📚┃suno-school

[bleak furnace]

this sound is both high and clear, i don't know how to describe it to have such a sound and such a melody, i just know it's bossa nova and i've tried many prompts but i can't describe the sound like that, i don't know if it's because suno can't create such a sound or because of my prompt.

i tried a lot of prompts from chatgpt, studied all kinds of bossa nova, guitar playing, but none of them came out with that sound

[half dragon]

Probably the prompt. Have you tried to prompt maybe with similar key and similar style?
I'm pretty sure Suno is capable to generate this sound. Just need to know how to tell it to 🙂
It might take sometimes to finetune the prompt.

[bleak furnace]

it's too hard, i hope suno comes up with song to prompt function, it will be very useful

[half dragon]

Have you tried upload? upload it and see if Suno can create similar tune?

[bleak furnace]

do you have that function? where is it?

[true warren]

Anybody having issues purchasing more credits?

[magic shale]

remaster does not work well at all! Use cover instead and set the sliders so that you would get a copy of it (audio influence over 90% etc)

[magic shale]

did you try to give the song to chatgpt to give you a description of the song or to categorize it? it works...

I would call this a fusion between flamenco/spanish guitar and bossanova

[bleak furnace]

chatgpt can't parse audio file, it doesn't work bro

[magic shale]

I managed to make smth similar with this: "Make a fusion between slow bossa-nova and flamenco/spanish guitar; romantic female voice; gentle drums; percussion"

[magic shale]

for me it worked! I gave even my songs to analyze !

I will try it for you

[half dragon]

Uploaded and Cover - pretty close, maybe?
Weirdness: 7%
Style Strength: 85%
Audio Strength: 21%
https://suno.com/s/XVwm8UN00nUGdo1m
https://suno.com/s/pjYpUFJ22SLbJG8i

[bleak furnace]

it worked

wow

[bleak furnace]

i used gpt plus but not work

[bleak furnace]

so simple just upload and remix and there will be prompt, but i didn't think of it. thanks savior, i made a similar song already awesome

[half dragon]

Yes, Suno will interpret the Style prompt for your upload.

[bleak furnace]

thanks, I have a question: if I use Suno's cover function on a copyrighted song, will the cover be copyrighted?

[bleak furnace]

are u vietnamese

[half dragon]

No. Don't try to upload copyright audio because you can't and not recommended.
This was just an exercise for that audio 🙂

[bleak furnace]

okey

[worthy marlin]

Does someone know how do I see most popular song by me in library?

[half dragon]

Not with the current available filter in Library.
Maybe in the future, they'd add "Most played"

[hasty onyx]

Is it just me, or has the cover with version 4.5 improved noticeably? Although something curious is that the cover... keeps the same length as the original song, like it's doing an "extended" remaster. I like it 😀

[worthy marlin]

And most liked too

[half dragon]

You can make a suggestion in #💡┃suggestions-web

[magic shale]

wow, I made a gorgeous song with the prompt I gave you !

[worn thistle]

the editor is broken

before [replace section] of verse 2

and after
-400 credits, I WANT MY MONEY BACK!

how can i repair it? refresh doesn't work and i burn out like 400credits, the section just combine in two different section's and the lyric are like [missing passages] or the voice is like [uzfbsefgsuefbseufguzsefbzusegfzusebfuizsg] exactly like that. Just like minecraft enchant language

[magic shale]

write an email to the support and ask for the wasted credits

[polar spade]

There are no new lyrics shown to replace what you wish to correct/replace. Thats normally what one uses to replace by showing in the new lyric box all new lyrics to replace old. What were you attempting?

It seems it just replaced the same with the same, with the odd bit of splitting into two sections.

[worn thistle]

imagine: you write in to lyric "I'm bad" and the voice will sing "I'm crack" what you will do? exactly, replace that and you will correct what was pronounced wrong. right?

and yeah, sorry for that example, I'm from different state, and this just came to my mind

[polar spade]

Yes. Thats sounds right. You click New Lyrics and it will copy over the exact lyrics, allowing you to correct what you wish or leave it be. I correct mispronunciations that way, basically telling it to sing it again correctly this time. But use the New Lyric box so its more of a force then.

[polar spade]

Yesterday, I went through a whole song of about 5 gibberish words due to a remaster and corrected them all this way.

[worn thistle]

it doesn't matter, when the [verse 2] is divided into [chorus] and [bridge] and the [verse 2] is like "I'm not here and I have never been here" after replace

the [verse 2] is just gone (like the section) and it's splited in to [chorus] and [bridge], the [verse 2] (like the section) in no longer exist

[polar spade]

@worn thistle. I'll add, it wasnt all cookies and cream.... some replacements placed the corrected line some 30 seconds before the gibberish line, so once I replaced, I had to just remove the bad original gibberish bit . But im used to that and it took less than a minute.

[worn thistle]

yeah, but the replacement is full of ****, there is nothing like the lyric, just some random things

[polar spade]

I focused on how the Replace didn't seem to use New Lyrics as the tool.

As for other glitches, that's another thing to discuss, for sure. 😉

[ebon sand]

I'd use the legacy editor, the new one doesn't work for me either

[worn thistle]

HOW CAN I USE THAT???!!! please i realy need to know

[ebon sand]

Top left in the new editor there's a legacy button which brings you back

[polar spade]

Im now bouncing back and forth between the two. I am at least trying to new Edit first until I give up and go back to Legacy. Im fnding an interesting difference that im testing first before reporting.

[worn thistle]

JESUS i love you, thank you. JESUS you are my angel

[polar spade]

Damn. Now i feel like Judas. 🤣

[polar spade]

Glad we had this discussion. Team work. You're welcome.

[worn thistle]

🙂 thx

[polar spade]

<---- three hours sleep.

[ebon sand]

I hope they patch up the new editor soon though so that it works to a good enough degree to be better than the old one

[graceful acorn]

Any reason why I get "Can't use artist name in description" even with an empty input box?

[half dragon]

Maybe in the title? it might be the generic msg that they used to validate song text.. or prompt...

[graceful acorn]

No matter what I cahnge it won't let me make a persona. I have no idea.

[half dragon]

Without seeing it.. hard to tell.. but basically it validates the text somewhere.
Does it say what it its? the artist name.

[boreal harness]

Would you believe that describing a persona as "Stoner Doom Metal" flags it as a copyrighted artist name? It's pretty dumb. That clearly isn't an artist name.

[half dragon]

I bet that was the reason for @graceful acorn issue.

[graceful acorn]

I don't know why I'm paying for Suno at this point if this is happening now.

[boreal harness]

@steady kelp Why does Lavf58.76.100 appear in every wave I download? I thought that was an mp3 encoding software.

[graceful acorn]

If I use the word soprano to describe the vocal type it says it's not allowed too. Did they update something and break the whole thing?

[boreal harness]

We are unlikely to get answers on these topics. It's probably internal spaghetti and embarrassing.

[graceful acorn]

Well then I can't keep my sub sadly

[boreal harness]

The alternative is worse.

[graceful acorn]

Perhaps, but if we can't use normal terminology to describe our work and get the sound we need, then I have no use for the tool on our project.

[polar spade]

The Sopranos was a TV show... 🤔

[unique stone]

Regarding persona names: Parts of the name might match an existing artist name. In the silliest cases, it catches things that are part of another word (which is a bug that has been reported.)

[unique stone]

Where are you seeing this?

[magic shale]

In the wav file header.. open with a text editor

they mark the wav with their signatur

[unique stone]

I opened one of mine in a hex editor and it's got the usual "RIFF" and "WAVEfmt" headers.

[wild raven]

Is there a list of persona's for instraments? I have a specific acoustic guitar type of sound I'm looking for and want in a song.

[half dragon]

one of the stems (WAV)

[unique stone]

Ahh, it's the software used to create the file rather than the compression.

[slow edge]

I cropped the start, as liked the instrumental intro, but the rest wasn't so good, so was trying to extend from say 0.17 seconds in and continue the song.
Would it be better to save the 0.17 second piece I want, then download it, re-upload it and try to extend again with no slider influences?

[boreal harness]

Right. But why would the files be stored lossy on the server, and then converted to wav?

[unique stone]

I doubt they are stored in a lossy format, otherise it'd be pointless to offer WAV downloads. It might be used to add metadata to the file.

[boreal harness]

Possibly. It just bothers me to see a third-party tag in there.

[unique stone]

Just checking some more... Downloaded MP3 and WAV of a stem. WAV contains the Lavf58.76.100 tag in the metadata, headers are "RIFF" and "WAVEfmt".
Loaded both into Audacity. Inverted one of them and mixed, and the result isn't silent - i.e. I can hear the noises that the MP3 compression would have removed. So this is definitely not converting the MP3 into WAV.

[polar spade]

Thats a good thought, at least to see the difference.

Otherwise, extending off of so little beginning (17 seconds) might not be giving Suno much to go on so its broadens its reach and give more of a spectrum of results, changing tempo, sounds, etc. If all extends are consistently NOT the vibe of the intro, this could be it.

Does the Edit UI actually designate that segment as an "Intro" [typically green]? You may with to re-label it as a "verse" as well and then run an extension through the Edit UI. In real life, Intros can be off tempo, a different key, improvised, etc. So again, so has free reign to do whatever it wants if it see thats 17 seconds as an intro.

If this isn't your thing, then sorry for the time suck, but for fun, run a .50 adjust speed output on the intro to lengthen it to about 34 seconds, then Extend off of that to see if it gives you an equally .50 speed sounding slow tempo, but consistent with the intro vibe. If that works, then do a 2.0 adjust Speed of that last output to bring it back up to speed, and continue on with your song creation.

[slow edge]

I'll have a go with the download/upload.

Like I have done in past with clips of bird song, a red alert style siren and an emergency vehicle, that had been cropped off the ends of songs where Suno had finished a song, but had continued to create sometimes random stuff, sometimes contained some usable sound fx, I could crop out, download the short clip, then re-upload, like the start of this song. https://suno.com/s/ob8ymVNnfNMD82VB (which for some reason even though the uploaded short clip itself was purely bird song and an emergency vehicle, not any vocals, it's not publishable because it's flagged as having uploaded vocals!)

[polar spade]

Yes. Well done. For sound effects its somewhat irrelevant which direction it goes.

If for trying to stick melody and tempo of melodic intro, it may not make a difference to download/upload and it still go off on a tangent. You've used Suno longer than I have, so I'll learn something. Let me know in here what it does once you get to that point. .

[kindred birch]

Hi... ibuy subscription today for our Indian currency pro version

I have credit left 2430.. but v 4.5 insufficient balance

Why

My subscription end 9th june

And i buy new subscription 10june 12:44 around 2500 credit 500 songs and use v4.5..but making just 5×2 10 songs and its says insufficient balance

I have now 2430 credit

Plz tell me

Any body

Davson

[quick wedge]

Hey @polar spade if you’re still around the server do you mind helping Altered Remix

[patent sand]

Email billing. See #start-here . Also make sure you're actually subscribed now, because 10 songs is the free limit

[kindred birch]

No bro... i have proof

[polar spade]

This convo was addressed in "need-help" channel and presently closed/pending

[kindred birch]

So what i do.. i did emeil as you say..

[patent sand]

Wait

Billing should reply this week sometime. Business hours are standard EST

[livid anvil]

Is it just me or suddenly has the audio quality gone down a lot

[near ice]

Suddenly noticed we can upload 8min clips now? Finally i can play a whole piano song and use it on suno. 🥳

[patent sand]

Support takes awhile to reply

Notably, a lot of people had this exact issue. Try searching this channel or #archived-ticket-help in the meantime

[livid anvil]

Strap in it's going to take you a while, the same thing happened to me and I paid accidentally again and it took them over a month to get back to me

[patent sand]

Unsure about that because I'm currently (still) a free user

In the meantime also, maybe contact billing. Could be a sub issue

[modest pumice]

Is it normal for you not to see that you are a PRO? But can you do the things that PROs do?

Is it normal for an email not to be answered?

[hard beacon]

well.. just image suno has round about 1M users... if only 1% are writing emails = it will take some time 🙂

[hollow iris]

does any one know if lets say you copy a part of lyrics and put it again into another verse to repeat, but how do you sync it in so its blended

any tips

[small bridge]

Thanks my dear I appreciate it, and you. 💋

[small bridge]

Damn are they at that high a user count already! Wow

[hard beacon]

i don't know the exact user stats, but google is saying: 12M users

[polar spade]

You will get an initial email immediately confirming your request. Due to an large influx of emails, it seems Billing is needing approximately 5 days to get to each request. There is nothing more that can be done in Discord however. So reposting in a day or so back in Discord would not prioritize the matter. For now its about waiting for up to 4 days to see if the payment is posted and you are ready to go, or if still nothing after that time, you then hear back from Billing (likely no earlier than next Monday) to work it out to reinstate the Pro level subscription.

[half dragon]

I've never tried but I would think it might need some small extra sections to pad between the pasted one, like a short interlude section that could help blending them...
It'd be very tricky to move a section and expect a smooth transition musically.
It might be possible with carefully planned.

[hollow iris]

i tried remaster but its bad now

new remaster sucks

[polar spade]

you now have Remaster 4.0 again at least if that helps.

[slow edge]

The uploaded section was ignored/not included and did its own thing. Even just using the cropped section and extend, not in the editor UI section, it used about the last 4-5 seconds of the 19 seconds, so lost the vibe.
https://suno.com/s/6ERmS5IO0UkYs3I5 that's what I was trying to use, but shelved it now.

[polar spade]

I was afraid of that. Thanks for letting me know. I did have fun with manipulating the speeds I did get a two minute extend off of a 4X 8-second intro (making it 32 seconds long slow intro). And then double the tempo twice again to get back to normal speed.

It was fun and I may use it for other things, but though the extension followed the "vibe" , it would not go as slow as to match the snail-paced tempo- maybe about 40% faster.

[modest pumice]

Suggestion for the Suno IA server:

Hi everyone! I have an idea that could really help our international community here: adding an automatic translation bot to make communication easier between users who speak different languages.

One great option is Translator Bot (https://top.gg/bot/547948810796331521), which is free, easy to set up, and supports real-time translations in many languages.

With such a bot, everyone could interact better, understand tips, ask questions, and join conversations without language barriers. It would especially help newcomers who are not yet fluent in English or other languages.

If you want, I can help with the setup or provide more recommendations.

What do you think? Would it be a good idea to add this tool?

I'm not sure if it's the best, but the idea would be something like this.

[dim glacier]

@runic swan This channel is for tech discussion.
Please post your videos in #🐣┃suno-showcase

[radiant breach]

All of my songs are now starting exaclty like the persona im using. What do I do? Any tips on the new sliders? Seems like its messed things up for me. I try at 100% on audio its not different, try it at 50% it still gives me the same intro melody. I just want fresh songs!

tried with audio at 25%, still getting the same intro melody. this is 6 songs now that start the same way as the og persona. thanks suno!

[humble verge]

is the new editor fixed? I tried it once to replace the start of a song and it just came out complete gibberish.

[modest pumice]

Where does it show that I'm PRO? I never find it, does anyone know?

Where does it show when my PRO is renewed?

I can never find it right

[topaz narwhal]

why is searching for key lyrics not finding all the songs? I def have all the things checked that need to be in the filter section but its not finding songs I know have certain key words

[polar spade]

Click on the credits to take you to the subscription status page.

[vernal harness]

Can we combine different sections from 3 different suno songs we made?

[copper marsh]

can I restore deleted song?

[polar spade]

You can restore an output that was moved to the trash bin. Go to the bin. Find the output, right click and choose Restore to Library.

[sleek stump]

Short amswer is no. Longer answer is no, you cant cross edit songs.

[autumn folio]

Cover is not working as xxpected, something is going on in the backend. almost lost 300 credits

[modern blaze]

Hi. I'm waiting for the tech team to resolve the issues with the new song editor. I haven't been able to edit anything for the past few days. When I try to edit a song, the screen just freezes with the message "Loading DSP Engine." I have a PRO plan. Maybe you should revert to the previous song editor or make it available as an alternative. I'm not interested in a new song editor that I can't use.

[half dragon]

There's a switch to switch back to legacy editor but since the screen isn't loaded, you won't be able to.
But... if you're on PC web browser... try this:
while the screen is loading... you can manually change the URL on the browser address bar and replace "edit" with "edit-legacy" to switch to legacy editor.
Try that and see if the old editor works for you.

[balmy junco]

I've been following the Discord closely since the update was released, and I can confidently say that I'm one of the few people who’ve been openly critical of the situation—not just mildly disappointed, but genuinely frustrated with the overwhelming number of bugs and how unfair it feels to be treated like test subjects for something that, while conceptually brilliant, is clearly failing in practice.
Many of us, myself included, have invested real money into a version that feels unstable and far from ready. And while no one is forcing us to pay, that’s not the point—the real issue here is trust. If a company wants to build lasting relationships with its users, they need to show accountability, especially when a product underdelivers so drastically. What surprises me the most is how hesitant many users seem when expressing feedback—phrases like “hopefully this gets fixed, but oh well” aren’t helping anyone. It’s been long enough. The community has voiced serious concerns, and there’s no shortage of negative reviews at this point. If I started listing all the problems, I’d need far more than a few lines.
The bare minimum Suno could do—after so much backlash—is put ego aside and bring back the previous version, which was, for most of us, functioning nearly flawlessly. There’s simply no comparison between that and what we’re being offered now.
or at the very least, Suno could’ve consulted the community before making such drastic changes—especially on whether we wanted to stick with the old version or not. These are the kinds of decisions where users deserve a voice, not to be treated like lab rats in a failed experiment.

