📚┃suno-school

[midnight plume]

That second one sounded to me like it started on C in every stanza.

[polar spade]

Then it's in the key of F (the root) 😄 you hear beginning after each stanza in the music riff.

[midnight plume]

Sometimes things in this software get WAY deeper than my old brain can handle thinking about! With a BA in Music (from 1976) and years of strong theory roots (started lessons in 1956 at age 5) this SHOULD be easy! Alas...

[polar spade]

Maybe I didn't get it. Between the first note in the Waaaohhh and the first note (root) in the riff.. seems like a 5th to root drop to me.

Otherwise, the one word lines may help distinguish the notes... and hey, the "mellinnial whoop" , seems Suno knows what that was.

Hope that gives some new ideas to try.

[midnight plume]

I should restate as maybe I wasn't clear earlier. I want a verse to start on the 5 CHORD, so if it is in C, the verse to start on a G.

[polar spade]

Oh. Chord. I was aiming for the 5th note to root, so yea, mine was in the F chord still. You're right. I get you now. Gosh. Hmmmm... what would start on a 5th chord. Think . Think. Think.

[glossy cargo]

I'm trying to end a song, it's never ending, I spend 2000 credits for an outro and I still don't have it

[midnight plume]

I also wish they'd add a fade out option to the edit menu.

[polar spade]

I always suggest extend using v.3.0 for song ends. Don't split/extend on an instrumental interlude. Find the end of a phrase and extend from there.

[flint merlin]

Suno ita

[polar spade]

Oh last one. I'm cheating as it still starts on the 5th NOTE (and 6th) , but sung in harmony, to sound like a chord, then down to the root. More of a Greenday feel. Lol. https://suno.com/song/eab610cc-d3ce-422b-b9a9-4393288c327a

[midnight plume]

That sounds like a bad Nickelback tribute band. LOL!!! When they hit the chorus especially.

[polar spade]

Ok. I tip my hat and give up. Haha. 🫡

[midnight plume]

Yeah this was a tough ask!

Like that!! And you exposed an aspect of the lyric engine that deals with context where you say "Ripped it open, read its over so" because "read" can be either "red" or "reed". It chose "reed" where you likely meant "red". I have had issues where the program picked the wrong option and had to go in and edit it to show as "red" or "reed'. Not with that exact word, but I have had to do things phonetically to get it right.

[buoyant cairn]

Ok so i was playing around with comfyui creating the lyrics and taught the AI how to add soun effects... its overdoing it but I have found that some sound effects dont work and one of them is hilourously wrong!
With prices so high, we’re just left to *moan*.
resulted in With prices so high, we’re just left to BEEEP.
in both versions of the song it did the same issue!

Example: https://suno.com/song/3e2eb196-5ec1-4a8e-9f45-2fe785e48c45

the beep makes it sound like its censoring a word lol which just sounds really odd.

hmmm With prices so high, we’re just left to *groan* has the same issue.
*growl* resulted in silence

Its going to be awkward to tell the AI which effects work and which are broken, the *moan* and *groan* glitch makes it sound like i have used a naughty word lol. Certainly an odd glitch.

[sudden pike]

After using almost 1,000 credits, this song finally came to life.
Video created with LTX Studio (super basic but incredibly easy to use).
https://www.youtube.com/watch?v=gfCpXq2VzJA

[neat field]

Anyone else experimented with Remastering 2x? Ie ->

Create song 3.5 (or even 4.0) -> remaster v4 -> remaster the remaster (remaster 2x)

Finding good outcomes there....

[sharp pecan]

Again, has anybody had any warnings indoor editing videos from Sony music?

[tiny falcon]

Yeah. I tried 3x too, but sometime it doesn't help much if the first remastered generate hallucinations

[neat field]

Regenerate section works pretty well in that case, and I agree I have had to redo some sections after remaster hallucination (probably about 20% of the time when Remastering final time)

[tiny falcon]

Thanks for the tips, I haven't tried regenerate to remove artifacts. I used fadr to get the sterm and sometime it can get the artifacts and I removed it in fl studio. Takes lot of time though 🙂

[sudden pike]

Come on, guys!
I was so happy after creating my Christmas song in Spanish, and I was even happier when SUNO AI allowed me to create my own persona: https://suno.com/persona/f1aee147-993f-43a4-a4c8-726bea02d521.
But now, the fake app issue is back: https://suno.com/song/ae87e7de-a799-4287-ae70-4be268db7a1c.

[indigo fog]

Hi there, I am new here. Just a technical question. I uploaded an audio file, but the "extend" button is always not accessible. Why?

[sudden pike]

you have to write something inside those options "Style of music " and "Title"

[unique stone]

Yeah just the style should be enough

You have to tell Suno what you want it to generate based on your uploaded audio. Unfortunately it can't work out anything from the uploaded audio apart from lyrics (if there are any) so even if you just want to continue in the same style, you need to describe it.

[naive vigil]

Hey, I am summoning the masters of Suno here cuz imma need all the help in the world here

How I make a consistent duo? Like a male female duo

I tried putting it in styles

In lyrics

Both styles and lyrics

But Suno only makes one voice

Any ideas? As I am out of them

[indigo fog]

Thank you!

[indigo fog]

Thank you!

[mint nexus]

What I read (and worked for me): pick a genre that includes duets, add "duet" as a tag, and I tagged each verse [MALE] and [FEMALE] but that could have been dumb luck

[naive vigil]

It doesn't work for me as I try to make a rock song. That means to select other genre and then cover, but it's useless as it reverts to monovoice

[mint nexus]

Perplexity suggests that "duet rock" and "power ballad duet" might be genres, but I'm guessing you already tried something like that

[naive vigil]

I haven't. I will try and come back

[naive vigil]

I am back, I can confirm that I generated 10 songs and none had any duet particularity.

[mint nexus]

I guess I should admit that I do not even vaguely qualify as a "master of suno" 😬

[naive vigil]

No worries

I was joking when I called masters of suno here :))
Any help is good help

[hollow hare]

Here is a playlist of examples (https://suno.com/playlist/11ff3827-bdcf-4ef5-bd04-da8647c6cdb6), most being the different methods (#📚┃suno-school message) I try to prompt the genders for duets

But it doesn't work all the time. In my experience, when it does work (~60% rate), half the time the AI swaps the gender roles

You can predefine a label and use it later on in the lyrics too, usually the AI is consistent with that

[Lucas: Male, tenor]
...
[Verse X] [Lucas]
...

And that worked relatively fine too

[mint nexus]

Such a pity you can't bookmark playlists

[worldly ruin]

Are they going to fix the shimmer/static anytime soon? Have they said anything?

Really getting frustrated with how many credits are just being wasted.

[wanton panther]

Hello, is anyone have issues with adding more credits with monthly subscription ?

[silk hill]

Hi, I have a question about song lengths. Can you specify or limit the length of a song? Can you, for example, tell it that the maximum length of a song is 1 minute?

[autumn vault]

Well they said you can fix that with DAW but dang those tool are confusing for me who's not an audiophile 😔

[worldly ruin]

I'd much prefer if they could fix it on Suno's end.

[autumn vault]

The V4 in the beginning was excellent but then it starts to turn worse after a month or so... idk what cause it.

[daring slate]

I'm trying to replace a lyric in Suno. The verse repeats, and when I edit it, the change only applies to the end of the song, not the beginning. When I use 'drag your selection' to highlight the part where the lyric is sung, it doesn't detect the text/lyric even though it's clearly being sung. Why is this happening, and how can I fix it?

[naive vigil]

For me it seems that rate is less than 1%

Since when I joined Suno, more than 6 months ago i got only one song with duet whenever I asked for it, then nothing else.

I thought V4 will be better but it's not.

[hollow hare]

Hmm, it also depends on your prompts. For example, if you list a gendered vocal (esp.female) in the style prompt, it tends to overpower the male vocals - or so I found during my test with 3.5

Do you mind sharing your latest prompt attempt? You can DM it to me

[autumn vault]

Well I got duets almost everytime I generate songs with v4 these days. I specifically type on the prompt the gender of the singer but still got duets most of the time.

[solemn lintel]

I haven't even touched v4 since I wasted a whole month of credits generating useless songs that either shimmered or just didn't sound good.

[high helm]

Hi how do I add new vocal on top of Style of Music for Persona

?

[neat field]

i swear every time i get back to Suno after weeks/many months away, https://suno.com/@yolkhead new songs are worth listening to and reading his notes for

always learning from that person

[radiant aspen]

Did v4 just go off the rails in the last few hours? lol 😂

[dark spindle]

i don't know how suno came up with the original melody in the chorus, but my ex-co-worker and his wife love this version^^

[neat grove]

Considering suno does voices so well, is there a capability for podcast creation - if not, it would be great to have. I tried prompting with no music - but it is coming out as a rap not as spoken language. Any ideas?

[neat nexus]

Offered for your consideration:
[Removed links]

[Removed links]

First one was intended to be a PBS type thing, second one was intended as a horror track, which works as a voice over.

This third one was a voice over for a music video I did:
[Removed links]

So technically it could probably be done with some work.

But you're going to need to optimize for voice over and try to prompt the lyrics and styles to avoid music. And then you're still going to fight the AI as it tries to break out into music or singing.

And you're going to have to script the dialog as lyrics with back and forth. It will not be a simple process.

[charred merlin]

v3 seems to be the most apt at doing spoken word -- v3.5 isn't far off but is more likely to transition to music.

I haven't been able to get v4 to produce more than 30s of spoken word before transitioning into music

[heavy kayak]

Is there a guide on iterating and augmenting prompts?

[heavy kayak]

same is there a way to limit other vocal types

[polar spade]

I'll be honest, I don't know what that means. I may be reading it wrong but do You mean a guide to coming up with the right prompts to get the outcome you want?

[heavy kayak]

Yes the "Style of Music" section

what tags are actually used, the prompt seems to ignore most tags i add

[polar spade]

I see your similar questions in general channel. When I started out I read through suno.wiki it also has a decent list of basic prompts to experiment with.

[heavy kayak]

magic, thx

[polar spade]

And for ideas, I search through members songs on the home page and see lots of cool ideas and techniques both in prompts and lyrics tags. That's about the easiest means of learning arrangements, tags which seem to work and prompts to get certain sounds. I played around with those and experimented on my own. Certainly asking in here as well.

[heavy kayak]

i like it, very 1984

[polar spade]

And looks like you joined suno discord today. So ... welcome!

[polar spade]

Keep in mind, it's not to be an on/off switch. Members have different tastes, and those tastes can induce different issues in the prompts they use.

As for many, the noise effects have lessened considerably according to anecdotal comments recently.... through the past 8 weeks. If you're still having issues, it can be noted for them to please continue adjusting what's needed to cover the full spectrum of genres. Keep clicking the thumbsdown as often as you have issues. If not, it will take longer for suno to recognize and pinpoint the problems.

Personally, I don't have any issues, others report having continual issues. Other mention though the shimmer has subsided, there is now is a more tinny and often too bright an effect with 4.0.

[glad mauve]

is anyone having static issues when extending a song?

[raven turret]

do you guys know how to make the songs scream when you want it to scream?

[knotty ore]

As, Im a paid subscriber. Can I upload my tracks in Spotify?

If I cancel my subscription, still have I the rights of those tracks?

[untold briar]

Quality problems that get worse as the song gets longer has been around for awhile. There's no easy fix, but if your song is very long, you may have to try using shorter extensions. If the song is short enough to be Cover, that is a good way to fix it as well. (And Remaster for longer songs)

[untold briar]

Yes, as long as you made the tracks when you had a subscriptions.

[glad mauve]

I mean only after 1 extension like a 5 min song

[untold briar]

There are some ways to fight it, but it's easier to backup the song and try to get a better version that doesn't degrade as quickly. If you really want to make a long song, the nuclear option is to use a series of very short extensions (without doing Get Whole Song) and then only doing Get Whole Song again after the audio quality clears up. But that means the segment will not be generated with full memory of the song, so it doesn't work for every situation.

[glad mauve]

Ok interesting it seems to have started at v4 launch but dont get me wrong the quality is much better this just seems to be a hickup in it. thanks

[untold briar]

Oh, it might be worth trying to use Replace instead of Extend. If there is a quiet (non static) section of audio on the left and right side of the Replace, you might be able to keep audio quality better, while still generating new verses to make a longer song.

[knotty ore]

I noticed that, After Paying, I can download WAV file of those tracks which I made on free subscription.. Can I use those free subscribed tracks now?

[leaden lark]

anything you make on a free plan is sunos even after you payed a sub

you can use it, but only for non commerical things and you have to credit suno

[knotty ore]

Can I whitelist some free plan tracks by contacting Suno?

How Suno detect a free track?

[leaden lark]

you can try to contact support, but i don't think suno tracks that when you made a free track

but i'd say contact support with your situation and see what they say

[honest violet]

I just upgraded my package on suno, can my song that I made before the upgrade be commercial on social media?

[true night]

yeah yolk is a great human he also intwerview alot of specia head cases too , its a joyt to see everytime he has a new interview listed.

[true night]

i diont think so but you can still post them on yt and add that you made it\them with suno. only thing that you retrain the rights for is the lyric if you wrote it.

before you had an paying subscrition

[raven meteor]

Can I use the same voice of a generated song over and over ?

[vivid ore]

Yes, use personas

[strange mirage]

Where or whom can I contact about a billing issue? I’ve tried to reach out via e-mail but I’ve had no luck. TIA

[bronze pecan]

are you re-encoding from AAC or from WAV?
Because if you re-encode aac or mp3 files then it would be no wonder the quality drops^^

[polar spade]

Response, due to heavy traffic (and holiday schedules, likely) have been slow, it appears. What I hear, up to 7 to 10 days or so. That's the only avenue is [e-mail]. I'd expect up to 10 days is a given (given the holidays this and last week.)

[polar spade]

So, as just a member, I'd ask, what's the general billing issue?

Some issues we are hearing are about time delays, as banks posting charges are delayed (again, likely holiday related).

Others I read, where it simply seems charge cards, debit cards, unrecognized accounts, and so on are not going through. In an international perspective, that can be a number of issues, which are par for standard bank initiated issues. Banks or your card issuer usually know whats going on first. It's helps to contact them first. Thanks in advance.

Those issues seem better handled by contacting ones bank or card issuer first to rule out a locked or blocked card - i get those alot when i click more than twice to pay for something- certainly might be a faster route, while waiting for Suno billing to get to you.

Otherwise, this channel doesn't really go past the discussion of new trick and techniques of using Suno.

[vague inlet]

It seems that Suno is poorly defining the language of the song's writing. Because he uses some words or sounds in Russian, even though the word is in Ukrainian, and the letters in Ukrainian and Russian sound completely different.

[polar spade]

Make sure you post those issues in the Language channel. https://discord.com/channels/1069381916492562582/1308617723328204880

Suno is interested in all related language pronunciation issues there.

This included mispronunciations in English as well, by the way, for anyone reading.

[polar spade]

Hi all. Good morning. I can't remember, but if doing a first extension off of an recorded audio, can you use the exclude toggle? Mine stays on for a couple of seconds, then turns off. I don't recall if it was ever a thing one can do?

Update: I extended a 2nd time off the first extension and exclude works. I guess you can't exclude when extending the upload file. Huh.

[north lintel]

Hi I need help. I send a link to a friend so he and me could get 250 credits each. And he did 10 songs and we should get 250 credits and we did not get that.

[twilit cloak]

ive got this exact same problem, i got invited via a friend and invited a frined my self and havnt recieved any credits

[fading forge]

I am trying to replace a section of a song I made, but when despite selecting a specific selection, it is replacing the whole rest of the song rather than the specific section. So instead, it's kind of extending the song rather than replacing.. Anyone know a fix?

[tardy hatch]

So I've just made my first persona. I want to use it to remaster an older song I have, and use the new persona's voice to replace the old one, but keep the same melody and tune of the music. Is there a way to do that? I can't seem to find a way to get that output/option...

[untold briar]

If you just want the voice and not the overall style, it can be tricky, you may to try Cropping the original song and making Personas of the different pieces.

[tardy hatch]

Dang... maybe they will make it an update for Suno at some point 🤷‍♂️

[small dirge]

The android app keeps burning up my credits with a Spanish song. I don't know Spanish. It did it once yesterday and twice in a row just now.

I deleted them. And I don't want to share something that I don't know is.

[gloomy sable]

Do y'all have a weird problem where a persona will always generate the same exact song length? It doesn't do it on every text I send it. I don't get it lol

[north lintel]

Very weird most be a bug some kind

[untold briar]

If the length is over 3:00 it may simply be the maximum length you can generate at once with that particular Persona. (You could try cropping the original Persona and making a shorter version, if this is a deal breaker.)

