#📚┃suno-school
1 messages · Page 31 of 1
Will Inpainting be in v4
Yes
georg has spoken directly about it yeah
How soon do you think v4 will be out
The way Suno has done things in the past is just kind of.....ALPHA
They say they are close, but when is it close enough? Will paid people get it first again, will it be alpha? idk
Probably
I think just like v3 they will test it on the paid people first
If only to slim down the amount of people with access to deal with questions and bugs and things
It's what I would do
I wish they would add a vocal only option
YOu can kinda easily do that already outside Suno
Yeah.. that's reeaaaaally problematic with copyright
I know it’s possible because they already have instrumental only
How
I mean instrumental music is common
If you add a voice that is not yours?
Vocals only is like, acapella? But it's specific
You are right about that
I think they just mean stripping the vocals out of a song
But still sang like it's apart of the music
UVR can do that
Ah yes.. well stems will come too when they have the right quality so it will solve that
You use it to rip the vocals out, then you can just 'have' the vocal track
Yea I tried uvr but it did not work for me
I could hear a ton of particles
I used the right settings to
If you have a noisy song then that makes sense
Also would be the reason why we don't have 'stems' from Suno, since they have to get them using the same sorta thing that UVR uses
Well it’s not that noisy
I don't have a lot of experience with UVR besides going "neat" so I wouldn't actually know
Yea
Alright.. time for me to find my bed.. have a good one 🌼
Oh damn...damn...
You know how we're always telling people to interrogate ChatGPT (or other GPT type systems) all the time about the music you like and the styles involved? It's because even if you never quite reach the destination with some cleverness you can still learn some really interesting stuff and improve your sound in general.
I was watching Rick and Morty the other day and encountered Eric Clapton's It's in the way that you use it song as a cover on that episode. Really liked its sound, so I pulled up OpenAI and ChatGPT and started asking questions about the playing styles and music itself and after a few minutes had mined enough details to try some sample prompts.
Now this is nothing at all like that song. I know that.
But damn if it isn't a different kind of sound than I usually get and that's fun too!
https://suno.com/song/b458a05e-b2c9-408a-819c-4cea109659a8
https://suno.com/song/080f13e3-6a53-4155-a99c-d5d14f523a2e
https://suno.com/song/f861f201-ee03-4d7b-be3d-3729d677de89
https://suno.com/song/3a09c7e7-7f0a-402c-8170-808852035ecb
These are cherrypicked from about seven submissions. So fourteen total attempts. So interesting stuff here I think.
melodic instrumental, catchy repetitive rock riff, bluesy guitar riffs, rock chord progressions, warm steady groove song. Listen and make your own with Suno.
melodic instrumental, catchy repetitive rock riff, bluesy guitar riffs, rock chord progressions, warm steady groove song. Listen and make your own with Suno.
melodic instrumental, catchy repetitive rock riff, soft rock guitar riffs, rock chord progressions, warm steady groove song. Listen and make your own with Suno.
Why wont this stupid thing play my verse? I skips half of it, plays the last line and goes into the 2nd verse. Its even skips the guitar solo.. Come on man.. I've literally spent 700 coins on this last part of the song. I feel like I'm wasting my money here.
Have you tested if continuing from a few seconds earlier will help you get there?
I'll try that. But it might collide with another part
Insufficient information.
- Are you signed in?
- Do you have enough credits? (Need 10+)
- Do you get an error?
Hi 🙂 is there an api available to use suno?
Nope, no API. Suno can only officially be accessed by https://suno.com
All mobile apps are unofficial and mainly targeted to scam others out of money by imposing charges
arf 😢
yes im signed in i see my credit left but when i create it just spin forever
Try going to your library and refresh the page, that often works for me.
still spinning im checking like every 15 minutes lol
well now it work .... its just unstable i guess
Hm, there were several issues this week
i heard there v4 comming soon anyone has more info ?
I just subscribe this week maybe it was because i was having too much fun lol
I too would like to know how a vocal only option would be more problematic.