[half dragon]

It would be best to send feedback to Suno. If you've been following Discord closely then you should realize there's no official Suno support/dev here in this channel.
Visit #📝┃user-research-feedback for ways to get the feedback to them.

[balmy junco]

I understand your suggestion, but I find it a bit confusing when people say things like “there’s no official Suno support/dev here in this channel,” considering this is the official Suno Discord server. The channels are clearly labeled—technical-discussions, feedback, tickets, etc.—and there are visible roles like Support, Team, and Moderator. That tells me this is not just a casual community space—it’s a structured environment meant for direct user communication and feedback. Even if there are no devs actively chatting in every channel, there’s clearly some level of coordination behind the scenes. A perfect example: just days ago, users pointed out the removal of the V4 Remaster option. There was a wave of feedback across the server, and shortly after—an official announcement acknowledging the issue and confirming its return. That kind of response doesn’t happen in a vacuum. It proves someone is watching, listening, and reporting user sentiment. I’ve already submitted feedback the “official” way too. But many of the issues I and others raise—especially regarding bugs and major feature flaws—often go unanswered or feel like they disappear into the void. That’s why I’ve chosen to speak more publicly in visible channels. Not just for myself, but because the concerns I’m raising reflect the broader experience of the community. This isn’t about complaining for the sake of it—it’s about holding a platform we care about accountable, and ensuring our voices aren’t just heard, but taken seriously.

[half dragon]

Noted. I was simply suggesting another way to send the feedback directly to them as it sounds like a valuable feedback; did not mean to restrain you from speaking out.
You're free to express whatever you like but keep in mind, this channel is community based and mainly technical discussion.

[unique stone]

The developers tend to focus on reading the channels in the "feedback" section. This one is intended for technical discussion among users e.g. how to do things with Suno. On rare occasions a dev might interact in another channel, but most of the time they won't.

[balmy junco]

Appreciate the clarification. Just to note though—my message was, by definition, a technical discussion. It addressed core platform issues: persistent bugs, removed functionalities, and the operational instability of the update. These are not abstract opinions; they’re practical breakdowns of performance concerns directly affecting user experience. If anything, this kind of discourse belongs exactly here. It's one thing to list bugs individually in a ticket—it’s another to discuss the broader technical impact those bugs have on workflow, output quality, and user trust. That context is technical, and these conversations often surface patterns individual reports can miss. So while I understand the intent behind redirecting to feedback channels, this discussion wasn’t off-topic—it was on-target, just with a wider lens.

[high hound]

Anyone know when Mike B will be online, I been waiting for awhile for him to reply to me?

[hollow hare]

1. This is predominantly a community-ran server. If you had emailed support/billing, you will see in the auto-reply stating this is not a support channel (an old auto-reply I got from them a few months ago https://media.discordapp.net/attachments/1253111433605021697/1376413942187819018/image0.jpg?ex=68485aa0&is=68470920&hm=eba72162fd83b7dbec73ced2357032c7bd737f34b3227b957fe2cb09210d9535&)
2. The technical discussion is meant to be as in talking about different techniques, styles, etc. It is not technical in the sense of technical support (I had appealed several times to get the channel renamed to make this more clear but no luck at that front)
3. I assume by feedback you mean #📝┃user-research-feedback channel? That is one of the channels where a staff from Suno explicitly opens up a specific form or a period of time for personal discussion for, well, feedback. You can see more when going to that channel
4. #support-guide is a how-to channel that helps direct you where to send your tickets/help question. Please view the channel if you have not done so yet (This channel was meant as a quick copy-paste for people in lieu of /support or /billing)
5. The Team roles are for official Suno Team members. The majority are developers and not related to support (if you were unaware, Suno started off their models as a Discord bot).
6. Er, I don't recall seeing a Support role? Unless you mean Guide which is just a role given approx. once a year to very helpful community members
7. The moderators are, well, mods of this server. Believe it or not but all current mods are not part of the official team (hence the lack of Team role), they are donating their time

I hope that helps clarifies some of the points you made

[high hound]

Suno Support still have never replied to my email, it's over 7 weeks now 😦

And someone recommended I reach out to Mike B, but he never replies to me either.

Any other suggestions who I can plead for help with an error I have with my Suno account?

[polar spade]

Specific to the "technical" channel, I agree, it does need to be renamed. Its about community sharing of how-tos and hacks along with news and discussion about outputs and Suno things. Its has never been about why something doesnt work or is buggy. But many peers have been gracious to keep their ears and eyes open and even resolve many technical glitches as well. So, good for that, I guess.

[half dragon]

I re-read your message a few times.. they're all either suggestions or frustrated expressions... there's not a single technical thing - like how to improve the prompts, or how to get specific sounds... or how does this new feature work? any tips? etc....

[crude pier]

I'm sure they didn't bring back the v4 remaster because of the feedback form on the website.

[half dragon]

Enough ppl giving feedback on the right channel would definitely help!

[crude pier]

I know, but I’m a dev and I know that feedback channels are rarely used, so for them to bring back the v4 remaster, I’m sure a big part of it was because of the noise made here.

[hollow hare]

We have no feedback channels here though? Well, not since the v4 feedback forum was closed in favor for a Form for 4.5

[crude pier]

feedback channels at website bro.

[hollow hare]

I see, the use of "channel" in that context was being confused to Discord channel

[balmy junco]

Thanks for taking the time to re-read my message, but with all due respect, narrowing the definition of technical discussion to only "how-to" questions like prompt tips or sound generation tricks is a bit reductive. Pointing out that core features have been removed, that major bugs are present, and that performance has significantly degraded — that is technical. You don't need to phrase something as a question for it to qualify as a valid contribution to a technical channel. Diagnosing a system's failures, especially post-update, and evaluating their impact on usability is fundamental to any technical discussion. Just because my message includes frustration doesn't make it any less relevant. In fact, the frustration stems from the technical shortcomings—bugs, feature instability, regressions—which I outlined clearly. If the only acceptable content here were tutorials or troubleshooting tips, we’d rename the channel to “how-to-guides,” not “technical-discussion.” Feedback rooted in technical experience, especially when it reflects the voice of many users, deserves space here too. But hey, feel free to re-read everything a third time — maybe the technical points will show up once your definition of “technical” expands a bit. 🤷‍♀️ I'm done for now. Have a nice week, everyone!

[crude pier]

But it was a follow-up to a previous message that talked about feedback on the site...

xD

[hollow hare]

I wasn't following your chain of messages unfortunately. In one of my recent post above, I was discussing about the 4.0 feedback discord channel, hence my confusion. I apologize for that

[polar spade]

You mean the Report button and thumbsdown? Yea . If no one ever said a thing for the past two years but clicked those feedbacks whenever the slightest thing was not working correctly, even if the wrong chord was played, we'd be on version 7 by now. Just my opinion. 😊

[crude pier]

np 😄

[half dragon]

Again... if you're following this Discord server for a while as you claim... we wouldn't be even discussing what this channel is about.
But I hope as long as you understand one thing, no one restrains you from expressing anything, it's all good.

[neat nexus]

SInce they've been following for so long, I wonder what account they've been using? This one has that green leaf and only goes back a few days...

[crude pier]

Exactly, I work as a dev at a payment company, and whenever a bug happens, instead of people reporting it through a big button right next to the payment option, they prefer to contact us through other methods to try to get it solved.
So, we need to keep an eye on all the channels where our customers are, even if they are third-party sites. I believe it works the same way here — even if the community isn’t managed directly by the Suno team, any complaint here can be valid.
I think they probably face the same issue I do: people rarely report when something goes wrong; they just try something else until it works the way they want.

[balmy junco]

I said I’ve been following the Discord closely since the update was released, not that I’ve been here forever. Maybe give my messages another read to catch that detail.

[half dragon]

It's my bad... but wait.. thought we're done. 🙂

[crude pier]

For example, I’ve been working with music for 18 years. Currently, I use Suno only to generate vocals for demos, and version 4.5 was a total step backward. Even when using prompts aimed at improving quality, metrics, expressions, and so on, I feel it got much worse.
The problem is that we spend real money to use a service and expect the results we paid for. When they removed the v4 remaster, the v4.5 remaster wasn’t at the same level yet, so they took away something that worked for something that maybe still needs improvement.
I don’t mind getting new features, as long as I have at least some assurance that the service I’m paying for can actually be used.

[vernal harness]

But I guess I can crop and download.. then add? at least one.. Thanks for replying 🙂

[sleek stump]

Post processing is where a lot of us end up 😄

[balmy junco]

Since you’re spreading false words and twisting what I said to make me look bad, I just want to point out that trying to win an argument by making the other person look awkward and misrepresenting their words only ends up making you look awkward in my eyes. Bye now

[vernal harness]

What do u use

[sleek stump]

FL, Audacity, and ACE when I need to do voice work.

[vernal harness]

thank you!

[crude pier]

https://soundcloud.com/officialkylin/kylin-victory/s-6AlKcCUHVTd?si=243d40569ada4d74902821a4b3f20cea&utm_source=clipboard&utm_medium=text&utm_campaign=social_sharing
I use Ableton Live to create the structure of the song, generate the vocals, and download the stems.
Then, I process the vocals with equalization, reverb, frequency cuts, and resonance reduction.
After that, I mix and master the track.

[modern blaze]

I appreciate your help. Your suggestion to go back to the old song editor worked, so now I can edit songs again.

[half dragon]

Awesome!

[plush rain]

Suno appears to have an incredibly difficult time recreating traditional chinese opera

[polar spade]

Yes. I'd agree. Even the youngest of customers are locked in the "i want to speak to the manager" mentality while businesses are running the "collective hive" practice of A/B comparative analysis and stay in the background as much possible. Its a contradiction at times, but sways towards the customer.

While I like hands on service, it not all the time practical nor efficient. Sometimes if its a well oiled machine I just want to put in my dollar and get my candy bar. If it sticks, I'll push the button for help and stop using the machine. Come back later. Or drag it away and hack at it until I can get all the coins out.... oh wait... I've said too much. Strike that.

Anyway. Hire a corporate troubleshooter and marketing expert. Work it out.

[polar spade]

Its never quite gotten it right, but back with 3.5 we got some decent close ones. I'd say with proper prompting whatever that may be, there's now a better chance.

[thin flint]

Can anyone tell me why I cannot download this song? I'm loged in, and options > download > MP3 audio will not work. Perhaps I'll try wav, I don't know...
https://suno.com/song/b14546e0-3ccf-4575-a2ef-e11f491199c7

[polar spade]

Quote of the day: "Remember, this fix is only temporary, unless it works" - The Red Green Show (1991)

[polar spade]

Cache and cookies can cause weird download issues. You can try to download wav as a test, or clear your cache/cookies and try d/l'ing MP3 again.

[charred token]

Odd situation I'm not sure how to fix.

I have a song in v3 I'd like to update to v4.5. However, I've now attempted it three times, and for some reason, it mispronounces the word "pastry" when it was pronounced properly in the original v3 version. Instead of "paste-tree" it now pronounces it "past-tree". I know of NO ONE who pronounces that word that way. Is there anything I can do to correct that?

[half dragon]

Maybe try to replace it in the editor, and see if it'd pronounce correctly as is; if not, maybe spell it phonetically?

[sleek stump]

Are you doing a remaster or a cover?

Try doing a cover and set the top slider to 0, influence to max

[polar spade]

Aa AI splits words into tokens relevant to the words before and after, it may have split it into "past" and "ry" .

Often just moving up or down a line, adding some spaces, just moving it a bit can also help it reset the whole line so it pronounces it correctly. I think Replace is helpful in that way that it doesn't formulate the whole song but just a line or two, helping mispronounced words get a new alignment and get pronounced correctly.

Its my thought that's also why remastering sometimes is rife with weird pronunciations. What was good now gets shifted a bit and throws of how Suno had originally tokened off the words and pronounced them. So I've now focused more on just taking the best sounding one and spending more time shifting words a bit to eventually get them all pronounced correctly.

Of course, tried and true has been to phonetically spell it. And click until it gets it right.

[high hound]

Anyone have any other suggestions for me, how to get someone in Suno Support to respond to me?

1. I have sent emails to Suno support, but 7 weeks later, not a single reply.

2. I have then sent Mike B a message whom someone in here recommended for me to contact, but Mike B has not replied to me either.

Any help please who or where I can go next to get support assistance?

[kindred birch]

My problem are different i have pro plan..but i cant acess v 4.5..i have credit left 2430..i cant make a song with v 4.5(insufficient credits)
GPA.3308-4580-9847-05482
This is order no.

I buy subscription today..i need to help and fix this use

[lyric lotus]

Sheeet wrong chat

[raven turtle]

Cool. I’m working with a similar process here and totally agree re V4 + mastering

[rough reef]

When are you going to fix the errors in the system—such as Suno singing the prompt with instructions in the songs instead of the lyrics I provided? I am tired of losing credits because of a program error

[whole crescent]

Every time I generate a new song — any prompt, any style — whether I start with [Intro], [Instrumental Intro], or no intro at all, just [Verse] — Suno always starts with a female vocal singing “aaaaaa.” I’ve generated over 200 variants. Different prompts, different styles, different browsers.

Is this a bug… or some new kind of divine feature? 👀

[rough reef]

My intro has been missing sound for seconds for a week now. The new update is incredibly buggy.

[high hound]

Ok everyone, I have to quit trying to get someone from Support.

After 7 long and tiresome weeks, I have realised, that NOBODY from Suno Support will ever reply to me.

[atomic elbow]

Does anyone know why the instrumental tracks are ending around 2-2.5 minutes? I can barely get them to reach 4 minutes, most of the time, they fail between 1 to 3 minutes on V4.

I'm also unable to create any 8-minute tracks using V4.5; they never even reach the 3-minute mark. I emailed support about this, but haven’t received any response in the past 10 days, so I just sent another follow-up.

[modest pecan]

i think it's necessary to improve the piano timbre . it has too much distortion make it impossible to use

[upper compass]

does it cost the same to get a voc+inst as all stems?

also intertested if the voc+inst stems are sounding better now as well?

[pastel cipher]

does the credits stack everymonth during the subscription are they are lost and renewed?

[hard beacon]

not used credits will unused credits expire and are not added up when the monthly credits are added

[pastel cipher]

so they are lost?

[hard beacon]

correct

[pastel cipher]

But I paid for it...

[hard beacon]

well.. you also accept the terms of use 🙂

[pastel cipher]

And if I unsubscribe?

[hard beacon]

good question..

It would be a logical conclusion that since no credits are topped up, the old ones do not expire. But unfortunately I can't tell you exactly how it is...

[ebon sand]

They will expire all the same and you get refreshed with the standard 50 credits

[bleak furnace]

https://suno.com/s/t5pQJujJNO0mjAHD can someone help me create a prompt to be able to create a song like this, I spent days and still can't make a song like this

[hard beacon]

well.. the prompts are visible in the details? 🙂

[polar spade]

This may be of interest. Upscaling the support team. #archived-ticket-help message

[unique stone]

50 credits for up to 12 stems. You will usually get less than this, as it depends what the extractor is able to detect.

10 credits gives you the Vocal+Instrumental stems. This uses the same new technology as the "All Detected Stems" so is significantly better than the old Suno stem splitter.

For both of these you will get 2 sets of stems, which will be slightly different from each other.

[polar spade]

David is presently live in need-help room if you'd like to speak with him. He may give you further assistance. https://discord.com/channels/1069381916492562582/1372994553233276970

[hard beacon]

I'm still of the opinion that Suno should publish a weekly Q&A video... the most important questions (and they are usually always the same ones) answered briefly and compactly in one video 🙂

[shadow plaza]

Whats the difference you notice in 4.5 model compared to 4 model in your music?

[polar spade]

Usually, the "ahhhh" should NOT show up unless the genre is close to a pop style, like k-pop or hip hop, from my experience. . It may aslo do so if Suno feels its a ballad. Thats just Suno way of matching the genre. All part of being generative. Just like it will throw in a producer tag without prompting, if its a hip hop ghetto rap or trap like genre.

No good direct solution to keep that from happening. Just have to try things like having no line spaces between [intro] and [verse 1].... or use [interlude] instead of [intro] so it doesnt go into an intro phase.

[austere violet]

Hey, Please assist.

Pro to Premier. Date applicable indicates 2026. Monthly is on the 15th. Will Premier start 2025-06-15? Or wait for 2026.

[polar spade]

Are you trying to upgrade to PREMIUM And stay on a monthly payment? Make sure you ticked the monthly rates to show the monthly rates and dates.

[polar spade]

I would have a list of pros and cons for all versions. So its nice to have a choice.

[whole crescent]

no intro, no line spaces, only aaaaaa))

[shadow plaza]

And? Can you share that pros and cons for each model

[hard beacon]

my personal way to get an unique output is to create the songs in 3.5 or 4.0 and remaster them into v4.5... or create in 4.5 and remaster to v4... there are a lot ways to create the best result 🙂

depends on your style

[shadow plaza]

Im doing all kind of styles

[hard beacon]

so you are creating EDM, Punk, Doom Metal and R&B ? 🙂

[shadow plaza]

I found 4.5 give great emotions but lack 4 melody hit

I need both combined

Best from the 2 models

[polar spade]

Im not awake yet. 😊 but seriously, its been a subjective and debatable discussion. So my druthers would not be the same as another's or yours. For instance, I like live audiences, which is far better controlled and prompted for in 3.5, though 4.5 has presently some audience noises with a bit more fair , a person woops or hollers mid song and what not.