This works better in a Cover, as long as you are open to changing the song a little more than a Remaster does. I personally rarely spend time fiddling with Prompts in a Remaster except to fix a specific mispronunciation. There's a real ceiling on how much it can change.

[gloomy sable]

I sent a long text lol, it cuts before it's over. I ended shortening the text and the maximum length changed?

3:16 before. 3:23 now. With half of them being slightly under since it finishes. It's interesting but slightly annoying

[lean moon]

Hey all. Anyone else getting low bitrate at the moment? Anything I make is sounding muddy with static on the high end. Wasn't like this yesterday

[austere vapor]

same since days already

[lean moon]

Yeah sure, I'll post the first 20 seconds of a cover and my original track it is covering

Here you go. The clear piano is mine, the muffled is Suno. High end hiss is a persistent problem I have to wrestle down with EQ

[sudden pike]

please help
the fake app thing came back
https://suno.com/song/29cf5edd-bd2f-475b-a506-1c0595c31b5a

[lean moon]

Oh lol, my bad.

I've used a bunch, but currently it's post-rock, ehtereal, atmospheric and piano. I've tried using just straight rock and it's doing the same thing though

Just did a new one, Alternative rock, piano, electric guitar, experimental. Same thing, muddy as all heck, low bitrate: https://suno.com/song/bae65ef4-e487-40fa-ad99-71196d06972a

[hollow hare]

You already know the procedure. Most of us cannot help you, contact Mike B and/or fill out the form given to you before (if you need it again, let me know)

[sudden pike]

i tried so many times

mike b never answers

still same errors

[hollow hare]

Talk to Mike about it and what to do when it happens again and the potential cause behind it (*if he can talk about it)

[hollow hare]

Ah, that usually is not a good sign, it may mean that there is some significant concern detected on your account from what I had observed with others

[hollow hare]

Yeah, even I would not reply to a message like THAT

[sudden pike]

im just the lucky one

i dont even know if that the right mike b

so many mike b

@mikeB

[hollow hare]

That is the correct Mike you messaged but the way you contacted him definitely is not.

Take your time to write a formal message. He is the community manager for Suno and has limited free time.

Introduce yourself, give the email associated with your Suno account (and/or Suno handle). Explain the issue. If your account was fixed before, give the estimated time when it was fixed and when the issue reappeared.

Give the estimated time when you filled out the form again (messaging him will not push you ahead, you will still be placed in a queue, it is just that he has a more direct line of communication to you if he needs to use it), and so forth

[lean moon]

sorry, just bumping this for any thoughts

[hollow hare]

Remember, you are not contacting a random person or fellow Discordian. You are addressing a Support Staff member and want to give the appropriate information

[hollow hare]

Try specifying a more specific subgenre if you can. When you have piano, the issue is that piano music is not only broad but it has been used through centuries so the training it may pull up on would be old.

Try specifying a decade, a specific style, heck even Try using a CD quality like "high quality" or "studio quality".

If you want something like a rock melody and piano rhythm, you can try:
Fusion of [Alternative Rock Melody][Catchy 2000s Piano Rhythm]

[lean moon]

Aaah thats a good point, it does seem to be the piano specifically

[charred merlin]

I'd seen a comment that insinuated that his discord DMs have a queue in terms of getting in touch with him now, so it's not necessarily a matter of DMing and getting an instant response

[placid kite]

When I view a song in a browser (Firefox, Chrome, Brave) but from not my account the default seems to show “Similar” on the right column, but "By XXX.." (artist) is empty, then going back to “Similar” is now blank. Anyone else have this happen? I have about 20 songs now. Thank you.

[high helm]

Hi how do I add new vocal on top of Style of Music for Persona
?

[polar spade]

If you already have a song, you can't swap out one vocal for another and keep the same melody and music of the original song. That would be nice, then I could just swap my voice and be the singer for all my songs I've generated. 😁

You can start with a persona, then use the prompts and add in lyrics from that song you like and see what results you get. Hopefully similar.

[rain mist]

anyone finding remastering a song to v4 just... sucks? or am I just having incredibly bad luck?

[lean moon]

Just to say man, you're 100% right on that. I think it needs to be trained on modern piano recordings.

For now I'll use it to make a backing track and record my own piano again over it. Cheers

[hollow hare]

It heavily depends on your song you are mastering. It cleans up the vocal a lot so if there was a certain quality to it (e.g. ethereal), that will be lost. And I think the sound becomes a bit... more faded per sé?

Though Remastering did help fix my old screechy v3.0 which the degradation was horrible on the ears

[sudden pike]

i found the error . Suno ai dont like the word sassy XD

suno ai allows me just to create https://suno.com/song/3a7b8181-d78e-4778-8c23-885e2bb40014 instrumentals
but when i tried to create xmas songs with word sassy included i get flagged https://suno.com/song/8361b572-399a-4f4c-9f20-e47676ef29fc

[lean moon]

You can do that. Right click the track, click "Create" and then Stems. It'll give you the music and the vocals separately

[hollow hare]

That will not be the cause of flagging your account, seems more like a coincidence. Typical reasons why your account would be flagged (even as a false positive) include using a fake app, API usage, VPN, and payment issues.

There are a lot of scam apps on the app/playstore. Ensure you have the official apps by going to
https://suno.com/Android
https://suno.com/iOS
(You will be redirected)

The official website is https://suno.com

To fill out the form for investigating the flags on your account, please go to https://sunomusic.typeform.com/to/zGsJSzsf

[sudden pike]

i have been paying this thing since began and basically create 4 songs that i really like

[hollow hare]

You have nothing to prove to me. I am unaffiliated with Suno the Company. I am providing you to some potential causes based on previous conversations observed here

Fill out the aforementioned form for the team to investigate your account.

If it is deemed to be a false flag it will be removed from your account. If the investigation deemed a breach to the TOS at your behest, your account will remain flagged (or banned, depending on the severity of the activity).

[polar spade]

Run a new song with the instrumental stem and a persona? Cool. That's news to me. thanks.

[lean moon]

I mean after you've got the song, right click on it and "Create" on that new song you made, then stems. It'll spit out the vocals and the instrumental

I just tested it myself and it works

[high helm]

Dasvon : I try generate whole new songs, with same music style and strangely it generated very similar vocals again.

[polar spade]

From my long chat with you before, you had issues specifically with your Christmas songs with , in part, reusing prompts of another member. Your Library shows a list of many of these songs, all similar, having similar prompts with just the prompts rearranged or the same style typed in multiple times.

I suggested a review of those prompts might be need as they can often get flagged. Also you're "reusing prompts" of another member. Honestly, I don't know if that would have any effect.

At least two times now, when creating totally different song from that similar group of songs which were getting flagged with the audio marks, those two times, it worked fine.

There seems to be a pattern and connection to these specific songs and prompts. I've read the lyrics and they seem pretty innocuous. It's great you're testing things out.

And cool you got a $10 discount on one month already for the 10000 credits that month. Share the love!

So thanks for your patience, at least from me. But you're only preaching to the choir to further ask for help when every member experienced in things like that have exhausted their resources.

Monday morning in the USA is about 20 hours away. I doubt there is further anyone can do to address this further for you until then.

Have a great evening and a Happy New Year!

[sudden pike]

I’m sorry if you’re feeling exhausted... that’s why we’re going to be replaced by AI robots XD

[polar spade]

That's my great fear. Haha. You're right! 🤣

[sudden pike]

We have until 2035, according to David Shapiro.

[untold briar]

Not using verse structures is probably not related to audio problems, but it could be related to audio hallucinations like repeating a verse when you don't want it to. But also Suno is pretty smart and can just figure it out, usually, even guesses the chorus accurately. I often prompt with and without structure tags just because the music style varies a bit.

You can reset or just change audio quality by using short duration Extensions, which means choosing a time near the start of an audio clip. When you do this, without using Get Whole Song, then Suno generates using only the memory up to the start of that clip. The benefit of this is that sometimes you want Suno to not be too stubborn about keep the audio quality or music the same, so if you remove memory and it only has memory a short segment (ideally a higher quality segment, but just being shorter itself also increases Suno's likelyhood of changing thing), then you can get a better quality less degraded segment. Practically speaking this usually easier to do with two or more short segments.

The tradeoff you make is that Suno is generating with an empty mind, it doesn't have access to the full song, and the melody and voices can be inconsistent. So to balance that, what you can try to do is preserve the key elements in the shorter segments, this is usually not too difficult for a voice for example. But even if you don't preserve it, the next step is to Get Whole Song and then Extend again. This will add back the musical history into the generation, and then you can choose a time that also includes enough fo the modified no-memory segment, that Suno has both memory and improves audio quality, because the recent audio quality is not as degraded.

Now, I should say that I used to do this a lot, it was an intellectual challenge to pull of a quality reset that sounded perfect. But I rarely do this now, instead I prefer to keep my songs shorter which avoids degradation by itself, and by being shorter, the song fits with a Cover and you can Cover the song to improve audio quality issues.

Really though, this sort of fiddly manipulation needs extra tools and optoins on the Suno UI side. It is awkward to do as things stand.

Is the song maintaining the same melodies or going in a new direction? Prompting an [end] in the middle of a song is another approach that can remove degradation, but it generates music that is mostly not super consistent with the verses that came before. (Suno is basically starting a new song, so it doesn't pay as much attention to the past) I like the result, but I know most people are aiming for more typical songs.

[lean moon]

Can you do structure in a cover? Do you put it in the styles section?

I'm going for instrumentals so I don't have the lyrics bit

[hollow robin]

in the simple prompt for the Song description, is it better to separate stuff by commas or does it not matter? (i mean separat stuff like this: vocals, pop, chorus)

[untold briar]

I think commas work a bit better, but whatever Suno is doing, it works pretty well if you skip the commas entirely too, so it's not super important.

WAIT, you mean simple mode. I don't know the answer to your question actually. I answered the question as if you were talking about Style prompt.

[untold briar]

Just in case you didn't realize, you can generate lyrics in non simple mode too. Simple Mode is just like skimming ahead and generating right away.

[hollow robin]

yeah thx i know, i mostly use the simple mode to let it surprise me

[untold briar]

If you want to be surprised, I suggest use Custom mode, and use Remi. Just close your eyes and click Create!

[hollow robin]

one i made that reammy surprised me was writing the styles as if it's a cinematic soundtrack, but let it write the lyris randomly really love the result, hav'nt gotten something i liked that much with random lyrics tho https://suno.com/song/77caf61b-e11a-4037-88e9-5a311e5d5102

[terse tulip]

Anyone know why the intensity of the music gets toned down when female vocals are used as opposed to male vocals?

[polar spade]

Across the board or certain genres? Heavy metal, screamo, punk, goth, musicals, I'm sure more seems about the same either one from what ive heard. Which genres seems consistently off through a few varied outputs?

[terse tulip]

Male vocals + deathcore = deathcore
Female vocals + deathcore = power metal

[polar spade]

Toned down in volume or the amount of instruments at one time?

[polar spade]

Ahhh... then that's interesting.

[terse tulip]

It's consistent for me

[polar spade]

Yea. Sounds like your saying Suno changes up the vibe to a power metal feel than a death metal when female vocals are the lead vocal.

That would be an interesting test to compare the same songs, changing only the gender of lead vocals.

Maybe Suno just relates female voices to be more Melodic?

[terse tulip]

That's the problem, occasionally I want clean female vocals over a deathcore song. If I remember correctly, it took 17 or 18 attempts to just get it in the ballpark.

[polar spade]

Sheesh. That's rough. Have you been able to set a persona so you can get more predictable results. You get one to work? If not, worth a try.

[terse tulip]

Personas don't work for me - when turned on, they cause the song to never get past the initial start phase - i.e. when the two tabs appear with the busy indicator on them.

[polar spade]

I hear you. Sorry, mate. Maybe other death metal creators can say more. Europe's asleep and late in the USA.

Suno does seem to be biased on how it treats vocals, but it's also training on death metal, I'm sure, so it overrides a bit when using female vocals.

All I would do it add more styles that might break up the female voice from being higher, melodic and , I guess less on tempo and even lower pitched.

Id search through the home pages and find songs with female vocals and see what styles are added. Maybe you can keep it from happening that way.

[terse tulip]

My attempts right now are with 4 genres combined:
Deathcore
Visual Kai
Industrial
Orchestra
I am wondering now if it's the Orchestra that is affecting it.

[polar spade]

Husky, deep clear female vocals, , brutal, aggressive, raw. Regardless of its connotation , satanic can also give a good death metal vibe in whatever vocals used.

Yea.. orchestral never gives me an epic symphonic feel, it's always more cinematic when I use orchestral.

[red bison]

I think maybe General was the wrong channel for this, so asking here.

Suno keeps skipping the (Chinese) opening line of my lyrics. How can I force it to sing them? It has no problem with Chinese embedded within English lyrics, just the opening lines.

[polar spade]

Oh. I don't do death metal BTW, but just from chats here, I get an idea of the vibe.

[red bison]

Hsi zu jan
Spare words: nature's way.

Violent winds do not blow all morning.

It usually starts at "violent winds".

[polar spade]

Are they in pinyin or character form?

[red bison]

It's exactly as quoted. Would going through Google Translate to get characters help?

[polar spade]

As most Chinese mainland songs are ballad in form, poetic and metaphoric, just a simple "Oooo" or "Mmmmm" at the beginning might get Suno prepped for singing?

[mellow moat]

Can someone help me on how can i remove a certain instrument from an instrumental?

[terse tulip]

When I mix Japanese and English, I always keep the languages in their natural format - i.e. Japanese in Kanji

[polar spade]

I think "industrial" is a good choice too, you have.

[polar spade]

Before you run a generation, you can always try the [exclude] toggle. Toggle it on and add all instruments you would like Suno to TRY to ignore.

List then with a comma between the instruments.

Then in style prompts, pick an instrument you want out front and table it as "prominent" . Such as violin-prominent, or piano-prominent, so suno will, hopefully keep other instruments in the exclude box from playing.

Will this all work? I don't know. It's there to be useful, so it's the best we have to help.

[polar spade]

Did you try one out with "Mmm" in the beginning?

[mellow moat]

Thanks man, i have turn that toggle on and added the instrument and others to ignore but in this case drums is not what i want and yet it still generates them

[polar spade]

I've mixed chinese/English and Japanese/ English and Greek/English. Using the native characters each time. I've not had any issues with dropping, so trying the actual characters is worth a shot.

[polar spade]

I don't have as much luck with drums or "percussion" in the exclusion box, but i always try. . You may have better luck adding in style prompts of genres that don't use drums. If just an instrumental, wow, I wonder if "a cappella" , Mbube, or even quartet or orchestral are often without percussion.

What kind of instrumental sound are you going for.?

A cappella: nope, it just gives back a cappella vocals ... not what we're looking for im sure

[mellow moat]

Something mexican (corridos tumbados) but with a mix of psychedelic preferably psychedelic guitar melodies

This is what i have in my prompt

[polar spade]

How did those work? Looks like a nice mix. Remember to add commas I. The style box as well to distinguish that each style should be weighted equally, otherwise Suno thinks only alteño style is being modified by all the other styles.

For example you can have ...." soft, guitar, piano" and suno weighs each separately or "soft guitar piano" and Suno tried to make a piano sound with a soft guitar vibe to it.

I did a few. They are all plucky guitar sounds. But I had pluck in the styles list too. Not what your looking for, I'd think.

https://suno.com/song/ae766688-336b-4f27-97ef-b8469ef2108d

[mellow moat]

Well corridos tumbados are mainly acoustic instruments like guitars and acoustic basses and upright basses, but also electric basses, bajo alteño is a style of electric bass where its low bass low mids but high trebel

[polar spade]

At least it seems the drums are missing. 😊

[polar spade]

Good luck with that then. Looks like you're going in the right direction.

[mellow moat]

Yeah man i appreciate your help🫡

[polar spade]

No worries. I'm off to sleep then. For that 5am alarm. Later all.

[polar spade]

Let us know what worked best to give female vocals an edge to it.

[red bison]

It took a while to find the right Chinese characters, but that made absolutely no difference. Mmmmm didn't help either.

[red bison]

Combining "mmmm" and using characters almost worked, but it still skipped a full line.

[untold briar]

Remember you don't have to make a song in one single generation. You could generate just a Chinese intro, and then Extend with the rest of your song. You could try Replacing the Chinese into the intro of a track that skipped it. You could try doing the same with a Cover of the song. And so on.

[red bison]

OMG. I finally got it after over a dozen tries.

That's smart.

[untold briar]

If you want to do it one shot, you could also do it by taking stuff away. For example add more Chinese in front of the skipped Chinese, then Crop out the beginnig of the song with the extra Chinese.