[intro]
[Instrumental]
[guitar solo]
would that be good
or all on the same line
im trying all king of setup but cant get it right
*kind
question about vocal tags
does anyone have any tips on how to get the AI to generate music like for horror video games? Like so far my thoughts are generally "music box playing in the background" "minor key" "creepy children singing" and possibly "lullabies" but I have only ever managed maybe two of those at once. (Never creepy children singing, alas. It's so iconic, though!)
I may have a biased sample of these things, since I don't actually PLAY horror games, but I find their soundtracks interesting
I am always trying to pick apart "what makes this unsettling?"
I realize this is pretty niche, so if nobody has particular suggestions, that's hardly the end of the world; I'm just brainstorming
Im not getting the chirp bot
Chirp bot is retired, use the website https://suno.com
Some helpful starting links: https://suno.wiki
https://youtu.be/lj8utBd_IUM
THERE ARE NO OFFICIAL MOBILE APPS. Mobile apps promoting Suno merely redirects to the main site for a fee/subscription. Do not give your critical info (e.g. passwords, financial info) through these apps.
Okaaay thank you for replying 😁😁
Someone on Reddit recommended adding dissonance, which I agree, and something you may want to add that as part of the style. Apparently the D key is popular in this genre
Choosing an instrument associated with creepy films should help as well, like an Organ or other unusual instruments.
Basically it comes to subverting expectations. What do you think about pop songs? Upbeat, melodic repeats.
To make it into a horror, you destroy the harmony.
I tried to make one in Suno but failed 😦 This was probably the closest:
Can i get the sheet music to a song after it has been generated?
Oooo an organ is clever; music boxes are just what went in my head with “creepy children singing” lol
Dissonance is also an A+ idea
Thank you so much!
You're welcome 🙂 I highly recommend looking back at some classical horror scenes and listen to the music, pick up their tricks
Crea una canción de tu eres la bomba
My brain just jumps to “violins shrieking like in Psycho” small children with music boxes, silent hill soundtrack, evil within 2 cover of ordinary world (another creepy child singing for sections of it) hmmmm I should pull out more examples to look at — you are absolutely right. (Thank you again!)
Xd
Google translated:
El bot Chirp está retirado, utilice el sitio web https://suno.com
Algunos enlaces de inicio útiles: https://suno.wiki
https://youtu.be/lj8utBd_IUM
NO HAY APLICACIONES MÓVILES OFICIALES. Las aplicaciones móviles que promocionan Suno simplemente redirigen al sitio principal por una tarifa/suscripción. No proporcione su información crítica (por ejemplo, contraseñas, información financiera) a través de estas aplicaciones.
Chirp bot is retired, use the website https://suno.com
Some helpful starting links: https://suno.wiki
https://youtu.be/lj8utBd_IUM
THERE ARE NO OFFICIAL MOBILE APPS. Mobile apps promoting Suno merely redirects to the main site for a fee/subscription. Do not give your critical info (e.g. passwords, financial info) through these apps.
Has the AI been updated? I'm making a fun little witchhouse jingle and I've genrated 10 part 2 extensions and they are all perfect I don't know which one to chose either that or im one hell of a lucky duck right now
It's luck
Awww.... :c
And if you are getting a bunch of hits, keep generating 
and I'd just make a playlist
and save all the gens you like to that one
If you can't decide now, look at that playlist later
theres 1 that i like the most and its the only 1 that the instruments cut out on the last few words of verse 2 before bouncing back in for the chorus. But usualy i'd have about 200 duds and 1 or 2 that fit.
Either way im pretty happy with this right now!
suno's being weird. i'm using english lyrics and have gotten back multiples in gibberish... its almost like the lower you are in credits, the weirder things get. I have one render that the song just completely cuts off in the middle but the timer keeps going... im very confused. It's also not going with the theme of the one i'm extending it form and has yet to do the chorus like it did in the first clip, not quite sure why
https://suno.com/song/c3614e4c-9bc8-47c3-a16d-c08eced34b8a
https://suno.com/song/d87002d2-5e82-4424-9b41-405307a3a1dc
female hard rock grunge angry song. Listen and make your own with Suno.
female hard rock grunge angry song. Listen and make your own with Suno.
hi all i cant pay for the Subscription $10 a munth just comes up as
Your card was declined. Please try a different card. i have tred other cands and the same out come all so had to phone the bank due to this pament been flaged on my accout so is there an other way to pay
You will have to contact billing@suno.com about your situation
I was just thinking that suno should add one time purchases instead of being locked in monthly or yearly subscription
but in return that one cost a bit more then the %20 monthly or yearly but not by a huge margin
Call your credit card company and tell them to approve the charge in advance.