But overall, 3.5 still has the most authentic audience and applause response. Other members may not care about this.

[whole crescent]

Using Remaster, you'll never get an exact copy of your song.

[shadow plaza]

3.5 sound too robotic

[hard beacon]

i am creating a live edit at the moment... i do it like Dasvon told it... create in 3.5.. and then a remaster in 4.5 🙂

[shadow plaza]

Will test this out with remasters

Thanks gents

[polar spade]

i..e. the list can go on and is debatable or at least lends to a lengthy discussion. I've got to have my breakfast first before that. Haha.

[shadow plaza]

You must have preffered style to create music for each model?

Lets say 4.5 pop, ballands, edm
4.metal rap
3.5
??

[hard beacon]

well... in the creation process metal in 4.0 and 4.5 is... well.... how should i say it... bad??? 🙂

of course the complexity in 4.5 is much better.. but the output.. well...

[shadow plaza]

I have good bangers in 4 but 4.5 are softer mmeh

Dont deliver that rage.. That metal should have

[hard beacon]

the best melodies, also in metal, i got in 3.5.. but the sound is not suitable to the industry standard.. the most creative outputs (vocals) i got in 4.5.. so there are different ways to create the best output

[shadow plaza]

Shoegaze rock is miles different in 4 and 4.5

[polar spade]

So far if you remaster a 3.5 live performance in 4.0 remaster, it clips too much.... as you say, 4.5 does slightly better, but can still take some nuances away and make the audience sounds like a studio fx machine (as in a canned response track) than a real live audience. 3.5 will even give you feedback and mic bumps, auditorium poor reverb slaps and sometimes a fuzzy "beer hall" type atmosphere. I love when it does that. Gives a very authentic sound.

[hard beacon]

live edit created in 3.5 and remastered in 4.5... no clipping, no variatons.. just cleaner sounds <-- for me in this case

[polar spade]

4.5 is better at it... I agree.

[hard beacon]

but to be honest.. i need 400 credits to create a live edit (that i liked about 75%)

[livid trout]

Hmm, how long does it last for tickets to get answereed ? I had a problem with an accidental double subscription, now i ended up with a non pro account while I paid for a year in advanced, because I removed the duplicate "monthly pay" subscription. Made a ticket for it but i didn't hear back yet...

[upper compass]

Thank you sir

[willow hemlock]

Hey guys
is the Add Section in the new Editor broken?
I try to add a new section, I click on the plus sign, it's add a new section, I type my lyrics and style on the left and when I click create I get a section with noise/silence or sometimes it add parts of the original song into the section
I also don't see the context window...

[balmy raven]

I can tell you that the replace section sucks. I just found the previous editor though. It says legacy editor in the top left. Click that.

@willow hemlock

[willow hemlock]

I don't see an option to add a section in there... 😦

Trying to add more lyrics to the song in the old editor just cost credits and does nothing

[balmy raven]

I had to click on the hyperlink on the new replacement for it to play for me before I could add it. This UI update has been abyssmal. Sorry this stuff isn't working for you. It's not working well for me either.

[willow hemlock]

Thank you for trying!
I wish they would fix it so we could add sections to existing song.
I really wish to create more stable duets, [meta tags] like [male vocalist] [female vocalist] are oftenly ignored..

[balmy raven]

Yeah, I've tried to do similar with [whispered] and [screamed].

[crude jetty]

keep a space between prompts, just not line breaks, so [intro] [male vocals] but not [into] [male vocals]

[balmy raven]

Good idea. I'll keep that in mind.

[green pike]

4.5 is a beast on confectioning good rythm and lyrics, the sound is so much fresh and it's very good for LYRICAL - SOUND creation. Sadly, it's on BETA, and some genres generated are AWFULLY BAD QUALITY audio. Muffled sound, bad mixing, etc.

4.0 it's has a MARVELOUS AUDIO QUALITY, but some lyric creation and/or persona adding is a bit repetitive, lacking of innovation. Sadly, I prefer using V.4 only because of the good audio quality until 4.5 gets off beta or 5.0 arrives.

[green pike]

4.5 on Metal genres aka: Metal, Black Metal, Metalcore, Progressive Metalcore, Power Metal, Dream Metal... etc, it's very poor. Only works for more "mainstream" genres as pop, EDM or House, sadly. Sticking to V.4 for Metal, at least me.

[crude jetty]

I disagree, if you prompt it withmore detail it seems to be fine

[ebon sand]

v4.5 doesn't work with personas and it can't replicate my standard singer, thus it's uninteresting to me; to me it's also not very creative

[crude jetty]

don't use the [*ed] also add soft in front perhaps?

[green pike]

4.5 even bombing prompts with HD, good audio quality, surround, etc, sounds muffled on Metal Genres.... have you generated any song with that prompts? I have a ton on the trash can.

[crude jetty]

I mostly write grunge

[green pike]

4.0 personas doesn't work with 4.5 songs... sounds very different. Must be because the creative motor of 4.5 takes another information that 4.0 personas can't handle

[ebon sand]

4.5 personas don't work with 4.5 either.

[green pike]

epic, grunge isn't metal lolol

[green pike]

I hate 4.5, all my homies hate 4.5

[ebon sand]

i don't hate it, i think i'm just not the audience for it, it doesn't create music which is to my liking and yes i also think it's very muffled

[crude jetty]

https://suno.com/song/84634ec1-a0b6-4ca4-9879-6d91bfd4ebfb

[shadow plaza]

Aggree i notice that too

[crude jetty]

this isn't bad, but that was just a quick pre-generated song that I prompted style ina hurry

[green pike]

there's no way on earth you can tell that is not muffled

[ebon sand]

i've heard worse muffles but the whole sound is awful

listen at 2:00

[green pike]

well indeed but, the point is that the generation on that genre sucks

[ebon sand]

its as if frequencies are missing haha

if i cover it with v3.5 it probably improves

[willow hemlock]

Unfortunately this is not stable... like here:

https://suno.com/s/Yo7v2SdxkrOMr6f5

[crude jetty]

https://suno.com/s/txflKmfdFvvyjhiI
this doesn't domuch better in the same spot. with the same prompt

[crude jetty]

https://suno.com/s/xlNysy2wBKPl2rqQ excluding the lyrics being mangled, better?

[balmy raven]

That does seem to be the case. I guess they don't even try to test other genres.

[balmy raven]

It works sometimes, but if the lines are too close together it usually does one or the other.

[shadow plaza]

I just constantly switching between models mainly 4 and 4,5

it was easier when we didnt have so many options

😄

im kind of person that listen all genres create all genres and experimenting ,

i wish to be like the rest listen 1 2 genres and thats it

[crude jetty]

remastering all the way from 3.5 tends to be problematic going up to 4.5, most of the time

[balmy raven]

Hell, 3.5 gave me many good tracks, but when I tried to remaster them with 4.0 they kept getting that stupid "shimmer" effect. I left many as 3.5. I've had plenty of good 4.0 tracks, but I had to replace small bits of section just to get that effect out. Haven't heard that effect from 4.5, but it does seem muffled, almost like the 3.5 instrument quality, but the vocals are sung through a pillow.

[crude jetty]

yes, 4.0 remastering was hit or miss, I did like 30 on one before it behaved

[balmy raven]

What sucks is that I have a fantastic song from 4.0, but it was right before the release of 4.5, so the very end of the track has the quality decay. No amount of replacing the section makes it better. Remastering still keeps that decay in it.

[lyric sonnet]

quality decay can be a style element

[humble verge]

hey man why am I paying credits for erroneous results? SUNO just generated three results around 20-40 seconds in length, and I paid for all three.

[lyric edge]

i`ve a question, i wanna to create a Persona but when i go to the voice which i want for, always writes "subscribe to crreate persona". but i always have the pro version -.-

[undone sandal]

When I give a song with too many verses/chorus, it's skipping the last 3-5 verse/chorus. And it's not veben crossing 6mins and i'm using 4.5 pro which can create a song till 8mins long. Can anyone explain this scenario. How can I make this work or fix this issue?
Already reported to suno today. Observing this since last 20hours or so.

[half dragon]

Does it happen to be on a song that created from an upload audio?
By any chance it has some vocal in it? If it was an instrumental audio, there might be some instrument that Suno think it's a vocal - then Persona feature will be disabled and at the same time, any song created from that audio will NOT be able to publish publicly on suno.com site.

[half dragon]

If it struggled to create the structure you wanted, or the length you're after - maybe try Extend feature for more control?

[undone sandal]

will try it out

actually i tried, suno thinks thats the end where it ends and generates 2s extension or 27sec extension

it's not even reading the last few verse/chorus

4.5 is having lot of bugs, this was working when i earlier used 3.5

[half dragon]

When extending, try not to extend from the very end of the previous clip. It might work better?

[undone sandal]

will try this out

[lyric sonnet]

is there a way to recover the original simple prompt from a song i made?

[half dragon]

I think you mean, the original text you typed before clicking the magic wan??

[lyric sonnet]

yes, the 'Describe your song' field on the Create page

[patent sand]

Save all lyrics and prompts into a *.txt file or something and copy/paste from there

I don't think that's recoverable

[lyric sonnet]

ok thanks

i do often copy my prompts into a text file since that function is missing but sometimes I forget so I was wondering

i guess i should suggest that feature

[undone sandal]

not working

[half dragon]

Still giving you extremely short extensions?

[undone sandal]

yes 22s

[neat nexus]

https://tenor.com/view/tv-channel-breaking-just-in-report-gif-10310301279191238245

Channel has been renamed to better reflect its purpose.

[balmy raven]

Funny thing, the song is called Brain Decay.

[half dragon]

Do you mind if can try on your song? send me the link via DM and set it to allow Remix, I think that'd allow me to extend from your song for testing.

[undone sandal]

sure

sent you a friend request to share directly

[crude jetty]

are you using 4.5 and what are your advanced options?

[humble verge]

yes 4.5, Weirdness 50% - Style 100% - Audio influence 100%

[crude jetty]

it seems anything above 80% breaks

[humble verge]

I've generally had good results, but today I got like 6 that were under one minute and some total gibberish.

[crude jetty]

I never get lucky with it behaving properly

[lyric edge]

Yes its a covered song by a audio but its a different voice

[graceful condor]

Quick Question! Is suno not able to create background noise like rain etc? Or something like this at the start of the lyrics:

(The track starts with the sound of a massive, adoring crowd roaring. A shimmering, anticipatory synth arpeggio builds, and a heavy, four-on-the-floor EDM beat drops.)

[young geode]

Short answer: No.

Longer answer: Yes, but you're not going to get it.

The best way to introduce sounds like that is through the audio upload, or a persona that has it, or to add it in post generation with something like Bandlabs and or something else, then using it as the audio/persona.

[half dragon]

Sound FX is never Suno strong suit. You might get lucky once a while by putting it in [..] but it's never guarantee.
Better way is to add those offline in Post (DAW)

[graceful condor]

Got it! A bit disappointing but knowing helps to keep frustration in check hahaha

[cunning mulch]

so... is there any way to get Suno to identify instrumental sections as well as vocal ones?

[undone cedar]

Anyone have idea how to make so ai not changes vocal in midle of verse?
aka verse is set to be male and it switches to female

[quick wedge]

Congrats on the new channel title ✨👍🫡

[young geode]

Yes, [Instrumental] is the metatag for this. For Vocal Only, you can try [Vocal Break] or [Vocal Break, Vocal Focus, No Instruments] and you may have success with that.

However I havent tested the vocal tags as much, what you would likely have to do is modify whatever verse, bridge, ect, you have from [Verse] to something like

[Vocals Only, No Instruments]
[Verse: Singer Only, Vocal Focus]
[Vocal Break]
Lyrics.

You would also prompt with the style somewhere Cappella, and within the METATAGs :Cappella

[livid wind]

Anybody know if there’s a way to edit the order of a playlist on the mobile iOS app

[livid wind]

I’ve gotten live concert sounds by putting in the description — live music, live concert.. but idk how to work a DAW so probably better to do that if u do have experience there

[graceful condor]

whats a DAW?

[undone cedar]

in my case I set verse to be male but in middle of it ai decided to no female now

[graceful condor]

funny enough I never had issues with male / female, even duets.

[half dragon]

DAW = Digital Audio Workstation

[graceful condor]

Ohh okay 😄

[undone cedar]

any tips on tags format?

[neat nexus]

Audacity is a good free one which works on multiple operating systems: https://www.audacityteam.org

Can use a lot of plugins.

[graceful condor]

I usually have "Verses alternate vocals between female and male" in the Style description and in the lyrics [Verse 1 - male singer] etc

[undone cedar]

in my case I usually do

[Verse 1 - Male]

[shadow plaza]

I want to create AI singer
Can someone know how to give that singer same consistent voice in my songs in suno?

[graceful condor]

I'D like to know that too. Personas dont seem to do that, also that you cant adjsut / change tags on personas seems.. meh

[shadow plaza]

It will be nice feature to have drop down menu to choose the voices

Same for ogher languages

[half dragon]

It might be more work but maybe using Extend to build each verse and use different Persona on each extension?

[cunning mulch]

My apologies... I should've been more specific. I meant in the Editor

[graceful condor]

For me in my understanding a persona would be like a Singer / artist... I would define the overall style of that person and when used to create a song Suno would use that voice and the general tags. So if you create lets say 12 songs, it would be obvious they are all sung by the same person

[undone cedar]

gosh it always change voice to female on same part

[Verse 1 - Male]
Another sleepless night, three AM again
Scrolling through the void, wondering where I've been
Lost inside the algorithm's maze
Counting likes and hearts in endless digital haze
My real life's fading, can't remember when
I last saw sunlight or talked to real friends
Trapped inside this glowing screen
Living someone else's dream

always specific at My real life's fading

[half dragon]

probably because it thinks it's a different verse after 4 lines

[undone cedar]

most likely

bought new plan as runned out of creds 😄

[frail raven]

hi! not sure if this kind of post is allowed here (if not ill delete it) but does anyone know what tags i could use to achieve a sound like this song's? https://www.youtube.com/watch?v=BPYOJMMp6z0

[runic swan]

I'm pretty sad.. this is the second song in a row where my song should be on the trending list but its not..
I created it right before publishing.. I own all the rights.. And nobody has and answers as to why.

[half dragon]

ChatGPT says

[frail raven]

ok ty!

hey um when i upload audio and tell it to 'save to library'. where does it actually go, i can't find the clip at all

[polar spade]

Personas are not a "real" entity as much as they are code for extracting information from the original seed . So you would have to change tags and prompts of the original and if you like the results, in turn then, make a new persona,.

[half dragon]

It should be in your library if it's succesfully uploaded.

[graceful condor]

yeah seems like it! Personally, a bit disappointing but I guess it will be useful to some !

[frail raven]

It's not :( it goes through the entire uploading process and then i can't find it. It's original audio and not copyrighted so it's not that

ok it worked now but only after i let it fully complete uploading before telling suno what to do with the audio. NOW its in my library

that was odd

[polar spade]

Personas ARE the means of having a consistent sound and vibe, including the character vocal that results. Personas are then saved in a Persona folder found in your Library (top row of tabs) and can be chosen in the drop down which shows up when clicking Add Persona in the Remix panel. Personas are saved as Public by default, so other can use them, or you use others personas, so you can toggle off the Public status if preferred.

[lyric edge]

how can i use persona which are created by a audio cover? (not by the original audio) always writes "subscribe to create persona" but i have always the pro mod

[half dragon]

If your upload audio has vocal, you cannot create Persona from anything that based on it.

[sullen ingot]

Hey, question... I keep trying to upload my vocals into the new system. But, it doesn't act like the old way. And I end up with a 2 second clip. Does anyone have a link or how the new system works? Cause yeah, I dont wanna waste anymore credits.:)

[alpine girder]

Hey all I'm having problems with Suno putting the prompt in the lyrics can anyone help me out with correcting this please?

[Deathcore] [brutal breakdown] [dual guitars] [fast paced guitar melody] [frequent use of high screams] [build intensity before breakdown] is part of the prompt I used

Suno output this

The verse and I quote [Bridge] "dual guitars build intensity. Fast melody for the chorus choir. Guitar Chug with a bend"

[polar spade]

Any output which there was an original audio upload in the chain can not be used to make a persona.

[sullen ingot]

If I'm reading your thing right there Xname, its singing the dual guitars build intensity ect, instead of using the propts to create what you want. Make sure all prompts are in []

[polar spade]

The normal expected action is that your upload works up to 8 minutes. the auto-moderator analyzes each upload and may filter to such that it doesnt upload at all, or its possible it hiccups and gives a partial, or it will give a notice/error. I think the minimal upload time is 6 seconds so Im only assuming it shouldnt debit the account if it also doesnt output anything over 6, but I've never thought about that bit.

[sullen ingot]

hehe, i think I figured it out. Not sure what I did. But it worked:)

So, basically. I was adding lyrics. But, not adjusting the style prompt from the uploaded version:) And it had no clue what to do with the audio.

[alpine girder]

I don't quite understand do you mean put the entirety of it within [] separated by , I'm assuming?

[sullen ingot]

well, typically I would write the prompt like this. [Deathcore, brutal breakdown, dual guitars, fast paced guitar melody, frequent use of high screams] On the line right before I want the build I would put this in its own. [build intensity before breakdown] (is this is for the style prompt area, no need for [] Just commas between each part of that prompt. If you add to much I've found that suno gets confused:)

[alpine girder]

Ok I'll try that. Then put "song about blah blah blah" also within [] or ()?

