📚┃suno-school

[summer dome]

I have the premiere plan, what does "Re-record a song myself?" means? does it mean I can record in real life with real singers and make money?

[tender compass]

Yes if you have Premiere or pro plan.

[lusty spire]

Hello, how can I edit a song that I already made without losing the instrumental voice characteristics, etc.? thank you

[wind viper]

extend the song

https://www.suno.wiki/faq/making-music/how-do-i-make-the-song-longer/

[bold trench]

thanks !

[celest umbra]

Hi. I am Anna, a blind musician Fromm Germany
I am new Here and have some experience in suno. By the way it's very accessible for blind people using screen readers and braill Displays
Well i came accross and issue, and i think this can be also important to others well to the admins, If this topic does not match my question. Please push it to the right topic. I have Written a song in German for my father who's. I Haven 'ttalk to for seventeen years and now got to have a conversation with him lately
I have written the text and created that Song in suno
Now i want to push the voice away. So i only have the backing track of the song and friend of mine pushed IT through Filter, but the instrumental sounded very scratchy after filtering the vocals out, so it would be nice to have the possibility to just have the backing track of the song, We have just made in suno, to sing the song with our own voice
Unfortunately, i'm not using reaper or having a possibility to create Similar instrumental like this one because i definitely don't have such instruments on my synthesizer
Greetings, anna

[fathom basalt]

I think that, this is not a good enough quality of songs generated by suno to upload them to spotify.

[tender compass]

don't expect to become a professional artist 🙂

You can only improve yourself and rise thanks to Suno. Think on the bright side.

[fathom basalt]

I see a lot of bright sides, that's why I'm here 🙂 It would be nice if they improved the quality, even for an additional fee.

[tender compass]

the quality of the songs will improve as the models improve 😉

we just have to wait.

[fathom basalt]

I hope You are right.

[celest umbra]

Well the song I wrote for my father in the verse has some slight Synthresizer Instrument and after filtering the vocal Out, IT soundedlike a hiss or a scratch. I can post the link of the song, If you like

[sly grail]

do pro credits renew at the end of the month or at the month anniversary of when I signed up?

[tender compass]

Free users' credit is renewed daily.

If it is in the premier or pro plan, it is renewed monthly.

[neat nexus]

On the date which you signed up.

(At least it does for me.)

[sly grail]

thanks!

maybe I'm greedy but it would be cool if pro/premiere plan also got 50 credits daily 🙂

[meager moon]

Hello. I'm Maureen from South Korea. I'm new here, can I ask something that I want to learn about suno ai here?

[broken lotus]

yes. I will help if I can

[meager moon]

I appreciate it. 😍
Can I change the lyrics that my suno already created?
I tried "extend" feature with the Custom Mode, but suno changed melody which was really good for me.

[zealous stratus]

is there any other ai? like suno?

[broken lotus]

editing inside a song is a potential feature for the next model (v4)

Developers are calling this 'in-painting' (like image-gen)

[meager moon]

Super cool 😁😁 Thanks for helping me. I can't wait to meet that upgraded Suno ;D

[spare talon]

is there a way to regen a song with slight variation? that's probably my most requested feature

so many of the songs are so close to being great

[tight canyon]

How come when I try to extend songs it's not reading the lyrics how I want. Like it's starting on the last line in the verse every time, or doesn't sing the whole thing and starts halfway through the verse.

[sly grail]

Make sure you're starting from a good break in the piece you're extending (note the timestamp). Also let it cook for a bit, I noticed if I press Play too fast it gets me weird results while it's still extending

[edgy vapor]

Hi

can anyone tell me how to add a guitar solo to a song, please?

[sly grail]

you can try meta tags but it's like rolling the dice whether it works -
[Break]
[Guitar solo]

maybe [long guitar solo] but short meta tags work best

[gray violet]

A cover can ruin a good song, that one has nothing to do with it. Can't I really change it with something as simple as changing the avatar on a profile of any SaaS?

https://suno.com/song/497bb3e0-7292-4f04-8947-6bc465447132

[edgy vapor]

Thanks. But whats more important I cant extend song properly(((

[silver kelp]

why cant i make another song?

pls someone help

[edgy vapor]

0 credits(((

[silver kelp]

where do i see it?

how can i get more?

ok got it

thanks

[bitter tendon]

/chirp

[bleak dragon]

HI, I have 4 songs I created a few weeks ago and some older songs in the library. I have links to the songs which I saved back then, but I can't seem to access the account. I think I signed in with the wrong account. I deleted the cookies out of the browser but I have no option still to sign in again. Little help?

[north rock]

on each prompt suno generates 2 results, is there a way to change it to 1?

[broken lotus]

no, each 'job' returns 2 clips.

[north rock]

ok thanks, I wish there were more control on how many clips are generated per 'job'

[hollow hare]

Click Library and the Sign Up screen will popup. At the bottom left, there is a link that says Sign in

(You can also click create but it might start a prompt when you sign in)

[jolly minnow]

well... that was a ez awnser...i counter with... Why?

[fathom basalt]

When v4 will be released?

[broken lotus]

we have not heard an ETA, but devs have said it is the priority.

v2 November
v3 March

[fathom basalt]

Thanx.

[rough flume]

Just getting a constant stream of 'credits refunded' when trying to extend to V2 over the past few days. Very annoying

[tropic barn]

I'm becoming increasingly frustrated with songs that are cut off in the last second or so, even when there's still time left on the 2:00 minute clock. This forces me to use extra credits to add in music that should have been included. Similarly, I'm tired of wasting credits on songs that start partway into the lyrics, missing the first few words or even an entire sentence. There should be a system to flag these issues and have them reviewed for a possible credit refund.

[broken lotus]

There is.

[tropic barn]

Thanks

[bleak dragon]

Thanks I'll try that!

[lime cypress]

Anyone know how to set the language for your generated vocals?

[hollow hare]

IIRC, you must have the lyrics in that language (e.g. Kanji for Japanese vocals). Tags may help but usually more useful when you are alternating between similar languages in verses (e.g. between French, Spanish, and English... or the tags could be a placebo effect)...

You can see this in the Hello World song
https://suno.com/song/225b52b1-a2d4-4c73-acdf-a6bfb1b391cd

Though you may need to phonetically spell out the pronunciations for some words

[rough flume]

Client Side Error. Site seems to be down. Not a good look

[hollow hare]

Pretty sure devs are playing with profiles atm, come back to it in like 5-10 min (we have seen the profiles work and stop working a few min ago)

[rough flume]

Hmm...judging by the problems I've been having today, I would suggest they play around with much more important things than profiles.

[final hazel]

They are working on it.. will be fixed soon.. It is only effect profiles who have no public songs @rough flume

[broken lotus]

the Suno Generator seems to only make English now...
I use ChatGPT for lyrics in other languages, then paste them into Suno's lyrics box.

[final hazel]

Should be fixed again!

[alpine venture]

whoa

https://suno.com/song/4907cce0-49e7-4f3e-be31-458cab9eba92

never had anything like that before

[neat field]

cross post with general, but seems a few of us whom enjoy screaming in our music have stumbled into one way that seems to generate "female growling" more consistently in songs

#💬┃general-chat message specific tags in the test performed by another community member - worked for me in some other private tests too, but only combined with other genres

[sour knot]

Can a free-generated tune be converted to a licensed commercial-tune by simply upgrading account? thank you, or does it only count for music generated after upgrade?

[final hazel]

Their FAQ says no. But you can always try by sending them an email to [e-mail]

[sour knot]

thanks I appreciate that response, it's just that I'm worried to make a good tune in free-account and then wish I'd been in the paid version haha

[final hazel]

That happened more than once 😹

[sour knot]

it should be a simple case of the music generated by you is now a licensed copy?

because you've paid

and I think that moving forward the devs/team/ceo should consider that

unless there is a technical reason that can't be?

[final hazel]

They are aware, but abuse is the problem

[sour knot]

You mean generating a good tune, paying 8$ and leaving?

[final hazel]

yes

So if you pay for a full year, you might have leeway.. But I am a moderator.. Not staff

So I would just send them that email and ask if that is the case

[unique stone]

Even if you generate something great and end up subscribing, you'll make other/better things in future so don't worry too much about it. Could just sub for a month on the lower tier to have plenty of time to try stuff out (that's what I did, not limited to a few tries per day then) and avoid this concern but understandable if you just want to play around for free.

[potent steeple]

Hello. Where I can find information or text tutorial about using tags in lyrics? Seems like I'm doing something wrong, because when I'm giving to lyrics tag "male vocal" I'm getting female vocal.

[neat field]

it's stronger in the prompt to add "male vocal" - i find tag in lyrics does not over-ride genre in many cases for this specific use-case

opera, male vocal, pop etc

some sub-genres may also have very low gender representation in the original dataset, so may favour one or the other

[potent steeple]

wait a sec, I thought I should use tags in []. Is it correct?

[broken lotus]

are you adding any 'male' or 'male vocal' in your style prompt?

might make it 'stronger'

[potent steeple]

yes, but I've added it like this "heavy metal, nu-metal, male voice, female voice". I wanted to make female to sing verse and to make male to sing chorus. So I'm using this construction in lyrics "[verse] [female voice]"

[broken lotus]

ahh, have a look at this page... some tips to try (but this stuff is very unreliable)
https://www.suno.wiki/faq/style-and-lyrics/duets/

[glacial scarab]

thats a bit difficult.
If you really want it. You need to extend before the chorus with the other voice type.. ah just read the link that was posted while i was typing

[green grail]

im tryna generate a deathcore song
they have no vanilla voice chorus sections and the vocals are simply screams vocal fries or false chords etc
how do i remove the chorus section which uses vanilla voice
and maybe add a guitar riff and breakdown
I have specified all of this and the ai still gives a normal chorus
is there a specific way to use the tags
also how to specify the bpm of the song

[potent steeple]

Got it, thank you all, guys. I'm just starting with this thing and just can't find a proper documentation. The last question - is the [] has effect on something?

[hollow hare]

https://www.suno.wiki/faq/metatags/

[potent steeple]

Thank you. I found only this https://suno-ai.notion.site/Using-Metatags-34944efe09ec41d693e314f13af44695

[neat field]

could also try to drop gender in prompt and specify "duet" then use gender tags for each duet part to your preference

[potent steeple]

Got it, I'll try

[neat field]

https://suno.com/song/5a9c7d44-a85b-49e9-8a43-fe4364905ed4 this is my go-to example but not quite the same

[potent steeple]

Oh, it's good to have an example

[alpine venture]

A new observation: adding fuzz, antifuzz etc in a continuation that has abberant noise growth and manipulated voice appears to reduce this affect on the continuation

IN general, fuzz seems to suffer less, ironically, from noise growth

likely because this genre uses it as an instrumental tool

and thus it has a specified range

[neat field]

oh

[alpine venture]

So, I actually think, broadly, using genres that have BAD audio quality, but like, purposefully so, produce better results

In terms of controlling the audio

as weird as that is

since nobody who is presenting a normal audio file will tag it as "denoised"

they'll just not tag it as "bad quality"

if that makes sense

Otherwise all professional music would be genre tagged as "denoised, well produced, etc."

this is why I often use cassette, anti-folk, indietronica, fuzz, slowcore, midwest, trap, and so on

all of these feature at times extreme noise progressions

and they all seem to deal with audio abberation pretty damn well

[brittle minnow]

i wish i could just make the AI create soundspaces, what i want i the sound of a crew on a spaceship laughing as they are celebrating the captains birthday. XD i have this idea for a very sad dystopian synthwave about a lost spaceship with a computer replaying all the memories it has stored on its harddrive of the crew that died hundreds of years ago xD

[glacial scarab]

oh I will experiment with some of these next week when my credits refresh. I collect tags that "improve" audio quality. Improve might be subjective so it's more like tags that change the audio in general

[brittle minnow]

will we have more control of a project in ver 4? I sometimes want to keep sertain songs but dont like the beginning of it for example.

[tender compass]

yes there will be nice updates for v4. stay tuned.

[maiden trench]

I think it would fall under you using their licensed prodcut to create so they can determine it's use. The commercial thing is incentive to get you to pay. Also someone could just use it for free and record their tune and not pay? Where would the control be there

[sour knot]

to be honest I am more than likely going to pay the 8$ and gen

i just wasnt sure

getting a feel for it atm, but wanted to also address legalities of using it in a commercial game

since AI is new, we're all worried =]

if i am assured the content isnt ripped from Sony, and so forth

we're all good?:D

[sterile trellis]

Does anyone have any tips for shifting from one style to another mid-song?

[maiden trench]

oh i completely understand! that's why we are all on here, to chat and figure this AI shit out lol

[maiden trench]

easier if you repeat keywords I've found. For Example if I go Flute Solo, Electronica, Downtempo, Dubstep, Dubstep, Drum and Bass. It will start with flute, then a slow electronic part then a dubstep/drumstep part. sometimes.

I've found it's better if you just write out a complicated genre. and then it will change itself in creative ways. rather than you forcing it too, it can force it into jarring transitions

For example I could get the same effect by typing "flute intro, Hybrid of Electronica and Dubstep"

THe second will sound better usually too

all these differences could be in my head i know half the shit i type in the prompt it doesnt recognize

[sterile trellis]

I was thinking that using the “Extend” feature, and changing the style for the new part would affect things more than it does - it does introduce elements of the new style sometimes, but it seems to have a very strong memory of what came before, when I do that

[unique stone]

@sour knot there is no info available on what the AI was trained on, if that is a concern. That's up to Suno to figure out IMO. So long as it's not producing chunks of existing songs I don't see a problem. I'm not an expert in the legalities of anything AI related though.

[maiden trench]

Extend but from an earlier point in the song before the Genre is set.

[sterile trellis]

Hmm ok will try that! I have been wondering whether it’s better to extend from a bit earlier than the actual exact time I want to continue from…

[maiden trench]

Not only up to Suno. Up to platforms too.

[unique stone]

Hmm true.

