be puttingh the influence, style, and weirdness?

[polar spade]

The short version is, if you just signed up, the subscription is set up (credits show, tier noted as Pro) to be activated as soon as the banks release your funds from any security or fraud review. Typical time is less than 5 banking days (not including sat/sun and holidays). At that time funds post to Sunos accounts, the "set up" subscription is activated.

[finite folio]

If I buy more Credits in Suno AI, I didn't have to enter my details again, will this money be deducted from my account later with the booking and will I get 2500 credits again if there is an update and what will happen to the ones I bought in addition now?

[half dragon]

Additional credits that you purchased will carry over, you won't lose those.

[quiet solar]

Even if the hymn itself is public domain on both melody and lyrics, any specific arrangement and recording is still a copyrightable work. It would be an impressive feat for AI to correctly spot that subtlety and correct copyright detection accordingly when comparing against a library of copyrighted material - and risky for Suno, since an angry artist could lawyer up and blame them if the AI made the wrong call.

Agree it sucks that you did everything right and still triggered the copyright detection though.

[quiet solar]

Suno will reference everything it can hear, but you can change the instruments being used through what you put in the style prompt. That's probably your biggest influencing factor, rather than the slider settings - but if you want to ensure the apple falls far from the tree, keep audio influence fairly low (it will still have a significant inbuilt weight as an audio/persona prompt) and weirdness anywhere but low.

Default settings of 50/50/25 isn't a terrible place to start, to be honest.

[polar spade]

There are hundreds if not more of copyrighted covers of Jingle Bells, among Classical pieces, hymnals, papyrus scrolls, covers of songs, etc. Theres a lot to ask for to retain 1237 copyrighted versions of Jingle Bells to compare and see that our upload is not one of them. So thats one reason they just blanket block any version.

Secondly, its Suno. A "generative" AI music composer. One composition of Jingle Bells to have trained on was enough. Suno makes efforts to not duplicate training on anything it already has thats not one of its own creation. If it censors the hymn, then it (the uploaded audio) is not creatively transformed into something unrecognized yet by Suno. SUNO is not there, at this date, for us to Cover an existing piece, but there to create an original, unique and exclusive compositions.

[scenic bane]

I agree, like seriously theres fair use for parodies too

[onyx sparrow]

yo guys whats up how i add fx glitch oder reverb delay to the vocals?

[sly mesa]

how can i make this ?
lyric..
instrumental music not voice 30 secs
for example, suno always sing al the time, how i make that instrumental breaks?
i tried using {}}

but no work

[polar spade]

[Interlude] always seems to help. An interlude for me always seems to go about 20 seconds before it starts singing the next verse.

But you can always stack them.

[Interlude]
[Interlude]
[Interlude]

[agile plinth]

Go crazy with the style Prompt glitch, reverb fx, glitch-step, etc, etc....

Then in the lyrics section add the words reverb, glitch , delay to the section meta tags

Song song song sing
I am singing song 
[Verse 2: Clean Clear Vocals]
Sing Song
song song 

This track of mine has glitch vocals, I just went crazy with my prompt which I share in the lyrics and details. I have a bunch of other tracks up with glitchy vocals as well.

https://suno.com/s/22vZTm2ZLxMsr36L

[onyx sparrow]

ah cool u chopped it

yeah didnt know that suno can do this

but i cant controll which word he glitch or?

also i would need more space in the style section and in the lycris section

1000 signs? are too less

[raw oasis]

if we put -glitch in the negative prompts will it get rid of the gens that glitch on us? 😆

[onyx sparrow]

yes i think so

if i understood the question rightxD

so but if i write after the lyric word [glitch] he dont udnerstand this right?

i need to write it or i can only write this in the style section right?

and if i wanna hear the fx effekts better?

if the fx effects maybe a slow down

or a stutter or distoration

sounds too weak

how i can increase the effekt

of the effekt