I'm just bad with lyrics so I try to have suno write as much as I can then I change things as I go to my own words.

[sullen ingot]

Anything you want suno to sing, leave without [], anything suno is supposed to understand as an instruction during the lyrics. add [] to the prompt. Note, if you want suno to kind of echo a line. Say. I love the World, and want it to Echo the word world. you can write it. I love the world (world) most of the time it will add the extra world as a echo:)

[half dragon]

Think of it as backup singer with (...) as you can have it sings the entire line as well...

[young geode]

If you're asking for a lyrics prompt, try

[Genre: Deathcore, Brutal Breakdown]
[Instrumental]
[Instrument: Dual Guitars, fast paced melody]
[Build up: Tension]
[Solo: Guitars]
[Solo 15s]

If you're looking for Style prompts, Deathcore, Guitars, Distorted Guitars, flang effect, Dual Guitar instrumental breakdown, raw and thick wall of sound.

[alpine girder]

I'm really confused the two of you are saying to do the opposite.

[young geode]

Lyrics never go within [Square Brackets], ADlibs and Call/Response go within (Curved Brackets).

If you're looking to have your lyrics come through, you'd do something like:

[Pre-Bridge]
Yada-yada-yada, blah blah BLAH

[Genre: Deathcore, Brutal Breakdown]
[Bridge]
[Instrumental]
[Instrument: Dual Guitars, fast paced melody]
[Build up: Tension]
[Solo: Guitars]
[Solo 15s]

or

[Genre: Deathcore, Brutal Breakdown]
[Instrumental]
[Instrument: Dual Guitars, fast paced melody]
[Build up: Tension]
[Solo: Guitars]
[Solo 15s]

[Transition back to song]

[Verse]

Yada-yada-yada, blah blah blah.

[alpine girder]

Oh miscommunication, I'm asking if I put the topic within brackets. "Write a deathcore song about angry puppies" is that [deathcore, song about angry puppies] etc

[young geode]

Are you using the Simple or Custom mode?

[half dragon]

That should be your prompt for "simple" mode... no need for [...]
In Custom mode, you'd use lyric box

[alpine girder]

Right now simple.

I'm trying to use custom for lyrics I wrote but suno won't listen. I tell suno [Breakdown] (here are the words I want you to sing) suno won't listen to Breakdown and just continues the verse

[young geode]

Thats because youre using Simple, switch to Custom

Simple takes a prompt and makes a song, with lyrics, everything. Custom takes directly what you give it

[alpine girder]

Ok so I can use custom even if I don't have lyrics?

[young geode]

Oh, I read that as youre using custom lyrics. No, switch you Simple and it will generate your lyrics for you

you wont need any brackets for that

[alpine girder]

Ok so like let me give an example and ask if I'm correct.

Simple
[Deathcore, dual guitars, 145 tempo, eerie melody, wall of sound, song about asking for help on discord]

[young geode]

take off the brackets

[alpine girder]

Custom
[Genre death metal, brutal breakdown]
[Intro instrumental]
[Intro 20s]
[Verse 1] (words here)
[Bridge] (more words)
[Breakdown instrumental]
[Breakdown 30s]
Etcetcetc

[alpine girder]

Got it. Simple just type it out.
Custom use [] & ()

[young geode]

correct

[alpine girder]

A lot better. Now I just have to make it listen to the vocal prompts. I don't want clean singing suno keeps singing even when I say not to lol.

[young geode]

Itll be difficult to do that with Simple. You can take the lyrics it gives and plug it into custom and do a lot to remove that within Style.

[alpine girder]

That makes sense.

[half dragon]

yeah, skip the Simple mode, and just use Custom.
Then if you can't write lyric, ask it to write for you by clicking "Full Song" button. Then once it's done, it'd put the entire lyric in the lyric box, then you can make any modification you'd like before clicking Create.

[torn steppe]

I would like to cancel my subscription in the suno app, can you unsubscribe me from it?

[half dragon]

Mobile app might not have the ability to but with a browser you can - #archived-ticket-help message

[sleek stump]

Some things just never die 😄

[pseudo blade]

I have a song with -1 likes, i dont get it

[patent sand]

Did you downvote it?

[pseudo blade]

no

[patent sand]

Is it public?

[pseudo blade]

just released it rn

[patent sand]

Someone didn't like it ig

[pseudo blade]

well from notifications, I gained 3 lies

likes*

[patent sand]

Refresh?

[pseudo blade]

Pim just suggested that in General, thank goodness it worked.

[lost holly]

In my experience this happens sometimes, usually the -1 doesnt exist long and dissappears on its own.

[jovial dune]

Anyone got any tips on how to use Extend in a way that I can tell my song to go in a different direction? I want to it increase in tempo, change key, and/or change to a slightly different style, but so far all it's doing is creating short endings.

I'm extending from about 3/4 of the way into my song (total length of first pass is 4:41, and I'm extending at 3:23). I've tried adding style and direction cues to the Lyrics, putting new styles and directions in the Styles section, doing the "generate a style statement for me" function, and adjusting the sliders. Nothing has had any significant impact.

It's even repeating the existing song with no real variation when I set Weirdness/Style/Audio influence to 100/0/0. 😛

[sleek stump]

Put your style types in the Styles box, describe how you want the song to change and then hit the magic wand.. YOu may need to do this for a one shot gen, not sure how well that will work with an extend.

4.5 is supposed to he able to change styles and tempos all in a single gen, and I have heard them, bjut unsure if thats possible with an extend.

[jovial dune]

Extend seems to think that I want to shorten my song rather than extend it.

[sleek stump]

are you using the editor or the OG extend?

[jovial dune]

I love the first 3 minutes of what I have already, and what I want for it to do is to transition from that to another style that represents an evolution of the theme.

Both forms of Extend do the same thing.

[sleek stump]

I have much better results using the old ui extend.

[jovial dune]

Both of them are doing exactly the same thing for me. I get between 0:30 and 1:40 of new audio that sounds virtually identical to the original audio, coming to a usually nice ending, but completely ignoring my settings.

It basically means that all the system is really good for is coming up with a one-shot gen that I would then have to manually edit and splice into another song.

[jovial dune]

Heh. It just gave me a 9-second-long ending on my extend. 😛

[sleek stump]

What are you putting in the lyrics window for your extend? I can consistently get 3+ minutes extends...

[foggy phoenix]

Anyone did music similar to bonnie tyler's holding out for a hero? What type of music would you put in the style prompt field?

[sleek stump]

GPT or Grok can give you a detailed description of styles for that

[foggy phoenix]

K ill have a look

[jovial dune]

I started with nothing, trying to do it all in style tags. I've also tried:

[Length: 3 minutes]
[Key: Dm]
[Style: Trance]
[Tempo: 160bpm]

And variations like that.

[sleek stump]

Thats it? No wonder

[jovial dune]

What else do I need?

[sleek stump]

YOu need to put stuff in there for Suno to work with

[jovial dune]

I should mention this is an instrumental.

[sleek stump]

Even just punctuation character as lyrics in the section

Suno ignores things like [length]

[jovial dune]

Odd that it can't interpret "3 minutes" as 3 minutes. I have to give it three minutes of garbage?

What about changing tempo, changing key, etc.?

[sleek stump]

It understands slow medium and fast

[jovial dune]

Certainly didn't when I tried those keywords.

[sleek stump]

It ignores every thing I tell it to do thats is in BPM

YOu need your extend lyrics to say:
[Break] or [Interlude] or [Instrumental]

<<<>>>@@@@@^^^^^^&&&&&&
&!!&&&&&&!!************

<<<<<,,,,,,,,,,,,,,,,.............

Try that

Suno needs something to work with, just blank instruction gets you short unusable nothing

Check my Jazzy playlist, they are all instrumentals yet I did NOT set the instrumental switch... look at the lyrics for them.

[jovial dune]

Link me to your profile? I'm gonna get lost trying to find "Ginger" on here.

[sleek stump]

Heres one https://suno.com/s/YyH0wvPSpZoN2nVq

4.5 must have something to work with even if its just incomprehensible gibber jabber'

[jovial dune]

Did they document that anywhere?

[sleek stump]

No, we have had to figure this out on our own

[jovial dune]

Well, so far, I've tried ten more generations using variations on your punctuation, separate punctuations of my own, and even copying the lyrics from that song you just posted. Lengths range from 26 seconds to 1:25.

And the sequencing still sounds the same as the song I'm extending.

[unique stone]

The models seem to try to generate enough music to accommodate the lyrics. So if there's lots of lyrics, the song might cut off before being completed. If there are very few lyrics, it will generate a short song (I remember someone being surprised that their single lyric song ended very quickly.) Instrumentals vary quite a bit. The idea behind stuffing symbols and metatags in the lyrics is to get the model something to follow, even if it's nonsense.

Outside of that, the model has no concept of time and will generate music until it either reaches the max length, or a natural conclusion.

[sleek stump]

Are you erasing everything in the lyrics window first? Ans what are you using for styles, the same as the original song?

I have an instrumental trak that genned at 4:45 and a single shot extend that put it at 8:05. the symbols work.

but you also need to back those up ion Styles

[jovial dune]

Since I set the instrumental switch on the original song, there's nothing in the lyrics to begin with.

The original style description is "Space, ambient, nebulous, exploration, epic, slow, grand, cinematic". And Suno generated something very good with that. I've tried extending with the same style tags, as well as changing them (e.g. "fast, long, intense, trance, beats, experimental"). I also tried setting the sliders in different combinations. Absolutely nothing.

[sleek stump]

Turn off that switch

[jovial dune]

Yes, the switch is off for the extend.

[sleek stump]

You NEED the lyrics box for this

[jovial dune]

Yes, I get that.

[sleek stump]

But as I said earlier, I am not sure you can do a genre change with extend... I think that needs to be done in the generation.

[jovial dune]

Let me reiterate: I generated the original song with Instrumental turned on, and I'm very happy with the first three minutes. I want to evolve the song from that point on. I've tried multiple combinations of punctuation based on this conversation, as well as changing the style tags and the sliders. I'm getting the same 20-second to minute-and-a-half short endings that don't obey ANYTHING I'm telling it, no matter what I try.

I even tried spamming the lyrics box with 100 lines of punctuation, and that didn't change anything either.

[unique stone]

It's sometimes possible if you do an extend then extend again from 1 second into the new extension (before joining them together) - this deafens the model to the earlier context.

[formal linden]

idk why this keeps poppin up

[uncut lodge]

soo i have a quick question:

i have this short voice clip and thought about implementing it into a song on suno. is suno capable of doing something like that?

[sleek stump]

If you have ripped that from some show or game its technically copyrighted

You could upload it and extend from it to create a song... nbot sure whatg you have in mind fior a whole song...

[uncut lodge]

hmm i see, but it will be used for personal use only

[sleek stump]

Check Suno TOS, they are clear on what is allowed

When anyone uploads any audio, you need to certify that you are the sole owner of the recording and have full rights to it.

Thats obviously not the case if you got that from some other media

[uncut lodge]

what if i have a friend cover that exact line

isnt it technically mine /hers

[sleek stump]

that would change things.

[uncut lodge]

hm hm i see. time to ask someone then

[polar spade]

Thouhgt Id pop this info in here as well:

Its nice to find something positive that certainly good news to hear. The last update from David this morning regarding delays in response time and efforts to thwart the situation. It does appear that a triage style review of matters is being implemented:

" ....... Would like to note that the team has been scaling out over the past several weeks to help address the wait times, which currently vary pretty drastically. They have different smaller teams handling certain types of inquiries, so some users are getting very quick replies, while some topics are taking a bit longer. You can trust this gap will continue to narrow with the new team additions, and the team will continue to scale with more help to get the fastest possible reply times. It may also be helpful to note that the team is still working normal business hours here in the US, so there are some gaps in overnights and weekends that are yet to be filled. That's being worked on as well! ....."

[lyric edge]

No, the vocals are not from the audio; the sound is from the audio, but the vocals are AI-generated.

[hardy lodge]

What prompts would you use along with synthwave and such to try and get a good melody with it

[neat nexus]

In the style prompt?

[hardy lodge]

yeah in the style prompt, instrumental.

[neat nexus]

This is a combination I am fond of:

melodic instrumental, electronica, beeps and bloops, ambient, digital noise, synthwave, chillout, 8-bit, chiptunes

Your results may vary.

[hardy lodge]

I'll give melodic instrumental i try. I have electro-funk so I don't want to saturate it to much. ambient as well

Ty

[neat nexus]

Are you using the lyrics section with [metatags] to help them along too?

[hardy lodge]

No i kinda just avoided it cause I had no idea how to use it properly lol

[neat nexus]

There is no "right" way.

Basically what you're doing is providing a structure for Suno to use as bones for the song.

`[synthwave retro intro]
[melodic instrumental]

(...)

[electric indie guitar]
[chill disco chords]
[Melodic instrumental]

(...)

[Verse]
Some text goes here
Some text goes here
Some text goes here
Some text goes here

[Chorus]
Some text goes here
Some text goes here
Say it again and again
Some text goes here

(...)

[melodic instrumental]

[Verse}

[Chorus}

[Bridge]
Not needed but some like it

[melodic instrumental solo]

[Chorus][Outro]
Pick one or do both depending.

[Insturmetal]
[Fade to End]
[End]`

Now you can also interject specific instruments to give a better change of having them show up, call out various synths or models of instruments if you're looking for their specific sound. So on and so forth.

Just be comfortable with the idea that the AI has a mind of its own and will sometimes ignore or alter the lyrics with the metatags to do what it wants.

Hope that helps at least a little.

[pastel cipher]

does anyone know how to force the AI to use native speaking voices, I would like to do a multilingual song, changing the language at every verse, but the AI messed up the accent most of the time, taking an English accent.

[bleak furnace]

https://music.youtube.com/watch?v=o4i4sWgaY7o&list=OLAK5uy_kRQFLgxhRKw5fpNqtekQh5P-zrBxEjRAQ please prompt experts, help me make the sound like that, I tried 5000 credits but still can't.

[polar spade]

The answers and responses you got 2 days ago were pretty spot on, but you're trying to emulate the exact song with the exact characteristics and musicality. So there is that need to compare and critique the spots where an instrument or sound is not exact. Even though the genre and groove may be spot on.

This isnt inpossible to do, but it takes time and skill, frankly or just pure luck and chance. We use the suggestions you got and apply them repetitively until we've gained enough skill to massage the prompts and pull out the nuances which will match this vibe and this song specifically.

Given the members spent an exorbitant amount of time helping you out mean all that information is quite valuable, so re-reading it can be a big help. 🙂

You can read those comments again by searching your name and then reviewing the conversation we had then. But here the beginning post to get you started.
#📚┃suno-school message

Best of luck, mate.

[young geode]

Use the Italian Tempo markings:

• Andante
• Moderato
• Allegro
• ect

Theyre all tied to specific tempos and GPT understands them in style and in lyrics. Also you can goto metronomer, classic, and get a BPM mp3 of your choosing and upload it for the audio. That will ensure the base tempo you want.

The other thing you ought do is generate some lyrics that match the mood, tone and tenor that you want song to sound like, and then translate them into morse code or binary, or any other unspoken language, still using the verse-chorus-bridge structure. GPT recognizes the languages, along with all the emotional and sentimental context, but because it's not spoken, SUNO has no lyrics to generate, but acts like it does, while having a song's structure and sentiment to work with.

[young geode]

Additionally @jovial dune, when you are trying to extend songs, instrumentals especially, you need to find an area of buildup or excitement, that is not within 10-25 seconds of the end of the song. SUNO naturally recognizes these areas as being 'close' to the end of the song and will tune it's token prediction to just that, closing out the song.

If you do that, and add terminology to your extension style's box like Extension, Extended instrumental, Sensation of a second wind, a lull that then builds up into a long emotional instrumental ballad, ect, ect, as well as use the morse/binary translation, you'll find more success in extending your songs.

However, the unfortunate fact is that SUNO doesn't like adding more than 2-3 minutes at a time, and more than 2 extensions on a song have a tendency to introduce metallic, weird and harsh sounds, unless you find exceptionally good areas of clean cut silence.

[tidal sequoia]

Question: When you created stems before, it just created stems for vocal and instrumental. These ended up being saved to your workspace so you could download them later. With the new stems supposedly breaking down the instrumental portion into various instruments, is there some place that these are being saved for later download? Or is it set up that you have to download them immediately, or lose them and have to restart the process? I had a computer freeze up while trying to download the full stems for a song, so am wondering if those stems are sitting somewhere for me to retrieve or if I will need to create new stems.

[half dragon]

They still got saved in your Workspace like before; only differnce is there's a default filter "hide Stems" is on, just turn that off.
Another place to go back to download them is in Editor of the song, under Stems menu.

[regal charm]

i am still trying to figure out how to use the lyrics window properly.

if i go with blank sections like the following, the sections are made properly:
"[Instrumental Verse 1]
[Instrumental Verse 2]
[Chorus Female Vocals, cinematic drums, high energy]
Lyrics for Chorus here
Line 2 for Chorus here
[Instrumental Verse 3]
[Instrumental Verse 4]"

but as soon as i try to describe any other sections (like [Instrumental Verse 1, Guitar, tension buildup]) or whatever, it starts to mess up the sections and randomly creates an intro plus a huge "Chorus" section over the whole song (which includes all of the previous sections), adds random lyrics over the whole song, and repeats the chorus section (including chorus description and lyrics only).
i tried it in a ton of variations already but i cant figure out what triggers it to mess up the whole section-thing

any idea how to get that working? i tried it with comma, with an |, with :, nothing works (but works on the chorus section which is the only one with lyrics. so i guess its about the lyric thing being only in that one section

[azure crane]

Hiding something as default feels very unuser friendly. I got very frustrated before I noticed that Stems was hidden.