[maiden trench]

my best songs have been all patches of 20-40 second "extensions"

[sterile trellis]

Also, been wondering about the lyrics when I extend - should I leave the end of the previous lyrics there so it has context?

[sour knot]

if some lawsuit was to happen is it theres or my fault?

[maiden trench]

Spotify apparently doesn't care. However I've made a track or two where you could def recognize the underlying music and I wouldn't be comfortable posting that. And I know a lot of music but not all of it so I'm sure all the songs I've produced have other "recognizable" parts that an artist could pursue you on. No artist has done it yet.

[sterile trellis]

There’s no way Suno wasn’t trained on copyrighted material - zero percent

[maiden trench]

For example I've made plenty of songs that sound just like the artist Yellow Claw. One of the easier ones.

And Hannah Diamond

I can recognize the actual songs it's drawing from too.

[sterile trellis]

Blink 182, Queen, Eminem…

[maiden trench]

Thus the artists definetely can.

And their lawyers can if asked.

[sterile trellis]

Made a grime track that sounded stupidly similar to stormzy “shut up”

The question is, is training an ai on copyright music “fair use”? Only time will tell…

[sour knot]

all new territory

[maiden trench]

I think it's being handled on single case basis now by artists and their labels. I've heard of a few cases of AI music being removed and I'm sure legal threats were made behind the scenes. But you won't get flagged anywhere unless you become mildly popular and have a bunch of songs that sound exactly like another artist. Just vary your genres slightly and you're find.

Just credit Grimes if you aren't sure. Not kidding.

I was semi kidding. But she is only artist completely down with AI generated music right now and her likeness/sound being used.

[broken lotus]

check this page, at the bottom are tips for a sudden style change
https://www.suno.wiki/faq/making-music/extend-from-time/

[broken lotus]

Seeing as how Suno takes MANY steps to prevent us copying any specific artist..., I'd say a court trial is going to take a hard look at YOUR intent, does your prompt or the lyrics (maybe your marketing) indicate the artist in question?

That would be very clearly on you.

remember agreeing to this? 👆

[sterile trellis]

Amazing, thank you!

[stoic stag]

I've tried to recreate this style with 'actual' prompts, but I can't figure it out. Been trying for days before I waste credits trying to find it in the sub version. Does anybody have any idea how to recreate these vocals and/or instrumentals? Any help would be immense, thanks.
https://suno.com/song/fdb6b9e6-bdfc-4d19-8621-ccd9240bd5e5

[sour knot]

only making instrumental for a game (bg music) like medieval

so not worried about song lyrics

[broken lotus]

lots of tips for game music here on discord, and orchestral soundtracks if that's more your flavor.

[marsh agate]

Can we include instruments in the prompts? We got a list what lands? Was trying some Asian instruments to mix into Synthwave and nothing was hitting.

[jade field]

I'm trying to figure out how to make the Mumford and Sons sound from the mid 2000s. I've tried entering separate instruments, etc. to no avail

[unique stone]

Yes, though there's no guarantee the AI will use it (sometimes depends on whether it "fits" with the genre.) What I try to do if the AI is ignoring my instrument, is extend the song and just replace the entire style prompt with that instrument. Sometimes even turn off "instrumental" mode and put the instrument name in the lyrics e.g. [piano solo] just to try to encourage it to use it.

It also depends whether it knows about certain instruments. I've had a koto randomly turn up when I didn't ask for it before and it seems to recognise that. The wiki at https://www.suno.wiki/ might give you some ideas to try. There's nothing concrete though - gotta try stuff to see what works.

[marsh agate]

Thanks, I'm pretty solid at Stable Diffusion and have some LLM experience, so I can figure out the prompt structure easy enough.

Oh for sure, these meta tags are the way.

[unique stone]

Sometimes an instrument or genre has a lot more weight to it, for example any time I try "music box" on its own I can't get it to do anything else later, but I think you can also repeat tags or put the most important tag at the front to give it more importance.

[marsh agate]

Yes, likely. It's like that in stable. The start of the prompt gets weighted heavier than stuff later in the prompt.

[sterile trellis]

Got it working really well, thank you again!

https://suno.com/song/b6545537-313e-48a4-a638-f991ef96c511

[broken lotus]

AMAZING!

[sterile trellis]

Been making music going on 20 years… this AI shit is wild 😂 and it’s going to get so much better…

[alpine venture]

Yea, I'm surprised you havent used them before

Cassette is quite strong in its impact

You'll hear that heavily in my next album

But also some random things like anti-jumpscare

have strong impact when combined with sound fx

or rather, anti-sound anti-fx

it usually leads to multiple cassette breaks

and I mean, literally you hear the person change the cassette

Which is super interesting in a dance album, at least to me

[tidal anvil]

I got broken audio where there's a part where the sound was like interrupted streaming audio or similar. But the other parts weren't like that... Glitch?

[dusky finch]

I am not sure if I'm going to word this correctly, but is there a way to make Suno stretch a word when using custom lyrics?

Sometimes it abruptly cuts to the next line when it would make more sense to draw out the word or pause before going to the next line.

I don't think [break] is appropriate in this instance, but maybe someone has better advice?

[analog jungle]

https://suno.com/song/12ad99d7-8158-4d22-824b-8ed870123a02

[tender kayak]

can someone tell me if its possible to subscribe while not having a visa/credit card or similar? I usually use paypal to pay for stuff online

[atomic isle]

Has there been any technical with suno lately? I went from **8800 **points to 600 And i did not use that up in a day

Tbh i dont think i even made a song

Any mods that could point me to how to submit a ticket? Cant find any info how or where to do it

[verbal bison]

Does suno have an open source api?

[daring wing]

A few of my songs don’t display the text lyrics when downloading video. Any known fixes

[zenith depot]

When drafting hard, I would like to mass delete, not one by one. Please fix.

also, some songs that have lyrics without instrumental checked, it doesn't sing

[neat nexus]

Please do not doublepost. It's in the #📜┃rules

There is no API available at this time but it is on the roadmap. I do not have details about what kind of API it will be.

I also have no ETAs to offer you.

[hollow hare]

Did you tried refreshing the page? Sometimes there is a visual glitch that shows the incorrect amount of credits.

If the issue remains, I believe the FAQs have an email address listed that you can contact

[neat nexus]

[e-mail]

[pale lichen]

just got hit with "Couldn't generate that. Prompt flagged for moderation". any way of finding out what part of the prompt causes problems?

[fallow gull]

Ask ChatGPT or Gemini why the song may have been flagged for moderation.

[buoyant cargo]

Do the Discord mods get secret roadmap access? Asking for a friend

[stone python]

Same issue, even if I don’t have lyrics and just generate Instrumental

[lofty merlin]

Asking for a friend named thatiodeguy who would like to become a mod if so? 😂 I feel you. I've been wanting/needing an API for my project ever since Chirp's first beta released... One day, friend... one day.

[analog jungle]

https://suno.com/song/dc93096d-8032-4cde-b668-12867c6a8af4

[sleek escarp]

Hi there! 🙌
I'm a software dev and would like to create a little project that allows music producers to generate tracks based on a input and text instructions( Like: "House Bass Line, 90 BPM), Then the 2 tracks will be generated and the producer can listen which one to add into the track in Abelton Live. And can continue iterating from there. The idea is to use AI to produce original samples/tracks instead of using an existing library.
I have an idea on how to do it, but I would need to have access to the API. Can anyone point me to the documentation ?

Thanks!

[unique stone]

A few messages above there is discussion about API. There isn't one available yet.

[merry pollen]

Is there a way to add more silence into Suno songs? Also what's re all the commands we can use example
[Drum solo]

[Jazz Ska Vocals Solo]

Could we have a list so we can explore the true capabilities of Suno ai?

[sleek escarp]

Thanks for the update 🙂

[bold current]

Can any of the Devs confirm for me if Suno supports the Solfeggio Frequencies?

[spring estuary]

Any plans for MIDI or Stems export (Maybe split into Drums, Bass, Instruments, Vocal etc)?

[broken lotus]

a primer for [metatags] is here
https://www.suno.wiki/faq/metatags/

[gray spindle]

Is there a way to just generate these album art images? I'd like to get a few more without spending all my credits....

[broken lotus]

There's a hack... to extend on a full song after it has ended...
add [end] to the lyrics (to make sure it stays ended)
then below
(a white unicorn in a lush green forest)

or whatever short image gen prompt...
This maybe costs credits but usually gets refunded as too short, so 'free' but with a risk of being charged if it decides to sing the image prompt...

[marsh agate]

You.. is that an open pose avatar come on ! 🤣 Im sure you can replicate a similar style.

[summer dome]

When will PayPal be available as a subscription option?

[drowsy tulip]

Has something changed recently? Song generation is like seconds instead of minutes, but the overall quality just doesn't seem quite as good - like my lyrics are being sung oddly and there's a lot more hallucination?

[broken lotus]

If you are getting hallucinations, try adjusting your Extend From timestamp 1 or 2 seconds.

[drowsy tulip]

Oh no this is from initial generation

[broken lotus]

if you are getting hallucinations in your 1st clip there is something wrong with your prompt(s)

[drowsy tulip]

I'll try some variations, but I've gotten some excellent results from a few weeks ago.. generation is definitely WAY faster.. I just wonder if generation time has been reduced to help with server load.. at the expense of quality.

[broken lotus]

I hear hundreds of songs posted every day.

No, when there are server load issues Suno drops the free generations.

[proven kestrel]

wow wetcircuit listens to all our songs confirmed 🤯

[broken lotus]

I try...
sometimes they go by too fast, but I try to hear a little.

[proven kestrel]

Damn impressive. That's genuinely nice, thank you.

[broken lotus]

(You never know when some has a prompt worth ste.., uh learning from. 🤣 )

[proven kestrel]

:<>

[broken lotus]

I wish there was a Suno Radio...

[proven kestrel]

Nah I'm doing that rn with felines tester voice so- who am I to talk... lmao

[proven kestrel]

y e s

[nocturne flame]

is the thing where it starts out one voice and then like slowly multiplies until it sounds like several voices a bug or what is that? Is it avoidable?

[broken lotus]

it is a 'bug' but it's an AI bug so it's weird and existential.

try minimizing your style prompt, remove redundant things (like electric guitar and drums in 'rock' they are unnecessary) It seems to hit certain genres much worse than others.

make a few songs with the 'simple' mode, and notice they do not suffer from the choral voices, and their prompts are very short (2 or 3 words)...

[mild yew]

Is it possible to assign weights to the the elements I put in the style prompt? For example, does order matter, in your experience?

[glass scaffold]

is there a way to keep alternating singers, like the first line of this song it goes from female to male. can i instruct suno to switch between voices like

[male}
<lyrics>

[female]
<lyrics>

[both]
<lyrics> etc

https://suno.com/song/a2124015-2795-4150-b096-1b51d31eee22

[green sierra]

https://www.suno.wiki/faq/style-and-lyrics/duets/

[glass scaffold]

thanks

[green sierra]

more info, also you can search to "duet" on top right corner

[strange ember]

Do you also have the problem that when you extend multiple songs, there are some noises/frequencies added?

[broken lotus]

yes, often... when vocals have lots of reverb, or distortion guitars... all around the same frequency... Rock and Metal are probably the worst because a lot of the song is around those high-pitch frequencies.... RnB and Electro will also degrade over many clips but a fraction of the issue so it usually isn't a problem... (not so much treble)

when that noise crosses the Extend into another clip it becomes like an instrument that is copied and keeps playing.
🤕
(it can turn itself off too, sometimes a [Break] can interrupt it)

[zinc island]

at the end of my full song everything was great then it cut off just before the last line of the song which would have been the title no way to edit I guess

[manic vapor]

just extend the full song

you can usually use short, short outro in the style tags, and try using

[End]

in the custom lyrics

after you extend you have th get the full song again from your ext

[zinc island]

thx I will give it a try, this is really freaking me out, My son is a songwriter, had 4 albums out and this is scary

[flat shell]

How do I make the voice that of a boy? I only get girl voices

[final hazel]

Specifiy it in the style of music, like for instance female vocals

[violet helm]

i cant extend my song anymore longer than like 2 seconds. Its a bug or whats going on? this has been going for a few hours now

[broken lotus]

have a listen to the end of the last clip you're extending from...
it probably has a dip in volume, or an abrupt ending with a second or 2 of silence after.
the Extends are continuing silence, then quitting.
(It's a know issue, that's why we get refunded credits)

the solution is to Extend from Time... and set the timestamp a few seconds ahead of the problem, before the music ends.

[violet helm]

i see thank you. I'll will try that. I think I actually tried it with another song today but then it was glitched and weird

hm

i continued mid-beat now and it seems to work, thanks

[broken lotus]

if it happened a few times on different song (and the same day) suggests maybe your prompt is triggering really short songs (shorter than the 2min max)

I get that sometimes when I am trying to invent a sound by mashing up genre descriptions... seems to throw all it's punch then go home early, LOL

(again, extend from time, and picking an earlier timestamp might help)

[fossil jolt]

i have some problems regarding the payment issues when buying pro. can a teamster contact me pls ?

[broken lotus]

ahh, we are just Discord mods... the only way to get help with a billing issue is through the email
[e-mail]

[fossil jolt]

i know, thats why i didnt ping mods but mentioned the TEAM guys.

ill email them tnx

[broken lotus]

(we can maybe answer questions, but have no access to a user account)

[fossil jolt]

all good

[tired basin]

anyone has recommendations on where to learn more about the types of models that are used by Suno? any info about the origin of the models being used? completely from scratch, derived from music gen / audiogen / audiocraft, based on publicly available research?

i've been wondering if it would be possible to have something akin to what Controlnet does with Stable Diffusion for example. control is really the huge missing piece, whether through weighing in the prompt or other means

[fossil jolt]

so no help with my problems here? i really dont like waiting days to get a response on an email tho. who are the TEAM people here if not direct contacts?

[violet helm]

if i have to "part 2" of one song, can i combine those ?
So that i have the main part with both of the part 2 as one song ?