Hello everyone, in your opinion, in custom mode is there in prompt a setting to impose the timing of the music between the lyrics? on most of the creations, he tends to string together the words too quickly.
I just subscribed, and then cancelled my subscription two seconds later. If you do it this way, then your plan will simply end at the end of that 30-day period versus it being on a subscription basis.
But do u keep the points ?
I did.
I work a minimum-wage job, so my income fluctuates between paychecks, so I don't know if I'll have the money at this time next month or not.
better to be safe then in debt
Exactly.
When I "Canceled" it, it changed the wording as well.
Instead of saying "Renew Date" it just says "Cancel Date", and I still got to keep everything.
It would be nice to not have to cancel to get just a month though.
I totally agree.
Slightly off-topic, but no one's bio has ever disappointed me as quickly as yours did. 🤣
I was excited for like two seconds, lol.
Love tha pfp tho!
Sorry! 😆
I spent like 2 minutes laughing at myself, lmao. 😂
"Yes! There's another! X3"
checks bio
"Dang it!"
I've no objection to furries though. I kinda like the irony of my name 😄
I love it, lol.
Happy Rhythmical beat, Vocal old style, whistle, Humming, Big band, Nature wild
thats one way to get whistle into the song whitout adding it into the lyrics
Got it!
i am using suno ai web, does anyone know how to prompt suno to get a cleaner natural voice, instead of one that sounds way too synthetized or like is in a chorus all the time
i noticed that clear voice,clear audio helps but not all the time
Where would you use that?
In the Genre or in the Lyrics []?
in the style of music
Gotcha. Good to know.
Thanky ou
try adding "clear voice", or "studio quality" , or specify the kind of voice you want "baritone male voice", soprano, tenor, whatever.
What are the current best practices for maintaining a consistent voice for longer songs?
I tried extending from both full songs and from just the extension and it seems that no matter what I do, the voice tends to:
- Gets fainter,
- Moves from a distinct soloist to a generic group/choral one, and
- Sounds more and more like autotune
One of these is bad enough, but it seems like all three of these keep happening. I even tried coming back a day or two later to see if timing had any effect but that didn't work.
The first extend usually is good, the second starts diverging and a third extend is almost impossible to enjoy.
Problem whit the ai reading musical tags how to solve it ?
Can you give more detail?
Graveyard screams, Thunder storm, Mystery, Haunting,skeleton,bombastic, Rain,Winds,Clawing. clear voice,clear audio song. Listen and make your own with Suno.
it reads this [Soft piano melody sets the mood]
[Gradual build-up of percussion]
I try to keep my metatags to 3 words max or it tends to ‘break’ the metatag and sing.
I don’t think Suno ever really talked about their LLM, but it loses the plot if the phrase has too many tokens… by the end of the sentence it doesn’t always remember it’s a metatag…
HI! Please, does anyone know the prompt to create a few seconds track to finish the song?
you need to extend the song
so even with the 2 minutes song without any extension the longer the song goes the quality of instrument is degrading a lot
anyone notice that?
for the last part use the style "end, short" and [end] in the textbox ... extend where its quiet and you should just get a super short part
use the extend... and in the new lyrics box type
[End]
with luck you will get just the last notes (or the reverb fade of the last notes) and the clip will end before 7sec and you will get refunded... 😎
yes... AI starts strong and loses coherency the longer it goes... Chatgpt with coding..., LLM with character roleplay....
It's not a universal thing on all clips, but it is often enough that I hear the voice degrade within the 1st clip sometimes.
We can always use Extend from Time to pick a point in the song that still feels strong..., but it's a wet blanket on people who want to hit 1 button and get a 4-5min song.