It's not ideal if you have to choose whichever track happened to not skip a line, instead of picking your favorite based on liking the music the most.

[radiant nymph]

I created a song utilizing a personal persona I created in Suno. Everything is sounding ok but Suno will not update any lyrics with that persona. I mean I can even write an entirely new song, but if I apply the persona, it just recreates the original song and lyrics. How do I get Suno to "unlearn" parts of a song and allow me to replace that section. The normal edit:replace section is doing what I'm talking about.

[fiery lagoon]

@daring lava build.a.better.world Vielen Dank für die Erklärung mit den zwei Songs pro Generation. Das Konzept ist verständlich, aber ich stoße auf ein Problem: Wenn die generierten Songs unbrauchbar sind, muss ich immer wieder neue kreieren, was jedes Mal Credits kostet. Das führt dazu, dass ich viele Credits verliere, obwohl ich nur brauchbare Songs suche. Gibt es eine Möglichkeit, dies zu kompensieren oder eine Lösung für dieses Problem zu finden?

[storm mesa]

Hey! I wanted to register with my gmail id, but I logged in with my personal id. And that id is a phone number. I hope you understand. Here is my library link:
https://suno.com/@nirmalkumar74

Actually what I want to say is that you should understand my problem. I want you to think about it a little.

I have been a long-time user of your platform and deeply appreciate the musical experiences you provide. However, I have noticed the absence of Bollywood or Tollywood classical music beats blended with Western culture music on Suno.com.

I strongly recommend adding youthful Indian-Western classical blend film background scores, featuring influences from both Bollywood and Tollywood. This fusion would not only create a unique listening experience but also attract younger and more diverse audiences.

To support this idea, here are two inspiring examples:

1. The announcement teaser of Salman Khan's upcoming movie SIKANDAR.

2. The song Kar Har Maidan Fateh from the movie Sanju.

Additionally, incorporating Tollywood-inspired tracks, such as Naatu Naatu from RRR, would further enhance Suno's library by showcasing the energetic and melodic symphony that defines this genre.

By combining Bollywood, Tollywood, and Western beats, Suno.com can revolutionize its music offerings and provide users with an unforgettable experience. I hope your team considers these suggestions, and I look forward to hearing your thoughts.

[patent yoke]

did suno push a new update? remasters sound more creative and more emotive with v4

[hearty verge]

Hello, there is a major problem with suno at the moment. I have wasted hundreds of credits, on songs that are broken, and suddenly it spits out some weird jazz instrumental tracks, that have nothing to do with the prompt. How do I get refunded for the broken songs? Out of 100 versions only one kinda worked, the rest were straight up broken. With sloppy sound quality and messed up my lyrics and leaving parts out and suddenly speaking a made up language.

[fluid gale]

I'm having a problem with one of the users I use.
The songs I create are cut off after about 1:45 minutes.
I noticed that with the other user I can create songs properly with the same prompt.
Can anyone help me understand why this is happening?

[shrewd turtle]

can we use a song made by suno if we mention that they made it ( for a game)

@pizza for everyone

[storm mesa]

"Apologies! Bollywood songs are indeed created here, but I didn’t explore the platform properly at first. When I reused the prompt for one of my songs, I realized this feature truly works. I even heard the sound of the tabla in a song. If you'd like, you can listen to it as well through this link:
Listen to the Song:

https://suno.com/song/d3b4dc8b-1925-4067-bc53-d80c53834125

Technically, the song is fine, and the tabla effect is clearly present. However, I faced an issue where I couldn’t make the song energetic enough. My goal was to infuse motivational upliftment into it, but that element is missing.

While the lyrics aren’t followed completely, the app still delivers incredible features. Suno.com is my favorite app and website for creating songs, but if I seek improvements, I believe it’s important to share my feedback.

Thank you for providing such a great platform!"

[polar spade]

I don't speak Hindi nor am raised in a culture where I grew up on Bollywood, but I use Bollywood singularly and in group prompts often to get the sound or vibe of what I have considered Bollywood from just my watching movies and listening all the time on Spotify.

It likely takes one who is more engrossed culturally to understand better the nuances of those two genres Bollywood and Tollywood to see a difference from what seems to be outputs in Suno.

That's cool to know.

I've enjoyed what I hear, and even compared to Har Kar Maidann Fateh, it's seems the outputs ive had are quite similar, but know my ears may be missing something only a true fan of the genre would hear.

In the end, honestly, my attempts are more to add the Bollywood vibe into other genres to morph into a new sound. So I wouldn't always hear a "pure" genre alone to be able to critique.

I haven't tried Tollywood in the mix, though I know the sound. I'll have to throw it in prompts and see what I get. I use "edgy" which helps turn bollywood a bit darker or rock out more, so Tollywood might just do the trick, instead.

It's even fun when Suno hallucinates in what I would guess is in Hindi-ish language. I have a Japanese/Hindi mixed gibberish song which is fun to listen to. Granted there are dozens of dialects and languages besides Hindi, so yea, I'm stereotyping the sound and just guessing. But fun , anyway.

[storm mesa]

"Thank you for your response, as you highlighted the discussion about Tollywood and Bollywood songs and talked about the genres. However, to be honest, I didn’t fully understand all the English, but that’s okay. At least my message reached someone."

[polar spade]

Using Tollywood really brought an edge to the song. Glad you suggested it. I'll use it more.

https://suno.com/song/c1c68105-6105-418a-bb52-f66533576ec4

[mortal barn]

Hello does anyone know why I can create song with male voice, I'm creating and I describe that I want with male voice and it doesn't work

[polar spade]

Sorry. Everyone says I talk long and write walls of useless words. 😄

[mortal barn]

Help

[storm mesa]

Oh, no problem! I like to listen and read in detail, only then I will understand everything clearly. I have translated this chat of yours, and I have understood everything you said.

[polar spade]

By placing [Male Vocal] as a tag before each verse can often help.

At minimum, [Male Vocal] or [Male Vocals] at the beginning of the lyrics.

Include "Male vocal" in your styles prompt.

Those two things can have a good influence.

[mortal barn]

OK I will try it

[storm mesa]

Absolutely

[mortal barn]

[Verse 1] [male vocal]

Is that the right form?

[polar spade]

That would work, but better may be to put [male vocal] a line above verse, so ti doesn't actually/incidentally sing/pronounce "Male vocal".

[Male vocal]
[Verse 1 ]

[mortal barn]

Thanks Mr.

I wii try it right now

[storm mesa]

I got it!

[polar spade]

If it works, you're welcomed. If it doesn't work, you can curse me. 😅

[storm mesa]

I heard the song, now I am a little inspired by it.

[fluid robin]

1 Theatrical. 1.1 Noh. 1.2 Kabuki. 1.2.1 Geza. 1.2.2 Shosa-ongaku. 1.2.3 Ki and tsuke.
2 Court music (gagaku)
3 Shōmyō
4 Jōruri.
5 Nagauta.
6 Shakuhachi music.
7 Sōkyoku.
8 Traditional music in modern culture. for japanese songs in style box

[polar spade]

Looks interesting. I'll save for reference. Thanks.

[fluid robin]

gives you a more Traditonal style instrumental instead of the modern one

[mortal barn]

Yeahhhh it worked, thanks

[polar spade]

Alright! 🤘

[harsh radish]

Did someone had manage to have the sound of a guitar bass?

[storm mesa]

You should try creating a song in this "meta tag style":

Epic Motivational Triumphant Symphony: Tollywood-Bollywood Background score with An Uplifting Inspirational Fusion of Electro-Orchestral Synth Pop, and Heroic Pop-R&B Dynamics with Melodies

You can customise it accordingly!

[lofty quiver]

Trying to add Suno's API into Bolt.new but keeps giving errors

[unique stone]

Suno do not provide an API. Any APIs you find are unauthorized reverse-engineered hacks that may put your account at risk of being banned or restricted.

[strange stump]

On web browser I can use a persona with v4 but on phone browser whenever I apply a persona it defaults to v3.5 and I can’t change it. Is that right or am I doing something wrong here?

[fading tulip]

I’ve been exploring Suno AI as a free user, particularly the V4 model, and I’ve encountered some challenges that I believe could be improved. Unfortunately, I’ve wasted 10 V4 songs due to the 4-minute time limitation, as the tracks end abruptly before reaching a natural conclusion. As someone who’s eager to experience the full potential of your tool, this has been quite frustrating.

I understand the limitations for free users, but it would be greatly appreciated if there were options for longer song generations, or at least a smoother way to manage or extend the song's length. I believe this would significantly improve the user experience, particularly for those of us testing the capabilities of the AI.

[lofty quiver]

What third party API could you suggest?

[unique stone]

None. Suno don't allow API access.

[lofty quiver]

Thanks for you help @unique stone
Looking for an API as good in terms of music, do you know any?

[charred merlin]

Stable Audio Open might be more what you're looking for -- https://huggingface.co/stabilityai/stable-audio-open-1.0

[crude tundra]

I Can Make ChatCPT API Keys If You Need Help

[stable barn]

I just cant work out how to use replace

[polar spade]

That's the crux behind having abilities to extend. Have you found "extend" yet? I'm surprised 10 songs would be considered "wasted". So perhaps you have not found "extend" as our ability to extend or extend to a natural conclusion.

Songs in real life can be anywhere from a minute to a 45 minute symphony. There is no standard. But there has to be a limit per song to be economical, marketable and not strain the system as well.

Online streaming is tailored to songs under 4 minutes (3:15 and under, specifically for Spotify) as well, for having priority play over other songs.

So, a shorter song to end before 4bminutes without further effort is often coaxed by having less lyrics, if that helps. But "extensions" are our friends.

[lofty quiver]

Nice one @charred merlin thanks

[lofty quiver]

With lyrics?

[crude tundra]

yes if you like

[lofty quiver]

sure can I see sample?

[crude tundra]

iv been using ChatGPT 3.5 and 4.0 for SUNO sent it came out its pritty good still needs some work tho

and yes

[fading tulip]

Fine.

[crude tundra]

all make a small video

[charred merlin]

...you mean you're offering to help people create their own API keys, right?

They've made it incredibly easy to create your own API keys

[stable barn]

what is an API key?

[crude tundra]

yes

[charred merlin]

Nothing that you'd need for Suno -- it's something you'd need if you're builing your own software, or using a program that requires an existing account from elsewhere (some services have all of the UI/UX stuff handled, but require you to provide your own OpenAI API key so you pay for your own AI usage, instead of the people who built the platform

[charred merlin]

DUDE you shared your org key @ 0:17 👀

You might want to delete the video or change your keys, if web security is something you value 😅

edit -- on second glance, just the org ID -- if you're just fiddling around with personal projects, likely not going to be an issue -- hard to shake off enterprise-level web paranoia

[stable barn]

oh ok Im messing with AI but not on that lvl

[lofty quiver]

thank you @crude tundra. Much appreciated! Im already aware on how to get API keys from OpenAI.
How to generate music with lyrics using the API is more like what I'm looking for...

[lofty quiver]

Thanks for the video!

[crude tundra]

Have Fun Maybe This Can Help

[vivid cedar]

Hi Happy New Year. New Member here. Anybody know how to extent an intro in the beginning of af song i Suno?

[crude tundra]

Yes

[hollow hare]

Hello, please see https://help.suno.com/en/articles/2409601

For more tips, see #1294418290348921007

[vivid cedar]

I am trying to fit a airy song with video and would like to start the music some seconds before an maintain synk

[hollow hare]

Ah, you cannot prefix/attach a track to the start of another track in Suno, only continue.

You will need to do that externally with an audio editor like Audacity

[crude tundra]

True

[vivid cedar]

Thanx! thought so. I am using Audacity for now😆

[charred merlin]

It's by no means credit-friendly, but I think it was Opti who'd suggested using "replace section" to replace the intro section of a song -- since it can "glitch out" and return much longer segments than what you were requested to replace, you could feasibly select the first 10 seconds of your song and get a 4m replacement 😂

[hollow hare]

Oh, very good point

[fading tulip]

Devs, why Suno counts my own plays of my songs more than once?

[vivid cedar]

Maybe better than nothing. Subscriber, so sufficuent credits. Any efficient help on endning a song without unnesecary prolonged. Tried Outro, End, etc often without luck and some frustration😆

[charred merlin]

There'd still be a number of uncertainties to get that method to work the way I described, so I bring it up more as a proof of concept than a route-one suggested method, but it's a curious possibility for sure

[polar spade]

Fine, as in, you've used extensions before?

Many of the seasoned members here even prefer extending songs a little at a time to get a more creative, varied instrumental or lyric song.

If you listen to enough generations, Suno has a faulted quality of starting out great in the first verse or 1 minute, then it starts fading out the fill-in and accompaniments as the song goes along. Extending in segments can help keep a livelier song by its completion.

Hope you find your groove and have fun!

[charred merlin]

v3.5 and v4 seem to have issues winding down a song in the same way that v3 did -- while v4 leaves me with enough sections where I could crop & fade out in post, recently I've been swapping to v3 to extend with just end as the style prompt, then remastering the resulting song to bring it back up to v4 quality

[fading tulip]

Yes.

[crude tundra]

[vivid cedar]

Regarding New intros I have experienced Suno 'looses style wareness' if making a new start. (Cant read style forward)

[charred merlin]

As in entirely different styles/genres, or "similar enough" genre, but feeling like it belongs in a different song?

[crude tundra]

V3.5 Was Better But V4 Still Needs Work But Reminder SUNO Better Then Other AI Music Making APPS Just Needs Time

[vivid cedar]

Not entirely lost, but specific singer not the same

[charred merlin]

That's a common report -- from the examples I've been shown, it usually sounds like a similar vocalist, but the audio quality is different, as if "recorded on a different microphone".

If you have any results that sound drastically different, I'd love to hear -- the devs had been asking about Replace Section segments & Extensions that didn't match, so I've been on the lookout for examples that demonstrate this clearly

[vivid cedar]

properbly deleted them. I will have it in mind next time

[polar spade]

Not singers, but spoken voice, is what i focus on. That's never changed much that with each generation or replace sample the dialects or inflections/ emotive qualities change within the parameters of being male or female. I just assume that's not t he same as a singing vocal which you can sometimes recognize the same female or male over hundreds of songs.

Persona's help a bit, but I don't know that I've done an all spoken voice and then made a persona out of it to apply to another spoken voiced song or prose. That can be a project to embark on.

[storm mesa]

"I have never understood what 'Persona' means in terms of 'Suno'."

[strange stump]

On web browser I can use a persona with v4 but on phone browser whenever I apply a persona it defaults to v3.5 and I can’t change it. Is that right or am I doing something wrong here?

[polar spade]

It's use will maintain a consistent quality of voice between or among the newly generated songs using it.

I've had fun using persona, creating a whole album of songs, old and new for a fake artist.

[strange stump]

Thanks - also can you use Personas on app? I can’t see the option

[storm mesa]

"Alright, I will try it once."

[storm mesa]

"I have observed this multiple times, but when we select the v4 option, the 'create' button below turns blue, allowing us to generate v4 songs."

"And I have tried this several times, and it was giving the correct results, even if it showed the v3.5 option."

My response was a bit late

[charred merlin]

I can confirm I'm also experiencing this bug on mobile browser, but isn't the case on desktop browser.

[daring lava]

Hi @fiery lagoon 👋

Thanks for reaching back! If you'd like to, we can work through a prompt exercise together to see if there's something about the prompt that can help get the generations you are trying to achieve. Feel free to ping me if interested. Additionally, you can always reach out to [e-mail] and explain to them what is going on and see what they have to say. They're a very helpful team!

[mortal horizon]

Anyone else having issues logging in with Google's OAUTH?

It could be just me.

[charred merlin]

Just tried it without issue

[mortal horizon]

Awesome, thanks for confirming.

[storm mesa]

"I had some issues with the tabla effect, but I tried it. I even attempted to email the Suno team and am now just waiting for their response."

"But this issue wasn't resolved easily, as I checked. I took a prompt from an existing song, added my own genre, and then generated my song. Only then was I able to see the tabla effect."

"There was just one issue with this—the song did not turn out to be energetic."

[placid kite]

Looking at a song from a different account on different browsers (Firefox, Chrome, Brave) the default seems to show “Similar” on the right column, but "By artist.." is empty, Anyone else see this? Thank you.

[near timber]

Is Suno broken?