[tired basin]

right click: get whole song

[violet helm]

no no its not part 2 and part 3

its two times

part 2

but i like both and want to connect both

[tired basin]

i think you're going to have to export

and edit them together.. unless i missed something ..

[violet helm]

ah damn i figured i had to do that.

Why is there no import feature to work with custom songs or an intern editor

[crisp jewel]

I am a paying user of suno from mainland China and would like to seek some help:

1. Since I paid to become a Premier Plan user, the lyrics of the songs I generated with Chinese lyrics have all changed from Mandarin to Cantonese. How can I restore the pronunciation to Mandarin? This is very important to me.

2. After I paid to become a Premier Plan user, I did not enjoy the benefit of generating 10 songs at a time. Currently, I still generate 2 songs at a time. How should I solve this problem?

[violet helm]

click on generate song

5 times

then you have 10 songs at a time

[desert wren]

anyone else down? full access to library and generate but no songs load to play

[violet helm]

nah

[violet helm]

for the pronounciation you might be unlucky. Maybe write in the style "chinese pronounciation" or "mandarin pronounciation"

maybe the ai can handle it in style

[crisp jewel]

not working

[violet helm]

whats not working

[lusty violet]

i was doing some tests earlier

I was able to get asia voices from the following styles

[crisp jewel]

I have marked the song to be composed in Mandarin, but the system still generates songs in Cantonese by default.

[lusty violet]

[c] (cantonese, china] [k] korean [j] japan

I think it just makes it cantonese evey time :(

[violet helm]

hmm that sucks :c try pointing it out more clearly. Even in the lyrics [mandarin accent chorus] the ai understands that something. Try multiple times pointing it out in the style. "mandarin, mandarin accent, mandarin pronounciation".

Just generate the lyrics in custom mode so you can use song sections and it doesnt make lyrics out of the style prompt

[crisp jewel]

I operate it in the custom mode. Before I pay, it defaults to Mandarin. After paying, it has always been Cantonese.

[jagged yoke]

Hey has anyone found a way to render a duplicate track, just with small minor changes - like instead of changing the entire song?

[crisp jewel]

I am a Premier Plan user. Logically speaking, I can generate 10 songs at one time, but now I can only generate 2 songs.

[final hazel]

it will always create unique songs.. so no

[crisp jewel]

Can only be regenerated or extended

[final hazel]

@jagged yoke Suno team is working on in-paint to change just minor things.. so in the future this will be possible. We just don't know when.

[crisp jewel]

OMG I'm going crazy, I spent $288 and I feel like I might as well not have paid it

Our company is planning to use Suno for commercial monetization. The songs generated now are all in Cantonese and we will not be able to proceed.

[final hazel]

I am sorry to see your problems. Pronunciation is a general problem with all languages. This is something that cannot be changed over night. It takes time. It has nothing to do if you have paid for a plan or not. One of the devs who is actively working on this Tony who actually speaks Chinese. And he always says it's more luck if it works.

[crisp jewel]

But the actual problem is that the songs I originally generated were all in Mandarin by default. After paying, they all turned into Cantonese. I was a little confused.

I have clearly marked the use of Mandarin, but the system still generates songs sung in Cantonese. I think there must be a problem somewhere.

[final hazel]

It's luck. It has nothing to do if you paid for it. You are still working with the exact same model.

[crisp jewel]

Yeah, so I think it's weird that it's going to delay our business process

[final hazel]

Everyone has this now and then. AI is unpredictable.

[crisp jewel]

I have contacted Suno's official team by email and hope they can solve the problem for me as soon as possible

[broken lotus]

Hi,

1. I remember this is an issue with dialects and also accents... I don't know a specific fix, maybe there are some tips in the language channels?
   https://discord.com/channels/1069381916492562582/1154772025940922438 <-- fixed link

#2. you can hit the button 10x in a row, but it is optional – only if you want to...

[maiden trench]

no its already super hard to even just get two singers lol

[unique stone]

Link for 1 is to an inaccessible channel btw

[crisp jewel]

OMG

This is actually the case. It seems that I misunderstood. I thought it was to generate 10 songs at one time. It turns out that 10 songs can be generated at the same time.

When I use Chinese to generate songs, sometimes the first sentence of the lyrics is always ignored. I also feel very strange.

[glacial scarab]

@crisp jewel btw. did you try to put "mandarin lyrics" in the style prompt already?

[broken lotus]

I have heard that Brazil and Portugal users sometimes start with a specific phrase or local spelling..., and that can work to change the way the bot pronounces words.

(I am not sure why the 1st lines get ignored sometimes..., that is probably an AI bug)

[crisp jewel]

[chorus mandarin]
[chorus - mandarin]
[chorus，mandarin]

all used，not working

Still pronounced in Cantonese

[glacial scarab]

oh no i meant in the bottom field where you also put genre

[red wadi]

Male Vocals

[unique stone]

I've seen this mentioned a few times so I've just made a suggestion to make the wording clearer 🙂 #💡┃suggestions-web message

[crisp jewel]

I've tried that one, it only affects the genre, not the lyrics and language, but I understand what you're saying, and it still doesn't work.

[red wadi]

https://cdn1.suno.ai/0e62c46e-d777-4d44-b387-8f874cf67f64.mp4

[glacial scarab]

ah that's sad, then it is probably not tagged separately in the Dataset and it is random how it will be pronounced

[jagged yoke]

a work around would be: use custom song generation - then use gpt to write the lyrics in your desired language - bring those lyrics in but make sure to use [brakets] for the format

[crisp jewel]

Yes, this is the normal Mandarin pronunciation. It was like this before I paid. After I paid, it all became Cantonese.

[crisp jewel]

[chorus mandarin]
[chorus - mandarin]
[chorus，mandarin]all used，not working
Still pronounced in Cantonese

[jagged yoke]

no that's not what I mean

bring in lyrics from gpt, and use suno just to generate the music.

[crisp jewel]

The lyrics are in simplified Chinese characters

It was normal during my trial period, but after becoming a paid user, it became Cantonese.

[red wadi]

https://cdn1.suno.ai/58cead8b-b556-4a03-bbae-354e8be139d5.mp4

[final hazel]

Again.. it has nothing to do that you paid.. And everybody is trying to give you things to try. We have to do that too. We also have that.. It's the same model. Nothing changed.

[crisp jewel]

I understand what you are saying, but there must be something that triggered the song I generated with simplified Chinese lyrics to change from Mandarin to Cantonese. I also only realized that it might have something to do with payment because I didn't do anything else.

I just hope the developers can fix this problem for me soon

[red wadi]

https://cdn1.suno.ai/121370c9-d629-4d86-9f92-80c0cfa0713c.mp4

https://cdn1.suno.ai/8109b364-f50c-47e2-8135-da59f996f6fe.mp4

https://cdn1.suno.ai/8109b364-f50c-47e2-8135-da59f996f6fe.mp4

[crisp jewel]

I'm so angry, why is yours normal and mine is Cantonese?

[balmy crown]

lately all of my prompts apparently secretly say "start out with the entire fuzzy digital chorus" so I'm wondering if it's my lyrical choices, my music styles (Though I've tried everything from symphonic metal to bouncy euro-pop) and I'm wondering if the music is being generated differently...?

I keep trying to guess what it is that I do that causes the AI to hallucinate lines/skip lines/repeat lines/sing with a thousand voices at once

this kind of stuff is really interesting, but I can only run 5 samples a day to try new things lol

[unique stone]

I don't know much about Cantonese/Mandarin, but as they are related I think the AI might be choosing whichever one it thinks is right. Are there some words or phrases that are used often in the language you want? You could try putting them in square brackets, maybe the AI will see these and take a hint to use the one you want.

So instead of writing [mandarin] put a phrase in there, using Chinese characters, making the AI see them but hopefully not sing them.

(This is just a guess though)

[final hazel]

Look back into history. People mention pronunciation problems way back last year 2023. This is not new.

[broken lotus]

next time (tomorrow? lol) try a 'simple' prompt, with like none (or almost none) genre/instrument description...

should give you a style with 2 or 3 words, and i'll bet they will be very clean...

Once I tried this, I now restrain how much I put in my style prompt... there is still a matter of weeding out problem words and finding good ones... but my vocals have improved – I barely hear any chorals, just normal solos and occaisional 2-voices in chorus kind of pop vocals.

[woven ridge]

I cant get it

[balmy crown]

I tried "symphonic metal", which I didn't think was that egregious. (I definitely am guilty of making longer stylistic prompts, but today I did a few very basic ones.) This isn't an issue I ran into in particular before, so I'm trying to figure out if it's literally something in my lyrics themselves that make the AI decide it needs to be sung with max voices

[woven ridge]

I cant get in

[balmy crown]

get in to what?

[woven ridge]

what next?

Icant see the how to chirp channel

[broken lotus]

everything has moved to website now:
https://www.suno.wiki/faq/getting-started/

[balmy crown]

ah, I've got nothing for that, sorry; weirdly, despite the fact that I'm here on discord, I'm just generating stuff in my browser using a different account. I just came to discord to see if anyone had tips on how to avoid some of the things I'd been running into with song generation

ah, looks like somebody else had an actual answer, good lol

[woven ridge]

ok... thanks!

[graceful condor]

im 100% not trying to steal work from musicians, right, so when i prompt "atb, tomorrowland" etc - i dont want to use any of their music, just their style - before doing this, is this correct? is that possible to do?

[lofty fjord]

I cant see the words of lyrics. Everything is black. Please help.

[graceful condor]

happened to me a lot - press CTRL + f5

[lofty fjord]

Thanks, It is working now.

[broken lotus]

the name may be blocked. Try genre and style descriptions.

[hollow hare]

Using existing artists or song titles will likely be ignored, if not censored. Very likely such information was stripped out from the training to avoid lawsuits (assuming if they were included in the first place).

You will need to use style keywords in custom mode to get as close to the genres you want.

E.g. you may want to try Ibiza Trance as one of the styles for ATB-like music (https://rateyourmusic.com/artist/atb)

Though I recommend looking at a specific song for more accurate styles. There are other sites as well to identify song genres like Sonoteller.ai and Chosic's music genre finder

[graceful condor]

just sucks ive spent.. 4,000 credits so far and got maybbbee 15 songs i liked out of it, the rest were just.. well most of the clips were really really bad quality, so i had to regenerate lol

[hollow hare]

Sadly that is how it can be, especially if a genre is diverse 😦

[lofty fjord]

How to generate lyrics in a particular language? I get an option of only English language.

[broken lotus]

switch on Custom mode and add your lyrics in your language (you might need to get ChatGPT to help write lyrics, since Suno lyric generator seems to make everything English)

you should also be able to prompt the style in your language.

[lofty fjord]

Okay. Thank you.

[graceful condor]

any way at all to get cleaner sound? theres quite alot of noise on 100% of the generated music'

[final hazel]

Please stop posting Chinese songs. You can use #🐣┃suno-showcase for that @red wadi

[jagged tree]

is there a list of languages that Suno knows somewhere?

[proven peak]

Hi everyone, I was watching this clip from Michael and I am so curious on how this was done. I think he used the extend feature but so far I've unsuccessful creating several genre styles and have them switch over. Anyone else? Thanks!
https://twitter.com/MichaelCarychao/status/1768988419462672597

[haughty flare]

Hi folks, is there a feature available to edit the lyrics after making a song? I fudge a lot of the extensions on my tracks and as a result the lyrics displayed on a finished track are all over the place and incorrect. Is there a way I can ‘clean’ this up so it accurately reflects what the song is actually singing?

[broken lotus]

this is an old language list from v1
https://suno-ai.notion.site/Supported-Languages-16550b00a3f04ee6bab541d135eaf713

I think there are more languages now, but I don't know a source that lists them

[broken lotus]

the feature is being called 'in-painting' by the devs (after image gen)
Probably a feature for v4 (in development, no eta)

[haughty flare]

Ahh superb, thank you for the reply! Been desperate for this feature

[broken lotus]

see the techniques described at the bottom of this page, especially 'sudden style change'
https://www.suno.wiki/faq/making-music/extend-from-time/

[autumn heron]

I just now have bought a subscription, and i want to know. Is there a difference between a free and paid version, except more credits and 10 running jobs at once?

[broken lotus]

paid subs get a commercial license and all rights to the clips in perpetuity... paid subs can upload on spotify, Apple Music, sell on Bandcamp, iTunes, and use in professiona projects (commercials, animations, videogames)....

the songs keep their license even if you unsubscribe later

but the features and the AI are exactly the same.

(and sometimes free use is suspended when the servers are overloaded, but that hasn't happened since they moved to a new provider))

[autumn heron]

Thanks for answer

[proven peak]

oh great thank you!

[sharp current]

Ask for developers. Can you implement track ratings not only + and -? Add please emotional responses. Like funny, real, great, bad, cry, disgusting. well and other categories.

[autumn heron]

I personally tried, an failed 😅

[autumn heron]

If i understood you right

[rotund jay]

Can the devs do anything with the hardbass? I managed to generate something like hardbass once, and when I tried to do it again with the same prompt, I never got the same result. It was very disappointing and didn't sound like hardbass at all. I had to spend 600 credits to get the same result as the music below, but I never got a similar result. The donk bass is non-existent in all the generations I've tried to make after this music. And yet it's the very essence of hardbass

https://suno.com/song/1c48fc9f-1065-4b1f-94b8-52eeb2accbe0

[regal ginkgo]

where can i create a song?

[broken lotus]

https://Suno.com

[tight berry]

Sorry for the late response. Could you give me an example prompt cause when I tried this it didn’t seem to work. Thanks again for your reply!!

[final hazel]

Go back to my message, and click on the link of the song. It will give you my prompt exactly.

[lone oak]

Why are rock songs always created without an intro?

[tight berry]

Alr! Thank you for your response :)

[lone oak]

And in addition to English, the first sentence of other languages ​​always has errors, which is really crazy!