Guys, can anyone help? I want to insert an instrumental before the vocal, but I can't do it, even by copying someone else's prompts
im able to make full complete 2 minutes song no extension only problem is all song after 30 secondes or so start to sound like the intro of Rge Against The Machine "Killing In The Name" go listen to the intro of that song and you will ear the same kind of effect lol
the ai do what it wants we can only 'suggest' or try to force certain behaviour
Dam, so i just have to try to make more versions of the same song and see which works?
well.... if you like the first 30 seconds of a song you can extend it
but its really difficult to know where the merge of the structure of the song will fit
i made a 'funny' test Metal song with 5 or 6 extension to get guitar solos etc wanna ear lol?
i really like my song but would have love more precise ajustement
melodic, brutal, hardcore song. Listen and make your own with Suno.
5 extensions seems extreme, i wanna see the result
its preatty heavy lol
Yeah, quite fine for my taste. I hear some distortions, i guess these are the parts of extensions.
the more the song goes it sound like synthetiseur
Well, that's a useful experiment. Seems like going over 2 extensions is overkill for this Ai
i have another one i like juste because of the scream of the singer in the intro LOL
melodic brutal hardcore song. Listen and make your own with Suno.
i laugh everytime i ear it
https://suno.com/song/2525f972-7ee5-49c4-b891-4749da545814 here is a 30 minute song
Pop, glitch, rock, Americana, indie, Rococó classic, EDM song. Listen and make your own with Suno.
The issue is that the AI does not remember the original style used. What it does, for each extension, is study the style ~1 min prior to the continuation point to learn/mimic the styling you had (which is why extensions are 1 min long as the other min was reserved for studying).
So if you had a loudness or degradation issue within that minute, it would start becoming emphasized as that is what the AI picked up on and all future extends builds up the issue as long as it exists within the 1 min of continuing.
It also explains why the style can change when you vary the starting point by a few seconds
i get it maybe it will improve with future update
Yep, figured it out by listening to the 30m song. It started to lose it's identity after 3m. I hope this AI will get some "brains", but even in this state, it's really impressive.
another thing to improve .... sometime you like a kind of singer it take me 10 test to get a voice that a like sometime so instead of just create a new version and wishing to have the same voice.... if we could have a ... re-use song tone .....
time to bed lol see you all!
"the identitiy" was switched by about 10 people intentionally on each part. its a community project.
Is there a list of editor keys I can look at for lyric structure? I know there's [verse] and [chorus] but I was wondering if there were more extensions.
Song structure entry at https://suno.wiki
Perfect. Thanks.
how can i make duets
Getting duets... that is more luck than anything else. Try to use male and female sounding names in tags, that might help, but it is still hard to achieve.
A few people were able to do it but at an expense to their credits:
3Daizy's duet
MrTree's duet
Wiki entry
Again, getting it working on purpose is hard. Alternative methods:
-
It is easier to make a duet by doing something like generating instrumental part 1 then Extend it twice (one extension where you specify for male vocals in the styles field, the other for females), or something similar for each alternating part, then edit them together in a DAW.
-
Even easier (and cheaper) is to make a full track then swap the voices gender using other AI tools like kits.ai (see #🗣️┃chill-chat message and #💡┃suggestions-web message ).
You can use a DAW to edit which voice track plays when.
thanks
Hi y'all, anyone got a fix for the clipping problem, where the vocal is clipped off when Suno begins creating?
add [instrumental bridge] after the lyrics and it will make more music after the text. (most of the time 😄 )
Yeah, I kinda sorted the ending Mr.Tree, and I'd like to stop putting Oooooh, as the first lyric line to stop the real start getting clipped. But Suno sometimes starts from halfway through the first line, and, you know, it's a brilliant take after that, but it's got no opening. Waaaaah!
Does Suno's creator respond to any symbols or punctuations from the key board, that you know of?
yopu could also try [intro] before the lyrics with just some ... after them to get just music
what do you mean?