Bc when I try to make a song it says I have 0 credits but I have 2,500 premium credits

On website version

But when make a song on web it says can't make song

Not enough credits

Even on the app

[charred merlin]

That usually indicates an issue with the payment details on https://suno.com/account

[near timber]

Yeah but I don't understand 😔

[charred merlin]

It should be noted that while other users can offer advice, Suno's Discord server doesn't operate as an official support channel -- you'll need to contact [e-mail] to get official support on the matter

[near timber]

yeah

[dense tiger]

Hey guys. Why i can't download a wav file from suno? it's shows ''Download failed, please try again.'' I'm using desktop.

[charred merlin]

Not sure if that issue has ever been identified, I always assumed it was a temporary server issue, but a common workaround is to use Firefox to download the .wavs from Suno instead of Chromium-based browsers (assuming you're using Chrome / Edge / Brave atm)

[warped furnace]

Have you tried another browser? Sometimes that helps. Just now, I have no problems with downloading WAV. Just video regeneration (has to be done up to four times).

[dense tiger]

tried. didn't helped

[charred merlin]

Which browser did you try? Lemme know if you still have the issue on Firefox.

[dense tiger]

edge, firefox, opera

[charred merlin]

Okay -- since it's happening on non-chromium as well that paints a different picture.

Are you having issues with one song in particular, or are you unable to download any .wav from Suno?

[dense tiger]

second. actually, when i just subscribed, all was good. Problem started a week ago, mb less.

[charred merlin]

I don't think there's been anything changed system-wide that would have impacted your ability to download .wavs -- if you're able to use a mobile browser to access Suno, I'd be curious if it errors out similarly.

Typically it's a browser issue, which we've ruled out, but it still could be something related to the device or the network, if you're on a VPN. Easiest to test would be whether it persists on a new device

[hollow hare]

Do you have "SUNO" appearing on some of your thumbnails or an audio saying to use suno.com?

[dense tiger]

works on mobile, and i don't use vpn

[dense tiger]

sorry, don't get it

[hollow hare]

Do you have SUNO logo on your song picture when the song is made? Like this

[dense tiger]

no, never saw that logo at any song pictures of mine

[hollow hare]

Okay, so that is not the issue then, hm

[hollow hare]

Can you DM me the link to the song?

[dense tiger]

well, i just closed all browser windows and relogined again in suno, it's working

idk what is that kind of magic or what, bc i did everything from relogin to clean up pc, but ok. now it's working. thanks you all for your time

[hollow hare]

Ugh, technology. Great to hear it is working again for you 🎉

[rough reef]

Hello. Merry Christmas. I'm subscribed to suno and I'm afraid of a problem. When mastering the songs to V4, he swallows words or doesn't say them at all. Always with the same words, especially "amanecer" Suno says "amanene" or makes up words. Have you thought about adding some kind of button to recover the credits of songs that have gone wrong? I'm losing credits unnecessarily due to AI errors. I think the problem only occurs in the Spanish language.

[urban sonnet]

have you tried phonetically? i.e. a.ma.neˈseɾ or ah-mah-nay-sehr

[junior spruce]

Good day, I'm a noob here. Have a question--wanted to regenerate video while adding some of my own images, is that possible?

[crude tundra]

Yeah SUNO Ben acting weird lately for me too

[rough reef]

Sometimes it works, but not always. In Spanish, it's always the same words. "Amanecer" always makes the same mistake 90% of the time.

[crude tundra]

SUNO ﹌﹌﹌﹌﹌67%

[charred merlin]

Merry Christmas to yourself as well 🙂

There's a megathread for posting language & pronunciation related issues with v4 https://discord.com/channels/1069381916492562582/1308617723328204880 -- as far as refunds for the credits, that isn't something they've typically offered. https://help.suno.com/en/articles/2550209

[rough reef]

THANK YOU!!! I have more than 100 songs with pronunciation errors, made-up words and crazy things from Suno in Spanish.

[warped furnace]

If you have a subscription, you can change the cover pics. After that you have to regenerate the video. But beware, it takes 2 to 3 times because it tends to crash the first time. Keep your eyes on the screen for the notice. 😁

[charred merlin]

You're welcome -- to note, reporting the issues in the megathread doesn't guarantee immediate resolution, but since it's a common enough issue, it allows the devs a place to see related issues -- it should help get relevant eyes onto the issue, but you might also find other users who have found clever workarounds to similar issues 🙏

[charred merlin]

They updated it a while back so free users can change the picture, but it only stays updated on suno.com -- but paid members can use the AI to generate new pictures, and ... > Download > Regenerate Video to change the background image for the discord embed/downloaded .mp4

[warped furnace]

Ah, thanks. I didn't know. But still. If you regenerate, you have to keep your eyes on the screen. 👀 It never works the first time with regenerating. At least not for me. 😅

[fading tulip]

Devs, why Suno makes profile pictures blury/low quality?

[charred merlin]

I've only had it fail once or twice, but sometimes it takes a while for the video to actually regenerate -- if I go back to ... > download, it's still spinning up to a couple minutes.

[charred merlin]

🙏 the devs don't usually answer questions in chat

But if it helps, the pictures that get autogenerated with the songs are 1024x1024, but images generated afterward are capped at 360x360, but you can always add your own images, as Pantera had just been suggesting 💪

[fading tulip]

I use 1024x1024 1MP images and they are blurry after uploading.

I tried webp/png/jpg... Same issue

[charred merlin]

hrm -- I was wondering if they downscaled images, but I just uploaded a 1024x1024 image and Suno seems to be using the exact same picture.

https://cdn2.suno.ai/27c63fb6-e1ac-425c-a78c-b6144f807a5c_2edfc227.jpeg

@fading tulip Can you share an example of a picture you've uploaded, and where you're seeing it degraded?

[fading tulip]

It's picture on my profile. Not for songs.

Your profile picture is blurry too.

[charred merlin]

Ahh -- the profile picture seems to be 256x256

[neat field]

i suspect that Exclude Styles and v4 training may have a flaw in the training set and therefore some genres are undertrained for V4 or do not populate the tokenizer correctly

when using "Exclude Styles" in the style description, afterwards there is a "-" hyphen in front of styles from "Exclude Styles"

my sample size is low so far, but v4 subsequently does not seem to handle hyphenated genres well (such as "nu-metal")

so when "nu-metal" is used, it never sounds like v4 gets that genre right so far

oddly, remastering a v3.5 "nu-metal" goes very well

not sure if anyone else has similar hyphenated styles that have the same behavior, leaving here in case it helps someone

[bronze pecan]

Can I somehow remove the vocal out of this otherwise cool instrumental?

https://suno.com/song/52154c3c-750f-4514-9fc9-0fafe9495512

With the choir prompt I assumed more something like this
https://youtu.be/ho_R0yUxENA

but it implemented these vocals instead apparently^^

[near timber]

Weird I can't even upload anything I'm not sure if it's a bug but it shows I have 2,500 but says I have no credits I contact Suno support but no help 😑

[tiny falcon]

The Gregorian Chant is cool. Suno can do it? I've been trying to prompt Gregorian Chant but all I got is musical interlude 🙂

[near timber]

I feel like I'm only one with this bug

[unique stone]

This can indicate that a subscription payment was not successful. Try editing your billing details on the subscription page to see if that fixes it.

[charred merlin]

If you just emailed them today after our chat, you might have received an automated reply informing you theres a higher than normal support ticket volume, but they'll get back to you as soon as they're able 🙏

[near timber]

Oh I tried too

I don't were to fix it

[sudden pike]

I’d like to share with everyone that Mike B managed to help me fix my flagged account.

We came to the conclusion that maybe my account was possessed by some spirits XD—just kidding! The issue might have been that I was using Suno on Microsoft Edge (Version 131.0.2903.112, Official Build, 64-bit). The simple solution was to switch to Google Chrome.

Now my account is back to life and ready to produce!

[red bison]

Is there any trick to get just one line of lyrics? Suno keeps improvising. I'm considering repeating the opening line several times.

[sudden pike]

be more specific ? example ?

[red bison]

I'm struggling to create songs with an opening line in Chinese, the rest in English, so I'm trying to do just the opening line and extend it.

[bronze pecan]

conclusion: Microsoft Edge is a fake app
and now you got a real webbrowser 😛

[haughty falcon]

Hello there, I ask the question a little out of spite, but is there an ETA for the correction of shimmering / artifacts?
It appears on the majority of the generated songs and it greatly reduces the listenability

[sudden pike]

mixing languanges is quite challenging for SUNO so maybe you can try using metatags [female chinese language ]

[neat field]

maybe the model was trained with a "polyglot" style tag, although more likely there is "Bilingualism" - just speculating here

[red bison]

Hmm. That's a good idea. I haven't tried bracketed tags.

[wild spade]

The shimmering effect is gone?

[hollow hare]

Can you provide examples to some of those songs? As far as I know, shimmering was significantly reduced

[warm stag]

Hi guys, I'm new to suno and I find it to be an incredible platform.
I have a question and I'm hoping someone can help me.
I recently created a song and sent it to some of my friends to listen to it.
I created it with a free basic plan and I was wondering, if some of my friends, without telling me anything, decided to use it for commercial purposes, would I also get into trouble because I created it or would only they have problems with their respective social channel/account?

[charred merlin]

Hrm.

I'm glad it's up and running again, but I've been using edge this entire time without any issues, both Edge & Edge Dev -- I'm noticing edge dev is on v133(...) and edge is on v132(...) and I've never had any of the watermark issues, so maybe it's because you were on v131, but it still sounds like there's a degree of mystery in your instance.

[sudden pike]

you are lucky

[charred merlin]

[charred merlin]

You might still find it in occasional generations, but it's been improved upon

[red bison]

Dangut. It took so many tries to get an opening I could extend, and then the extension ended up singing the right opening line. I should have just kept banging the "reuse prompt" button.

[wild spade]

niceeee finally

[charred merlin]

"getting in trouble" is a relative term, and it would depend on what kind of trouble they're running into.

If a youtube channel makes money, using Suno tracks in their videos would require commercial usage rights.

[fading tulip]

I invited friend. He created 10 songs. But I don't received 250 credits.

[warm stag]

totally right, but to get in ''trouble'' will be the person who post it right? not me that i created it, just wondering. thank you

[charred merlin]

Oh heck yeah -- remastering old tracks that were really impacted by shimmering seems to "tidy up" most of them, with some exceptions still giving me trouble, but it's a noteable improvement in the frequency of "hits", imo

[charred merlin]

If "getting in trouble" means having the youtube post taken down, it's likely to just be them.

As Lumine is about to suggest, if they're involved in a lawsuit or other legal action, you might get involved.

[hollow hare]

Depending on how far the legal entity that is suing you/them wants to go.

For example, if the song was a "cover", "remix", etc. of a copyrighted song from a major recording label (like Sony), certainly they and/or RIAA will pursue all parties involved, which can include you from information obtained by sending a court letter (well, subpoena) to Suno.

The likelihood of that happening? Low. But a possibility all the same

[charred merlin]

If they post your content without you knowing, you'd be a lot better off, but if you know that they're uploading content that they/you don't have the rights to, and you're providing the songs, it complicates matters.

[unique stone]

Also, all Suno songs contain an inaudible watermark. There is no information available about the information contained in this watermark and as far as I am aware only the Suno team have the ability to examine the watermark.

At the very least, it could identify that the song was created using Suno. It might also indicate whether the song was created on a free plan or a subscription. But it's possible that it might contain the unique song ID (this is what I would store in the watermark if I were designing it.)

I would simply not share songs with friends that I do not trust. That goes for things I have written/composed myself.

[warm stag]

I understand perfectly, my only fear is that without telling me they would publish it and I would end up in trouble. It's obvious that I would never give consent as I would only like to use the songs on a personal level.

[charred merlin]

I think it's unlikely that you will face issues unless you are intentionally being sneaky -- you cannot control whether others act badly, but if you know they are acting badly, you can protect yourself by asking to not be involved.

The bigger fear should be that they would use your song and make it famous, but people would not know you wrote the song 😂

[warm stag]

ahahahahah you are right. thank you guys

[charred merlin]

@fading tulip I think it might take a certain amount of time before the credits are added to the account, but you'll need to contact [e-mail] for billing/credit related issues, since the server is basically for other users to chat with each other -- if it helps give perspective, most of these issues seem to be resolved shortly after emailing (whether it's because the support team solved it, or it just took a day or two to update, not sure 🤷‍♂️)

[fading tulip]

Does it mean 250 credits for v4 or v3.5?

Total 2500

[charred merlin]

The credits can be used for any model, and the total just means you can only earn credits from inviting a maximum of 10 friends (250 x 10)

[fading tulip]

I invited 2. They both created 10 songs. But I still didn't received 250 credits. Could be an issue you talked about.

[red bison]

How do I remove a song from a playlist?

[hollow hare]

Allow up to 24 hours. If you do not get the credits for the invite after that, send an email to [e-mail] and they will credit you the missing amount when they view your ticket

[hollow hare]

If it is your playlist, view the playlist entry then click the options of the song listed and it should say Remove From Playlist.

If you are in Mobile View, there appears to be a bug in the context menu which does not show it. Instead, you will need to be in Desktop Mode to get that option

[red bison]

Ah, that's it then. Gonna see if I can change view.

[bronze pecan]

Can I somehow force a minimum length for instrumentals? It did a 1 minute song^^

amazing 😂

[gusty cosmos]

hi does anyone know how to remove "made with suno" from the top left corner if you are using a video you created then adding music with suno in the create / upload picture

basically suno watermarked my image with made with suno

[charred merlin]

@gusty cosmos like this?

Or do you mean "Suno" is written on the actual image itself on the website?

[gusty cosmos]

hey fam

check it the issue is i uploaded a video

then said create a song

when i download the video

in the top right is this

[hollow hare]

All videos generated by Suno has the "Made with SUNO" tag, like how theharber showed you above.
If you do not want that, you will have to make your own video which plenty of people do

[gusty cosmos]

i got it - i made the video already i just will need to separately add the audio
thanks team

[neat nexus]

Using simple mode and your text completely unchanged:

[Removed links]

With simple mode it tries to extrapolate and build something based on the request. It may or may not get close to what you want and it will edit the style and create lyrics for you.

@storm mesa

Here I just reused the lyrics that were auto generated above but did some editing.

[Removed links]

For style I used your prompt for inspiration and rewrote it myself:

Bollywood Background score, Tollywood style musical interlude, uplifting electropop, heroic theme, pop R&B, dynamic melodies

Then in the lyrics I added some [metatags] to indicate what I was looking for.

If I'd put some time in to search for specific instruments and knew more about the style you were looking for I could tweak more specific metatags but I think you get the idea from this example?

@storm mesa I hope this helps you figure out how to get what you're trying to get from Suno.

[storm mesa]

Got it!

[eternal topaz]

so found an interesting issue. i gave up as I couldnt solve the problem but the AI seems to not know how to say Strawberry XD i doubt theres a way to fix it but if there is i figured I post the issue here

i tried even straw-berry and straw berry and even straw berry but it always goes "Strawberberrish'

[charred merlin]

https://suno.com/song/db60a134-1669-4804-94b2-0cce48e5ba99

Doesn't seem to be an issue with that word in particular, but it's definitely fallen into some sort of groove if it continues to mispronounce it in your case.

[soft pawn]

I'm still getting shimmer in basically all my v4 generations. Is there ANYTHING I can do?

[west yew]

Try again tomorrow. Seriously. Some days the shimmer is essentially gone. Three days ago I had so much Morse Code going on I could have sworn they were broadcasting the secret location of the Ark of the Covenant to anyone who could decode it.

[hollow hare]

Can you please provide a link to one example? As far as I am aware, shimmering has been significantly reduced and you may be experiencing another artifact or likely something related to the prompts used

[soft pawn]

https://suno.com/song/66adc16c-c535-4df1-8b8b-d8a9977cbef4

You can hear it immediately.

But honestly, even if it's "another artifact" it shouldn't be there.

[hollow hare]

Well, definitely shimmer in the first few seconds but I didn't hear it anywhere else. Did remastering fail to get rid of it?

[ripe spear]

Hey guys, can anyone help with this issue. It seems to have allocated a persona to a few songs, the persona definitely wasn't used to create them. is there a way to remove the persona?

[hollow hare]

Not really since it is baked in, you will have to recreate a new one from scratch to remove it I think.

Also verify if you actually removed the persona. Happened more than once that I forgot to remove it from Create as it was persistent

[ripe spear]

Ok, thanks for the advice, appreciate it. Will have to pay more attention next time

[abstract mural]

can I upload the songs I made in the free version on spotify

[hollow hare]

As per #1294423389515284552, no from the free plan. You do not have the commercial license for it (and you will need that in order to distribute)
https://help.suno.com/en/articles/2410177

[eternal topaz]

lol rock on strawberry XD

[abstract mural]

Can I get a license for the song I made on the free plan

[hollow hare]

Also in #1294423389515284552 :
https://help.suno.com/en/articles/2425729

You could email [e-mail] after subscribing the first time to ask and they might give it

[bronze pecan]

ok 3 , you won xD

[eternal topaz]

https://suno.com/song/95b63436-b1ca-48c3-85de-cb614fd9193f it seems to be sort of a hit or miss. I rendered four songs two did fine one did ok and this one XD

[delicate bloom]

Help please.. I am constantly getting system errors trying to create a song.