[unkempt viper]

Why after I extend a song does the extension only go on for like 10 seconds? i want a proper extension adding like an extra minute or two, how would I do this??

[graceful condor]

i cant create songs no more can someone help me???

[lone oak]

It takes 2 minutes to create the first paragraph with long enough lyrics, but the officially designed extended second paragraph only takes 1 minute, so it is really difficult to generate a song with lyrics that lasts longer than 3 minutes.

[lone oak]

It takes 2 minutes to create the first paragraph with long enough lyrics, but the officially designed extended second paragraph only takes 1 minute, so it is really difficult to generate a song with lyrics that lasts longer than 3 minutes.

[unkempt viper]

im trying to extend an instrumental, is there a way to make it around a minute more without lyrics?

[primal nymph]

How do you create a song and fade it out in the end or make a powerful end, not just sudden stop?

[unkempt viper]

click extend

[final hazel]

You need credits.. Do you still have those?

[primal nymph]

I have tried, then I can extend 60 sec, but end is still sudden

[unkempt viper]

mine just makes an instant stop

[primal nymph]

same here

[unkempt viper]

no i mean

when i generate more

it doesnt have length

it just stops

[graceful condor]

no

[unkempt viper]

it makes around 10 seconds of an outro

[lone oak]

The official obviously underestimated the number of bugs outside of English. The second paragraph can take up to 1 minute. According to tests, the success probability of generating a complete song over 4 minutes in some languages is very low. The official should provide a method to correct the lyrics bug while retaining the sound and melody.

[final hazel]

Then there is your answer.... You will get 50 credits every day.. So you can try again tomorrow

[primal nymph]

Mine doesnt make an outro, how can that be added?

[lone oak]

For songs with lyrics, you can get a good melody by trying at most ten times, but the lyrics are prone to bugs.

[final hazel]

https://youtu.be/lj8utBd_IUM might be able to help you guys @primal nymph @unkempt viper

[lone oak]

suno wiki knows all the details

But songs created by rock usually have no intro, and the first line of the lyrics in other languages will be wrong. The lyrics in the song I generated today have been wrong no less than 300 times, which is really outrageous.

There are many songs with good melody and sound, but the lyrics are wrong from the beginning, which is terrible.

If the official cannot reduce the bug, it is recommended to generate only one song at a time with lyrics, otherwise the success rate will only be in single digits.

[final hazel]

My video shows how to create that intro

[lone oak]

I understand, but when rock is generated, it will not be executed according to the lyrics settings.

Tried a lot of rock genres, with or without [Instrumental Build] to no avail, the intro is all about luck

[final hazel]

I am providing solutions, so that you do get an intro. AI is unpredictable, and if metatags do not work, you will have to find methods that get the result you want.

So if you want that intro? My methods works.. 99,99% of the time.

[neat field]

add random punctiation around the [introduction] tag like this

..
..
..>..>>
..
..``` seems to help some of my gens

[primal nymph]

is [bridge] a meta tag?

[neat field]

i seent it work many times, not 100% success ratee

[lone oak]

Currently, the songs created by rock have no intro, and the first lyrics are prone to errors. I browsed the rock songs on the official website and found that Asian users have such bugs. The probability of errors in the first lyrics here is nearly 9/10. This is Two bugs are really sad. Please help me.

[primal nymph]

is there a list over these meta tags?

[final hazel]

As said.. my video provides a solution. You are expecting full control. Suno is not there yet. This is not going to be solved in the current model. I am an offical moderator and guide here. I know what I am talking about. 😐

[broken lotus]

a primer on metatags:
https://www.suno.wiki/faq/metatags/

[balmy crown]

what does it mean when it says "Couldn't generate that: prompt flagged for moderation"
like, if it's a song i've generated 7 times before, why is it suddenly now a problem? And I'm just trying to get it to actually sing the last lyrics, so they aren't even NEW lyrics, they were just ones cut off...? If this is a glitch, is there a way around it?

[hollow hare]

Most likely got censored from ChatGPT not liking something. Ask chatGPT itself about "what was not appropriate about the [following content]" and it should tell you.

Usual causes are using something copyrighted and swear words

[balmy crown]

that's so weird. it's definitely not copyrighted, because I wrote it, and there weren't any bad words, either - at that point, I would understand the response

I'll go ahead and drag up chat gpt and see if it thinks it can explain it to me?

but again, these are normal, uncopyrighted lyrics, with no bad words, and i've made other versions of this song with no issue

so I'm wondering what's flagging it NOW

[hollow hare]

ChatGPT can be weird. I remember someone mentioning that bones and ribcage got censored in their lyrics

[balmy crown]

I've definitely had the word bones in one of my songs lol

[unique stone]

Sometimes ChatGPT doesn't like the intent rather than the words

[balmy crown]

but -- why would it let it through before, but not now?

[lone oak]

I've calmed down. The tags used before were all from suno wiki. Is this official? The style tag seems to be only partially displayed

[unique stone]

I'm not sure if ChatGPT's behaviour/understanding of stuff evolves over time as people talk to it. Never really looked into that but I have no idea :/

[broken lotus]

if you think it's a glitch it is probably a glitch.

if the words cleared before, but are now blocked... and this has happened to me but not recently, I add blank spaces before the first lines, and possibly a [Break] at the end

you want to preserve the beginning of the song for Suno to sing, but change the way ChatGPT is tokenizing the text...

usually (for me) just a several blank spaces work, but you can also edit the end of that lyric with something that can't be sung

[balmy crown]

you're a genius, thank you Orz

I'm gonna try that

[lone oak]

Are there any friends in the chat room who have created more rock songs? Hope to communicate and learn from each other

[balmy crown]

alas, no

[balmy crown]

when it says flagged for moderation -- does it ever get reviewed by a human?

alas

[broken lotus]

rock not really my genre...

[broken lotus]

no, no human will see it (that I know of)

[lone oak]

I'm going to cry

[balmy crown]

if you want to do rock songs, you can also try Udio? it gives me actual intros for rock songs

[lone oak]

But I still have 6,000 points, and they will be gone by the end of the month.

😓

[broken lotus]

Did you understand 3Daizy's advice?
She showed you a video how to build on an instrumental opening clip.

[lone oak]

I know

[final hazel]

https://tenor.com/view/linasanad-daisy-flower-romashka-lina-gif-18760010

[balmy crown]

ah, that would make a difference. you could have a very short set of words, then say [guitar solo] or something like that, followed by the actual song content, and then cut out the extra words at the front manually? sorry if this seems overcomplicated

[lone oak]

I understand how to use suno. I'm not a new user. It's just that the bug only occurs in the rock genre. Even if the instrument introduction tag is added to the lyrics, it is difficult to really take effect.

Test results show that only some rock genres create intros

[broken lotus]

Yeah, We heard you the 1st 8 times
DO YOU WANT HELP
because we all have better things to do Babe.

If you can't take advice... there is nothing we can do for you.
No one is going to your house to do it a sensible RELIABLE way.

[balmy crown]

well, chat gpt told me which line was probably triggering something, so I went in and just changed the words to... sounds... that sounded like the words... and it worked? now I feel like I've done something evil to slip past a rule, but the line really wasn't something that I think a human would've censored

[broken lotus]

When I make rock I get a little intro 🤷‍♀️
https://suno.com/song/86aa6e31-0bea-4f11-92f8-37dbdfae9bbc

[balmy crown]

(thanks to everyone that helped me!)

[broken lotus]

if it works it works. 😉

[lone oak]

Yes, I know that the method you gave requires 2 minutes of the first paragraph to create an intro of more than ten seconds, but the extensions except for the first paragraph are all 1 minute. Suppose you want to make a 3-minute song with lyrics. When it comes to the second extension Sometimes, the music style will deviate.

[broken lotus]

Weeeell Weeeeelll that goal post sure did move somewhere else... the moment I started posting examples 😏

I think you've wasted enough of my time.

I'm going back to making music.

[final hazel]

6000 credits left? Then keep burning them.. Still plenty to burn it on a intro.. Just stop whining if it is not working. lol

[lone oak]

You got the question completely wrong

[final hazel]

Mine don't deviate.. but you are not willing to listen or try.. I give up..

[unique stone]

AI does do a derp sometimes. The moderation seems to be there to weed out lazy attempts to e.g. insult, bully, threaten or otherwise deliberately cause offence. I mean, it's a tool that generates music at the click of a button and some folk are just unpleasant! 😄

So when it's getting in the way of you trying to do something actually artistic/creative with it, it's reasonable to try to fool it. ChatGPT police won't knock on your door in the middle of the night demanding answers 😉

[lone oak]

I have explained that the lyrics with bugs are in songs other than English, but you still gave an English song to prove that there is no problem, and I also mentioned some types of rock. Even if the lyrics indicate that instrument creation does not work, you But thought I was talking about all types of rock.

https://suno.com/song/a92b8380-f365-4ae4-9205-30b7e3db36c0

You think there is nothing wrong with this song, but the first line of the song is wrong

I'm here to report bugs, but you think users don't know how to use it. Think about it calmly, is suno a complex tool? How could you have such a funny idea?

[final hazel]

I told you this is not going to be solved in this model.. And this is not the bug channel..

You just don't want to listen. 🤷‍♀️

[lone oak]

OK THX

[daring void]

Hello, any option to download the voice track and chord progression from a Suno AI created song? to produce and finalize in my studio? Thank you!

[final hazel]

You can only download it as MP3 and MP4 as a whole. Any suggestions can be posted here: #💡┃suggestions-web

[daring void]

ok, thank you!

[sudden garden]

how do i switch male voice to female voice when extending song?

[final hazel]

https://www.suno.wiki/faq/style-and-lyrics/duets/ has information and you can use mine as example to learn from.. this is a duet I was working on: https://app.suno.ai/song/5045e5f8-d213-4cac-9d87-99b370e89adb/

[knotty bronze]

Hi, Nice to join this discord 🙂
Q: Is it possible to generate only vocal songs, without instruments/music? I've tried "Vocal Only" in the music style, but it didn't work.

[broken lotus]

'a cappella' works, but can create a doo-wop song
'spoken word' also works

[knotty bronze]

Thanks, I'm still experimenting with, but I've run out of credit 😦 I'll try more later 🙂

[crystal sable]

Currently encountered website bugs:

• Create button gets transformed into extend but there's no extend display to clear the "extending from" song away, may have to do with multiple Suno windows, clearing cookies temporarily resolves this new issue

[kindred prawn]

Is there ways to edit the lyrics on a completed track. The system keeps putting in incorrect lyrics or merging lyrics from multiple parts together creating incorrect lyrics

[analog jungle]

https://suno.com/song/9592df21-6c5e-4f69-95d9-8655c256681a

[elder parcel]

Working on a new song tonight, and I just had a great example of what I was talking about in my post over here #💡┃suggestions-web message

Short version is I get A LOT of songs where it just wants to give me this high pitched schreeching over and over again. I only use V3, haven't tried V2 yet. I absolutely love this AI, but my primary frustration is that it currently REALLY seem to like locking in on high pitch tones. Not even just in this example provided below, where it's just terrible, it also likes using high pitches in melodies and harmonies way too much in general.

Here's the song where it locks into the high pitch. Normally, I just delete these to clean up my library, but I wanted to keep one as an example

https://suno.com/song/9ba86c91-f258-44e4-9a8b-bc38012ff2bb

[analog jungle]

😆

[eager topaz]

Not good! Hurts my ears. There is also a low frequency noise generated. I get both in some genres and try everything to try and get rid of it. Hopefully Suno will improve.

[mighty pier]

Hi everyone! Is it possible to edit the lyrics in an already generated song. I will use the same style same voice with different lyrics.

[hollow hare]

That is planned for version 4 with in-painting. For now, what you can try to do is Extend and hope it picks up what you want (aim for extending around when the section you want starts or ends)

Check out the Wiki links related to this topic: https://www.suno.wiki/faq/making-music/how-do-i-make-the-song-longer/

Or Daizy's video if you haven't seen it: https://youtu.be/lj8utBd_IUM

[analog jungle]

https://suno.com/song/b2137923-df97-4753-a844-e8f02044a972

https://suno.com/song/fe3409ba-f43c-4fd2-b8f8-440d56969731

[acoustic bough]

Why does it say "prompt flagged for moderation"? I haven't changed the prompt from before when it worked, and I have adjusted the lyrics about 20 times already.
Would be helpful if it gave any reasoning.

[hollow hare]

Got censored from ChatGPT not liking something. Ask OpenAI's ChatGPT itself about "what was not appropriate about the [following content]" and it should tell you.

Usual causes are using something copyrighted and swear words. Even if it worked before, it could be a small change caused its interpretation to become different.

Worst case scenario, you can phonetically spell the word it has an issue with, though mind the ToS of Suno about what is and is not allowed (e.g. no copyrighted song lyrics)

[acoustic bough]

I get the AI is sensitive, but some genres should allow for more controversial lyrics.

I already asked ChatGPT, and it gave me lyrics with the same word in it, and it still does not work.
I've written songs with more explicit or more depict imagery, and they were created.

This is the first time it just won't let me continue the song.

[graceful condor]

stop using chatgpt and start using openai over your own system, it´s free and not illegal or smth im using llama3 right now and it´s the best multilanguage model i ever tried, its compact and even understands sarcasm

[waxen topaz]

帮我写一首赞美春天的歌曲，用中文写

[acoustic bough]

It's about what Suno uses to check the lyrics, if I understand correctly.

[graceful condor]

ahhh so you mean you inserted some lyrics and it was flagged?

that would be because of some swearing words (fuck not included) and some song lyrics from famous songs are not able to be generated too, as far as i know. But when you change the text to make it just similar to original you can still use it

[acoustic bough]

To wrap them up, it's about a worm that's eating all the roots so the trees all die and someone tries to find it and when he can't, he gets a digger, drivers over it so it gets smashed.

I can write them metaphorically or as vivid imagery, it doesn't matter. The AI won't let me kill the worm.