'''' ,,,, .... !!! is it progreammed to read pauses etc?
sometimes it does. i see people using dots to give it structure but without singing.
have you had any success with that?
my music is more of experimental nature 😄 i dont normally try to be very accuarte with texts. if something doesnt fit i just regenerate or extend from where it got wonky.
largely, I am too. but certain genres lyric are a particular rythm and meter, and if the first line doesn't get recorded, the take is basically wasted. I'm a bit surprised Suno hasn't listed symbols or keyboard thingy's to allow pauses and breaks
i think suno assigns stuff dynamically. its hard to describe. sometimes stuff does exactly what i want it to do. and sometimes suno is like : no. you just get a piano piece 😄
Heheh, I know what you mean. I was diggin deep in the solo jazz trumpet, when Suno wanted me to listen to solo piano 😅
have you had any contact with Suno tech/developers?
or know how to contact them?
they are super busy coding on v4. they dont give out any infos that go deeper. they even say the community can use suno better than they 😄
Thanks, Mr Tree, I had a feeling it might be like that.
thats the weird thing with ai. if you tickle it at the right places super interesting things can happen that were not even intended. as i understand it.
Absolutely, the Suno LLM depths of nuance and authentic individuality poke their heads up occasionally, like a gopher on a golfing green.
to wit, I just posted a piece in #1133478404222623804, a church choir piece, Suno added an ending so exquisite it brought me to tears.
for me its my personal challange to let suno do the weirdest stuff you can imagine 😄
to be fair, we can't stop it, and if you're in the right genre at the right time, Suno hits A grade sympatico
if you want you can have a look at my suno profile. even if all the stuff is not your type of music it could be interesting to see in wich ways suno can break 😄 https://suno.com/@mrtree ... i listend to your choir piece and its really amazing that suno can do peaceful and beautyful music and also super broken stuff thats more like a art piece 😄
Ok, I'll check some out
i tried to sort everything a little bit in playlists. "delicious food" and "game on" and "suno songs about suno" are community playlists with even more awesome music 😄
Have you got any favorites I should start with?
maybe "my musical journey" ... it has a bit of everything.
👍
Soi soi soi soi, could use a few less soi interludes, and u'd capture the Daft Punk minimalism well. 'Suno does what you want' is a banger - did you edit that on your computer? If not how did you get the breaks and pauses from the spoken word, talk, presentation, sound fx prompt?
everything is done in suno. you cant reupload into suno. sometimes its just luck that it works but metatags like [spoken intro] or else help suno to get what you want. its super smart. in one track i just have posted a rough script for a commercial , and it made the whole commercial. several speakers, sound effects, music. it just understood what scene1: cut to XYZ : ... ment... its totally crazy
Yeah, if Suno is in the zone with you, it does its thing. And I feel like we're just tapping the tip of its potential. Whatever th logarithims they take some time to penetrate atm.
i use suno more like a tool to enhance my ideas. i have a rough idea of what i like in my music... i throw that at suno and see what it comes up with. For me the main goal isnt the song at the end... its more the way of making it. experiments, slow transformations with different styles, hard cuts, funny vocals or sometimes just a regular song.
Yeah, I understand. the process can be the fun. I'm at the point where AI can give me better production and content than I can create, so an authentic, polished piece is my aim. Mr Tree, can I have a look at your ai produced commercial video?
i just make music 😄 . i made 1 song for a video ... but until now i have not made the video 😄
" i just have posted a rough script for a commercial , and it made the whole commercial. " so you're saying Suno can read script cues and direction?
ahhh that was ment for the song itself... one second i see if i can find it
your 'House is House' is getting into Skrillex territory. If you ever get a prompt that can Skrillex, let me know 😎
here is the thing i ment https://suno.com/song/2815b276-8b11-43a9-b3a3-656b6b432070 it did all that by itself just based on the [metatags]
spoken word, commercial, voice song. Listen and make your own with Suno.
Hahahah, yeah, ok so Suno is quite self aware as well. Once it got the gist of what you're doing, it said, 'here , let me help you with that."
where is the place to get news about what's comming up in v4
Just released my Interview with the Vampire themed song!
https://youtu.be/QPsWxpCymnY?si=Aq3m7Yh0rREpJfAO
Thank you so much @AnneRiceDotCom for making such an amazing character.
To Support My Work:
Join my Patreon: https://www.patreon.com/AuthorKatherinePhillips
Subscribe for more music: https://www.youtube.com/@NarrativeDrivenArt
About Katherine Phillips:
Katherine Phillips is a writer and indie artist. Explore her narrative-driven music, fuele...