[polar spade]

Wonderful! And thanks for posting an update when things work out. If more like you could do that we all as memebers and peers might be able to help others more efficiently. Glad you're up and running. Have fun!

[polar spade]

Here are those same prompts shortened to 1-word prompts:

Motivational Symphony, Orchestral, Synth-Percussion, Uplifting , Triumphant, Tollywood, Bollywood, Melodic, Synth Rock, Heroic, Inspirational ,Rhythmic Progressions, Contemporary R&B

I ran a few outputs and they sounded well enough similar to songs with your "phrased" prompts.

Have you used one word prompts instead of phrases to see if there is any difference?

But in the end, they all sound pretty good, so it may not make any difference between the two.

[crude tundra]

PLZ Fix

i cant go past 3 minuets for any song i make im pretty sad i pay for this shit for this to happen plz fix devs Thanks

[upper fox]

I am having the same issue as Bush Cat as well I wrote an email. Please help when possible

[crude tundra]

@upper fox Yeah Its Been Like This Sents yesterday for Me ): I pay a lot for suno too sad days

Hopefully They Fix It

[upper fox]

mine also cuts mid way on my last hook and dosen't finish the song. its been 2 weeks for me

[crude tundra]

Shit Im srry there Here that

that sucks

[upper fox]

they are on their holiday break so I hope they see my email and this once they get back.

[crude tundra]

They Never got back to my last email and its been weeks so i thot is i get in there dc they see my Bug reports Im A Small Dev But I Could Fix It My self in the api but it be agents there tos

Sad Devs Just Want Money All Day

Really Dev Look Out For The Community

My VRC Dev Get Back Too Me Faster Then This Team lol

[upper fox]

I disagree they are probably working on other bugs and other feature and on top of that its the holiday its understandable

V4 was probably a big project

once they get back it will be worked on

[crude tundra]

Yeah Same Just I Guss I should

[crude tundra]

you right you right

[autumn vault]

I hope they fix the audio degradation soon

[crude tundra]

not saying there not good at there job it is a lot of work to do on there team there busy bees

yeah fare fare

they did fix one of the bugs i emailed them but no reply back is all a reply back to my email would be nice of them

but i get it there big team so they got a lot on there palter

[steady kelp]

can you DM me with the email you used to reach our team? happy to check on it for you 👍

[crude tundra]

Thanks @steady kelp Yes i got you

[upper fox]

Thanks @steady kelp for the support by any chance can you follow up on mine to if possible?

[steady kelp]

shoot me a dm!

[runic elk]

Hey, how's its going. I am new to Suno and I love it. I do have a persona question is this the right space for it?

I made a persona and I have one song in it. I made a few more songs from the same persona but they do not appear in the persona even though I used the persona. Is there a way I can just add songs to a persona without redoing the song?

[eager condor]

A while back I had a couple songs with the suno watermark on the art, and I've heard stories lately of audio watermarking. I was browsing the homepage, and one of the people I follow has about 40 new gens ALL with the watermark on the image, so I listened to one. You can hear the watermarking, and it does sound like "download the app"
https://suno.com/song/8bfef312-fbd4-4f61-a542-1a34ff76468f

[smoky dagger]

whats the point of doing a cover if it just changes the sound? I got it to make a great song with an accordian at the beggining but when i do a cover it completely gets rid of it.

[half dragon]

for me, the biggest gripe is the degrading of the sound on extension... like some kind of filter was added... creating this ringing, air whooshing sound at high pitch... the 2nd biggest gripe was the Replace Section issue.. it's always messed up the section, as I wasted so much credit trying to fix a word or two in the lyric...
I DM'ed @steady kelp on this but I'm not sure what I should do? live with it or wait for an update
https://suno.com/song/e05b0663-ce57-493b-b8c4-fcaff904a051

[steady kelp]

how's this? (maybe crop to 1:03?) https://suno.com/song/38d2326f-d70e-479f-9463-de050ff619d1

[novel creek]

Hello is there a way to create a ai voice based on our own voice and use it to create our music songs?

[glacial bobcat]

Hello, i wonder if it's possible to recreate the same instrumental of a song that i generated with lyrics, i freakin love the instru but i couldn't find any tips to do it... Any solutions ?

[latent lagoon]

hi every one! ia have a problem with personas 😦 I created a persona of a song I had created previously but which unfortunately was sung badly. The problem is that if I reuse this person, he doesn't consider the new text I give him at all and he always makes the same mistakes. Even if I delete entire sentences he always sings the original song. what can i do ?

[soft karma]

Hey, does anyone know if you can get a response from [e-mail]?

[storm mesa]

"Not yet."

[hollow hare]

They have a high volume of email to go through, like 8 business days worth.

[brittle hemlock]

Where do I go for troubleshooting? I have 1500 credits but whenever I try to make a song it says I need atleast 10 credits... I havent used the website in a few days.

[hollow hare]

Sounds like a payment issue, did you check if you got a message about that on your account?

[azure wharf]

@unique stone ?

[brittle hemlock]

ill check my mail

[azure wharf]

where is the "SUNO" text on the images?

[hollow hare]

Should have it mentioned when you view your subscription status iirc

[polar spade]

Likely, either you've passed your renewal date and your payment hasn't cleared, or your renewal date is today (past 24 hours) ?

[brittle hemlock]

could be, ill just renew it manually again, its only 10euros

[unique stone]

I would expect to see it in this part if it was there.
But it's not, and I see the song there is just under 4 minutes long.

There's a problem that can happen if you have a "SUNO" text over this image and shorter songs but this looks OK to me.

[warped locust]

With various (easy) prompts, Im getting bitcrushed and squashed audio in actually literally 90% of the outputs that I have. Anyone have an idea to why this is happening?
These are just a few but I have about double as many that are the same, and they all are super bad quality and mushed in the last 60-0 seconds of the song. Why is this? See spoiler for list, it happening for 90% or more of all the songs ive made in 2 days.

hmm how can I link all the songs so I can ask for help?

[hollow hare]

(BTW, that assumption is based on this message: #🪲┃bugs-web message)

[azure wharf]

for example here, those are generated today and ARE under 4 minutes

[brittle hemlock]

appreciate the quick help, resubbed and now its working

[warped locust]

Ok heres 3 of them, but I have this problem in over 90% of my outputs:

||https://suno.com/song/b72ba951-6af1-4733-8bac-e503be6b5b27
https://suno.com/song/1d6f38e5-422e-4887-a19a-2bca2453fc5f
https://suno.com/song/99238bb3-16d4-4649-9dde-724291eb05b4||

I dont think theres anything bad with my prompts so how is it possible that so many outputs get all broken towards the end?

[hollow hare]

Just be sure you do not get double billed. If it looks like you did (though maybe needs like a day for an accurate update), be sure to contact [e-mail]

[brittle hemlock]

will do, either way the website is worth it, had lots of fun with it

[warped locust]

Heres another two outputs that are just purely broken. Is my account bugged ?

suno.com/song/61a6e299-62f6-4a8d-9088-000f879ad729
suno.com/song/06906247-52e3-41d9-8f45-fc223146e029

[hollow hare]

For the gg song, I can already see it has too many style prompts:

• Be aware of weights #💬┃general-chat message
• Be specific with the subgenre you want. Techno is a broad genre (I like using https://rateyourmusic.com for their genre hierarchy list)
• I am unsure if the AI knows about X/Y beats
• I know that the AI does NOT understand <#> BPM. Use keywords like downtempo, upbeat, slow, fast, etc. Maybe musical terms like presto might work?
• You can use metatags in the lyrics field to prompt for instruments. E.g. [reverb bass guitar]. I think ideal is three instruments (incl. Vocals) at a time/paragraph? Something like that for best results

Basically you are expecting too much without knowing what the AI limitation is (best to treat it like an ADHD child, NOT as a musical genius with a degree in music theory)

[warped locust]

As you can see I have various prompts in these only in a few later ones do I have a more typed structed. So weighting isnt the problen generally.
I just used BPM in a few of these examples, I have like 10 more with normal weight that all are qually broken

Dub techno is a specific genre and I describe it well.

But none of this is relevant the the quality, it shouldnt be the casue for distorted glitchy and broken audio. Its not my fault :O

see

suno.com/song/814b7883-fd40-47da-ac7a-e5e765ba5d82
suno.com/song/bee0fbe9-cbf9-454d-a29d-67ac100bdd3f
suno.com/song/c9bb5b9d-6fda-4266-92f4-f0204e0d945f
suno.com/song/5e1a9984-1952-4c61-b0fc-d9755bcae72b

[unique stone]

Those look OK to me (in terms of the problem I was suspecting.)
4 minutes is the maximum length for an initial part.

[warped locust]

Im not treating and expecting the AI to be a genious. But I expect the sounds to not be broken (as mentioned, I cant repeat myself more) in over 90% of the outputs

[hollow hare]

I also forgot to mention that 3.5 and 4.0 are trained towards more popular genres.

For non-popular genres, you will want to use 3.0 for the instrumentals (then extend into 3.5/4.0 for the song structure)

[warped locust]

But this shouldnt be cause for glitched distorted bad audio quality. Like its not I who is asking for that, it generates unusable audio quality

[unique stone]

I'd try it without the DubTech persona and see if that helps, in case something with the original song that persona is based on is causing this

[hollow hare]

Can you provide a specific track and timestamp to demonstrate what you mean? I listened to your gg one but didn't notice something like that

[warped locust]

The reason I used the dubtechno persona was because all but this 1 generation was without glitches and flaws.

[warped locust]

Its noticeable in all of the linked songs above, towards the last 30 seconds at least

[hollow hare]

Ah, the longer a song goes, the higher the chances for degeneration occurs. Especially from like every minute time point (esp. At the 2 min mark in the past)
There is a reason why there is a maximum time limit for generations and that is the main contributor.

People do recommend extending during earlier timezones to minimize degradation to retain HQ sound as the latter half of the song.... yeah... degradation builds on top of it self ya know

[warped locust]

https://suno.com/song/814b7883-fd40-47da-ac7a-e5e765ba5d82 from 1min and forward

https://suno.com/song/bee0fbe9-cbf9-454d-a29d-67ac100bdd3f 1:12 forward

https://suno.com/song/c9bb5b9d-6fda-4266-92f4-f0204e0d945f from 1:30 forward

https://suno.com/song/5e1a9984-1952-4c61-b0fc-d9755bcae72b abouit 1:50 forward

I can understand that it produces other things but its not really acceptable than I spend credits on having to work with distortions and extending from before that, theres zero guarantes that it helps and which means Im paying to get from below to normal audio quality.
Thats not alright

[hollow hare]

Can you add https:// to those please? I am on mobile so it is a pain to visit the songs like that

[warped locust]

sure, I wanted to make it easier on chat so I remove em

[hollow hare]

Thanks

[paper grail]

hello

any update on nasheeds / acapella?

[hollow hare]

I do not think I am able to help you because I cannot hear the issues on my end 😦
All I can say is that it is known that degradation occurs the longer a song goes on so you may want to consider extending early on if you want to retain high quality sound.

I do not think there is any magical keywords you can do to make a great HQ one-shot generation. Though maybe try adding in high quality or similar CD quality terms and see what happens?

[hollow hare]

What do you mean? The AI should be able to work.with both of those (though the latter is typically spelt as a cappella and can have a different effect, see the following comment)

[charred merlin]

https://suno.com/playlist/3dd7bdf0-5f6f-4ee7-adb0-8f42d9cb8b01

[paper grail]

i keep getting isntruments

so i isolate the vocals with another tool but then i have no background melody

[hollow hare]

Oops, forgot about that test playlist. Teehee (ugh, I cringed so hard saying that)

[charred merlin]

a cappella is the traditional latin spelling, and seems geared toward what you'd expect from vocal harmonies only.

acapella is the anglicized way of spelling, still completely valid way to spell it, but seems to also incorporate instruments on top of vocal harmonies & vocal instrumentations

[paper grail]

i lsitened to some of them

i dont want the barbershop quartet style

i want like a modern song but with vocals only you know?

[soft karma]

Yikes.

[paper grail]

ok so the spelling makes a difference

[charred merlin]

Note that the ONLY style prompts I used in that playlist were acapella -- you'd still need to add other style prompts if you wanted a different sound. You're not stuck with barbershop, but it was easier to show the difference between the two without other styles on top

[hollow hare]

Probably a little longer because of the winter holidays 😦

[paper grail]

i just did a capella

song

both of them have isntruments 😭

[charred merlin]

Can you share the song you're working on? It might be handy to see what you're using as style prompts & how they're arranged

[hollow hare]

I would not expect a no instrumental / vocal-only song from Suno unless you are doing comedy show or some similar style to that

[charred merlin]

Unfortunately, no prompt is a magic bullet that will overpower everything -- they all work together and influence on another.

[paper grail]

"a capella song, modern pop song, male vocals" @charred merlin

thatss it

my prompt

custom is off, istrumental is off and its v3.5

[soft karma]

Problem is I'm unable to login with my paid account (can't remember how I managed to get in there in the first place). Think there's an option to use a non-gmail email address and not a phone number?

[warped locust]

Its very odd you dont think you hear the distortions and I cant say I believe you in that sense, they're as clear as the sky in my ears and I really mean that, theres literally no question about it.
I could dabble in high quality tags like you mention

Do you not think this is low audio quality? Im a little bit shocked:

https://suno.com/song/811a9d8d-509e-4b78-b0b8-30bf3348470a

[hollow hare]

Well, I am using a mobile phone to listen to at low volume so not to wake up the kids

[warped locust]

hah okay well explains it ! maybe not the best setting to determine quality :P

Like I feel sad and helpless, Literally every song is absolute broken. I just want to make audio thats at least something, I dont want to have broken outputs 95% of the time (because its a higher percentage now since EVERY output since I talked here has been disrtorted and bitcrushed to smush)

https://suno.com/song/d4a8864d-6532-44af-af1b-cbef035344b6 its just crushed noise. its incomprehensible to listen to

and I have really good specific prompts

[hollow hare]

Ah, I see, the dangers of using a phone number without a SSO assigned. Especially if it gets reassigned through like using the app by accident

Sadly I think the only course of action would be to wait for a response from [e-mail] with that situation as I doubt @steady kelp can help with that (maybe? Doesn't hurt to ask for your scenario)

[soft karma]

Well thanks anyway

[warped locust]

Like if I get 1 listenable song thats not even spent credit to extend, ONE per 100 credits. Its not acceptable in anyway. Its full out scam. And Im not talking based on subjective style, I mean bottomline decent audio quality that can be called music

[charred merlin]

@paper grail with v3.5, I'm finding it responds to typos -- there's 1 P if it's all one word, but 2 Ps if it's 2 words -- I've been getting what we expected from v3.5 a cappella, but I keep getting guitars when I use v3.5 a capella with just a single P

[hollow hare]

A good rule of thumb is to follow the KISS rule. Try seeing what happens when you reduce the terms used.

E.g. "dub-techno, minimalism, soft bass, modular synth, -arp, -arp eggiate"

[warped locust]

I mean I do all of this I haha I cant even, this stuff just doesnt want to work

Im at the mercy of a really dumb AI

and Im paying money for this like I feel so stupid

[hollow hare]

That is why I like to encourage people to stay on the free plan until they feel like they got a hang of how to prompt (to a degree) with Suno before paying

[paper grail]

interesting

[charred merlin]

https://suno.com/playlist/c23a5769-e947-44ea-8197-a334ab55bf85

[paper grail]

@charred merlin any more discoveries i can take note of?

[warped locust]

I did that, but the quality has gone downhill since after christmas. but I also expect a baseline quality as a paying customer. There should be consumer rights for this

[charred merlin]

I shared a playlist just above with a couple tracks I generated -- it seems like v4 might be a little less picky about how it's spelled, but I haven't really explored it too much yet 🙂

[urban sonnet]

lol you mean since the beginning of december 😂

[paper grail]

im curious, how do you and people here know what a nasheed is?