[graceful condor]

maybe try to write "he passed away" or "he dissappeared"

[acoustic bough]

Somehow killing it is not okay. I've written far more controversial lyrics and they all get through.
It should be possible to write lyrics in the style of bands like slipknot or trash or death metal, as these songs won't even appear on the new or trending pages!

[acoustic bough]

Wouldn't fit the song and the rest of the lyrics.

[graceful condor]

i had this problem too, but sometimes it be like that, you are right death metal bands can´t really be copied here, but i think thats somehow good cause they stay really special

[acoustic bough]

I was just aiming for something like Mastodon. They would write songs like that.

[graceful condor]

i was just having fun with fame songs at the start too, but i really got into suno when starting to write my own textst even if there were many many bad ones, i found some gold too haha

[low stream]

I know about using the [verse] [chorus] ect tags to help keep the words in order, has anyone found something else that also helps to keep things in order and or not repeat lyrics?

i know you can extend the song before the lyrics repeat

[graceful condor]

[verse] etc. are called meta tags, there are way more than just the basic ones like: [Breakdown] or you can add adverbs like [Happy Ending]

[low stream]

oh meta tags, thanks you! I figured there had to be a word for them. perfect! That is verfy helpful, I didn't think to look for more meta tags.

[graceful condor]

You are very welcome my friend

me neither just discovered it few days ago haha

[low stream]

nice!

[steel night]

[male female voice] not working in Style Tag when extending this song. Any reason for this? I tried out many credits on this 🥸
https://suno.com/song/f82de541-0e54-4cc2-acf8-69b06695603f

[low stream]

what are you trying to do?

@steel night

[jagged gulch]

I write male voice in the style of music and the result is a female voice... why

[low stream]

how did you write it?

what was the song style?

[graceful condor]

depending, how many things did you write in the style, if you use a lot of different ones, sometimes the ai only picks out a few

[jagged gulch]

I wrote "1940s, romantic, swing big band, strong, emotional, slow, male voice"

[graceful condor]

Try "Male vocals"

[graceful condor]

when i used vocals it was always working for me, cause it can´t be that theres too much

[low stream]

put male vocals first, so it will prioritize them

[acoustic bough]

I write most of the lyrics myself and I do not specifically aim for a certain band's style 🙂

[steel night]

I tried to extend the song but this time with a [female voice]. After that I wanted to continue a duet of both. No way 😔

[low stream]

oh gotcha. What I Would do is put in the style of music at the bottom "female vocals" then also use the meta tag [Female voice] see if that works

[jagged gulch]

thank you

[low stream]

welcome!

[steel night]

I’ll try. Thx 🙏

[low stream]

let me know if it works

welcome

[jagged gulch]

yes it sounds perfect now https://suno.com/song/b12713d2-1e1a-44c4-8849-94e2f76dbfb4

[untold oak]

I was pleasantly surprised when adding "[vocal solo]" to the beginning of a line actually worked. Suno is a pretty smart cookie

[summer dome]

please someone put paypal on the subscription option! I really need it since last week! I want to create music!

[cursive rapids]

buy a prepaid visa or use zipmoney if you have it but yeah paypal be great also 🙂

[steel night]

I tried it many ways. As you said. But always I got a [male voice] when extending 😤

[low stream]

damn... for style try putting "blue grass, female vocals, female vocalist, duet," without the quotes in it

[lusty violet]

you willhave to experiment with genresspaces

blue grass may be coinsidentally overfit

for male voices

[acoustic bough]

Finally got it right.

It's definitely not suited for kid's ears

https://suno.com/song/99114b73-b673-4b36-b537-34ee5961abec

I don't listen to this genre personally, but I like to try em all out with Suno to see what's possible

[graceful condor]

same

[steel night]

May be. A new Bluegrass is producing the [female voice] but cannot switch between gender 😁

[lusty violet]

maybe like folk alternaitve or something

[graceful condor]

the lyrics are good, but i feel like the distortion sounds are really heavy and the ending is too long. Overall still a nice try keep up the good work!

[lusty violet]

not sure rly

folk probebaly has lots more women than bluegrass

[lusty violet]

maybe rhythm and blues?

[low stream]

hmm can you try putting duet in the first rendering of the song? like male female duet and then put [male] for the male parts and [female] for the female parts?

[acoustic bough]

Yeah but the first genre prompt I set was sludge metal and this is definitely common in sludge, although I don't like it very much

Most songs by bands like Mastodon can reach from 2 minute slaughterfest to 7 minute doom metal snoozefest

So I consider 3 minutes alright

Overall it's really hard to make a well rounded track if/as you're not making it yourself

[steel night]

I tried. Not working. That was my last try as my credits melt like Ice in the Sunshine 😁

[acoustic bough]

Also I tried to not use metatags this time, I wanted to see how Suno went with the genre

[low stream]

I am out of credit too, just finished the wong I was working on with my last credit LOL

[acoustic bough]

Just landed smoothly at 0 token airport too
Now we gotta pay or patiently wait 🙃

[low stream]

yup lol want to hear my newest song?

[hoary mango]

I have the same problem, my second Verse ist different to the first. Is there any soulution with a special tag, that they sound same?

[knotty bronze]

What time does the daily credit renew? Which timezone?

[acoustic bough]

Post it

[thick berry]

hi im new in using suno but, can you use it infinitely? or there are cooldowns such as 5 songs per day or something? if no why cant i create a song even though it isnt flagged?

[final hazel]

As a free user you get 50 credits per day. Which equals 5 x hitting the create button.

Paid users get either 2500 or 10.000 per month (depending on the plan you choose).

[knotty bronze]

What time in GMT do the daily credits renew?

[final hazel]

Free daily credits refresh at approximately 10pm EST, 3am UTC.

You'll have to calculate GMT yourself 😛

[prisma wraith]

UTC and GMT are the same luckily

[knotty bronze]

Thanks for the info 🙂 However, my credits have not renewed today at 3am UTC. What could be the reason?

[final hazel]

Maybe it's not 24 hours ago for you?

I am a moderator and cannot see beyond Discord 🤷‍♀️

[knotty bronze]

Yes, you're right. I've registered less than 24 hours ago. But I thought it the credits will be reset either way! I didn't know it's that strict 😦

[final hazel]

Just wait a day longer, or consider paying for a plan (it's just 10 dollars for 1 month if you don't take the year plan). I had so many people who were disappointed because the made a banger with their free credits and then realized they don't have full rights since it was made on the free plan. 😉

[runic dock]

I am getting more and more, suno starting singing with the music. Usually i add intro in the lyrics and in the style. Now suno ignore them and as we dont have inpainting feature i send all the clips in the trash. Is there any reliable workaround. Maybe start with an instrumental first ? @final hazel

[unique stone]

It's always worth starting with an instrumental. I find this helps to establish the foundation of the song at least. You can then extend at the point where you want vocals to start.

[final hazel]

I actually address that in my video https://youtu.be/lj8utBd_IUM
But it is what @unique stone says.. Missing intros are very genre-specific, and starting with an instrumental is a reliable workaround.

[rotund geyser]

Please be kind to me because I haven’t understood the policy properly. Please, if I make a song on Suno (paid for), can I remake the song with my voice and release it on my album or can I reproduce the beats and use the lyrics to make a song and release as my song?

[chrome mesa]

How can I add like metal scream to the song?

[lunar sundial]

I forget how to use this to create a song.
Please help.

[lyric sonnet]

to create a song please go to the suno webpage at suno.com

[inner steeple]

didn't see if it got suggested but you could try [Verse, female vocal] and continue with your verse directly after the tag, don't put new tags between the lyrics

[foggy jasper]

I want to immitate this style, but I haven't been able to figure it out yet. I got one track that sounded great, but since then I haven't gotten anything like that, even though I used the same prompt.

https://youtu.be/G5oIUsPkigQ
https://suno.com/song/899eb533-4960-44c3-aa4a-40e6d2679af5

[broken lotus]

that's a great homage to Carpenter!

[foggy jasper]

the master of synth

(imo)

but thank you. I have been working from scratch to make beats similar to his style ever since I started producing. I was just curious if there was a certain prompt I am missing and if Suno could make it. It would be such a trip to listen to nonstop Carpenter-like music, completely original.

[broken lotus]

I was hoping you had some tips...

what I would add is 'sparse'

[foggy jasper]

muffled Fender guitar (oh & Linn LM-1 drum machine). That's all I could find for now aside from the obvious (analog synth, 1985, vintage)

[broken lotus]

(reading your prompt that worked the 1 time)

[foggy jasper]

I will give it a shot

[broken lotus]

Interested in what you come up with...

[foggy jasper]

thanks for being a chad🫡

[junior lava]

Is it possible to generate only one song instead of two?

[snow maple]

i made some songs before i bought a subscription does that mean i cant use the old songs that i made before i bought a subscription

[broken lotus]

you should remake the songs, so you own them.
afaik Suno does not change free songs into paid songs.

[thick berry]

Noted thx for the help

[tacit mauve]

Hello guys, maybe exist a prompt guide for suno? I feel sometimes that it don't read properly the prompt and make songs in the same style

[neat nexus]

You could try the unofficial community wiki? That's at: https://suno.wiki

[west oar]

I havent found a good solution yet. My compromise is the follow structure: verse 1, verse 2, chorus. So that I get 2 verses that sound the same and end the song shortly after chorus.

[broken lotus]

🤔 I need a prompt word with the mood of 'suspenseful' but more energy...

[neat nexus]

thriller?

(Just off the top of my head.)

ChatGPT suggests:

Anticipating
Electric
Thrilling
Charged
Exhilarating
Vibrant
Dynamic
Pulsating
Stimulating
Impending
Invigorating

[mild yew]

I spent some credits for you, after some generations this the best result I came out with. I used https://sonoteller.ai to help me identify genres and moods

https://suno.com/song/495de7d5-d3f7-40c7-bccd-f55889d4620c

I love Carpenter 😍

Admittedly, mine sounds a lot more "retrowavy" and less bright in tone than the original

[gentle willow]

Hi guys, i'm new here, i just using suno ai, can anyone help me about the style of music ? , i already make some but still kinda confuse on using style of music prompt ? can anyone help me correcting my style of music prompt ? 😄

[mild yew]

it's difficult to suno "tell me you want John Carpenter without telling me you want John Carpenter" 😄

[broken lotus]

sure we will give opinions (aka advice)

[gentle willow]

Japanese symphonic metal , heavy guitar riff, melodic electric guitar solo, powerful , tenor and emotional male vocal

That's are my current style of music, but i think i have a mistakes cuz it doesn't sound like i want

I'm kinda confused on the style of music, :/

On the lyrics too, will send here , wait a sec

here's the lyrics (ignore the japanese lyrics xD)

i put the song structure on the lyrics textfield, am i mistaken ??

[Start]
[Intro]

[Verse 1]
時が薄れて　夢の中
孤独なる庭に　佇む
蒼き月が照らす　夜空
運命の舞台　始まる

[Instrumental Interlude]

[Verse 2]
朽ち果てる時の　中で
永遠の誓いを　捧げよう
美しき哀しみ　抱いて
荒れ果てた世界　彩る

[Pre-Chorus]
風がさらう　愛の調べ
闇に舞う　刹那の花

[Chorus]
華麗なる　貴族の交響曲
誇りを持って　舞い踊れ
紅く染まる　地平線に
我らが旗を　掲げよう

[Guitar Solo]

[Pre-Chorus]
風がさらう　愛の調べ
闇に舞う　刹那の花

[Melodic Interlude]

[Chorus]
華麗なる　貴族の交響曲
誇りを持って　舞い踊れ
紅く染まる　地平線に
我らが旗を　掲げよう

[Instrumental Interlude]
[Fade to End]

[neat nexus]

I personally never bother with numbering the verses. Not sure if that would help or not.

[gentle willow]

sorry a lot of question, i just use suno ai for likely 2 hours xD

[neat nexus]

Have you tried using Japanese words for the style too to see if it influences things? I found that it seemed to do so when I did it.

Not sure if this would suggest anything to you but here is one that I played with before:

[Removed links]

[broken lotus]

I would take out anything (my opinion) that seems redundant: Metal is all guitar riffs and guitar solos... you don't need to add instruments that should already be there.

but what you might do is describe the playing like
virtuoso electric guitar
or
stadium guitars

I would similarly shorten the singer's description (if even necessary with a metal genre)... see what the genre gives you by default before you add 'corrections'

[gentle willow]

uhm not yet, yeah i kinda confused, i wanna create a japanese symphonic metal like versailles , moi dix mois, malice mizer, and gazette, might try that later

[gentle willow]

Ohh i see, so i need to put the 'bigger picture' like what's inside on metal is heavy riff and guitar solo ?

[neat nexus]

One thing that I have found helpful when trying to get a specific sound is to try interrogating a GPT about the music, playing styles and vocal styles of an artist and their song. It can offer suggestions like the ones that wetcircuit gave above.

[gentle willow]

yeah i already do some , here's some example

i wanna recreate kamijo's singing style
here's GPT answ

Kamijo's vocal style primarily falls into the category of tenor, which is one of the higher male vocal ranges. Within this range, he often employs techniques characteristic of operatic singing, such as vibrato and dramatic expression.

also the versailles musical style

Versailles's music style, particularly in songs like "Masquerade," can be described as a fusion of symphonic metal, visual kei, and neo-classical elements. The band combines intricate guitar work with orchestral arrangements, creating a grand and theatrical sound.

[neat nexus]

Good hint.

[broken lotus]

some of the 'expressiveness' in the singing can be added to the metatags

[dispassionate verse]

[emotional chorus]

so it varies

[gentle willow]

yeah but i still confused what to put on the prompt.. is it ' Tenor, Operatic singing dramatix expression male singing ' ?

[neat nexus]

You should also look into genre parent genres to see if a mix of those will draw out something closer to the sound you're looking for. Think of it as painting. You can mix styles and genres to get child genres sometimes. It takes experimentation and patience.

[broken lotus]

simplify:
operatic tenor

[neat nexus]

You only have so many characters in the style prompt. So yes, simplify.