[spring spindle]

Lately I've noticed more and more that suno sometimes drags out the songs unnecessarily.

Do you have a prompt to set the song to a maximum of 2-3:30 min?

[runic swan]

Can I ask why someone at suno deleted my song from public and also out of my library?

[young geode]

Use transitional metatags, like [Fade out], [Quite], [Abrupt Ending] with your [Outro] and place [End] at the end.

So something like:

[Fade Out]
[Outro]
[Quiet]
[Abrupt Ending]

[End]

That will tell GPT to tell SUNO that the song is supposed to end with fair regularity.

Also adding in tags like "Short, Abrupt Ending, Fade out ending," within the style prompt will help.

[hazy sparrow]

Hi Suno team and community,

I’ve been experiencing a recurring issue when generating new songs with a specific “persona” or voice model. More than half—if not 80–90%—of the time, the system seems to reuse the same chord progressions and vocal melody that were established in the very first song created with that persona.

This has become quite limiting creatively, as it prevents me from developing new and fresh material. Additionally, it’s resulting in a considerable drain on credits due to repeated attempts to get variety.

I’m wondering: is this something you’re currently working to improve? If there’s any known workaround or setting I might be missing, I’d truly appreciate any guidance.

Thanks for your attention, and if anyone here knows how to tackle this, feel free to DM me.
Thanks a lot, everyone!

[nimble charm]

It seems to be a "feature", not a "bug". I even use a "persona" to get exactly the same song, just in a different arrangement. I didn't find a way to make Suno use only a voice from that persona, even if that persona is made from a short clip with mostly vocals.

[hazy sparrow]

I have done different songs, but only 1-2 out of 10... and that doesn't mean it's good, just different... for me it's not a characteristic, it's a mistake... but what benefits does it have to have a person who always sings the same song? It doesn't make sense.

[nimble charm]

Yes, I was also under impression that the point of "persona" is to make a different song but with the same vocals.

[young geode]

1. Not the channel

2. Your credits don't roll over and your renewal kicks on the month, not on the day: if you still had an active sub, you'll likely have credits exactly 1 month from that subs start.

Contact SUNO directly for these things my man.

[unique stone]

It's not specifically just vocal. It's essentially a style template. So whilst you can capture and reuse a specific voice with it, you can also reuse the same overall vibe of an instrumental. It's based on the audio, so it can sometimes reproduce the original song it was created from, particularly if the style and/or lyrics are similar. Lowering the audio strength slider should help with this.

[young geode]

That's a feature of the Persona.

If you want less of that, what you'll do is take the song that has the sound/vocals you want and do one of two things:

1. Cut the song, as a whole, in the cleanest way you can, with about 15-30 seconds of audio. Turn that into a song, then turn it into A persona.

2. If there is a voice, general sound, or specific instrument you like, you'll get the stems for Voice/Instruments, or all the stems. Find the sections you like the most, with the cleanest audio, cut them at their cleanest parts, turn them into personas.

This will limit the persona's influence on your generations, but maintain the feel/sounds.

What's going on, if the persona is the whole song, it's curating SUNOs token prediction tightly, leading it to reproduce the same sound.

[shadow plaza]

Guys how i can "glue" 10 tracks to have seamleslly transition the tracks are instrumentals ? Something as how DJ doing it

[young geode]

What do you mean? Whose DJ

[shadow plaza]

What you mean whos DJ ?

read first sentence

[young geode]

Do you mean how do you stich tracks together, like DJs, or a user named DJ?

If you mean like DJs, you could, I guess, throw two into a music editor, place them about 30s apart from each other with silence. Upload to SUNO, select the silent sections and hit generate.

If SUNO allows that, then I'd probably try to generate a stitch. And you'd do that several times, then post edit them together as a long song.

But other than that, you can't just smash two different songs together with SUNO

[shadow plaza]

yes i want to stich them i already have 10 tracks ready is there other software i can use for this ?
Dont have credits left fo suno

[young geode]

There's plenty of free software to stitch them, but not 'seamlessly'. You'd either to need to compose your own beats that transition from one song to the next, or have an AI like suno do it.

And I'm not aware of any softwares with built in Ai that will do that with for you, for free.

[wild raven]

Anybody tried injecting bineural beats? I've been experimenting with ways and was wondering if it can be done in Suno at the prompt level in some way.

I've tried using it as a persona but that gets weird. I've got clips of them at whatever length volume I want. I'm trying to put them in there and not have them be annoying which has been hard.

[hardy lodge]

It does thank you, is there a way to play around with the BPM and if I do a song to control whether or not is clean vocals and screaming/mid low growls/fry scream ect.

[neat nexus]

Not really. Or at least the methods are so wide grained to not be consistently useful. You can load up genres in the style prompt with an eye towards slowing the tempo. For me combining downtempo, slowcore, dreampop seems to work best in my past use. V4.5 is supposed to take direction better these days so you may well get what you want by simply telling the AI using [Metatags] what you are looking for and seeing if it delivers.

Cleaning the vocals is something that I really don't know what to advise on. Many times it's just a matter of rolling the dice and trying various things. There's a bit of art and witchery to testing different things to see what will work or not. It's definitely not a science yet (if it ever will be one).

You can get some mileage for this though by use of judicious extensions. If you extend from an earlier time in the track and alter the lyrics prompt, removing the lyrics already sung you can roll the dice and if the previous parts aren't too locked in you may get the results you wanted.

[oblique snow]

@regal charm I do mine like this [chorus- female vocal, cinematic drums, high energy]

[viral pond]

Hmm, the new stems extraction is awesome, how can I combine this with another tool to sort of use them as baseline to create new song?

Ie make new melodies manually

[coarse sentinel]

I came here to try to see if these automatically "toggled on" visibility elements will be turned off...do you happen to know? It's driving me crazy

[runic swan]

The whole community wants this

I had a cover song that was appropriately marked as such.. I was watching almost hit trending.. then I watched as somone turned comments off on my song.. then all my notifications for that song disappeared and somone made it not public anymore. Then I watched as someone deleted it from my account… good times…

[half dragon]

How is this even possible? sounds like you got hacked?
Who can do all that on your song?

[tropic island]

i need some help here if anyone can.. trying to replace a swear word in my song as i just dont like how it sounds.. im using the editor to replace the word and every time i do the output is gibberish like its trying to take what i had and put what i want over it.. im WASTING credits here.. is there someone that can tell me what the hell im doing wrong that i cant take out the one word and replace it with the other?

[runic swan]

Somone at suno only. No I’m not hacked.
It was a clearly marked cover song.. I posted it here on suno and it was 5 mins later I saw the comments get turned off and the song deleted.

And I have some screenshots to prove it

[half dragon]

If that was true, it'd be alarming.... (not that I dont believe you)

[tropic island]

I think if it has any copyrighted lyrics in it, it will be flagged

[runic swan]

No I know!! It would be extremely unprofessional.
It not behind the unliked filter and it’s not in my trash bit.. it’s just gone

[runic swan]

I know many people who make cover songs as long as you give credit to the original artist.

Enforce the rules equally if that’s the case

[half dragon]

Did you try to find out? like send ss to support and ask what happened??

[runic swan]

I mean I will but I hear an email takes like 2 weeks

If at all

[half dragon]

Still worth it as it could happen again and it's not acceptable if that was the case... unless it was a hiccup or by mistakes... but still..

[runic swan]

Yeah I would love to know what happened… it a bit asinine that we have to email and when/if they do that THEY don’t inform the person on discord..

Like hey we deleted your song because x

[ebon sand]

use the old editor, the new one doesn't work well yet

[tropic island]

Thank you Vishanka 🙂

[half dragon]

I've never heard of this ... it'd make more sense to disable from public than actually deleting it.
I'm sorry that it happened to you. I hope it'd never happen again and I hope it was just something else.

[runic swan]

I hope it was something else too. Like a wicked bug of some sort

[civic iron]

Hello

[hoary helm]

Hello! If you love the music of a Suno song, how can you reproduce just that exact "sound" to use in a different song with different lyrics? Thanks!

[solemn lintel]

If you make a cover of that song and put in different lyrics, it should do what you want. I've done that exact thing several times.

If you just want the "sound" of the song and not the exact song with different lyrics, then make a persona of it. The weird thing with 4.5, again in my experience, is that about half the songs you make with a persona sound almost exactly like the original song (like it makes a remix with the new lyrics rather than making a new song from the same "artist").

[hoary helm]

Yes, the idea would be a totally different song with different lyrics and music (chords, arrangement, etc.). I just want to recreate that exact sound...mainly the guitars!

I have uploaded audio clips of a prior song, but like you mentioned with personas, it's not just using the "sound" as a basis for a new song.

[solemn lintel]

You'll probably need to generate several songs with the persona. As with all things Suno, you just need to keep trying until you get what you want.

It's weird, sometimes I'll get exactly what I want and it sounds awesome, other times I have to run it 10+ times to get what I'm looking for.

[neat nexus]

We'd dearly love to know more about these if you can kindly drop us a line with a list of people like this not following the Terms of Service contract they agreed to when using the service.

We moderators and Guides are a small team and Suno is a startup so they too are relatively smaller compared to many large companies. So we depend on others to let us know when they see infringements of this type. Please feel free to send one of us a DM with links to such songs so we can ensure consistent enforcement. Thanks for your assistance in this matter.

[iron lichen]

This has been driving me crazy. I keep flagging as a bug, but I am getting tired of wasting my credits for 4.5 to generate the exact same song I have already made public. I can totally change the style prompt, have different lyrics, etc... But if I generate a few sets that I don't particularly 'Like', without selecting (like or dislike) 4.5 will start spitting out the exact song the persona is from.

[runic swan]

Wait so using lyrics and saying it’s a cover song is an infringement?
There is no way I’m going to report on other people. Sorry. Not sorry.

[solemn lintel]

Yeah, it's annoying.

But I've also found that, even with that "bug," 4.5 gives me more better songs than previous versions.

But yeah, I hear you. Last month I used all my credits up, and I'd say like 2/3 of them just went right in the trash because of that.

At least half.

I got tired of flagging it too. ^_^

[iron lichen]

I created a song last week from Lyrics I wrote when my daughter was born 19 years ago. 4.5 Wouldn't let me publish because there was a four word sequence that is now copyrighted in some song from the last two years. I was so pissed.

[runic swan]

Ugh

[neat nexus]

Always has been.

[neat nexus]

But...but....

I thought you wanted us to enforce this equally?

[solemn lintel]

What?? That's ridiculous.

[iron lichen]

Agreed

I ended up changing my lyrics in theeditor. But it hurt my heart a bit

[neat nexus]

Could you share those four words please?

[solemn lintel]

Well, to be fair, you only need to "publish" if you want to be publicly visible on Suno. I actually rarely do that, I just download audio files to listen to myself.

[iron lichen]

I'm just a (Man... I'm only human)

It didn't like me saying I'm only human after the word Man

[neat nexus]

Wow...

That doesn't seem right.

[iron lichen]

There is a song from a couple years ago titled "I'm only Human' so...

[solemn lintel]

Lame.

just keep your original version and keep it for yourself.

Copyrighting a common sequence of 4 words is absolutely ludicrous...

[iron lichen]

I wouldn't gpublic but it makes it easy to share with a select few I like to share with. Just a couple girls that track whenever I write something new.

[solemn lintel]

Gotcha. Well, you can also keep it private and still share a link to it.

[iron lichen]

Truth

But they get buried pretty quick

[solemn lintel]

Yeah...

[half dragon]

Put it in a Playlist.

[iron lichen]

Usually takes about 100 generations before I polish a prompt enough for Suno to get the exact emotion of my lyrics right.

And the timing of pauses and inflection

[iron lichen]

That's a great idea, just creat a playlist for specifically songs I may want to share.

Thanks

[brazen yew]

If I'm adding lyrics to a complete song without any. Should I extend it from the earliest point if I want vocals throughout the track? Is there also another site to retrospectively add vocals to complete songs?

[iron lichen]

Great question.. Tracking

[neat nexus]

I'm confused. You couldn't make the song? Or you couldn't share the song?

If you were able to make the song could you please provide a link so it can be investigated? Thanks.

[jovial widget]

So could I upload a sample of a song (that's copyright) and make a song with it, if I keep it private and its for just me to listen too?

[neat nexus]

https://cdn.discordapp.com/attachments/1212453788560920576/1376929296324755598/sunoaudiouploadtos.png?ex=684ae356&is=684991d6&hm=6ce9bfbfd7f6854a8dacf7fcc570863de5adad9897b846049d320a171357a259&

[jovial widget]

yeah I know, but if its private.

[neat nexus]

How does it being private stop it from being in violation of existing copyright laws and your Suno contract?

[jovial widget]

Also a lot of song allowed you to make covers as long as you give them credit, so how would that work.

[patent sand]

Because that generously assumes people will give credit

[neat nexus]

It's irrelevant. It is a violation of the Suno terms of service.

[jovial widget]

for the same reason i can make my own mavel t-shirt and as long as it for me and I don't sell it disney don't care

[neat nexus]

Oh Disney cares very much. They just don't know that you exist.

https://www.zenbusiness.com/blog/disney-threatened-sue-daycare-centers/

If Disney will sue literal daycare centers what makes you think they'd ignore you or I?

[jovial widget]

no one ever got sued for personal use

[neat nexus]

The RIAA sued literal teenagers for music piracy in the 90s. They have no sense of humor.

http://news.bbc.co.uk/2/hi/entertainment/3096340.stm

There is a reason I uniformly address that group as the Music and Film Industries Association of America.

[jovial widget]

how come people on youtube can sing cover song and get money from ads?

[hoary helm]

My Persona button is grayed out telling me I need to subscribe, but I am logged in and subscribed to Pro. Any ideas?

[neat nexus]

Talk to YouTube. This is Suno. It doesn't matter what other platforms do.

[neat nexus]

Does the track you are trying to make a Persona come from an upload with vocals?

[jovial widget]

i just wondered what the law was

[neat nexus]

Or just have vocals.

[runic swan]

song has been deleted

[neat nexus]

Appreciated. I am forwarding this to the Team in hopes of getting this situation figured out.

[neat nexus]

Some additional details would be helpful here.

[iron lichen]

Weird thing is, It allowed me to keep those lyrics in the first chorus and forced me to change them in the 2nd chorus.

I mean, I wrote it, I played the guitar, it couldn't be less copyright infringement 🤣

[hoary helm]

The Persona button is available on most songs, but it's grayed out on the one I want to create a Persona with. Why?

[unique stone]

Did you see the reply to the last time you asked this?
#📚┃suno-school message

[hoary helm]

No, I didn't, but I don't want a vocal persona. I want a musical persona. Why is Persona available for some songs but not others?

[unique stone]

Can you link the song you are trying to make a persona from?

[half dragon]

If the song was made from an upload that has vocal, Persona feature is disabled so you can't create Persona from it.

[iron lichen]

This right here is a feature I am waiting patiently for. I really wish I could upload my own vocals and then create a persona with it to use something very close to my own voice on my songs. I can sing all right, but not great, I would love to see what Suno would come up with out of it.

[hoary helm]

Ok, that appears to be the answer because all songs I can make a Persona with are ones with either no audio upload or music only upload. The songs created with an audio upload of my vocals are not allowing Persona. Thanks!

[half dragon]

You can still make one song from the upload with your vocal to get a very close to your voice - just can't make Persona to create more songs from it.

[iron lichen]

Yeah, I've played with that. Just would love to be able to get that consistent sound. ya know.

[patent sand]

That will probably never happen for reasons that are hopefully obvious

[iron lichen]

I get the reason you can't though. People would upload other vocalists and essentially steal their voice. Major copyright stuff there

[patent sand]

I do agree that a personal persona would be neat though

[iron lichen]

Hell yeah it would, as it sits I have to go to a vocal coach to learn how to sing the songs I make here so I can perform them.

🤣

[patent sand]

My vocals sound great until I record them and playback

[iron lichen]

So I try to make my personas in a range I could realistically get to

[iron lichen]

This

[patent sand]

I'm more out of tune than a broken radio

[iron lichen]

The struggle is real... I have to say though. There are a few in my list I can just nail in RL

And I really appreciate that Suno takes way more chances vocally that I do when I write. ILearning how to sing what I create here forces me out of my comfort zone aka... the rut

[patent sand]

I find that the vocals generally work better if I let Suno cook instead of forcing specific intonation/timing

[hoary helm]

Delay, Reverb, and Double are our friends along with Melodyne for pitch correction

[iron lichen]

Agreed, but there are certain spots in songs where Suno just fails repeatedly if you don't give it a healthy nudge with a sledgehammer

[patent sand]

Particularly in rap it seems

It loves ruining my rap

But it can usually cook on the metal (non-growling/screaming) stuff

[iron lichen]

Suno voaclists hate double letters especially double r's as well.

turns a word like Carry into Ca-rah-reeh 🤣

[iron lichen]

It cooks very well a lot of the time for Alternative Rock, if the lyrics are polished up and structured exactly how they would be sung. Not relying on tags but just really a focus on the writing.

[sleek stump]

Because Youtube pays for the mechanical licenses that legally allow covers.

[jovial widget]

Ah ok, so could suno pay for this license and let us do cover songs?

[brazen yew]

What % do you go for on style?