Hello. Can somebody help me to create music for striptease? I hope I won't be banned for it because, well, how music can be nsfw? I've tried different styles, but everything I've got suit only comedy like Benny Hill show.
How do you end a song properly? even when I put [End] its always trailing off doing whatever adding in ghost lyrics or doing another interlude instead of just stopping.
maybe some downtempo jazzy stuff , slow swing, saxophone, soothing ... just try to describe the music to suno
have you tried the style "end" and then just [end] in the lyrics box... extend from a part where it looks like it has concluded or wants to naturally end
Define strip tease music.
Well, I've tried bluz, latin music, pop music and other things I've found in the internet, but in all cases I'm getting something strange
they all sound fast. but it depends on what you want like hepcat said... i was thinking of something like you can leave your hat on or so...
If I could do it correctly...
I can try that yeah, i never thought to put it into the genre/style box
what about lofi influences... pretty slow. or do you want it more energetic?
Sorry, I didn't get the last sentence. What does it mean? About hat.
"You can leave your hat on" is a song ... i think its called like that... one sec
Joe Cocker - You Can Leave Your Hat On
Energetic is more preferable, And I suppose it should have rythmical "cutoffs", to indicate sudden movements
oh, yeah, I thought about this as references, but I didn't know the name of this song.
thats the classic song for that... application 😏
it worked ty, now i just need to figure out how to manipulate the picture from these damn guitar birds D:
ahhhhhh... you are exactly right here!! ... let me explain. one sec
the picture is done by the last part. you can manipulate the last part to show whatever you want. my technicque is to look for a silent part at the end and extend it with the style "end, short" and in the textbox [end] followed by (the description of your picture) ... here is a example. sometimes it tires to sing the picture info then i add things like [mute] between end and the picture text... here is a example https://suno.com/song/9f305755-3111-473d-bcaf-e677e38e95f8
turntablism, glitch recordscratch, cartoontoon sound effects, nuskool, virtuoso scratch song. Listen and make your own with Suno.
all song that i make no extention the sound quality is getting worse and they all sound like the intro of 'killing in the name of' from rage against the machine will v4 adress this issue ?
this highly depends on the styles and the complexity of the song. you can also try to extend the song before it gets noisy to try to steer it in a cleaner direction if you really want to save this part.
am i doing something wrong with the link to a song? I link a song and when it plays the picture cause to something completely different and the title is not in the video.
If you are playing a video of an extended song, the video basically concatenates the video of part 1 (which was created upon the initial generation) with the video of the other parts being combined (also created ONLY during the generation of the part).
Upon reaching the extention time point, the image/background/title will change to the new part.
That means the song title for each part is specifically the ones used during their initial generation and changing the title does not affected the original generated video (as that would be a costly process with the number of users using the service).
if anyone is interested in vocal prompts for clarity (I'm guessing they've circulated by now, but some people may still want them) my two most recent songs use some of the more straightforward techniques and you're welcome to strip them off
to use
I also want to note that "the seminal work of [end] ----" where ---- in my case was piano, but can be anything produces more dynamic endings
just he sure to toggle on instrumental
Always interested in tips and tricks.
it combines "post-post-vocalization" with a general format I've found, for whatever reason, to generally produce vocals "first" sound
and has the end tags associated with "noise" reduction
they tend to tamp down that growing background muddiness
"polyrhythmic(anti-pop-punk lofi cassette, dynamic anti-sound anti-fx)anti-hyper-gothline trapelloized jewish nigun"
what a prompt 
you have to ne cautious though; frankly, 1,2,3,4,5,6,7,00 may work equally well without having as strong a style jfleujce
influence
anti-hyper-gothline aloje
I've been making heavy use of 'cassette' and 'anti-fx'
Seems to really do things
that and this whole thing post-anti-instrumentalized anti-hyper-gothllo trapelloized jewish nigun40404040404040, accidental
makes magic happen halfway though anything
I don't ask why
it was essentially random, but it continually pays dividends (although it will absolutely increase noise distortion)
but I like thay
So when you mention your post, are you talking about Southern drawl?