[gentle willow]

i see

[broken lotus]

the AI seems to loose track of the prompt if it is more than 3 words... (it is very specialized, and maybe not as smart as some other language models)

so I avoid long descriptions...

[gentle willow]

oh, about the musical style

is 'Japanese Symphonic Metal' is already enough yet ?

[broken lotus]

I would try just that, tbh

[gentle willow]

simplify yet big picture of what i want

thanks for the advice guys, it's really help me

[broken lotus]

yeah.
the AI is good at bringing in all the stuff that goes with the big picture... sound quality starts to suffer when we give it too many things to try to add in... sometimes it's better to let the AI breathe... (hard lesson to learn because I want to control it all)

[gentle willow]

thanks guys, it's really help me a lot!!!

[neat nexus]

日本のシンフォニックメタル

According to Google Translate: Japanese symphonic metal

[gentle willow]

thank youuu

[neat nexus]

Someone I used to be friends with online once described it as a kind of taxidermy. The skin is only so big and if you try to shove in bones too large they'll rupture everything, so the trick is to figure out how to give it just enough to go where you want to go and stand back letting it work.

Obviously I've not absorbed this lesson myself as well as I could..

[frozen gulch]

Hello fellas, I'm very new to the technology and I'm trying to learn

I have a question, is there a way to get a female singer? I'd been trying to make a rock song with a female lead but I can only get male voices

[final hazel]

In the "style of music" you can add something like female vocals or any other descripting depending on what you need.

[marsh agate]

I recommend adding her tenor of vocals if you want a deep women's voice. (I always use Contralto) and it seems to land.

[frozen gulch]

i did try but the result was still a male voice

[frozen gulch]

what do you mean ?

[final hazel]

What genre are you trying to create?

[marsh agate]

It's not perfect sometimes the Ai will ignore, but if you add to style and first tag before vocals. I have usually gotten it

[marsh agate]

Pitch. Women have 3 vocal ranges in music. I prefer the lower tenor voice. (pitch)

[final hazel]

If you are going for a genre that is dominated by males, you sometimes have to force it by repeating it a second time.

[frozen gulch]

rock, or nu metal

[final hazel]

Yes, that's male dominated. Just force it.. 😉

[snow maple]

what if you extend does it also give you copyright @broken lotus

[obtuse ridge]

When you love the song but the time limit cuts it Off and the Extension just doesn't fit. Right within the Outro its stops, three failed Extensions so far. I love Suno, but this two minute limit and Extension fiddling just sours the experience a bit: https://suno.com/song/b8a6bf9d-56ad-49f6-8c61-045d52bc6a7e

[elfin bay]

I’m having a huge issue with extensions creating new songs and also lowering volume randomly in middle of songs. So crazy

[final hazel]

Do you mean from a song that is particularly made with free credits? Nope, that's cheating.

[obtuse ridge]

Indeed. I think it stops nicely but then they whisper on or start again and whatnot, so awful.

[snow maple]

I have comercial use now but that song was made before I had the subscription so can I extend it right now with the subscription and gain Commercial use for the Extended song @final hazel

[final hazel]

Your song needs te be created with paid credits in full. So I personally advice you to send an email to [e-mail]

[snow maple]

ok but i think if i extend it, it will cost credits so does i mean i own it?

[final hazel]

No, because that means you are cheating. If everybody makes great songs, but just replaced the last few seconds with paid credits, does not give you copyright. Else it would be a loophole to abuse the system.

[slender trellis]

How does one know if a song falls under the subscription copyright? Specifically, is there a 'get info' command or similar that will tell you?

[final hazel]

You can see which date it's created

And on your account you can see when you joined

[slender trellis]

Got it. Thanks.

[gentle willow]

hey @broken lotus @neat nexus , i'm using advice from you guys, and now it's sounds better 😄 , thanks a lot!! really appreciate you both guys

it's getting closer to what i want 😄

Ohh, i have a question again guys sorry hehe
Can i edit the created songs? or i just can extend it ? , i didnt found command to edit created songs

like i just want to edit some part of the songs

[graceful condor]

is possible to get better quality of these songs or to get the stems ? I tried to extract the stems but it is not separating them very well

[broken lotus]

you can Extend from Time (over-write from a timestamp)

but we can't 'spot edit' yet... Devs are calling that 'in-painting' it's planned for v4...

[gentle willow]

ohh i see, thanks for the insight

[solar valley]

Hello there, I am dipping my toes into BARK, suno. I want to finetune it and make it support Bengali how would i go about doing it?

[broken lotus]

The quality from the AI varies, even within genres... but some genres have much cleaner voices and instruments than others...
😕

[graceful condor]

do you have some hints ? you think electronic music sounds better than clasical music or real instruments ones ?

[broken lotus]

'soul' voices are usually very clean (RnB, Electro)

'Metal' and 'rock' voices are usually very reverbed and occupying the same frequencies as guitars... makes them really hard to extract...

'singer-songwriter' is very clean too, and usually not much instrumentation...

do you wan the voice or the instruments?

[graceful condor]

do u think that will be possible to extract stems directly from suno ? sometimes i need just instruments but sometimes just vocal .btw its possible to use these songs or vocals for comercial use ? to publish them on spotify ?

I love suno it is very good to give ideas (im a music producer and i use it to get ideas and it is helping a lot )

[broken lotus]

music made with paid subscription you get commercial license and all rights. (Free has non-commercial license and must give credit to Suno)

Suno knows people want stems, but they've said the AI writes everything as 1 file (so demux is the only way) and they have held off – presumeably until the AI model has a higher quality output...

[gentle willow]

ahh, so when i want edit the same song, i go to extend -> and then i set the extend time to 00:00 ?? @broken lotus

[graceful condor]

do you use real vsts to process the quality ,mix and mastering ? maybe using some vst or real hardwares trough cloud for rendering can help im really interested to know how suno write vocals ,melody (elements of a track )

[graceful condor]

i think 02:00

[broken lotus]

if you just want to extend from the end, leave the timestamp as is (it will be the end)
if you want to extend from a specific time, set the timestamp maybe 1sec ahead.

if you set it to 00:01, you would essentially be creating a whole new song, Extend will only 'hear' 1 second...

[lyric sonnet]

does it actually make a difference if you extend part 2 or 'get whole song' then extend that?

[gentle willow]

okay thanks again wetcircuit

[hasty silo]

if any youtube channels stole our song and add to their channel what can we do? report to suno or youtube?

[broken lotus]

if you are extending from time on a part 2, like 00:32, the extend clip can only hear those 32 seconds....

but if you get whole song first, Suno can hear more of the song, and should match the earlier parts better (it will also hear some of Part 1)

(there is an upper limit how much it can hear, probably about 1min)

[broken lotus]

to youtube.

[broken lotus]

definitely report... 3 content strikes and Youtube removes their channel.

[gentle willow]

if i extends the song, should i use the same lyrics prompt , or change the prompt to new structure ?

[broken lotus]

pick up where it left off.
you can also use Extend from Time to pick a clean spot in the lyrics.

[gentle willow]

okay, thanks a lot wetcircuit, sorry if questioning a lot

[hasty silo]

Thank you. How can I prove to YouTube that the song belongs to me? It is said that your channel may be closed as a result of fake reports. Since the lyrics belong to the songs I created with a free account, can I propose that the video be removed?

[broken lotus]

I haven't done this. I'm not sure.
if you have uploaded the song to yt, that might be proof if your upload date is before theirs.
Usually I hear that people with music distributors (DistroKid, etc) are contacted when their song is detected.

You have a creation date on your page, I don't know what proof youtube would want. That would be up to youtube.

[final hazel]

If you release it via a distributor, it's registered in a global database. And YouTube also uses that one to check if someone uploads it. YouTube does the work here then. If the channel is monetized, it will mute the song. If it is not monetized they will allow it. And it will be beneficial for you.

[fathom plover]

I still had 410 credits, on April 29 my credits expired. I canceled the automatic payment for the next subscription because I'm leaving the city, and immediately the 410 credits disappeared. The Create button is blocked. Does anyone know why? I also sent them an email, but probably when they answer me, my 410 credits will have expired.

[final hazel]

Credits do not roll over. It's per month. But I see it's not the 29th... Yes.. then you will have to mail them.

[fathom plover]

I know this, but my credits were expiring in 2 days, that is, on the 29th of April. But because I canceled the automatic payment of the future subscription, my 410 credits in 2 seconds reached zero.

[final hazel]

We cannot help you with that. You will have to mail [e-mail] for that.

[patent jetty]

Is there a document outlining the syntax and punctuation for inputting lyrics? Trying to work using brackets and parentheses but sometimes the directions come out as spoken words in the song

[fathom plover]

here is my data from suno.ai .Your Pro Plan
Usage Details

Monthly Credits Left: -500/2500

Extra Credits Left: 410
Billing & Payment

Cancel Dates: Apr 29, 2024

Billing Period: Monthly

[final hazel]

See answer above

[final hazel]

Everyone has that now and then. That is the unpredictability of AI in general. You can check the community wiki for more tips: https://suno.wiki

[fathom plover]

yeah, thanks, i send it, but when i give response my credit will be gone because my bill expire

thanks anyway

[final hazel]

Sadly we are just volunteers so we cannot be more of help I am affraid. 😦

[patent jetty]

Thanks for the link, offers some help. But I've seen songs that used a SFX prompt to put a siren or other nonlyrical sound into the music, but I don't see what tags are recognized versus just having them say "siren noise" 😅

[mild yew]

Expanded the experiment

https://suno.com/song/8acb74d4-6b9c-4785-ade2-153c1c534228

[foggy jasper]

I really like this! It reminds me of Carpenter with that heavy arpeggio bass that is playing the entire time, but has its own distinctive elements that remind me of Tron

good work!

also @broken lotus I didn’t get to see your dope Carpenter-like song you made. I got a ping tho✌️

[mild yew]

thank you! didn't mean to steal your idea

[foggy jasper]

By all means, I hope we can get more people inpsired to make Carpenter synth. It sounds so much better than mainstream synthwave (which I don't mind)

[mild yew]

I agree. I mean, Carpenter is the OG here, retro/synthwave is a revival afterall

[foggy jasper]

yep! I just wish we had more producers pioneering new sub-generes for synthwave. its really crazy how many ideas you can come up with.

[foggy jasper]

also just saw this. tell me about it. the amount of times I have been tempted to put "John Carpenter analog soundtrack, sound desgin by Alan Howarth"

just won't happen

[deft adder]

Guys I have a problem, whenever I want to continue the lyrics of a song in the sun he doesn't hold the lyrics that were written, he invents words and the music gets the wrong lyrics, does anyone know how to fix that?

[final hazel]

The AI can lose its way in the song. One of the things is changing the timestamp 1 or 2 seconds back, or placing back the last sentence that it already has sung.

[deft adder]

Thx só much

[thick geyser]

CIRP

[hollow hare]

Are you trying to use chirp? Bot is retired, use the site https://suno.com

[deft adder]

Is it possible to separate a song after it is finished?

[hollow hare]

You can Extend the part you want to separate. See the topics at https://www.suno.wiki/faq/making-music/

If you want to clip the song for something like video editing rather than for Suno to generate songs, you can use DAWs or audio editors like Audacity and split it up (there are many tools available for simple splitting)

[thick geyser]

Thanks

[hollow hare]

Some helpful starting links: https://suno.wiki
https://youtu.be/lj8utBd_IUM

[thick geyser]

And while asking questions,
Is there a way to control separate tracks like in PT or Cubase etc in Suno?
If not - any other AI music generator ?

[hollow hare]

Suno does not do that currently but I saw the following two links given at one point I'm the chat.

Note that I did not use them so I do not know their efficacy. You should be able to pull ip many more using a search engine

https://splitter.ai/
https://vocalremover.org/

[trim pagoda]

Hello!

I am hoping to download a high Quality version of the songs I am making in Suno, rather than the default low quality mp3's. How can I accpomplish this?

I also have the Pro plan

[thick geyser]

Thanks much

[trim pagoda]

https://www.audiostrip.co.uk this one is good too

[hollow hare]

There is no difference between the paid and free version except for the monetization and copyright.

If you meant the lackluster vocals, try using clear vocals early in the style prompt. Defining a more specific genre may improve the quality as well as it would reduce the diversity that may lend itself to the quality degradation because of the broad pools (this is my speculation though)

If you meant for music quality like in terms of getting WAV, basically you cannot. The training used essentially MP3-quality music so this is the best bitrate it can generate (though you can use programs to remaster the quality)
https://www.suno.wiki/faq/polish-and-production/can-i-download-uncompressed-audio/

[trim pagoda]

Is there any plans to be able to produce 320 kbps quality mp3 files? As far as using this music for commerical use, having anything less than 320 kbps makes it essentially useless, as the audio quality is substanitally worse. So the usecase for Suno is significantly diminished, if the highest achievable bitrate is 192 kbps

[trim pagoda]

thank you as well for such a quick response

[trim pagoda]

Lastly, Udio, your competitor, is worse in everyway EXCEPT that they have 320 kbps

[hollow hare]

I'm just a random user so I do not know the answer for this question.

I suspect the answer is eventually the quality may get there but not anytime soon.

[trim pagoda]

Okay. ty!

[final hazel]

The Suno team is working on V4.. I assume and hope it will be 320 kbps

[trim pagoda]

Have you heard any whispers about a potential release date?

[broken lotus]

uh their mp3 is 320kbps
their AI is more like 128

[trim pagoda]

Exaclty. Suno's music is leagues better, it just sucks we don't have 320 kbps yet. Suno is better in litereallly every other way

[unique stone]

Yes, this!

I did some testing a couple of weeks ago and posted in off-topic to demonstrate this: https://discordapp.com/channels/1069381916492562582/1159929934270181437/1229239072510181552

[trim pagoda]

192 is audibly worse I've been making music in Ableton live and have 30+ releases on Spotify etc. 192 is a long step down from 320 - once the Suno team fixes this, this software will be unstoppable

[unique stone]

From what the developers have said before, 192kbps was chosen for now due to the AI model quality limiting things at the moment. I'd like it increased and as was suggested above, it's likely it'll increase with newer models.