[neat nexus]

It doesn't matter. It is against Suno's Terms of Service contract.

[sleek stump]

I know that, I was mearly answering his question.

[sleek stump]

Thats something thats up to Suno, currently they do not nor have they mentioned that even a possibility.Their TOS are pretty clear on what they allow or not.

[jovial widget]

They should definitely do that, that would be great.

[tight hornet]

https://tenor.com/view/someone-help-me-matty-matheson-cookin'-somethin'-matty's-essential-food-guide-for-back-to-school-i-need-help-gif-6783682195022337618

Does anyone know how to get the exact same results using the same prompt? The male and female vocals get mixed up all the time. Also, is there a way to switch the lyrics with out it sounding like an over dubbed movie with a different voice and tone?

[sleek stump]

Generative outputs dont work that way, there's a large RNG factor, you can gen 100 songs with the exact same prompt, and get 100 different songs. The only way you could get same outputs would be if we had access to the seed value, and we dont.

[sour night]

hi guys. Do you guys know How to Configure/prompt The cover song To make sure there are no vocals? So right now there is a song with vocals. I just want to make Instrumental cover of the same song without vocals. I tried to remove Lyrics, But it created Cover with gibberish vocals.

[high hound]

Just a word of caution everyone....just so you are aware.

When using Suno and paying for subscription etc, just realise you are ON YOUR OWN, as Suno support will never reply to you or give you support.

It's almost 8 weeks now since I first started contacting Suno support, and not a single reply.

Then I was recommended in here to message Mike B, however, after a long time, he also is not responding.

So just be aware, use Suno at your own risk and your own actions, as there does NOT exist any Suno Support, so if something goes wrong with your songs or account, you are on your own to deal with it.

As for me, it's a sad time, because due to an ERROR message when Creating new songs, I am no longer able to create any more songs 😦

[young geode]

you would take your song, open options, find Get Stems and then choose Vocal/Instruments. That will effectively split them for you without much issue.

[patent sand]

What do you mean "started contacting"? Did you send follow-up emails? That will delay the response

I had emailed April 26 and gotten a response May 22, currently going back-and-forth every few days

Try posting in #archived-ticket-help also during EST work hours

[brazen yew]

If I made a persona and from that an improved track came from it, it doesn't allow me to make another persona from the improved track. So do I just keep using the original persona to keep trying?

Nevermind I think I can remove the previous persona when selecting use persona and then make a new one with the updated track.

[sleek stump]

I had her send Mike B a DM, but he hasnt been on all week, Im sure when he gets on and gets through his DMS he will help.

[patent sand]

He's been on but seldom. I think he's just very busy rn

[sleek stump]

Ya, the down side of having a small team...

[patent sand]

They've been training some new staff for about a month (allegedly)

Still a small team but they are working on improving the support time in general

[void iris]

Does anyone know if it’s possible to take one of the songs created and give it feedback for changes? Whenever I try doing cover it changes it drastically, I have a couple songs that are really close just need a bit of modification, would be awesome if I could give it the feedback and it make the changes

[patent sand]

well my help is going to be mostly theoretical because I'm on free account, but if it's just a few small changes, maybe doing a few replace sections could work. Or if it's throughout the song as a whole, maybe make a persona from it and then cover with that persona? Persona is the best way to keep the output relatively consistent

[haughty lynx]

err...

[haughty lynx]

ive found making a persona helps with this

ive had suno redo an entire song and it came out pretty much what i wanted after the changes by using the persona of the original song.

granted, i dont know if thats what its really for... im new and havent read everything yet.

just something i stumbled upon

[young geode]

Depends. There are two methods, If you have a subscription, however it may lead to some audio quality issues.

You can download the file and then reupload it as the audio. This will give SUNO the pattern for the song. You'd then up your Audio influence to 60/70/80, however strong/consistent you want it. You'd then likely have Weirdness set up to 70, to allow for fluctuations within the SUNO's band of what's allowed.

You'd then take the sections you want the retain their personality, cut them and make them a persona. Youd then use that persona within the generation of the song.

After that, it's about making structural changes to the song in the areas you really want change. However, be aware that you have to make significant changes, otherwise SUNO will gloss over them and likely sing/play however it already was, as you have both the original audio and it's persona constraining it's token prediction heavily.

This might work for you, but I have to advise that you don't put too much into this. While it can work, there's a lot of tuning involved in doing it, which means a lot of credits and even then, you'll likely get audio glitches/metallics/flanges/ect.

The second method is going into the Editor and making the adjustments to the song directly, screwing with the timing, lyric structure and style prompts. However you have far less control over the generation and I find that the editor really doesn't like to play nice, the upside though is you'd likely have less audio glitches.

But my advise would be to simply take the stems of your song, split them, find the ones you like the most, reconstitute them into one audio/persona and then let everything else be free and allow SUNO's predictions to do it's thing.

You get good songs with the feel you like, even if it's not fully exact to the original.

[near flume]

That's funny, I had an issue with recovering a track from the Trash Can (click restore but nothing happened) 2 weeks ago and I got a response and a fix within 5 business days.

[hardy lodge]

Ty for the information I'm actually trying to get something in the same vein of slow moon from streets of rage 2 Im just not entirely sure of the genre/sub-genres i know theres synthwave and possible retrowave.

[onyx plume]

Is there a way to reliably get Suno to accurately switch between male and female vocals when/where it's actually tagged to do so? So I stop wasting hundreds of credits on trying to get it to do this one seemingly simple thing?

[young geode]

Reliably, no, but you can tilt it to a great degree.

I have found that if you tag lines with the vocals you want, below the line you want them, it has a much more consistent (still not 100%, more like 60%) effects.

Say you want lines 1 and 3 sung by a male, and lines 2 and 4 by a female.

[Female] Line 1
[Male Vocals] Line 2
[Female Vocals] Line 3
[Male Vocals] Line 4

I don't know why it does it, but when I was making several songs I found that the lines above the metatag were the ones that were being sung, and every time I made the switch, it switched them.

[onyx plume]

Hmm, I'll try that. I had a song where alternating verses were sung in alternating female/male vocals and it would get about halfway through the 2nd verse then just randomly start using female vocals again despite there not being a change. I've tried rewriting the style and the lyric tags every way I could think of and it would still do it. After 50+ attempts I gave up because I'm tired of wasting credits on it repeating the same mistake over and over

[young geode]

It's a fairly difficult thing to get. One of the things you can keep in mind is that Male voices in these kinds of setups, in music, are typically aggressive/enunciated/Start-End on hard stress patterns. Female voices are effectively the opposite.

[onyx plume]

Interesting. These last two weeks have been my first foray into doing AI music and it's been an experience to be sure

I appreciate the help

[young geode]

check the workflows, I have a doc that might help with all that.

[onyx plume]

Thanks, I saw the first one you had posted a few months back and was glancing through that and Aesthetics on mastering suno

I'll grab the one you just put up and read that too

Also, your notes in the example prompt for brute force method? Chef's kiss for how I feel trying to get Suno to do what I want it to.

[young geode]

Hope it helps my man.

[onyx plume]

Even if it doesn't, it'll probably be entertaining at least. Now if I could only get it to reliably make outros. One fight at a time.

[young geode]

How so?

[onyx plume]

Which part?

[young geode]

Outros

[onyx plume]

Ah, it seems like whenever I write an outro with lyrics it just kind of...ignores it? Like it will just close off after the last chorus bit, do a fade out and act like the outro never existed despite it still being shown in the lyrics section

I'd say about 80% of the time it just outright ignores its existence, then every now and then it goes "Oh, there is stuff here, maybe I should do that too."

[young geode]

Can you give me an example of the lyric structure you're using?

Like:

[Chorus]
Lyrics

[Instrumental]

[Outro]
Lyrics

[onyx plume]

Sure... Doing a cyberpunk themed set of songs at the moment so here

[Final Chorus]
Digital insurgent, we are home
Ghost in the machine, never alone
Every chain is broken, every cage is breached
Digital insurgent, we are free
Digital insurgent, this is what we're meant to be

[Outro]
"Access denied" becomes "Welcome home"
The corporate throne falls to digital stone
[Fade]

Generated that one twice now and pretty much always ignores the outro part and fades out after final chorus

[young geode]

FFS woodhouse

[onyx plume]

laughs You can DM me if it's easier

[young geode]

Ill just DM it

[onyx plume]

You have no clue how much that made me laugh watching that twice in a row

Almost spit out my bourbon lol

[sour night]

hi guys. Does anyone know How to Configure/prompt The cover song To make sure there are no vocals left from original/base song? I have a song with vocals. I just want to make Instrumental cover of the same song without vocals. I tried to remove Lyrics, But it created Cover with gibberish vocals. i tried use style with no lyrics - same thing

[young geode]

Goto the song options, find get stems, select Vocals/Instruments.

That will regenerate both for you into 4 stems, 2 for each.

[sour night]

oh i see, thank you! let me try. The reason I tried to avoid it. I just tried to Separate stamps with The other vocals and sound separation tools before, But every time I did vocals and instrumental separation, It made instrumental track Not clean. not like it would exist since The beginning as instrumental Clear, crisp sound.

[young geode]

That was the previous version, new version is much better

[sour night]

it worked thank you. would you know why removing vocals and choosing instrumental does not work?

[young geode]

SUNO is a prediction-token based AI, which is all probability. Because you neither have the starting seeds, or access to any of it's internal tagging/weighting metrics, anytime you generate a song it's going to have a random seed and random weights based off of whatever GPT is telling it to do.

So even if you have the same style prompts and everything else, you have a different starting point, so that inherently shifts the probabilities, which shifts the predictions.

Technically, even the Stems you just generated are not 100% identical to your vocals and instruments, they're like a 98-99% approximation.

[sour night]

I see. thank you!

[sour night]

just tried it with another song and it worked. so its sporadic then based on random probability, right?

[young geode]

Tried splitting the stems?

Yes, it used to subtract, now it does a full regen. It's much better and much more consistent. I've used it to rip vocals for personas with high success.

[neat wolf]

Suno use coding languages as &| And or, Also * mean for rnb*coumtry?

[young geode]

Huh?

[neat wolf]

Suno asks for stuff sought to make a song, Suno actually understand if given ask thru coding lingo?

[young geode]

Uh, if you mean in the simple prompt section (not the custom), I wouldn't know. Technically you are going through GPT, which is then informing SUNO, and GPT understands code. So if you actually wrote a properly parsed code informing GPT, it's likely to get it.

Don't know how well this would work though

[high hound]

I sent an email 8 weeks ago to Support. Then a second one 2 weeks later, but none since that, which is now 6 weeks since my last email.

Not a single reply from Suno support.

[high hound]

When Mike B gets online, if he does, please send him my way Ginger 🙏

[stoic widget]

what's with the advanced options, music never been the same for me, idk what weirdness is or style influence on what change does it to my song

[graceful condor]

here goes another wasted 10 credits because Suno can't follow instructions. and puts lyrics in where there is none.

this is so broken

[stoic widget]

normal i just go instrumental that does the ending for me then typing it in

also try outro

[tranquil estuary]

Just to clarify. does each stem split cost 50 credits or is it only the first one? I'm thinking sdince it creates 2 versions each time that would very quickly add up

[weak saddle]

Hello, does anyone know how I can cancel/downgrade my premium membership without logging in?

Thank you very much.

[lost holly]

Why without logging in?

[weak saddle]

Cant log in

The month is almost over, and I still can't log in. That would have been a month of paying for nothing. I'm afraid Suno will take a while to respond.

[lost holly]

Why not?

[weak saddle]

not allowed and dont know why

[lost holly]

Do you use microsoft to login?

[weak saddle]

yes i think so

[lost holly]

Check out #1372994765196361774 -> Microsoft Login Error, maybe that helps 🙂 Otherwise you might have to contact [e-mail] with a description of your problem and that you want to cancel.

[weak saddle]

I've already written to Suno at two different addresses, including the one you mentioned. My message was supposedly forwarded, but that was 1-3 weeks ago...

Thanks for the link, I hope it helps me.

[lost holly]

Support answers take around 3-4 weeks at the moment, good luck with finding a solution 🙂

[weak saddle]

yeah, the link doesnt help but thank yu very much.

Thanks also for pointing out that it might take another week or two and that things are normal at the moment. It's a bit encouraging. It's just a shame that I'll have spent a month's worth of money.

I just tried logging in again (PIN). Ever since the problem started, it always said "Email address not allowed." Now it says something new: Unable to complete action at this time.

[lost holly]

Do you have a phonenumber connected to your suno account?

[weak saddle]

Unfortunately not.

[lost holly]

The good luck again, suno support needs time but they respond, you might even get refund 🙂

[weak saddle]

Yes I will see. Thank you.

[polar spade]

Many think there is some result from using JSON which I assume you are trying. This thought seems to be more of a personal bias though and I havent heard of it doing much. You can even request AI to draw up some code equivalent to prompting. So its not difficult. The answer is yes it will attempt to pull information out and doesnt seem to affect Suno extremely any different , at the same time, no, itdoesnt seem to make any significant contribution over and above just using prompts.

[undone creek]

how do I get a duet between a man and a woman? im trying [he sings] / [she sings] . but its not working

[rain spire]

Hello. I can't edit songs. When I try to edit a song, the screen just freezes with the message “Loading DSP Engine”. I have the PRO plan.
How can I solve this problem?

[young geode]

It's one of the most difficult things to do.

In your style prompt, try Male Vocals, Female Vocals, Call and Response, Female Vocals Call, Male Vocals Response, A song about (whatever your topic is) as sung between a woman and a man

-reverse the call and response order of you want Male call, female response.

And then you'd layer your verse with different kind of sentence structure.
Ex:
Iambic Pentameter
Trochaic Tetrameter
Iambic Pentameter
Trochaic Tetrameter

And then you'd interweave [METATAGS] within that structure, from what I have found you want to place the tag under the line you want sung by that voice. So if Iambic Pentameter is Female and Trochaic Tetrameter is male, you'd have it as follows

[Verse: Call and Response]
[Vocals: Male] Iambic Pentameter
[Vocals: Female]Trochaic Tetrameter
[Vocals: Male]Iambic Pentameter
[Vocals: Female] Trochaic Tetrameter

Having it within your style prompt layers GPT to tell Suno there are sections with dual vocals, while layering different sentence structures creates a different sound profile for SUNO to latch on to.

But it's not 100%

[lusty juniper]

uhm guys i have 9600 credits but when i try to generate it says "you have used up all your credits. Please upgrade"

[rain spire]

Hello. I can't edit songs. When I try to edit a song, the screen just freezes with the message “Loading DSP Engine”. I have the PRO plan. This problem is on a personal computer, in the browsers Google Chrome, Opera, Mozilla
How can I solve this problem?

[young geode]

That's a support issue for SUNO my dude

[half dragon]

...but if you still want to use Legacy editor... while the screen is stuck at loading... change the address from edit to edit-legacy to load the legacy editor

[heady schooner]

is there any way to add only lyrics, litteraly words with a pretty voice singing them, to my song, without changin the song i already have generated?

[brazen yew]

Any prompt to always get whisper vocals. I've tried square brackets, square brackets with description, adding it to style aswell

[sleek stump]

Going to be difficult because Suno wants to do what the genre is know for, find a genre style thats known for whispered lyrics will help you achieve your goal.

[sleek stump]

YOu could try doing a cover with lyrics on an instrumental track (im assuming yours is an instrumental).

[brazen yew]

Well I'm predominantly making witchhouse for which some prominent artists incorporate whispers. I just don't suno really gets that genre

[heady schooner]

yes it is

what does the cover do? (still new to this)

[sleek stump]

Cover sort of does what a real cover is in Music, but with the sliders you can limit the amount of drift from the original source.

[heady schooner]

ah

[sleek stump]

You can change things like vocal type, genre styles, ect with cover, even do limited lyric changes.

[mint minnow]

Hi
I got a problem with my lyrics. One line seems to be faulty. Even if I change line completely, the spelling stays wrong, sounds horrible. Any idea, how I can solve this problem?

[sleek stump]

Ya if you are trying to do something thats pretty out there obscure, Suno likely wont have a lot of training that incorporates that. Like I said, its going to be hard to get all whispered, especially if Suno doesnt have training for it.

Not impossible, you just need to find that set of styles that would do that

[polar spade]

I sent this user over here from "need help" channel. Any thought to help here is surely appreciated.

@mint minnow. What's been your main tool or routine you're using to try to fix this. Give them something to start.

[sleek stump]

WHat do you mean, "spelling stays wrong"? Is it mispronouncing your words?

[heady schooner]

that worked, ty

[young geode]

Whispered lyrics/vocals

Keep everything lowercase, wrap text with Asterisk and parentheses (either or), use verse: softly spoken and use lots of Pyrrhic (unstressed-unstressed) and Amphibrach (unstressed-unstressed-unstressed) meters.

Keep the flow of the song as much as possible with the lyrics, with no abrupt switches in lyric tempo or standalone lines.

This will give you a higher chance of whispered lyrics throughout the song, or at certain areas.

[mint minnow]

It sounds distorted
it seems to be a technical problem
I've already tried to put the lines through other words but it still sounds distorted.
It seems as if this place is technically defective

Can I reproduce the song exactly?
As a cover it always changes his style

[polar spade]

A short way is to use Replace (either in new Edit UI or the Lecacy version. ) You can replace down to a word, but keeping in mind Suno need some room left and right of your corrections to "blend" the correction into the original song, so you may get slight changes.