[warped locust]

I think its reasonable to believe we are expendable beta-testers and bigger companies that use Suno have better more clear generations, but free and paying users are working with bad data

[urban sonnet]

ofc we are

[charred merlin]

Google & cultural context 🙏

[hollow hare]

I made a few islamic nasheeds when testing v4 during alpha along with a lot of other various styles

[urban sonnet]

trainingdata was most certainly pulled after black friday sales came in but most users dont seem to care i guess

[warped locust]

its very believable. This is how most AI sites operate anyway

I think I will refund my latest sub and continue to use free accounts. I dont want to be a paying tester

[charred merlin]

I mean this 100% wholeheartedly, and not meant as "smacking down" an opinion I don't agree with, but that's a super technical claim, and I'm curious what it's based on?

If it's a matter of Suno responding differently, I'm cool with that assertion -- but I'm not sure where the certainty of actions comes into play.

If training data were pulled -- we wouldn't be able to regurgitate training data content anymore, right? (like using snoop dog rap, snoop style, doggy style as style prompts, or getting DJ KHALED screamed in his voice, or the instances when copyrighted lyrics are used & Suno regurgitates the entirety of the song used in training)

I've seen a lot of these claims in #1308499132356362383 , but the source is always "trust me, bro, I've been using Suno for a while", so it's hard to find any common ground to discuss 😦

[urban sonnet]

dunno.. just frustations in users i guess, with the silence and lack of coms etc for a month now.. every user has their own experience im sure, some are happy some are less and most (like me though I should be spending my time on other more productive happy things) end up on here discussing and some even arguing over this.. I have about 200 daily gens and have a good library to look back on and for my own part there are significant degredation when comparing v 3.5 gens pre v4 and earlier (november v4 gens) with today.
But.. at the end of it all, every keystoke users put into this forum (unless for helping others) are pretty much wasted time and energy.
Anyhow. Happy NewYear and may 2025 be shimmer and treble free 🙂

[warped locust]

And generally AI-companies have shady practices as everything is behind the scenes, club goods membership but without any guarante for what you buy.
Like the report function and thumbs up-down isnt for my benefit, its for Suno to be able to train their data so they can sell better products to better paying companies. We are literally paying to work for Suno, they already have fundings because no AI service this popular exists without it.
Its sickening in a way :O

[charred merlin]

I understand a lot of the frustrations, to be sure -- I really appreciate you taking the time to share your take on the matter with me 🙏 While "trust me bro" might have sounded a bit adversarial, I understand that our subjective experiences of Suno are largely all we have to go on -- in that sense, there's nothing wrong with expressing your gut feelings, but it's a really relieving experience to feel like we might be on opposite ends of this opinion, but acknowledging that it's largely based on our gut feelings & personal experiences.

I'm looking forward to our interactions going into 2025 🙂

Happy New Year! 🎉

[paper grail]

can i listen?

@charred merlin do you have paid suno?

[urban sonnet]

Its not like its an investigation into the mechanics or even motives behind Suno, and if people really disliked it, they would just leave and go silent, the frustration when it all boils down stems from the fact that Suno at large is beautiful and amazing, and we are like the pushy parents that want perfection i guess... I hope they prevail in their lawsuits against Sony, Universal and Warner (though odds are not in their favor) and hope they are allowed to grow and evolve because AI music is here to stay, no lawmakers can change that, only the timeline of it.. And you are right the "trust me bro" arguement is not a good or valid one at all, but I feel that a lot of the speculations and frustrations comes from a mix of true experience of real changes in the product and the lack of communications going back to the users.. Silence feeds conspiracy. 🙂

[paper grail]

@charred merlin are you then technically able to let someone else use what you produced?

[charred merlin]

I'd imagine so, yeah 👍

[paper grail]

i think you know what my next question is then 😄

can i use one of the songs you sent on telpen tests

[charred merlin]

Sure -- if you've found one you liked, shoot me a DM & I'll work out how to transfer ownership 🙂

[charred merlin]

Similar to a conversation I was having with a friend regarding Suno -- they were at a breaking point after wrestling with it to get a track produced a certain way, and after he shook off his "this software is bullshit and I'm never using it again" phase, we talked about how you need to genuinely care about something to some degree in order for it to make you that angry. I feel like the longer we use Suno, and the more often we "hit the jackpot", the higher our standards become, whereas it wasn't long ago that if there was a weirdly pronounced word, you just posted a video with the corrected lyrics and gaslight people into thinking they just misheard it 😛

I'm not sure how frequently they used to communicate in-between major updates, as the #📢┃announcements and #🚀┃updates channels seem about on-par with what they used to be, but to their credit, despite being overloaded with support tickets, Mike B has been increasingly more present, which is a step toward open communication.

There's definitely a lot to try to understand -- I don't think everyone who says "all of my songs are great" are shilling, and I don't think everyone who says "all of my songs are terrible" are catastrophizing. The curious part is figuring out the differences in what they're prompting/generating

[paper grail]

"gaslight people into thinking they just misheard it" 😭 😂

[polar spade]

When people are given a mountain, they often demand the stairs to reach the top while others know it's best just to tunnel through to get to the other side.

-- that makes no sense but it just seemed very Buddah-like. Throwing a bit of kum-bah-ya vibes to everyone.

[polar spade]

When expecting chicken salad, you sometimes must realize your first issue is you're using chickens**t. Study why your prompts are not giving you the results you want.

"These are not the prompts you seek" -- Suno-wan-Kanobee

[raven turtle]

Getting a lot of shimmer today... dammit. Unusable.

[neat nexus]

https://tenor.com/view/just-lie-on-the-internet-gif-19885520

Do people really do that?

Gaslight people about the lyrics?

[charred merlin]

https://tenor.com/view/look-smile-smirk-breezy-moves-ornery-gif-3520545

[hard horizon]

Damn took me like over 40 prompts generating audio and a few cover generated variation to get even slightly a result in sound I was after.

I was more after the like loud chanting Jamaican vocalists that on some tunes hit their chant right before the drop or so..

“Maandaam let’s get dat in ere. We-be-destroying-di-dans-flop-yeehee!” 🤣 like that or so

Here’s my result so far from satisfied:

https://suno.com/song/71d4f279-97a0-43e5-b226-60a8ecf6066b

Earlier I struggled with getting other ethnicities as vocalist. Luckily enough chatGPT works better answer questions then Google these days. Insane to think this is actually happening today’s date though.

[hard horizon]

What’s wrong? I’m trying to discuss the struggle I have technically

Someone reacted the cross..?

[fading forge]

Anyone have technical issues recently with extend and replace section not transitioning properly (for extend) and not replacing just a rection (for me it replaces the rest of the song)?

[neat nexus]

Not the person so I'm only speculating but maybe they were reacting to the idea of ChatGPT as a guru?

From experience I can say that ChatGPT can be useful for inpsiration and give an idea where to look for things but I wouldn't take it too seriously as it tends to just literally make up things. There are a lot of people using a sort of pseudo-code that they've gotten from ChatGPT and think it allows them to "hack" Suno somehow for better results. As far as we can tell it really doesn't do anything ChatGPT claims it does....

So maybe that was what they were reacting to?

[hard horizon]

Oh oops my bad - didn’t know this was competition from one another?

[neat nexus]

Oh it's not.

It's fine to post about it here too if you want.

Just explaining why another user might have reacted that way.

[hard horizon]

Bo I just ask GPT to help me with how to formulate things and such, even with my work and company it’s “rewrite and improve x”

[neat nexus]

@dusk lodge Do you mind interjecting? What was the reason for your reaction?

[neat nexus]

I mine GPT all the time myself, I use it to figure out what instruments might be appropriate and what kind of playing styles would work for that genre or the artist whose work I respect. It gives me ideas all the time.

But I wouldn't use a direct prompt from it because in my experience it gets it wrong or makes up stuff too much. It's a good place to start though an dlike you said, rewrite and improve!

For one thing, those time stamps do nothing. Ironically for a music generator AI doesn't understand time stamps.

[hard horizon]

Sometimes it’s like “I want an African tone of voice vocalist” and it comes up with 100x times as good way of wording it

[half dragon]

Yeah, one of my old songs I wrote years ago

[half dragon]

I have no issue with the ending... and I can crop as well... my issue is that extra sound adding through out the parts... sound like low quality sound. is it because the length of the clip that extended from was too long causing the degration of the generation?

[gritty badger]

I many times expressed my feelings about the way Suno team handles the on going issues, here and in the feedbacks sections. Yes MikeB is doing his best to respond every single feedback, but in the end users prefer to see minor or major updates being posted by the devs, regularly. The silence doesn't help the process. Not going to lie, I do hear the work being done in the background each time I try to generate, less chances of getting shimmers, less ear piercing highs etc are definitely noticeable. But the reassurance comes from the acknowledgement. Yes they are aware of the issues, but I believe they need to be a bit more engaging. A weekly update of where they are at the point of problem solving (even if it is a minor one) should be told, so that the frustration eases a bit more. Right now it is like a constantly boiling water, eventually the water will vaporize entirely and it will be much harder to control the anticipation of a fix.

[neat nexus]

Also the style prompt is overdone. BPM is not something that is recognized either.

I'd reduce it to something like this and rework as needed: dubstep inspired, house, heavy drums, distorted morphing bass, Angry vocal, energetic, Jamaican accent, Jamaican Patois, chant, aggressive phrasing delivery

Do you mind if I play with the prompt and lyrics a bit?

[crude tundra]

for anyone of you that know that I like making music on Suno you’ve known that I’ve made a lot of cool stuff. Unfortunately I’m having some difficulties talking to one of the head guys in charge of the communications for your server. I really like the team that you guys have and I like the things that you’re doing I would just simply ask if my account can be fixed later today at some point when I get home from work, not asking for much just asking for my stuff to be unflaged so I can continue making music I don’t know the reason, behind being flagged maybe I had to deal with one of my old accounts. I only made one other account before and that’s just my second for a back up account. I’ve always paid money for Suno and I’ve paid a lot of money for credits. I’ve spent a lot of my money and time into the work that I’ve putso I just don’t want it to go to waste. That’s all you guys have a wonderful day. love all you peeps your family to me.

Cat 1559 signing off for the day

Last little note, my workflow is very fast pace so if I work too fast for your AI bot, I’m sorry I am not about myself. I am a human Ggs

[dapper kernel]

Why do I not have any credits? I havn't used any this month

[neat nexus]

If you're talking about credits expiring, please scroll down.

[final hazel]

Thanks for reaching out again. As previously mentioned, account flagging occurs when activities, such as unauthorized API use, breach Suno’s Terms of Service. This is not a bug, and our system flags these issues automatically.

Since this matter has been addressed both here and in DMs, further discussions won’t change the outcome. Suno’s support team cannot assist in cases involving confirmed policy violations. Please respect our guidelines and refrain from spreading misinformation in public channels to maintain a positive community environment.

[charred merlin]

@dapper kernel

[neat grove]

Thanks so much for the reply. You have found some amazing hacks to get the model to bend to your will.
But,
I was hoping more for a "Podcast Mode" instead of a hack that takes up a lot of time for each pass. ANy chance of finding such a mode or requesting for one?

[charred merlin]

I'd personally use the built-in podcast generator in https://notebooklm.google.com/ -- much easier than trying to force a music-gen to compete

Even includes the expected cadence, pauses, interruptions & extrapolations that you'd expect from a podcast

[neat nexus]

Took the words right out of my mouth.

[neat grove]

that started my quest to search for such an option. But their generator limits the output significantly. Does not stick to script and voices are limited - thats why searching for variety. Any open source models that can do this?

[neat nexus]

It's not a bad idea, not sure how easy it would be to do. It would be nice if such a mode were possible, right up there with the idea of a "DJ mode" which would allow a bit of a radio style to work.

[charred merlin]

https://labs.google/fx/tools/music-fx-dj Agreed, it'd be curious to see Suno's take on something like Google's MusicFX DJ

[neat nexus]

Great another tool to try out! 😅

I'm going to have a busy new year, I can see!

[neat nexus]

Thanks for mentioning it!

[spiral bane]

i really like this tag [end the song] for extensions lol

[whole tartan]

is Uploading broken right now? whenever i try to upload the box just get stuck on loading image

[half dragon]

clip too long?

[keen zenith]

I’m not seeing “edit section” anywhere - am I missing something?

[whole tartan]

no, first it complained it was too short, but was still stuck on image gen. then i made it longer, over 6 sec, same exact issue. tried different formats, restarting browser, logging out/in, no change

[polar spade]

Does "this month" mean since renewal date of December 20th? I see 2500 credits starting , I assume, on December 20th.

No credits except "extra credits" roll over or accumulate. Extra credits continue as long as your account does not get deleted or expire. You have "0" extra credits. Should you have more?

[polar spade]

Extend: located in menu under Create/extend
Edit lyrics: located at bottom scroll on lyrics page. Visually edits lyrics only.

[half dragon]

Under Edit/Replace Section

[keen zenith]

I thought there was a new edit section feature rolled out

[polar spade]

@half dragon VRKing hit it. Yes, under Edit ir "Replace Section" as well.

[keen zenith]

I only see Edit Details

[half dragon]

You should see 3 options:

• Edit Song Details
• Crop Song
• Replace Section

[half dragon]

mobile site?

switch to desktop site

[keen zenith]

Is it only desktop?

[polar spade]

Try clicking the menu button in different screens.

It may show better while viewing the songs lists in Library.

It may show when viewing the large , song only screen (with art and lyrics).

[half dragon]

seems like it

[polar spade]

It's been awhile since I've checked. Yes, for Pro and Premium users on desktop and generally tablet formats only.

Someone may correct me if I'm wrong.

[keen zenith]

okay haven’t tried that yet. I guess I’ll boot up the computer

[polar spade]

Thanks, mate.

[polar spade]

I feel your pain. I keep having problem just bringing up "Move to Trash" in my menu. I know it's likely a format viewing thing. I just keep clicking until it appears. Lol.

[keen zenith]

I got it on my desktop. Thanks for your help

Thought I was blind or missing something obvious lol

[charred merlin]

A lot of features aren't available on the app 🙏

[fading tulip]

I've noticed that the platform has stopped updating the play count on my songs. While likes are being tracked correctly, the play count remains fixed and doesn't reflect actual plays. This seems inconsistent and affects the accuracy of engagement metrics.

Additionally, I’ve come across accounts where the number of likes exceeds the number of plays.

@tribal remnant What you talking about? It stopped counting plays on my two trending songs.

It counts only likes.

[unique stone]

As play count goes up, the number of plays required for it to update the displayed number seems to go up. So the first few plays will be counted immediately, then it will lag behind/take longer to update. This seems to be how it's designed to work.

Bear in mind that the play count and upvote count may not be accurate. People can and do manipulate these. Personally I assign no value at all to these numbers.

[crude tundra]

My Stuff still broke

[neat field]

if i wanted to have 15 -30 second previews of some of my copyright songs on my profile, would the best way be to:

a) download the audio of one of mine
b) edit it in Audacity to be shorter
c) use "Upload Audio"
d) extend
e) crop the extension
f) make the crop public

?

[fringe surge]

how to use cuda GPU in the bark API?

[lean moon]

Afternoon. Can anyone think of a way to do a cover of an entire song? 120 seconds obviously limits that, but could I upload two clips and have them done in identical styles and then stitch them together in post?

[half dragon]

The video generated from the song doesn't display the non-English lyric properly... I usually don't download the video but I do once a while and this time... it doesn't display the lyric properly with the non-English words.

[rapid bone]

So, there's definitely no way of getting my account back by talking here? sent an email, but the response is like "New years! Come back later!" 😦

[polar spade]

I'm not good at interpretation. What does the screenshot show?

[polar spade]

It's like an ASCII code symbol or just blank, or doesn't scroll in the video. Any language? or which language you using?

[charred merlin]

Yeah, that should work -- except if the audio is flagged as having any vocals, then you can't set it public (but you CAN set a playlist as public with restricted songs in it).

Or you might run into the issue that @mystic flume had when Suno blocked the upload of her own tracks because it recognized the copyright

😅

Good to know that it doesn't just protect big artists, I suppose -- unfortunate though, because I don't think support was able to help her bypass it (though that might have just been the policy at the time, no idea)

[neat field]

maybe "profile song previews" is better as a long term dev request in #🚧┃suggestions-layout

[unique stone]

It'll be the Suno watermark on the songs and short generations. It's been discussed elsewhere and with Mike B. As far as I'm aware this should be resolved now.

[charred merlin]

While Bark technically was developed by Suno Inc, after they open-sourced Bark & spun-off Chirp into Suno.com, the server really focused on the website.