[trim pagoda]

Then "furrymattr" will be unstopabble

[indigo oasis]

I was just thinking this, I have a pretty good song, just need to adjust the lyrics slightly. Same number of syllables and everything.

[sonic harbor]

anyone have ay good 8-bit videogame type songs, want to get a better idea of the prompting for that

[final hazel]

ATM there is no solution, but the DEVS are working on "paint-in". With that you can redo a part. It will probably released with V4. I don't know when though.

[final hazel]

You can find some on the explore page. And here is how it works and how you can find them: https://youtu.be/kB3W5xLza_k?si=lF2sPwlQGIHYWanc

Keep in mind, if you have a paid plan, your credits do not roll over 😉

[sonic harbor]

I dont see any search bar on the expllore page just a giant wheel and a roll dice thing, maybe ill watch the vid

[final hazel]

Yes in my video I explain how you can search there

Yes, your credits are per month. What you don't use will expire each month

[sonic harbor]

oh you have to ctrl f to search lol

[final hazel]

That was my personal tip.. since there is no search 😉

[sonic harbor]

so there isnt a search basically

[final hazel]

Nope

[sonic harbor]

This is where I'm at on the 8bit stuff

[final hazel]

It's is on the website, but still people are not aware 🙂

I am Dutch.. so not a native either

[prisma wraith]

I'm on the lookout for a possible [breathe in] kind of effect, or a small pause within the lyrics

[sonic harbor]

Thanks, one goal I have is to make a videogame with ai, getting llms to decide on story progression, ai art, and the 8bit music suno does. All of the pieces feel like they are almost good enough to make it, but it might still be 1 gen away before someone cracks it.

[grave pasture]

Hey guys. Has anyone tried prompting for jersey club music and gotten good results?

[graceful condor]

Yeah

[graceful condor]

Yes

[signal pier]

Does anyone know how to influence the characteristics of the voice. I.e. male, tenor voice, etc.

[final hazel]

Just look up technical terms for what you need. Like for instance deep voice or baritone or vocal fry and add that in the "Style of Music"

[glossy chasm]

I can’t create my song for some reason?

[final hazel]

Please don't post double

You have to be more specific

[glossy chasm]

Yeah but I can’t tho

[final hazel]

Then it will be hard for me to understand. Maybe look at my video first, and then tell me what doesn't work for you? AI Music Mastery with Suno (model v3)
https://youtu.be/lj8utBd_IUM

[hollow hare]

Does an error appear? Do you have credits ? Are you signed in? Etc.

[glossy chasm]

Yeah I don’t have credit

[hollow hare]

It'll replenish in about 1.5 hrs if you are on the free plan

Free plan's daily refreshes at 0300 UTC+0 (<t:1713236400:t> in your local timezone)

[glossy chasm]

Ok

[deft adder]

I tried to do what you said, but it didn't work, it seems that Suno has difficulty understanding native ancient languages.

[graceful condor]

good video, ty

[astral osprey]

is the service having issues right now, im getting like all blank songs suddenly and sometimes with no credit refunds

Seee all the fails, and the two red circled ones are also completely blank but didnt refund credit

[hollow hare]

You can report it as a bug IIRC

[astral osprey]

i did, i dont want to keep hitting extend tho if im gonna keep getting blanks that eat my credits rn lol

yeah just did another and got 1 refunded and 1 blank that wasnt refunded i guess im done for now lol

[unique stone]

Are you changing the "continue from" time at all? Moving it around a little can sometimes give the AI something more to grab onto.

[astral osprey]

It was working great for like an hour then suddenly, just started giving me blanks

[unique stone]

Do you change anything between tries? Rearrange the style prompt at all? It might work better if you put the style like this: "glitchy, stuttered, choppy vocals, minor key, very glitched" - I'm not sure how well it does with a sentence for a style. Just throwing ideas out there on what I'd try if I got stuck in this way.

[astral osprey]

I didnt but i figured out what happened i think. Somehow the extend ui glitched out, it still thought it was extending but wasn't carrying over much information. I'd start getting peeks of the song but most was silence, i think it was just a bad bug

[crisp jewel]

Too many bugs, apply for a refund！！！

Too many bugs, apply for a refund！！！

Too many bugs, apply for a refund！！！

Since I paid to become a Premier Plan user, the lyrics of the songs I generated with Chinese lyrics have all changed from Mandarin to Cantonese. How can I restore the pronunciation to Mandarin? This is very important to me.

[lone oak]

hello

[hollow hare]

That is quite likely a coincidence. If you kept to the free version, you would still encounter the same issue, as free and pro are the same except for when it comes to ownership/monetization.

Did you try recreating one of the earlier songs you had in Mandarin and see if that was affected?

If it stayed the same, that means something in the lyrics is possibly triggering the dialect. If it changed, well, it means that the dialect is RNG (likely is the case anyways).

This issue has been happening for a long time (use the search button for Mandarin) and, as mentioned to you in the past, pronunciation is an issue.

You can try incorporating some Mandarin-exclusive words at the start to see if that forces the dialect change followed by [Intro] or [Break]? Hmm

[lone oak]

ni jia wo

[lone oak]

add me

Have you made any updates this month? Such bugs will not appear frequently at the beginning of the month. The current Chinese lyrics tend to fail 100% after several successes.

[untold briar]

I can only guess about dialects based on other languages since I don't speak Chinese. But as far I know the Suno model should not be performing differently since the last major update on March 22, 2024.

However there may be some specific bugs related to the 'Extend' feature. But I think if there was a change, it was to fix bugs on Extending songs.

[lone oak]

It’s often the first lyrics that are wrong.

first line lyrics

[untold briar]

Are the lyrics the wrong dialect or are the words completely **wrong **, meaning words that do not exist in any dialect, or words that do not make sense?

[lone oak]

At the beginning of the month, it was almost impossible to make a mistake in the first sentence of Chinese, but now the probability of making a mistake is very high.

It is very normal Mandarin Chinese, not a dialect, and the Cantonese dialect I was testing today and found that the dialect tag does not work.

[untold briar]

Can you show link to an example Suno song?

Is this the same problem as @letter__g "the lyrics of the songs I generated with Chinese lyrics have all changed from Mandarin to Cantonese. How can I restore the pronunciation to Mandarin?"

[lone oak]

OK

[signal pier]

Does anyone know how to add in a pause in a specific spot of the lyrics? or even add a sound effect like a laugh or chuckle?

[humble ice]

A few things, when I click extend it doesn’t work. It gives a 3 second clip of nothing. Also, is there a way to also download the instrumental itself if we want to re record the song ourselves? Also, can we implement a key and bpm stat to each song just incase we like the flow and lyrics and want to find a new instrumental?

[lone oak]

He is exactly the opposite of me. AI seems to retain memory function in order to eliminate repeated generation, but if it cannot be reset, it will still make people crazy.

[hollow hare]

I think ... and/or with several spaces can add a small pause but sometimes the AI ignores it

You can also try a [Break] variant as well (https://www.suno.wiki/faq/metatags/song-structure/)

If you have not done so already, try asking in General. Some people have played around with various pause lengths (you might be able to find it when searching Discord... but the wording may be too broad)

[knotty kite]

With the discord songs we made archived, how do we “get full song” on v0, v1 stuff? The feature is not available on the site

[hollow hare]

The option appears on the Extended/part _ clip
https://www.suno.wiki/faq/making-music/how-do-i-make-the-song-longer/

[knotty kite]

But it doesnt .. get full song does not appear on v0 or v1 stuff i made last year

[hollow hare]

Oh, sorry, read that as "like v0", my bad.

[knotty kite]

Yeah , thats what i said. V0 and v1

V2 and v3 are fine

Its the old discord songs made and continued within discord. How do we get full song for those ?

Theyre now archived

[crisp jewel]

Thank you for your solid advice. If I extend the song on the basis of Mandarin Chinese, the song is still Mandarin, but the new generation is still Cantonese, I feel very confused, I can not use the previous extension every time, I am still a paying user, there is no official person to give me technical support, I suggest that everyone do not pay now

Our company is using Suno products to complete the commercialization project. The customers are Putonghua people, but now the generated songs are all generated in Cantonese, I feel very anxious and helpless, and our business plan has to stop. Even though I have contacted the official suno by email, no one has given me a positive response. Now I really want to return the $288 I paid

[pure bay]

I tried generating songs with the same prompts at different times of the day. It turns out that at some times of the day, the phenomena of jaysinging is more severe, which means that if you get a jaysung song and generate the same way immediately, the next song is still jaysung.
So I hypothesize that when too many users worldwide are generating songs at the same time, making the AI too engaged, then it will jaysing more likely.
Does my hypothesis make sense?
And does anyone have the stats of the curve about the occupancy rate at each hour of the day?

[young prawn]

Hello there! I have been making songs for quite some time now, and I've seen a new bug arising constantly - the "Instrumental" feature which I turn on for almost every song I've generated for the past week gets ignored with every alternating prompt... may I please have any guidance on how to fix this?

[untold briar]

Try adding "instrumental" to the music style as well. For example, "indie folk, banjo, instrumental" or even "indie folk, banjo, instrumental, instrumental". Especially if your music style commonly has words, you may have to do this.

[humble ice]

Is there a way to get the instrumental of a full song? when i extended and tried, it was different instrumentals. Also watching this video to extend a song, I don't have a continue from song option.

[ocean salmon]

Anyone knows what flags the prompts for lyrics? Or what words? I used a prompt and it went through the first time but I generated the same exact thing again and it was flagged for moderation? What's going on?

[untold briar]

Wasn't really going for this, but pretty nice droning ambience. Feels like the dial tone of a modem connecting to your mind.
https://suno.com/song/a718bffb-0869-41a2-bcac-9b52da2675b1 Not usually my style of ambient but I really liked this one, extended into full clip in show case.

[broken lotus]

chirp has moved to https://suno.com

[hollow hare]

Got censored from ChatGPT not liking something as it is the moderator. Ask OpenAI's ChatGPT itself about "what was not appropriate about the following: [content]" and it should tell you.

Usual causes are using swear words or interpreting harmful situations. Even if it worked before, it could be a small change caused its interpretation to become different.

Worst case scenario, you can phonetically spell the word it has an issue with, though mind the ToS of Suno about what is and is not allowed (e.g. no copyrighted song lyrics)

[broken lotus]

👆 THIS

seems like maybe what passed in the larger context gets flagged on it's own...?

BUT..., I treat it as a glitch (because sometimes I can't see why it would object), and to work-around it I add a couple blank spaces to the first 2 lines, and possibly a [break] to the end of the lyric prompt...

I try to force chatgpt to re-process the tokens (change the text), but try to preserve what Suno will need to sing (don't change the actual lyrics).

(that's when I am convinced it is not the words or context)

[signal pier]

I keep reading that any content generated while using the free plan is only for non-commercial purposes. If I upgrade after the song has been generated, can I then use it for commercial purposes, or does it HAVE to be generated WHILE using the paid plan?

[unique stone]

Has to be generated whilst using the paid plan.
(Hopefully there'll be some way to "buy" commercial usage for a free song at some point in future, it's something often asked for.)

[signal pier]

Hopefully soon! Thanks

[novel nebula]

Hi. Is there a reason why song doesn't coordinate well when you put [chorus] and doesn't sing the same? and there are times when [Bridge] doesn't work at all and the AI just sings its own lyrics and end the song.

Like first two clips tend to follow the rhythm and same melody, then the next two starts getting all different and AI just sings its own lyric.

And there are times I want to just keep it solo singer but all of a sudden it ignores prompt and starts having group singing when Chorus comes in.

[pale hawk]

had that happen to me a few times, it actually made my songs better for some reason.

[broken lotus]

the AI can only hear back about 1min so anything you are extending that played >1min ago isn't going to come back.

If I understand you correctly, you need to catch whatever song elements you want to repeat within 1min of the extend (you may need to get whole song to get the full minute)

[novel nebula]

Is there any way to make extend go longer than 1 minute to prevent that? or at least make the first generated longer than 2 minutes?

I am only asking because most jpop/kpops tend to go at least 3-4 minutes and they usually have verse 1,2 and chorus in between them with bridge then chorus at the end.

[pale hawk]

doesn't seem like there is, probably just have to let it do what it wants to do, that's what I did.

[novel nebula]

[broken lotus]

no, these are limits of computer memory....
Separate tracks can be combined in a music editor, you could branch your 3rd chorus from an earlier clip where the chorus is heard, for example....

[novel nebula]

So , make first one, then extend that one first with 2nd verse. go back to first to extend it and make it as bridge one, and so forth - then link them together? is that possible?

I was under the assumption that you can only make song based on what you've extended from 1st , 2nd and 3rd etc. x_x

[broken lotus]

yes, in a music editor like Audacity... it is usually pretty easy since most DAW will cut on the beat...

[novel nebula]

oh, so you have to do it manually

I only asked because some folks who had their songs on Suno had melodies up to 3-4 mins and kept the same melody and had AI to read the lyrics properly. So I wanted to know their tricks.

[broken lotus]

yes if you want to take your music from meme-prompt to full music production.

[broken lotus]

I keep the chorus within each clip. that is the way to do it.

I keep instrumental breaks short, so they don't erase the vocal from memory.

Because I am lazy and don't want to edit my songs, LOL

but it is not possible for all songs.... like very slow songs are longer than the memory maybe...

or a song with a long instrumental break...

It means choosing the clips I extend from because they have the potential to extend more accurately...

[worthy ocean]

hello i have a problem where all my songs i create are instrumental even tho i dont check the option can someone help

[frozen gulch]

had you try to specify the type of vocalist you want? like maybe deep male voice, or something?

[shell loom]

you could also try clean you browser cache out.