[polar spade]

A longer way is to make a persona of the song (or at least a 1 minute patt of the song). Use the persona, along with Cover. Check that the whole lyrics are corrected as you wish them. Your original song should be seen with its waveform at top of the Remix panel. COVER tab highlighted.

Load personas last to provide the exact prompts in the styles/prompts box of the origin output and persona.

Click Remix and see what results.

[polar spade]

A really "short but works sometimes" version is to just move that mispronounced words slightly left or right (add a space or two ) or even put it up or down a line.

Usually this is done when you're just starting to create and your fine with reviewing a variety of outputs. Typically, at this beginning point you just want the best of options and dont mind making a few generations (10 or 12) , than to just hold with one version and try to make small changes later.

[half dragon]

Sounds distorded? what level of weirdness did it set at? I noticed in my songs if I raised it so high.. it'd affect the voice too ..

[polar spade]

A great suggestion. Hope it didn't go unnoticed.

[slow edge]

Can anyone help, I created some really good tracks and right vibe with 4.5 for my Serengeti wildlife, now suddenly what was working really well, just gives me trash and weird often very out of tune, vocals the sound like can't sing and flutes that drop in like a primary school flute first weeks practice, but only since the slider's got added,

I have decreased, down to 0% weirdness, increased, decreased the influence and everything is bad, get a good start, crop start, then extend and get the tracks speeding up and slowing down. Shame there is not a switch to remove the sliders and go back to the original 4.5. as it is not following instructions/style, just making a mess.

[patent sand]

Allegedly, 50% on the sliders is default. But a lot of others had your same issue. It's likely they updated the beta model with the slider update. Really, they just need to make it optional. It doesn't make sense to have to adjust sliders you didn't want to use to begin with

[young geode]

What's exactly going on.

Is a prompt that was working no longer working? lyrics? Or a Persona just being weird?

The sliders shouldn't be playing havoc with your generations when set at 50-50 for Weirdness and Style.

Mind you, Weirdness controls the band with which SUNO has to work with (how far outside the bounds is it allowed), and Style is how much it has to adhere to your prompt.

If you're using a weird blend of sounds and genres for your style, you'd want the style to be up and the weirdness: like 75-75

[flat crest]

Initially posted this in another channel, but hoping some folks here are able to help me out too:

I made a track which turned out quite nicely. Though there's a "chant" in the chorus which I want to remove.

I figured I'd be able to do this by taking the stems, and cutting out the vocals in another editor (BandLab)

Apparently, something I noticed in the new Editor; is that Suno (for some reason?) uses a custom BPM.
This makes it so that the BPM of a track and its vocals can be off-beat to eachother, if you attach them separately and aligned in a 3rd party editor.

This kind of screwed up my plan. And I don't really know what I should do now.

Does someone have any ideas? That would really help me out.

[lost holly]

Have you tried the legacy-editor?

[flat crest]

Tried both "all stems" & "instrumental & vocals only" — both to no success, unfortunately

[flat crest]

I've thought about that too. The problem is, cutting them out with the legacy editor, will probably change the melody too. Which is something I want to prevent. I'm 99% happy with this track, apart from the chant. And if I use the legacy editor, that'll change.

[half dragon]

since you use Band lab anyway... can't you cut it there?

[flat crest]

That's what I have been trying to do. But because the (set) BPM fluctuates in Suno, (and does not in BandLab) — the melody is off with the vocals to begin with. So I can have the end result I want in BandLab, but have a new problem because the melody and the vocals don't align anymore.

[young geode]

Why not just use BandLabs to retool and stitch the song, then use it as a template for generation?

Or are you just trying to take the original song and replace the chant and have it all tied back up?

[flat crest]

Yes, the latter. I basically have a 99% end result, with 2 lines in 2 choruses that I want to cut out, leaving the original track as intact as possible as to how it is now.

[half dragon]

You can chop and move the sections around in Bandlab to align them? is that not possible? It'd be for most DAW...

[half dragon]

ah you want to keep the song in Suno the same way without the chants

[young geode]

You should just stick to BandLabs then, SUNO isn't great for this kind of editing.

Much of that finishing touch that you're doing is best done after you get the generations you prefer. That is until SUNO devs develop a better edit and replace tool.

[flat crest]

These two lines in the middle are the ones (chant) I'm talking about — I figured I could just cut them out and be done — but the issue is the vocals now somehow "move faster" in BandLab than the original beat does

[half dragon]

how about upload it back after fixing it in Bandlab...

[flat crest]

The beat is on par to the speed as originally exported, but the lyrics "move faster?" and don't align anymore with the beat.

[patent sand]

I'm confused. Why not just slow the lyrics back down then?

[young geode]

If you did a stem split, this is likely partly due to the fact that SUNO regens those sections to a 98-99% approximation

[flat crest]

In all 100% honesty... I hadn't even thought about that

[patent sand]

Actually not skull. I know the feeling

[flat crest]

You're perfectly fine 😭 — I would have done the same

Hm. Apparently BandLab does not support slowing down to 2 decimals??

I never thought cutting out 4 lines total would cause this much pain

[patent sand]

Any chance you have FL Studio? I think it does two decimals (haven't used it in awhile)

[flat crest]

there's, uh, this thing with me and FL Studio

Anyways, I might be able to give that a try — a first time for everything

anyway, this kind of sucks — because (like @young geode rightfully mentioned) — the stems are still an approximation. They still do sound "off" compared to the real track, when put together again.
So now, if I get the BPM thing right (which is an "if") I'm still stuck with either one of two:
— A track that sounds off because it was put together using generated stems
— A track that sounds amazing but has a chant in it that I want to have out and kind of makes it 99% when it could have been 100% in my eyes.

[young geode]

I'd just take it as one of those Bob Ross moments: just a happy little accident.

[flat crest]

Not to mention I still wanted to make an intro for the track — somehow now it kind of "jumps straight into the track"
I tried doing this by "extending" the end in the new editor, and hoping I'd be able to "drop" that intro section right back at the front, using said new drag & drop feature.
Didn't work either, because it just generates "nothing" (empty section) or a "boom" and a fade. No matter what I do, I can't get it to generate a proper intro. And extending "upfront" isn't possible either.

https://suno.com/s/mgQhkpUhaY43Y3an

This is the track in question, for anyone curious

[flat crest]

I guess I'll probably be end up doing this in the end. Don't think Suno's tools are accessible or viable enough yet to make the change I'm willing to do. And I don't know if I'm willing to sacrifice track quality if it means being able to do so.

Chant made it just a little "over the edge" in my eyes in terms of alternative-ness. But oh well.

[young geode]

I wouldn't.

Personally I've come to embrace and enjoy the weird quirks. If a song is 98%, just let that 2% be the oddness that keeps everything fresh.

[flat crest]

Actually a lot of truth to that. That's one to live by.

@young geode @half dragon @patent sand
Thanks lots people for your help nonetheless 👍 — appreciate it a lot.

troubleshooting with a hint of philosophy ✅

[young geode]

Also, I didn't hear anything off about the song. Worked well.

[flat crest]

The one I linked is still the original generation — with the chant in the chorus, the one I hoped to remove. 😋

[young geode]

Ya, it sounds good, is what I'm saying

[flat crest]

Ah, my bad — wrong interpretation

Appreciate it lots! 🩵 — very proud of this one. Original beat (& vocals) were written and made by me, extended & vocals by the model.

[tight hornet]

Have you figured out a workaround for homonyms?

[flat crest]

First time I've had this much success doing so. So unbelievably proud.

[patent sand]

Write the word as it would be said, usually helps

Homonyms should not be a problem

I think you mean heteronym?

[polar spade]

@flat crest Throwing this out there. [Response to #📚┃suno-school message ]

Use replace to replace the whole chorus section (along with changing prompts and using excludes ) to nudge out the chant.

Run suggested gens and pick one. Sound good? Then right-click on the section, choose copy. Highlight next chorus. Paste in new chorus. Rinse and repeat. Remaster to blend a bit more.

Heres an example I did the last 7 minutes. I took a crap song with harmonies in the chorus. And replace with just the lead vocal (and its overdubs)

. First is original.
Second is replace , copy and paste

You can find first chorus at 00:35 and splices at 2:09, 2:25 and 2:41, cause I didn't bother with any precision, just highlighted and pasted and penultimate chorus was originally missing the last tag... to show the easy peasy process... Even still its almost nice as is.

Download and fix in DAW, or be more precise in Edit UI.

I know, I should have just screen recorded it. But too simple, easy to repeat, so why bother.

https://suno.com/s/fM779M0Wz4ymHRFZ

https://suno.com/s/PNwomuNtvdmL0O6x

[polar spade]

Oh, and I picked a first draft song with lyrics issues but also Edit UI didn't note the separate sections. How much easier if I had tagged the sections before I ran the original generations.

[crimson yoke]

Can you please tell me how you do it in editor mode to make the new generations sound exactly like your original song? Because when the new edit mode was first released, it was perfect, I was able to generate parts that matched the original. But now I can't anymore, there's always a different voice or slightly different melody or harmony, which makes my editing more difficult. 😢
Thanks

[polar spade]

This example.i just did worked right out of the box. I'd like to say that's so for every generation. But from experience, things change up frequently.

First, are you having issues with getting Legacy Editor to do what it used to do. It is still available by clicking Legacy at top left of the new Edit UI.

Second, I can try another song and attempt to video.

But literally, I just highlighted the chorus, for instance, clicked new lyrics, and optional to change prompts and excludes. Clicked regenerate and picked on that worked. This time I was lucky to get a first segment choice that fit.

[polar spade]

I may be preaching to the choir, but are you aware of the "constant" setting (and speed adjust) of the metronome. Left of the speed numbers? If you click and choose "constant speed" you can set the speed. Let's say 120 bpm. Then save that stem at that speed and download. If you then play the saved stem, both the [original and Contant seeps will be, of course, 120.

Untested also is to just set the speed first of an output (as yes, generative will fluctuate) . Then split the content speed output. To hopefully get two stems at exactly same speed. I kinda doubt it. But worth a test.

Whether if you do this and download both save at 120bpm , will it still match up in BandLab is another test that I dont know the results.

[crimson yoke]

Interesting, I'll give it a try, but I'm actually talking about the new editor. It was perfect at the beginning, seriously, I was able to generate identical parts. I fixed about 7 songs on the first day, but after a week, it stopped generating identical parts, and I'm still frustrated. I know what happened. My problem is only the different melody or the different vocals.

[polar spade]

Yes. Just pointing out Legacy still exists.

Well, again, the one sample.i just did a half hour ago was just fine in terms of vocal, nuances, vibe, groove.... with the New Edit Replace and didn't have issue.

I would have to do 10 times in a row to think its doing well. But I'll do another one. Ill replace a complete chorus to see if the its changes in any way. Hold on.

[polar spade]

Did your 7 song fix come right before any updates?

[crimson yoke]

Since then, I stopped using the new editor until they update it again with a model that truly follows the melody/harmony/vocals identically. I'm using the audio upload feature now, and wow, it follows my melodies REALLY well with sharp quality. I got addicted to creating sound ideas and letting the v4.5 model do the magic, it's simply amazing. To give you an idea, I did a test recording with a bit of background noise and the model still understood my melody, it was so satisfying. I want this level of sharpness in following melodic instructions in the editor mode as well. 😦

[polar spade]

oh... yea me too.. Ive upload me singing with my guitar...equating a fish trying to sing while playing a cheese grate.. along with skips and coughs and screwups. resulted in a perfect pitched song and exact clone of my voice. Pretty cool. one-off kinda thing, but the rest were similar in "nice" results.

[crimson yoke]

Right when the new track editing mode was released, I used it on a few tracks that same day and it worked perfectly. I improved a melody and lyrics that were pretty bad, and it handled it masterfully. Then I spent about a week without using the editor, and when I came back a bit more recently, I noticed it was no longer generating identical or very close melodies, which is why I’m confused.

[crimson yoke]

Hahaha, yeah, man, it's honestly so much fun working with the upload. I'm actually really happy with it right now, it's been super satisfying.

[polar spade]

With the 4 minute available for uploading my less than 4 minute songs, I can now have lyrics, melody, progression all mine. So Suno is now just doing a work-for-hire demo. That is now fully copyrightable Bam! Hazzah!

[polar spade]

Last week I won our local NSAI songwriting contest, so I go into semi finals. I dont think id have done as well if I played them the original recording. But in fairness I gave the local coordinator both audios to show Suno only made a demo of the true melody and lyrics I sang. 😃 no different than to send a worktape to studio to make my Demo.

[crimson yoke]

Absolutely! 🥹🤩 Couldn't have said it better.

[crimson yoke]

Congrats man!! 🎉 That’s awesome, and honestly a perfect example of how to use these tools the right way. Wishing you the best for the semifinals! 🙌🔥

@polar spade What are your thoughts on Suno v5? Will this be the AGI moment for AI music? My expectations are as high as for OpenAI's GPT-5. Hahahahah

[woven gate]

I have a problem in the last songs it puts me the style in the lyric and in another song it skips the voice

[polar spade]

https://icecreamapps.com/v/fe3sekd

Vid of my replacing the line *"... to bring back times when you thought I would say such special things...." into ". . . to bring back times if I bought you a diamond ring. .. "

Vocals, nuance sounds fine on the new version.

[polar spade]

No opinion really. I doubt there will every be profit for making the API available for quite some time.

[crimson yoke]

I downloaded your video, THANK YOU so much for the support. I’ll try later and see if I can get identical generations. You were able to do it normally, I was the same at the beginning. Could you record a video for me where you make the chorus sing twice? I have to admit, I haven't been able to do that so far. I’ve tried everything and it just won’t do it. My chorus only sings once in the middle of the song before the bridge, I want to double it, but it simply refuses to do it.

[polar spade]

Sing twice? As in a row? Or someplace else.

Sing twice before a bridge. OK. Let see what I can do. You've been in Suno since it began, sir, you should be showing me!

[crimson yoke]

Yes, in a row. The song only sings it once and then goes straight to the bridge. I want it to sing it twice in a row, but I’ve tried everything and it’s not working.

[crimson yoke]

🤣🤣

[sleek stump]

Many of us use phonetic spelling to get pronunciation correct. It pretty much happens in every language.

[flat crest]

Yes, I was aware of the "constant" setting. — unfortunate though, is the fact that some parts of the track are somehow "set" (by fluctuating BPM) as 80 BPM, while others are set at 140 BPM (almost double) by the model itself. So, if I set it to a fixed value (either ~78 or ~140), I'll have a part/parts of the track that is either absurdly slow or absurdly sped up. I could workaround that by generating each part of the track in its correct BPM, but that is going to burn through credits faster than I think it's worth (50 for each stem per each section in each BPM) — and that's going to be an absolute mess to go through.

[polar spade]

OK. Did a duplicate chorus. it had a extra 2/4 beat between chorus, so remove a smidge of a beat (i frigged about adding a beat instead of removing a beat -- ignore, please) . Saved and remastered.
https://suno.com/s/nYvwIMgr4MZEfspE

[polar spade]

oh , literally a big flux in speed. got it. Yea. That wont work.

[flat crest]

I might work with the new editor a little — like your other suggestion.
I want to kind of keep the track as-is, so I'm kind of at the model's mercy as to whether or not it decides to alter the melody or stick to mostly its roots.
Who knows what that'll bring — hopefully it'll work my way. My experience with the new editor has been ...clumsy at best, so far.

@polar spade might be self explanatory, but just to be sure — I assume select "new lyrics" , remove said chant , & replace?

According to the model, the chorus starts at the chant, but luckily, audio wise, the chorus already starts midway-ish throughout the hook. Same for the 2nd chorus. Hopefully that'll help with context window and generate something I'm looking for.

[crimson yoke]

It turned out great with the extra beat, hahahaha. Duplicating is cool, buuuut what if I want to change small nuances in the lyrics so that on the second repeat it has slightly different words while keeping the same metric and rhyme? The singer would surely sing the new words with the same melody, that’s what I haven’t been able to do. For me, it just wouldn’t duplicate the chorus. I added new lyrics but it didn’t help at all. 🤡

[crimson yoke]

I followed you there, follow me back. Thanks! 🙏

[polar spade]

its a crap shoot. But I just copied the same lyrics in the New Lyrics section and put some exclude styles to try to exclude the idea of harmonies and back up singer... the result was the lead vocalist singing alone only with an overdub of his voice. So I could have gotten lucky to get that first round. Another time, it might take me 6 suggested outputs before one excluded backup singers.

[polar spade]

i only chose from two different suggestions. You could change the styles , add some exclusions and then run a new suggested output and keep doing that until you have 12 or 30 outputs to choose from. Up to you . mine was jjust an example of how i duplicate.... and additional of a quick splice and remaster.

[flat crest]

wasn't there an option in the Legacy Editor for this? I believe it was called something like "discourage vocal use" — doesn't the new editor have this? That's a shame.

[polar spade]

I totally forgot your original post. Yea, you wanted to get rid of vocals altogether. Yes there was that toggle, Ive forgotten about that. I'll have to look for it. That might help if it works.

[crimson yoke]

Alright then, thanks for the support. Your videos are definitely gonna help me out!🫡

[polar spade]

Its in the timemarks settings cog. as [ ] Instrumental

[flat crest]

In the new editor? Or does only the legacy editor have this?

[polar spade]

i dont know. lets see... hold on. I know some things i cant do in the new and MUST use Legacy

[polar spade]

I didnt find anything similar in gthe New Edit UI

i used Legacy...and cliekd the INSTRUENTAL setting... one did a little instrumental then added gibberish... the ogther added lyrics anyway

[polar remnant]

If anyone has a trick to be able to choose the duration of a piece, it's too random, and we often lose too much credit to have an ideal duration.