The moderator Jonathan Fly is our resident Bark expert who has expressed being open to DMs regarding it, but I'm wondering if it'd be worth still having a server channel for Bark-related discussions

[polar spade]

One person's broke is another person's jingle. I didn't know. Lol

[charred merlin]

There's a feasible method that would require external editing & the use of the Replace Section feature, but there would still be a lot of luck involved in getting 2 covers that sounded similar enough (unless you disguised the change in style during a bridge or something)

[mystic flume]

This is the best way to do it if the second half has melody changes or anything interesting that you don't wanna lose. Otherwise, if the song is pretty simple verse / chorus / etc, the best way is to just upload the first 2min, then correct the lyrics (just the ones in the clip), and then do a cover in the style you want, making sure to add in the rest of the lyrics in the cover prompt. I've had very good results this way, especialy when doing "live" versions of my existing songs. Breathes a bit of new life into them, I've found. You can remaster the upload in v4 if you want, but I've had my best results by doing the cover first (in v4) and then remastering that.

But if you're a real perfectionist, doing a Bride of Frankenstein in post is valid; you just can't have that song on your Suno Profile anymore. But as long as you don't care about that, it's a great approach.

[stiff blaze]

Ist

[fading tulip]

Where are our v4 gift credits for Christmas? Such a shame. 😁

[half dragon]

yea not all but some ascii codes... it does scroll.. Vietnamese is the language

Interesting. Yea, I downloaded video before on v3.5 and it was good.. no issue at all... seems v4 broke a lot of stuffs

[polar spade]

Happy New Years to all! https://suno.com/song/66bd8bc9-8fbf-416d-890d-18acf64237f0

[muted trail]

Plz help

https://suno.com/song/f9dd0784-52db-42f5-bd33-6d666c03d951

I can't play or Download this song

Don't know why

It's loading from 24 hrs

Means I can't get the same song?

Fully same?

Cuz love that one

[glacial scarab]

its still available on the CDN for whatever reason
https://cdn1.suno.ai/f9dd0784-52db-42f5-bd33-6d666c03d951.mp3
But there is no way to get that fully back into suno.
Might be worth contacting suno with the song link because this seems to be a very weird isse (song is fully generated on the cdn but in error state on page)

[muted trail]

I was able to hear it once when it was just generated

How to contact suno?

[quick wedge]

Contact them via email [e-mail]

[muted trail]

Thanks

[quick wedge]

You’re welcome

[dusky moat]

Hi guys! While replacing a section of the song, the section generated seems to be okay and transitions smoothly to the "gray area" of the preview, that is outside of replacement. However, when I confirm my choice, the section has a rip there and does not merge into the main body as it was in the preview.

It happened 3 times already, and I am a little bit stumped, since it was never an issue for me before.

I am unsure how the Replace Section tool can know where the selection should fit and even show me the transition but is unable to repeat it when cutting into the original track

[untold briar]

Are you getting a popup menu when you click Replace that asks you to double check the red highlighted lyrics? If not, or even if you are, try clickgin Edit Song Details, adding a blank space, then Save Song Details, so you get an "aligning lyrics" message.

[dusky moat]

Yes, I do get it, and it seems to be about right (about, because there is a cut in the middle of a word. The model knows how to deal with it most of the time).

Thank you for your idea, I will do that, and I will get rid of the comments and mark-up while I am at it.

Related question: it is the first time I am working with this, the model decided to play around with the contents of the chorus, repeating some of the lines without any input. I actually like what it did, but am I correct in my understanding that I should transcribe its improvisation into the text body now, so that it does not confuse itself?

[untold briar]

Yes, but you only have to transcribe if you're going to be doing more Replaces in the future. Doesn't matter for extensions. Does matter for Remaster and Covers as well, which you probably expect (though remaster can work if you don't fix it, sometimes)

Maybe don't fix it if you're in a creative mood and open to Suno improvisations, I should add. But if you actually want to fix lyrics precisely you should. 😄

[dusky moat]

Yup, I am always trying to milk it, until I get exactly what I am shooting for 🙂 But in this case it was a really great improvisation

Thank you, hopefully it will clear up issues that I have and prevent some of those that were laying in ambush

[dusky moat]

Oh, sorry, a follow up, since it happened previously. There were cases when I tried to realign the lyrics after a replacement, but every time in this window I got the original text, which is not to hard to correct, until more than a dozen fixes pile up. Is there a way around it or it always defaults to the very first edition?

[untold briar]

You may need to refresh the browser.

[lavish breach]

I have a question

If I subscribe to the pro version, can i use the song for my monetized YouTube channel?

[dusky moat]

Okay, it did not get much better in the end. The words are all in correct order and there are now no non-lyrics lines, but it still messes up the replacement itself, even though the preview, again, is very precise and smooth. Which bugs me, actually, because it already knows where the transition should happen and does it great but then fails to place it correctly, it seems.

I am now wondering if this might be due to the non-English language, which might be harder for it to figure out and it misaligns by 0.2-0,3s. I, of course, have tried realigning again three times, and I double checked that both new lyrics and the replaced lyrics match, as well as checked, that the sound in the selection matches the highlighted zone exactly. 🤔

It's okay that I am wasting credits on this, I am confused, since the previous attempt at creating and replacing in this language went without any issues for another song

[untold briar]

When you highlight your selection, in the right column, is Suno accurately chaning the lyrics to red, if you listen to the audio?

[dusky moat]

No, it misses by about the same time that it messes up when merging. However, when prompted I moved the red selection to match what is really sung.

I am now making an attempt at using the highlight it initially gives me, since it might actually turn out better. But it is a shot in the dark, since measuring both offsets would be quite an involved task

[untold briar]

There may be some caching bug. I'm not sure, but maybe try Control Refreshing after you make any change, don't rely on the Suno webpage to auto update.

Non English is 100% much worse in Replacing, however if the red highlighted text is working, and if when you listen to the section (hit play when paused, it should skip to start of your selectoin) the audio matches, it should still work. The failure is more that those things sometimes are wrong or don't' work at all (you never see red text).

[viscid locust]

I love ReMi

Where tf does the last line come from

[dusky moat]

Can you please elaborate on what is Control Refreshing? Reloading the pages after realigning did not do much, as my last attempt at no correcting its selection.

I will try to make a full browser reload or maybe I will try using another one. But my little window of time for Suno is closed for today, gotta go to work. But I'd still appreciate any potentially useful input, so that I could try a few things when I am back.

Thank you again for trying to help!

[untold briar]

I meant control-R refresh, which on Chrome if I recall, refreshes thing a little more, uh, freshly.

[dusky moat]

Yeah, it does work in that sense. The only issue, other than the actual result of the merging, is that it selects two more words than are actually played. But, somehow, yeah, both correcting it and not correcting it gives the same result

[dusky moat]

Huh! Never knew about this feature. Thanks a lot!

[untold briar]

If the previous attempt Replacing worked fine, it may just be bad luck on this specific segment. Could try making the selection wider and replacing more than you intended to, but may change more than you like as well.

It's also worth trying to do the edit in a Cover. You may be able to get a Cover that is quite close the original but has your change.

[dusky moat]

Oh, I was not aware that I can actually change the lyrics in the cover in any way. I will look into this option, thanks!

[languid umbra]

The sound quality issue has not yet been resolved???

[hollow hare]

What sound quality issue? A lot of times, the issue could be addressed to the prompts you are using (e.g. not including quality terms or including broad spectrum that includes training of instrumental recordings from decades ago)

[languid umbra]

When I gen songs using ver 4 or extend songs the singing would be okay but instruments sounds is like it is damaged

feels like listening to a bad recording

I get this this issue with ver 4 mostly not 3.5

[hollow hare]

Can you link to one of your songs as an example? You can DM it to me

[languid umbra]

Sure one sec

[languid junco]

Okay something weird again.

[languid umbra]

DMed

[languid junco]

Most keyboards give us this symbol ' the Apostrophe.
When writing lyrics in Suno most users likely use Apostrophes in place of Right Single Quotation Mark, which is the correct grammar. ’
ChatGPT seems to switch Apostrophes to Right Single Quotation Marks.

Dunno, just one of those things where symbols might hold suno back.

Words like "They're, you're, they've"

Hm.

I think a lot of it depends on how AI models read those symbol

[velvet yoke]

Hey, can anyone help with this? I created a song using version 3.5 before the release of v4, and now I have a Pro plan. If I remaster the same song in v4, can I distribute it commercially on platforms like Spotify?

[hollow hare]

No as it would be a derivative of the free song (which related topics are covered at #1294423389515284552 )

A close example would be https://help.suno.com/en/articles/2871105

However, if you email [e-mail] after signing up, you can ask if they can grant a retroactive license to select pieces - I heard that usually gets granted

[young prairie]

Hello

I signed up for the 2,500 credits plan for $10

Now it's over, how do I put in another 10,000 for $30 without waiting to renew the subscription?

[worn orbit]

Hi all, I'm using Suno version 4 and I've noticed that on every song I create, the sound gradually degrades by going more or less high in the treble!?! Am I the only one who's noticed this? 🤔

[snow loom]

why there is always "this is a fake app" water mark

[rough reef]

I'm pretty disappointed with version 4.0, I'm losing credits unnecessarily because it generates 2.39 songs and hasn't finished the second chorus... In addition to making up stories, not respecting the lyrics

[blazing ingot]

Yes, SUNO's TOS gives you all rights to commercial use.

[rough reef]

I really don't know what's going on.

[young prairie]

Hello

I upgraded plan $10/month on December 14th, and now I have used up the credits. I now want to upgrade plan $30/month. What should I do to recalculate the date instead of expiring on January 14, 2025?

[tiny falcon]

Good morning and Happy New Year, everyone! 🎉

Thank you all for your kindness and for helping one another out. This truly is a wonderful community to be a part of, and I’m grateful for each of you. Wishing you all a fantastic year ahead!

[gleaming dirge]

happy new years as welllllllll, thank you! ✨

does anyone here knows what should be genre details to make songs in lana del rey style, i want to achieve dreamy vocals and music blended but when i put genre that are similar to her songs, it comes out very different

i want to go for siren vibes

haunting, dreamy, alluring, pop

[polar spade]

Maybe you're wondering why it has both 3.5 and 4.0 versions? I can't tell.

[rough reef]

I've been using Suno since the beginning. This has never happened to me before...

[polar spade]

The beginning was in 2023. Or your beginning? I don't follow what the issue is, mate.

The settings between 3.5 and 4.0 may change regarding doing covers or remasters. You blanked out most details, so that's unknown.

[rough reef]

It doesn't matter if I covered it or not. The problem occurs with both versions, same lyrics, same structure and it doesn't go beyond 2.39, remaining at the second chorus. It's a bug and I'm losing credits unnecessarily.

[warped rivet]

Despite subscribing pro version, i still get watermarked audio like this. I want elaboration
https://suno.com/song/cd1a0c50-ac84-4e8c-8df3-e7bf6170999b

[polar spade]

Then , as a bug, best start with posting on the bug channel https://discord.com/channels/1069381916492562582/1082775596767117312 From there, send off an email two [e-mail] and check back here for other responses. Because of the consistently shorter outputs, it could be an issue with the app itself which need a direct chat with support.

Keep in mind it is New Years Day and staff response may be limited.

[rough reef]

Thank you so much. I thought I should post here. Happy New Year too.

[polar spade]

Noting when you began using Suno and which format: app: browser: phone you're using may give a clue that others might respond.

[rough reef]

I've been using the same PC since the beginning. The bug has been occurring since I signed up and for the last 24 hours. I see that there are more users in Bugs having the same error right now. Thanks.

[rough reef]

Change the image manually and the logo disappears. You can put any image you want, I use my face with my own logo. song details.

[young prairie]

Which of the two options should I choose to recalculate the date instead of expiring on January 14, 2025?

@polar spadeWhich of the two options should I choose to recalculate the date instead of expiring on January 14, 2025?

[open sleet]

yea nah mate

[dim glacier]

Thanks for the ping ❤️

[warped rivet]

not the image, i mean the audio it self. as you can hear, there is "suno dot com" words pronounced in the lyrics

[charred merlin]

Do you want to start a new subscription starting from today — receiving the difference between the two plans?

Or do you want to wait until the next billing cycle and get the full 10,000?

[barren snow]

Please use @ reply, I'm going to bed after this question: How do I use SUNO to export all the stems, not just vocal / music, but every instrument: example: vocal, fiddle, banjo, guitar, drums as separate stems?

[raven turtle]

How's the shimmer for folks ATM? Any improvement or still causing issues for you?
I'm still getting it on everything.

[charred merlin]

https://sunomusic.typeform.com/to/zGsJSzsf

If you fill this out & contact [e-mail] they can assist with that 🙏

[charred merlin]

Not an option in Suno atm — https://bandlab.com/splitter can separate into bass, drums, vocals and other — paid memberships can separate into guitar and piano as well

[barren snow]

I was hoping for something better than RX 11 Standard or what I'm currently using Acon Digital Acoustica Remix with is currently better than RX 10 Standard rebalance.

[charred merlin]

Tremendously improved in my generations, with the current occasional issue being a progressive degradation in audio quality.

That said, certain genres seem more impacted than others, like Metal

[charred merlin]

The majority of services provide 4 stems — some paid services provide 6 stems. (If you have studio-level budgets, audioshake has like 11 or something)

[barren snow]

Was hoping that SUNO's A.I. could create the stems as it creates the master, (understanding it would be a slower, more expensive process). Maybe if not split the stems into WAV files, maybe split them into MIDI files with would be less CPU intensive.

[charred merlin]

It’s not THAT long ago that “we” managed to figure out how to retroactively create the stems from a provided track (Marvin Gaye’s “Grapevine” master track was found after decades and his isolated vocals are amazing to hear, but now ANYONE can do that with AI)

But creating multiple stems at the same time as it generates them would be a bigger technical problem than it might seem.

[raven turtle]

Thanks for the reply. Yes, I'm doing synthpop and electroclash and getting it on everything unfortunately

[raven turtle]

I use Fadr for the stem splitting. Works pretty well TBH and it's Free

[charred merlin]

(I understand it’s not 1:1, and the generative methods are different, but for analogies sake)

When you generate an image with AI, it creates it all at once — easy enough for AI to segment after the fact, but a bit trickier if you tell the AI that you want it to create everything separated by layer to be used in photoshop — suddenly it has to create MULTIPLE images that work cohesively

Not impossible, just way different than how it’s being generated ATM, so a lot of work would be required

[barren snow]

Generating the details for creating the music, doesn't it produce the notes needed for the instruments to play in the song? Those notes and note lengths could be recorded into a MIDI file without having to encode all the data required for like a wave file.

[charred merlin]

It generates an audio file — not metadata or musical notation 😦

This idea has been suggested before, and it’s a great idea — keep an eye on the AI field, because this is likely something companies are fighting to work on

[raven turtle]

I don't think it works that way. My understanding is it's more like AI image generation which is like an impressionist painting. Blocking in the major aspects and then refining it. It's a single work, as opposed to a layered Midi file like in Logic Etc

[barren snow]

The way it sounds like is that the computer computer's voice has all the instruments and everything wrapped inside itself.. It's kind of like a human voice. It's got like two tones I believe.. at one time it was hard to reproduce a human voice, but now we figured it out.

So which means the computer voice is got multiple tones going at the same time.

[charred merlin]

@raven turtle Is the shimmering still just as present if you remaster one of those generations?

[barren snow]

I will look into it

.
.
How about a unmastered version? Is that possible right now?

[novel dove]

There's still no way to delete a persona?

[barren snow]

.
.
Anyhow I should go to sleep.. thank you for your input. I'll be watching this field. I current use SUNO to get me out of writer's block.

[wide tendon]

wrong chat lmao my bad

[raven turtle]

It's less TBH. But usually effecting vocal reverbs and delays when there is risers and cymbal crashes etc.

[raven turtle]

It's not bad at all. And totally usable as a bed in a DAW.

[young prairie]

I want to start a new subscription starting from today

Are there 2 options?

1. Immediately with proration
2. At the end of the current billing period, on Jan 14, 2025

Which recalculation date should I choose instead of expiring on January 14, 2025?

[charred merlin]

Go to the persona page, top right you can click the 3-dot interaction menu and send them to trash. (there's no way to delete them from the trash, last I checked, and songs created with a persona are just stuck)

[graceful grove]

hey i have 3000 credits and a proplan but suno wont allow me to generate song plz advise

[novel dove]

Ok, there's no way to delete from trash. Why?

[charred merlin]

According to the support lead, there's a 90% chance that means you have a payment listed as "past due"
#🪲┃bugs-web message

[young prairie]

Are there 2 options?

1. Immediately with proration
2. At the end of the current billing period, on Jan 14, 2025

Which recalculation date should I choose instead of expiring on January 14, 2025?

[young prairie]

I want to start a new subscription starting from toda

[young prairie]

I want to start a new subscription starting from today

[charred merlin]

@young prairie -- you'd DMed me the same questions where I'd answered -- would you prefer to talk in DM or here?