[soft willow]

hi anyone experiencing issue on the web version where you tell suno to write a song and all it gives you is instrumental only?

im cleaning out my browser cache and trying again

nope its instrumntal only still

[neat nexus]

In the style prompt did you try putting in the type of vocal you wanted?

[soft willow]

yes, female vocals

[neat nexus]

Huh.

I’ll try testing when I get home from the store.

[soft willow]

but then thats from web version custom mode

I was trying to do the standard generative mode

[carmine moss]

Hello, it looks like my suno credits have dropped from 90 to zero! What could be the reason for this?

[neat nexus]

So testing with some old lyrics I had laying around and the bit of music I've been noodling with I was able to prompt male vocals and get them. Let me try with female vocals and then see what happens in a fresh song.

Was able to get female vocals in the bit I was working on. Now trying fresh song to sample.

[true carbon]

Need your input here, stumble upon someone that claim that he always delete librairy from Suno because if he keep files in it the AI will always use same voice, any opinon on that?

[neat nexus]

So I tested it with a fresh song (started completely fresh) and damn....

[Removed links]

I really wish I'd had more lyrics for this because I want to hear more!

[grave pasture]

what prompt did you use?

[unkempt viper]

How do I make the custom songs instrumental?? every time I have "Custom Mode" checked it always has words in it that I never put

[prisma wraith]

put Instrumental in the style it tends to help the AI steer clear from hallucinating vocals, not always tho

[unkempt viper]

seems to have worked so far, thank you

[brittle dew]

is there an API we can use to make long songs programattically?

[final hazel]

No api.. it's on the roadmap but we don't know when it's available

[prisma wraith]

if someone extends a "Part 2" that I've created, will this then make a part 3 for them? or would my "Part 2" become their "Base"?

[broken lotus]

it would be their part 3, I think

[broken lotus]

not even remotely plausible.

[final hazel]

It's already on the edge that you post your profile page, we normally would not allow it because it leans in into soliciting but posting double definitely is not allowed here. So I removed it from #🗣️┃chill-chat. Please keep our #📜┃rules in mind.

[round bobcat]

Hi, does anybody know where the Remix function went to? I can nt find it anymore.

[brittle minnow]

https://suno.com/song/9207745c-02ec-4c11-b490-ce7dd7ea2adc I hate so much trouble with this song...

its often when i generate lots and lots of clips i keep getting this very high tone in the background thats always there, it gets very annoying after a a while.

[hollow hare]

It has been renamed to 'Reuse Prompts' for more accuracy (at the same time, Continue had been renamed to Extend)

[barren vault]

Hi, is there a way to remove the strange noise in the background? Seems to be that there is a diffusion algorhythm running on KI. 🤯 sometimes it's wobbeling, too, at the beginning. Tried Audacity to clear up the mess. Like to get more clean results for the vocals. Any ideas?

[eager topaz]

The only way that I know of to clean vocals properly is RipX Daw Pro. It does a great job. They have a trial version if you want to try it.

[barren vault]

Thx, I will watch it out.

[cloud sphinx]

Hello

Can I reproduce a song suno ai generated for me in the studio

[cosmic venture]

Yes I think, i've just posted this - think it sounds better?

https://www.canva.com/design/DAGDvjtZAes/hiyvNmVTaNNtHhGiZGFaxg/watch?utm_content=DAGDvjtZAes&utm_campaign=share_your_design&utm_medium=link&utm_source=shareyourdesignpanel

This, takes to long as a file

You think this sounds different/ better afterwards?

[barren vault]

Thx I will try out. Otherwise I need a tool to rebuild a song.

[rugged crane]

Hi, was something changed over the weekend? From Fr. to Sa. my credits weren't refreshed and now it seems as if the "extend song" feature just does what the bloody Hell it wants. It doesn't follow the text or beat. I burned through 50 credits yesterday trying to finish a simple Deep Techhouse Track and just now on a NuJazz thing… 😢

[hollow hare]

Note that the free plan's daily refreshes at 0300 UTC+0 (<t:1713236400:t> in your local timezone)

Unfortunately, sometimes the AI gets stuck in a certain pattern when extending (as it first reads a minute to get the pattern hence why extensions are 1 min long instead of 2)

To force change the pattern, try changing your continue time by a second or two

The following talks more about it (and be sure to check out the other entries): https://www.suno.wiki/faq/making-music/extend-from-time/

[rugged crane]

Thanks @hollow hare. I did try to extend the song from a different time though (one second) and it didn't change the result as I had expected. The AI not following text at all nd just making up new hooks was really, really strange. I'll check out the videos and give it another go tomorrow. peace

[hollow hare]

I should also note that if that was a part 3 (extended directly from part 2 without making a whole song) you were working on that had changed, make parts 1+2 whole first then extend that so it has more consistency.

As get whole song does not cost any credits, it should be something to try if you haven't done it 🙂

Have a good night

[vivid charm]

what are some of the best prompt combos for some solid tech house?

[rugged crane]

I started out with the simple "Deep Techhouse" prompt, then "Deep Progressive Techhouse", added "Deep Bass", then "Roland TB-303" and now a 808. Not quite sure if I want to keep the 303 / 808 though, as the track seems to be going nowhere (currently 8 Parts).

You can listen to it here (7 Parts)… https://suno.com/song/6903db3a-eec0-43af-ab66-b25c31d54728 @vivid charm

[gentle willow]

hi guys, do you have any example or created songs that using vibrato on specific line on the vocal ?

[crisp slate]

General question, generated a song with silly lyrics but it turned out very interesting.
Is there any way to keep the same music and same vocal but completely change the lyrics? The extend feature could work but it doesn't give me the results I want

[spring haven]

I'd love the ability to use a seed, like you can with the art bots.

[severe reef]

hi! i have a question, why does the ai sometimes generate an instrumental track despite the "instrumental" mode not being switched on?

[hollow hare]

It can be from a certain pattern it had learned, especially if the genre is known to have long instrumental starts (e.g. certain subgenres of EDM)

If you keep encountering this issue, you can likely force a short intro by using a variant of [Break] or the more unreliable [short Instrumental intro] : https://www.suno.wiki/faq/metatags/song-structure/

Using search for other tips can help you too.

[severe reef]

Alright! Thank you!

[cursive patio]

Sounds cleaner indeed, what did you use?

[grizzled slate]

Working on Intro, Instrumental. Can you limit the time, for example [20 sec] ? BTW this didn't work.

[jolly loom]

Normally for generating short songs i do it like this,

turn off instrumental and put this in lyrics with custom mode on

[START]
[FADE IN, SLOW]


 
[FADE OUT, SLOW]
[END]

It sometimes spits out short form beats

[lapis sundial]

is there a way to force-regenerate the video? I've renamed a song but the video still has the original title.

[mild yew]

Before I commit an indeterminate amount of credits to experimentation, what's your experience with key/scale/tempo changes mid-song, either in the lyrics box or in the style box?

got any tips, gotchas?

[grizzled slate]

Thanks for this idea. Much appreciated.

Do you add anything in the, Style of Music, such as no vocal?

Is this just added to the Lyrics Prompt as it is:

[START]
[FADE IN, SLOW]

[FADE OUT, SLOW]
[END]

and then Generate?

[jolly loom]

You can do any type of style you want i don't think it matters

[grizzled slate]

So this is just for a instrument intro?

Or do you add these to your lyrics as inro and outro. Hope that makes sense

[sour stirrup]

Signed into wrong Microsoft account, suno won't let me sign out.

or remove account

[grizzled slate]

I think [Fade In, Slow] may have an impact, but Fade out didn't.

Do you have any more of these Lyric tags or a link to someone who has consistently working ones?

Do you find Captals has any effect?

[hollow hare]

I assume you used Sign Out first and it still logs you into that specific account? Then delete your cookies. You can search how to delete specific site cookies if you do not want to purge it all (or use incognito mode)

[sour stirrup]

...I swear I'm not this dumb 😭 😭 😭 😭

..it worked..

[brave gull]

lmao suno just hallucinated the n-word in a generation

https://suno.com/song/3720fc04-5e09-4a9c-92bb-a203136c5b66 (litterally the first words)

[brave gull]

yep done

[leaden echo]

I've found that ramping up the bass at 100hz and treble at 500hz, then scooping the mids, then cutting above 11k, dramatically improves tracks (with a little limiting of course)

[quaint cypress]

Hey, how to make Suno stop generating 2 songs at once?

[neat nexus]

You don't.

That's the way it's supposed to work.

You generate, select between option A and option B or you roll the dice and try again.

You keep doing this until you have completed the song.

[quaint cypress]

Option A and option B?
(Sorry Im new here)

[neat nexus]

When you input a prompt you get two options.

The AI makes two different segments.

If it helps you can also think of it as Song 1 and Song 2.

[unique abyss]

hello , i m fan of dangdut music. many times trying to use it , it became as pop

this is the 3 rd day experimenting wnad loging all seeds

[quaint cypress]

Thank you

[knotty yarrow]

привет

[hollow hare]

👋 To talk with other Russians, please head on to https://discord.com/channels/1069381916492562582/1154779864851812422

---

Google Translate: Чтобы поговорить с другими россиянами, зайдите на https://discord.com/channels/1069381916492562582/1154779864851812422.

[restive pewter]

yeah it sounds much cleaner, what program is that?

[frozen meadow]

Use extend after some of the vocals and melody. Write new text

[silver shell]

There has to be a way to make the vocals not just sing run on sentences it looks at punctuation and laughs someone help lol (put no punctuation on purpose)

[hollow hare]

You can use ... for small pauses.
You can enforce pauses between verses with things like [Break]

https://www.suno.wiki/faq/metatags/song-structure/

[elfin bay]

Seems to me I have to roll the dice many many times to get a good take

[neat nexus]

That can happen. It also helps to vary your prompting when a good result doesn't come out as hoped for.

[steep saffron]

Hello I am doing a voice clone with bark-voice-cloning-HuBERT-quantizer repo on my voice. I want to generate around 2-3 mins of audio and I find that there can be inconsistencies in the voice generated with suno-bark even for the same speaker. Any tips on how I can improve?

[obtuse ridge]

Help, my Suno site bounced so I accidentally disliked a song I enjoyed and its gone, how can I undo this

[desert wren]

any idea why this content is being moderated?

a song about a girl
she's a pediatric neurology nurse
her glasses fall off her face when she's changing diapers
between saving lives and saving grace
banners and crafts
a dog
a car she's proud of
she's too tall she says
i'm too out of her league she says
but she don't know
how dorky I am
and how in my league she is

[obtuse ridge]

I got it, was still displayed in the library but hidden from the create Tab.

[untold briar]

If that's a simple mode prompt, it might just be too long for the AI to deal with. if those are lyrics, not sure why.

[desert wren]

lyrics, even tried deleting each sentence one at a time and generating, no luck :/

[untold briar]

Well you can use ChatGPT by yourself to write the lyrics. The actual words generate fine BTW:
https://suno.com/song/43ab045d-1b87-4adb-8581-fbbd99a76e49

[desert wren]

moderated section was "banners and crafts", but we're grooving now. thanks mr fly

[untold briar]

Even funnier with the failed command in front https://suno.com/song/42b23fc3-81dc-4fe7-b67e-dbf86e30555b

[storm compass]

Instrument solos are very unreliable. I use the tag [guitar solo] but it takes several attempts to even get a solo. Sometimes the ai generate a solo but using a different instrument, mainly keyboards.

[true ore]

probably a known issue, but most times there's a droning sound that gets more and more audible as the song goes on, as if it's slowly collapsing

[storm compass]

True. The longer the song messier it gets. I try to keep mine about 3 minutes.

[true ore]

hope that can be fixed in future iterations, otherwise can't really complain

[cold zenith]

Hello everybody, did someone find a way to avoid the crashy voice when we continue ?

[novel nebula]

what lyric prompt is instrumental start?

[low remnant]

I tend to put in the genre, then check instrumental. Gen til you get an intro you like, then continue from that 10-15 seconds in with some lyrics.

Because I cannot seem to prompt reliably for it

[cold zenith]

solo man

[wary wedge]

https://suno.com/song/aa700107-d164-40c0-aad8-ad290007fcc9

[stable dawn]

with the way LLMs work, does doing several parts actually cause issues with the models memory over the several extensions? I am asking because I am wondering if it is better to "Get Whole Song" every 2-3 parts and keep extending and repeat in lieu of extending to 6-7 parts.

[forest moat]

hello the bot not work or what

someone can tell me how i can use /chrip on bot ?

[finite rover]

i think the bot may be out of commission right now, i tried using the help option and no response

[forest moat]

it same problem for me

[finite rover]

try using it tomorrow, maybe it'll be up by then

[hollow hare]

Chirp bot is retired, use the web app https://suno.com

Some helpful starting links: https://suno.wiki
https://youtu.be/lj8utBd_IUM

[hollow hare]

(Yep, See above)

[opaque cairn]

Came up with this chat GPT prompt that seems to generate full songs pretty well. Thought i'd share:

PROMPT:
Write me lyrics for a short song:
Topic: [insert topic here]
Genre: [write the style of genre]
Write each section as follows:
Verses: 4 lines lyrics with approx 6 words per line. Use ABAB rhyme scheme.
Pre-Chorus': 2 line lyrics with ~5 words per line that talk about the feeling. Dont Rhyme it.
Chorus: 4 line lyrics with ~5 words per line. Only the 2nd & 4th line rhymes. Dont rhyme 1st and 3rd line. If possible keep the 2nd & 4th line ~3 words per line.
Use the following song structure:
Verse
Pre-chorus
Chorus
Verse
Chorus
Bridge
Chorus
Outro

Also give the song a name.

Additional instructions:

You can copy and paste the lyrics from chat gpt into suno and it seems to work pretty well

[midnight gorge]

Hope I am in the right area...
Why do my songs get static at the end of the song and as I extend, each extension is more and more drown in static . (I did not mean for that to rhyme..)
I have had this happen with Synth-wave and Ballads I have either written or auto generated.
Tanks in advance