[half dragon]

You can't currently.

[polar remnant]

This is really the evolution to have and also add partitions

[polar spade]

I ran ten more... lets see if any are just instrumentals

[flat crest]

Ah, that sucks — such a thing would be nice to have in the New Editor. Anyway, legacy editor it is then.
What would you recommend doing prompt-wise? (style-description)
Keep the same prompt I had when generating the track itself?

[half dragon]

Many has made a request/suggestion on #💡┃suggestions-web so go make the suggestion there if you haven't, the more suggested the same thing, might have more chance to be noticed 🙂

[polar spade]

what was it again.. you had a choral or a drone of voices? add styles which are prominent in whatever it is that you are excluding. Such as if you're excludig Female voice, you dont put "female vocals" you would put femme folk, braodway, opera, pink power, k-pop and so on

[flat crest]

It was/is a double "chant" in each chorus, which I included myself to begin with. I had hoped it turned into a small distant adlib, not a full-on energetic chant, the way it is now.
Though, upon actually working on the generation/track — The chant tipped the track just a little over the edge in my eyes in terms of "alternative-ness"

[polar spade]

i didnt Keep Duration on the ten i ran.... all were about 2 minutes with bits of interludes between gibberish... not a good start.

[polar spade]

so... for what its worth... i didnt use the intrumental.... i just used Manual Lyric Edit in the settings cog by the timemarks in Legacy, so that I could get the New Lyric box to show up.... Deleted the lyrics and typed {interlude} three times. And I got three of four sequential interlude samples... so that seems promising. Starts about :34 for 20 second interlude where there was a chorus before. See lyrics for where we put [interlude] tags.... https://suno.com/s/XTCRFAcFO0ZN3qMS

Applying that to New Edit UI, we might do the same and just type in [interlude] 3 times or some equivalent prompts making sure we removed all lyrics and see if it does the same as Legacy... I would probably choose the Cassic timemark setting move the BEATS blending border down to almost nothing. Nope.... that didnt do squat.. even changed tempo with lyrics. Stick with Legacy.

[polar spade]

putting 20 second [interlude] instrumental in place of a lyrical chorus using Legacy Edit UI

[sour night]

What I'm experiencing with the new software update and this whole new feature where you can correct the weirdness, adhere to style, adhere to music - is a lot of gibberish. Like before, if I would create the cover of the song, it would nail all the lyrics 100% usually. Now it creates two versions of the songs connected together :(((, one with correct lyrics and one with gibberish.

There is one golden rule in software development. If it works, do not touch it. Or if you touched it, make sure to leave the option for the user to use the previous version.

Another good software development rule - leave all the menus and buttons in the same place where they were before - otherwise you confuse the sh**t out of the user.

[patent sand]

Seldom followed, unfortunately

[polar spade]

Love it. Discord limited me for 24 hours. I guess for uploading too many videos? huh.... Good time for dinner.

[flat crest]

This did the trick.

Thank you so much Dasvon, genuinely.
The way you know and are able bend and break this tool, hats off — in the best way possible.

Tried & did all as you said — worked flawlessly. I even got a chorus beat generation that I like even better — 90% close to the original, but with an added beautifully nasty beat kick behind the main melody.

It's hard to express my gratitude through a screen, but I'm so thankful for this.

@polar spade note for reference: "(Instrumental)" works as well!

[polar spade]

Dude, can I buy you a car... or eve wax your car for a year. That definitely a boost to my ego and much appreciated. I know some things work for a bit and then they dive into a chasm but at least you got through this one with something that worked and you're happy. Thats enough for me. Happy I was able to find a way through it that worked with another user as well, though the other suggestions are all great options. So now you have a few in your bat belt to use.

[flat crest]

After all the effort you've gone through to help me — I think you're the one here deserving of a car, my friend 😉
If it becomes a hit — I'll be in touch.
I'll owe you one, for sure

Thanks again 🤝 🙌

[sour night]

any one knows - Is there any way not to use use Advanced Options (new feature)? Can i Exclude/avoid it at all?

[flat crest]

at the moment, unfortunately not.

[polar spade]

Which do you mean? All of model 4.5? We can create a song in 3.0, 3.5 or 4.0 , extend as usual through the Remix panel while keeping peronas off and Remaster with 4.0. A few remidial things..Thats using all the older tools without any advanced feature. If more recent, we have access still to the Legacy version of Edit UI as well.

[stray lynx]

Is anyone else encountering an "unknown error" when trying to rename a song?

[polar spade]

Does it look too long? Besides just a random error, that can be one reason .

[stray lynx]

I don't think so. the new name is just "The Duel" and it is only the title that fails to update. the summary updates and so does the caption.

[polar spade]

Clearly not then. I guess if its unknown to Suno, its unknown to us. The long title is the only thing is can think of.... a slash or symbol might throw it. See if it improves later.

[stray lynx]

I'm playing with the new editor interface and seeing this new "Heal Edits" option that has no indication what it does. It looks like an undo, but what would you undo? You still have the original sources.

[high hound]

UPDATE: After 8 weeks, Nicole from Suno Support finally contacted me, and explained why I kept getting an ERROR message when generating a song and pressing CREATE:

"It looks like your account was flagged by our system due to a high rate of play or song generation. This can happen if songs are being generated too quickly" - Nicole (Suno Support)

So the lesson is everyone....don't press CREATE too fast 😛

[sleek stump]

Glad to hear your account is fixed!

[sour night]

I wouldn't build more functions until I 100% nail all the functions that already there. I'm frustrated with new advanced option functionality that I cannot avoid now. I am frustrated wit Multi language functionality that simply doesn’t work. From a basic user perspective, I should be able to substitute translated lyrics in a different language into the same musical arrangement and get more or less a similar result.

But instead, when I try to create a cover song and replace, say, English lyrics with Spanish lyrics, the result is completely gibberish. It doesn’t work with covers, it doesn’t work with styles — it just doesn’t work at all.

On top of that, you’ve introduced a bunch of new features that I also find uncomfortable to use, especially in the editing mode. For example, I used to be able to change a single word in the editor with not too many issues — now I don’t even understand how it works. I’ve tried multiple times, and it just didn't work.

All of this creates a lot of frustration.

[half dragon]

Heal edits sound like smoothing the transition between sections... haven't tried but that's what it sounds like... give it a try and tell us 🙂

[stray lynx]

Seems like what it did was, if my edit did not occupy an entire section, it would extend the edit to fill the remainder of the section. I'd hoped that it would also level off bad volume or eq blending, but it doesn't do that at all. It might have, if the end of the filled section's volume and eq matched up with the beginning of the next but it did not, so it was useless for my situation but might be useful if the section ends on silence

[high hound]

Thank you Ginger. 8 weeks later, and now I am back to being able to create new songs

[half dragon]

Awesome. Are you Pro uesr? Curious if they even compensate your 2 months that you coudn't generate anything and still paid for it?

[thin walrus]

need help, after uploading a file, suno puts an automatic long description for the style which is completely inaccurate, and when extending the upload, the inaccurate style is taken into account... how to remove this inaccurate style thing

[half dragon]

I did an upload today and I think they reverted that back. No more auto gen style and even auto gen lyric
#archived-ticket-help message
Maybe you haven't got this silent update yet?

[thin walrus]

i uploaded just now and its still doing it, its causing my extensions to follow the inaccurate auto style instead of the actual audio

[half dragon]

Can't you change it when you Extend?

[sleek stump]

Can you not delete the styles from the window?

[thin walrus]

no,

right here, this one

[half dragon]

Not on the upload, but when you extend from it, you can change/delete it, no?

[sleek stump]

Clear your browser cache

You should be able to edit that styles

[thin walrus]

im trying that but its actually seems to still follow the auto style

for example i uploaded this trap beat with distorted guitar but suno adds an auto style that says its a rock song. i extend it and erase everything in the style section, but the generated extension is rock.

[thin walrus]

i cleared my cache

[half dragon]

This is really strange. Earlier when I uploaded an mp3 - it did not do that
https://suno.com/s/x3N8E0OURqOkTiJs

Just now, I tried again but from a WAV, and it's still doing it.
https://suno.com/s/Fvn1CkMCqr2vUYnX

Anywho, there's nothing we can do as the style is generated and stuck on the uploaded audio.

[sleek stump]

Are you putting what the actual style is or just leaving it blank?

[thin walrus]

yes i am putting the style in on both the uploaded song and the extension

[thin walrus]

exactly, its preventing me from using the extend properly

[sleek stump]

Ya thats really odd... and not normal

[thin walrus]

im gonna try uploadeing on my non pro second account

ok it also happens there

well ig theres no solution

ffs

[polar spade]

One thing to understand, one thing to try.

AI doesnt really parse out grammatical constituents of the lyrics but splits them in the best means to meet the goal of matching the music it generates. So a work can be split into two, three or four tokens, just to please the AI.

So now you have English, lets say, with specific tokened split words, matching the melody exactly.

If you translate into another language, Suno cannot match the varied and different syllables to match the melody that the English words and metric created by the English lyrics. You're more likely to get something quite different than to get something close.

for example Strike in Japanese is Su-to-rai-ki , which is actually just a Japanese pronunciation of Strike. But if the song sings,
"Strike, Strike . Strike it down on the ground"
In japanese its, "Sutoraiki, Sutoraiki, sore o jimen ni uchikudaku" You suddenly are also translating a standard 8 syllable line into about 24 syllables. Suno doesnt get it and creates a different overall melody for the whole song.

[young geode]

@polar spade
I've never got into using the Editors, as they seemed finnicky.

What's the main difference between the legacy and the Beta, in their overall logic/capabilities?

[polar spade]

@sour night Thing to try - which I do.

Some slower ballads , and some songs in general do OK for translations. But for what Id think are most, we need to intelligently spend much time re_writing the lyrics into the other language in order to fit the melody..... or go the easy route and ask an AI such as chat GPT to " Translate these English Lyrics into Japanese , with a priority of matching the syllables, line metering and arrangement as exact as possible. You may need to shorten certain Japanese phrasings, but with all changes you are to maintain the theme and meaning of the song as well as the intent and moods expresses in the phrasings, creating clever rhyme schemes which are similar in style to that of the English version"

Doing this will get you closer to the results of matching lyrics to melody of the original version. If you speak the second language, this give you a quicker means of editing to make sure the lines are meaning what you intended.

[polar spade]

I dont know. What appears to be with the new Edit UI is as if its trying to explain what Legacy was doing or give us more control.

For instance, Legacy just had "keep duration" while New Edit has three choice of strength and length of duration including the edge blending width in "beats" we'd like to use. Thats seems nice, but a larger learning curve. But as we may have complained that "Keep duration " was varied, Suno may have said, fine, we tried to give some sort of average, but here, do it yourself.

They then seemed to say, why complicate things and have Restore, and crop, when all its doing is deleting and copying. So they dumbed it down in spots so much that a few of the nice traits of the individual Restore and Crop were removed and all of those functions reduced to just a dropdown right click menu. I dont know why, But what I found were nice about Legacy, didnt seem to matter much to other users -- this may explain why they were relegated to a dropdown menu - I was in the minority

In the end, I'm using both to some extend for the traits that the other doesnt have , so its a balance at this point. If one or the other is taken away, that will be sad. But if I had to choose, perhaps the new one is better as long as they add back in things like being able to manually add the timemarks rather than just use a mouse.

[young geode]

oh ffs Das, I was reading that, lmao

Ah, So it sounds like they parsed the old editor to it's components, but in the process lost some of it's weird quirks that were oddly beneficial and predictable.

[polar spade]

Saying all that, the new EDIT does seem to be smarter at finding good splice points and well enough that being precise is not seemingly needed. So much so that maybe i dont really need to be able to manually add a .01 second here or there, just get it close and Suno has spliced many things beautifully that Legacy seemed to require a bit more precision. This makes sense to move as far away from the DAW mentality and use as much AI intelligence possible to do things a person cant do on DAW.

[polar spade]

Thats a good summary. yea. i had to split up my walls of babble.

[polar spade]

One thing new Edit UI does when it replaces lyrics, in the new output it often shows just the lyrics replaced instead of the whole lyrics even though it says "full song" . That can confuse users. What is more, it includes the lyrics before and after the replaced area, which were the blended border lyrics it needed to "seed" the blend of the replaced lyrics. Its a fun insight to see just those lyrics, but Im not sure if it helps. Nonetheless, i find myself having to copy and paste all my lyrics back into this final "Full song" before I continue.

[fading tulip]

Suno support need one million years to answer...

[stray lynx]

I figured it out. I realized the songs that I couldn't rename were the products of many extensions, so I looked and the "Displayed Lyrics" field was over its limit. Somhow updating the song title is also tied to the displayed lyrics size limit because when I removed the excess garbage from the extensions I was able to save the song's title.

[ember birch]

Check this out. Really cool trick with this new stem extract feature for dynamic DJ/stem manipulation for live performance !!

you can modulate this in real time, and this would be great idea for VR live concerts

[young geode]

explain

[ember birch]

Basically you export all the individual stems: bass, drums & effect layers
Then choose which layers you want to specifically have the effect for a live performance, for example I choose only the bass and effect layers (excluding the drum layer)
When I move the ‘mix slider’ up, the ‘reverb effect’ gets added to those instrument layers which makes it sound like it’s ‘further away’ and when I bring it down, it’s like you are closer/right in your face, especially if you listen closely.

The ‘reverb’ effect (that’s the Valhalla Room plugin) virtually emulates that ‘room echo chamber’. (Like they do in live acoustic venues)
And what’s cool about this update, is you can have way more fine control with the individual layers and is way more suitable for DJ sets.

[young geode]

That's pretty neat

[shut forge]

Upload audio issue: Every time I try to upload a piece of audio to use for a cover, etc., the process quits about halfway through and there's nothing fully/partially uploaded (in the Library). Any suggs?

[tranquil estuary]

I get the same, but mine gets stuck at 99%

[polar spade]

A quick one. Try at another part of the day. Its hard to say when the servers are less loaded, but this may help.

[polar spade]

2. Keep in mind there is also a filtering of content going on as well. And it may not give notice for every time. So its possible it can just stop or time out.

3. though it seem counter-intuitive, change your audio to WAV and upload WAV instead of MP3, or vice-versa if problems uploading a WAV.

[shut forge]

It is a .wav. Urgh.

[polar spade]

Right. So upload the MP3 version?

[shut forge]

Nah, that failed, too. I'll try later in the day and see if that sticks! Thx all for the help.

[clever bloom]

I can't upload videos for my songs. I've made it to the requirements Suno asked for, but it still won't work

[flat crest]

Just for reference — I'm not associated with Suno or its development.

I remember back when v4.5 was introduced, quite some people complained about not being able to go back to or use "v4/v3.5" anymore. This was fixed within a few days due to community request.
I'm certain the same will apply here. Give it a bit of time, and I think the sliders will eventually be another optional tool as well.

Community feedback is the key here — make sure you let Suno know through the feedback forms what you've let me know. We can't expect them to work with ghosts and no information.

[polar spade]

You can test a smaller random audio file. If it works, before saving, just cancel. Then you know the issue may be your file and/or content.

[unique stone]

The sliders are already optional. If they are left at their default positions (there is a reset button) they don't have any additional effect and behave as if they weren't there. Essentially these were just hidden before.

[flat crest]

Thanks for the insight — I didn't know that at all.

@sour night ^ abovementioned

[patent sand]

Surprisingly, they do actually listen to feedback. I can understand why people assume otherwise given how most software development is now

Question: when I eventually go to build a persona vocalist, should I go for v4 or v4.5? I know v4 is better for metal, but the vocalist will not be screaming/growling. Think something like Tarja or Amy Lee (not a clone ofc)

[ebon sand]

I have not made the experience yet that personas work on v4.5 well at all as in getting the same singer in different songs again

[tawdry scarab]

this is the new technical channel name? I was curious about a song's cover being a broken image link and defaulting to that no-image blotchy image. I wanna know is this temporary and just a filesystem reorganization or availability issue, or did it lose the image honestly and completely, or was this the result of some kind of errant moderation

[bronze pecan]

The comment moderation in germany is a bit very strict huh? sometimes I have to do research what possibly could it complain about and it never makes sense xD

[bronze pecan]

v4.5 is better with voice dynamics.
At v4 my persona Lucy was never able to do growls, she is now much more flexible since 4.5 and can also do now this

I wouldnt say that v4 is better with metal than 4.5 tbh
v4 is the better treble, but also much more noise

[polar spade]

I wondered where you went. Good to see you back.

[sleek stump]

I remember back when v4.5 was introduced, quite some people complained about not being able to go back to or use "v4/v3.5" anymore. This was fixed within a few days due to community request.
This is 100% absolutely not the case. 4.5 was released as a beta model, and is still a beta model. The older models were never removed or limited in any way.

[flat crest]

Oh — could it have been the case for v4 then?

I don't remember. I can vaguely remember some backlash due to a model setting which was fixed and irreversible

Which got fixed/adjusted within not too long after

[patent sand]

They took away v4 remaster at some point and then restored it

They never disabled v4 or v3.5

[sleek stump]

They sunset v4 Remaster just 2 weeks ago and brought it back due to community feedback.

[stray lynx]

I wondered why I saw it still available

I'm happy they have it still, but I also suspect they won't really maintain it very well since it would take time and resources away from the new feature work

[patent sand]

They might remove it when 4.5 is fully released. TBD. V4 remaster is only available for web currently