[charred merlin]

[young prairie]

here

[charred merlin]

Option 1 -- you receive the difference between the 2 plans (instead of getting 10,000, you receive 7,500 -- this is "prorated")

Option 2 -- you receive 10,000 in 2 weeks when your plan refreshes.

[young prairie]

Thank you, I understand👍

[charred merlin]

Option 1 gives you the credits today, I just wanted to highlight that it likely wouldn't be the full 10,000 -- as I didn't want it to come as a surprise 🙏

Happy New Year 🙂

[polar spade]

I suspect, but haven't tried. One could do a whole-crop of the original song so you have an exact copy. Then delete the original song, then delete the persona?

That sound right?

[charred merlin]

You can't delete a song that a persona was made from

[polar spade]

Sorry, mate. I see they have now changed the options to allow for proration. So ive deleted my original response. Looks like @charred merlin has gotten you up to current.

[polar spade]

Oh? Hahah... can't delete the persona or the song. Would deleting every song you used the persona on help at all. I just realized why you can't delete anything.

[charred merlin]

Yeah, I was pleased to see the new dialogue -- 💪 It's a common enough occurrence

[polar spade]

I understood you can't delete the persona without deleting the song first. The pop up says you mist delete the persona. So is that do-able?

[young prairie]

thank you for your help👍

[polar spade]

I just haven't tried. I'm lazy and just asking. I could just go under persona and test.

[charred merlin]

You can send it to trash, and it just... lives there indefinitely.

[polar spade]

I only attempted to help. Just made things more confusing.

My new years resolution is to just be quiet. Lol.

[polar spade]

Well. That's sounds nefarious. But I can guess why.

[young prairie]

Happy New Year

[polar spade]

Happy New Year , my friend.

[charred merlin]

I'm not sure if it's an oversight or an intentional choice, but either way it's a little silly 😛

[polar spade]

🙏.... that I stay quiet? Ohhhh....OK. 😉

[charred merlin]

lol, no

[polar spade]

Oh... and Happy New Year. Long time. Been what?since last year?

[proper quiver]

Happy New Year!

so I hear that fadr.co finally works with MIDI?

[tiny falcon]

Happy New Year!
Yeah. You can download the sterm as Midi file.

[proper quiver]

can I upload my own midi file though?

[tiny falcon]

I just tried but I don't think you can. I tried upload to (Get sterm) and (remix) option but it won't let me.

FYI

[proper quiver]

ooooh i seee

[proper quiver]

thank you

[tiny falcon]

You're welcome. I am using plus account also

Not sure the free account have midi option though.

[naive quail]

It does, but it's quite cringe-worthy 😄

[tiny falcon]

If you want, I can help you get the sterm and download the wave. I can't really tell a big different with my ears 🙂

[naive quail]

It's not something that I need help with - don't really have any use for it. I've just played around with it a bit out of curiosity.

[tiny falcon]

I am not sure if bandlab is better thought. fadr, when you extract snares and drums, the drum has snares in it. it's not totally seperate everything.

[tiny falcon]

I think RX11 does a bit of extracting sterm. Haven't play with it that much

[lyric sonnet]

the "Public" switch is disabled on my newest songs; is this a known bug?

[naive quail]

Did you use uploads for these tracks? If Suno detects vocals in an upload, it will flag the upload as well as any tracks created from it (via extensions, covers, etc) so that they cannot be set public.

[lyric sonnet]

it is based on a short audio clip i made myself which does not contain vocals. the songs in question do have vocals (which suno created during the remix)

also, i have a LOT of versions of this song and until a certain point the 'public' switch is enabled, just not for the latest 20 or so

[polar spade]

To go further: Suno could interpret other sounds as human vocals such a didgeridoo or distortion. Anything you'd think might trigger it?

[lyric sonnet]

actually. it was a two-step process; the first audio clip produced a track with vocals in suno which i liked, so i downloaded it and cut out a section i liked and reuploaded just that and remixed it

so suno detected its own vocals as problematic?

[polar spade]

Any vocals are problematic. Yea.

[lyric sonnet]

no dodgeridoos. here's the song if you're interested. it's very good. https://suno.com/song/afd12eea-1c4a-4925-a46c-332b31c73704

but then why is 'public' enabled on other versions of the exact same song with vocals

[polar spade]

It wouldn't know the difference between suno , yours or Aerosmith vocal.

[naive quail]

I choose to view this as flattery...

[polar spade]

Not knowing your library, but older songs which "passed" the original lighter moderator filter might not pass on future generations of the same.

[polar spade]

My comparison would be "can't tell difference between my vocals and Randy Newman."

[lyric sonnet]

double-checking this.. you're right, exactly all songs derived from the suno re-upload clip are nonpublicable

i find that a bit harsh.

[polar spade]

Might be it then. It's hard to tell, just have to guess, but that's been common.

[naive quail]

Yea, I doubt that it's doing any level of checking other than simply detecting whether there are vocals or not, so it has no idea that it may have just created those vocals a minute ago.

[polar spade]

🎶 You got a friend in me 🎶 well, he's made millions with that voice. So I guess that's something.

[unique stone]

Anything uploaded that is detected as containing a vocal can't be made public.
This restriction is also inherited by anything created from that upload.

[lyric sonnet]

thank you for the clarification!

well, at least i can publish it elsewhere. it's still unique

[upper wren]

I can't see the lyrics a few songs, anyone knows why? please help. I am using ios app.

[runic ocean]

dumb question but how do you add sound effects like gasping, screams, pained noises etc to a song in the vocals?

[shell cradle]

Try meta tags in []?

[runic ocean]

will try that when I get credits back at midnight thanks

[naive quail]

Suno is very hit-or-miss when it comes to sound effects; mostly miss.

You can use tags like Fowks suggested, such as [beep], [gunshot], [laughter], [sigh], etc. Some work better than others (beep and gunshot, for example). Some may only work one out of every 20 times you try ([bong rip], for example).

When they DO work, it may not always be at the exact point in the lyrics where you added the tag. I've had some sound out at the beginning of a track even tho the tags were much later in the lyrics.

Here are a few that I've experimented with: https://suno.com/playlist/49a40042-a09a-421b-9df1-3f6d976221f9

There are TONS of sound effects in the training data, tho we can probably only directly tap into a relatively small number of them.

[quick wedge]

That’s correct meta tags are in brackets []

[lilac garden]

Suno doesn't know about a genre i am looking for that only uses vocals and drums. Anyone has any helpful tags to create music that includes only 1 voice and 1 instrument

[placid basin]

Can a previous (or export) of a song be uploaded and then can I add new/updated lyrics? One of the (many) attempts is exactly what I want, but I've rewritten the lyrics.

[half dragon]

I think you can use Cover feature for that, no need to download then upload it.

[signal summit]

So Why does the tracks get so scatchy/high pitched towards the end of every song. I have to Use DeEasser in Logic pretty heavily to get it to even get close to matching the beginning of the songs...Which start amazing. Hope this makes sense?

[charred merlin]

This is a known bug in v4, especially after the 2m mark -- beyond using remaster/cover to "reroll" the song, or editing the song externally, at the moment all we can do as users from the suno-sude is to downvote the songs or use ... > Report > Bug > Poor Audio Quality to give feedback to the AI model so they can try to avoid that type of pattern in future updates

[vocal cipher]

Has anyone else noticed that when using the replace section featuer, that when you confirm a choice , it is NOT splicing it togeather properly at the end of the new clip?

[charred merlin]

Are you able to share a particular egregious recent example, with the timestamp of when you can hear the seam?

[vocal cipher]

https://suno.com/song/1f517dcf-3a29-4db6-b3f8-c789ac702172 , at 1:07 / 1:08

[lethal pecan]

Hi, can I chose the voice of the singer? and use it for all my songs ??

[charred merlin]

Technically, there isn't a way to outright choose the voice by default -- if you're on a paid membership, you can use Personas to replicate the voice & instrumental style from another song, either making a persona from a song you made on Suno, or using a public persona that another user made.

[lethal pecan]

the system automatically chose the voice randomly ?

[charred merlin]

Ahhh -- I gotcha, the style & EQ matches up, but it's skipping / not lining up?

Is that happening with every replace section you do?

[vocal cipher]

Seems to be, another attempt at same time stamps https://suno.com/song/f6c13682-783c-4e1d-b15c-0652dbb49fcc . in the preview of the new section it has that 'overlap' at the end where its supposedly playing where its blending it back into the original song, which seems fine, but it doesnt seem to be actually doing that , its taking the whole end and stickign it in , so it doenst line up

[charred merlin]

The voice is largely determined by the lyrics & the style of music -- sometimes it can REALLY strongly push a song to have a male or female vocalist, but users experiment with different ways to impact this.

Sometimes adding [male] in the lyrics before a verse, or telling suno to use a male/female singer in the Style of Music section is enough to help push it back the other way if it already has its mind made up, but sometimes it is very hard to overpower and we need to be creative in how we approach the problem.

Personas can be very powerful because instead of asking Suno to imagine a new voice & hoping that it sounds the way you want, Personas let you choose that existing voice and use for other songs

[lethal pecan]

and if I like the voice and I want use the same in a different song is it possible? I read that in paid plan it is

[urban sonnet]

damn thats harsh and upthere in the FQ field (like everything else:) 🤣

[vocal cipher]

Yeah, super frustraiting burnign through a few hundred credits, getting a replacement that is perfrect, then having it just not line up the way it does in the preview

[charred merlin]

It's as simple as clicking on the options for a song you've made & choosing "Make Persona" -- then whenever you make a song, there's a box for "persona" that you can browse & select that vocalist/band to play more songs 🤘

Hey @patent pendant -- you're correct that signing up after the fact doesn't give you retroactive rights to previous songs, but if you've signed up recently, you can email [e-mail] with the list of songs you are requesting commercial rights for, and they've been incredibly helpful with that in the past 🙂

[patent pendant]

so i could ask them for the rights to it? they will actually do that?

so sorry again new so i just email them and do i just request the rights to the songs? or do i need to tell them why? would the subject be something like requesting rights to songs? sorry kind of slightly freaking lol

[charred merlin]

@vocal cipher Ah yeah, an audible seam overlapping "kept hidden" -- I know it's not a credit-friendly way of resolving the matter, but a Remaster is likely the most straightforward way of "smoothing over" any hiccups like that -- if you edit & realign the lyrics it might have an easier time pronouncing "kept".

"Replace Section" is seemingly something they'd like to improve upon, so I appreciate you sharing the examples -- I'm sorry I wasn't able to be of any more immediate assistance, but I'll pass them along as examples 🤘

[vocal cipher]

Hadnt considered a remaster, on the result, worth trying!

Remastering the result smoothed it over ! thanks! 😄

[charred merlin]

As long as you're currently subscribed, you're most likely in the clear;

Hey, I made some songs while I was on a free plan, and now that I've subscribed, a user on the discord server told me that I might be able to get the commercial rights to the following songs that I'd made? {list of songs}

Happens pretty frequently that people play around with it and don't realize how good the songs will turn out to be -- one of the founders commented a while back that if you sign up within the same day, it's never likely to be an issue

I'd imagine unless it seems like something fishy is going on, they're interested in helping artists -- but if someone said "I made all of these tracks over the last year, and then I signed up for a month, can I have all of them?", there might be a bit more pushback 😅

[patent pendant]

Thats awesome. i had no idea they could help with that thank you.

[lethal pecan]

ifi I find a voice I like in different song I can save it and use it ?

[charred merlin]

Only songs that you've made, not other suno artists -- not quite what you asked about, but still related; you can make personas from songs you upload if they are instrumental only

[charred merlin]

Bodacious! 🤙

I'd love to hear it if it makes its way to #🐣┃suno-showcase

[lethal pecan]

I can upload only instrumental song and based on that it will create an artist?
And I can save my artist? how can I do that ?

[charred merlin]

By "Suno Artist", I just meant other Suno members -- you are only able to make Personas from songs you have made, you are not able to make Personas from songs that other users have put onto Suno.

And yes, Persona will attempt to copy the vocals and / or the instrumental style -- if you upload a song of you playing acoustic guitar, you can use that track to make a persona that sounds like that acoustic guitar 💪

[lethal pecan]

and ser please, when I like one of my vocals, How can I save it ?

[half dragon]

As @charred merlin mentioned above, you'd be using Persona feature to do that

Go to your song, under Create/Make Persona

[tropic star]

"Hello Suno AI Team!
I'm Manup, a PHP developer with a passion for music. I've been using Suno AI to create amazing tracks and I'm blown away by its potential!
I'd love to collaborate and contribute my coding skills to help take Suno AI to the next level! Let's make music creation even more magical!
DM me or let's chat here about potential collaboration opportunities!"

[hollow hare]

Hi there! If you are looking for a job, please check out https://jobs.ashbyhq.com/suno for job opportunities at Suno

[vocal cipher]

Its posted there now if you're interested

[charred merlin]

Oh man, I was worried Woodhouse was going to take down the showcase post 😅

I see it, listening now 😄

[charred merlin]

Dear lord it must have been satisfying to hear the line sang perfectly after the journey it took to get there -- I'd be the one to make someone stop what they're doing just to listen to it.

"I don't get it, it just sounds normal?" -friend
"EXACTLY!" - me

Also, that remaster turned out really well, kept the bass steady throughout -- nicely arranged song, man 🍻

[vocal cipher]

Thanks ! I'm glad you liked it 😄 after 5 /pages/ worth of generations , section replaces, crops and remasters, it was a very satisfying result 🙂

[charred merlin]

I'd love to see a data-science breakdown of each finished song I've made, showing each of the alternate takes that I ended up abandoning; a breakdown & tally of the credits spent (though I feel like I'd go insane comparing the "one-shot generations" that I made for shits & giggles vs the songs I poured my heart & soul into 😛)

[vocal cipher]

Oh god , i can imagine . There's a few that have had SO much time / efforts and Credits put into them , then a couple that were just hilarious or amazing on the first try

[charred merlin]

legit what it feels like to start work on a new project after Suno just one-shotted the greatest song you've ever heard

[leaden lark]

That is so trueeeee

And then it takes forever to find that perfect gen again for a new song

[charred merlin]

Sara-san! Happy New Year 🙂

[mellow nest]

||Not me getting Suno to do covers of 50s rock and roll songs and getting consistent bangers that I mostly keep for myself to hear||

[unique stone]

If you've uploaded audio from such songs, or used the lyrics from those songs, that's against ToS just so you are aware (even if it's just for your own use.)

[mellow nest]

That's why I always save the best ones just in case

[west yew]

The "replace section" has definitely been borked for a while, at least in v4. It has a habit of deleting work several seconds after the splice to make things work. It worked perfectly in v3.5, but I haven't worked in v3.5 for about a month now.

[lethal pecan]

thx.. btw do I own rights to music ?

[faint kindle]

Testing out/training my own UVR model for AI based songs. Here is a current sample of what a output is.

The original is on my page.

Plan to make the model public once more kinks are fleshed out.

[hollow hare]

See #1294423389515284552

[faint kindle]

Forgot to include the vocal track

[cunning mulch]

that's... pretty damned good. I mean, sure there's a tiny bit of instrumental bleed on the vocals, but that can probably be worked out in a DAW

what we need the most is something that can rescue v4 tracks from shimmer and instrument degradation

[faint kindle]

It indeed can, its only about 88% trained. I just wanna test other genres. If anyone is willing to lend me a couple of their songs to test it would help. Cause there might be slight bias on my end for training

As per shimmer, I've been targeting that in samples for a 3rd track that removes it

[cunning mulch]

I've got a banger v3.5 track (that also degrades, unfortunately) that I'd love to be top-quality the whole way through

[faint kindle]

Send it over in dms, I can do a quick run though and see what I get

[neat field]

my current method for this problem
Start with v3.5 good generation
-> Cover in V4 with a slightly different Style tag setup than the original (3x on average), any slightly different style tag will do
-> Pick the best Cover from previous step and Remaster that one V4 (3x times on average)
-> Pick the best Remaster
-> optional step! Remaster the Remaster (again) - sample size too small to say if this helps or how much

[autumn field]

Hello, happy New Year to you all, I am new here, and I want to use my persona feature, but it's not available, I have to wait? and how long?

[cunning mulch]

Well, there's kinda one major feature I'd like to keep: There's an instrumental section after the final chorus that (melodically) lines up with a point nice and early in the song, which would make a nice looping point

[neat field]

i find the method keeps instrumental loops mostly intact, given enough Cover attempts (10-25% of the time) - i would love if someone found a more "credit efficient" version of this method

[cunning mulch]

I mean... I'm not sure if you get what I'm referring to