yes
the other one also has clarity
but it's more cassette
like, it's got that lofi
sound
Did you do any recursions on it?
continuation? tons
to get creative transitions
lol well, here is another anecdote for you, slapped "anti-hyper-gothline aloje" in and yeah it shakes things up
lol
it do be like that
Nah. I mean specifically where you get whole song, then continue from that more than once.
seriously, sometimes I just start stringing things in there
oh, yes, always
I rarely do continuation from non-whole song
Do you also shift the style prompt as you go?
because one of Sunos strengths
it depends. Like, I have one or two that I didn't
but almost always I do
I have some prompts I like to throw in midway to "shake it up"
and then my ending prompts
which for the outro. It's usually just [end] in style with a but of pizasz
and a cool picture prompt!!!
OK, the reason I'm asking, is that everytime you do a recursion it resets the style prompt. Also when you do a stitch the style prompt reverts to the very first part's style prompt.
So if there are other tags you wanted us to see, they're not in the song page. Only you know what you did and used.
(Also I suspect that this is what is causing so many issues with people having a demon choir. But cannot prove it.)
yea, it's annoying
but I just paste over it
I keep a list of various pastes
and just do 6-8 variations and see what's working
but yea, it didn't used to do that
as the default
Same but I keep playlists of clips specifically so I always have the prompt handy 
I do too, but if I'm not paying attention that can really mess me up. Then when I finally see what happened I have to roll back to the earlier part and do it correctly.
(I tend to do a musical intro with different style prompting and then when I'm happy I start my vocal style prompts until I like that and then roll it over. This means I have to be careful to save stuff in my notepad.)
I think it's just plain old overfitting as I've done more. I think the genres they're dealing with just didn't have enough varied data
if you know anyone who is having it, I would love to get the genre tags
so I can do some tests
as I'm interested to see how far I can improve it without polluting
the style
Clearly they should give us a bit more access to the model 
I'll keep an eye out. And let you know if I encounter it in anything of my own.
I want to break it
Any way I gotta go. Nice talking with you. Thanks for sharing.
always ✌️
you gotta, it's the only way
first thing always I do is basically red team. I cracked llama 3 day 1
although that was stupid easy
but that's off topic
Someone just found this
https://suno.com/search
suffice to say, two sentences will turn it into a groveling moron
that will tell you anything
yeah
https://suno.com/song/6116dacd-e81f-46f8-8927-c75b7c392b1a what is creating this problem before it worked just fine but now it dose this to every song
56s rock swing, Fast upbeat, Vocal Radio Voice, male song. Listen and make your own with Suno.
it shouldt be saying the numbers nor drums
what are you trying there? 😄
trying to make a rock swing song
Basic Swing Snare Pattern
Tempo: Moderate Swing (120-160 BPM)
Snare Pattern (4/4 Time Signature)
Verse and Chorus:
Measure 1:
1 & (2) & (3) & (4) & | Snare on 2 and 4
Measure 2:
1 & (2) & (3) & (4) & | Snare on 2 and 4
Repeat for the length of the verse and chorus.
thats not how it works 😄 thats all just lyrics for suno. you dont have that deep control over things
Hey yall, am I the only one having struggles to make a song start normal? Since like the past 2 days they dont have an Intro anymore no matter what I tell them to do
have a look at the wiki for general workflow infos and use of [metatags] https://www.suno.wiki/
for me often a [instrumental intro] just followed by some dots .... .... .... does the job
Lemme try
Its weird cuz it worked fine for months lol
And somehow now it wil lalways start in right away
thats the suno lottery 😄 ... works 10 times and then you have to tickle it to get it again 😄
Felt like something got changed
Weird weird
But seems to "somehow" work out of every 5th generation lol
Thanks tho
i have that all the time. i use weird styles. often they work, and then just do weird stuff for like 20 generations and then work again until they dont 😄
https://suno.com/song/c9913dc5-d449-4d84-8967-70263574c78e kinda getting there but not 100%
100% is a number that can be incredible credit intensive 😏 if you embrace small mistakes as creativity of the ai it gets much cheaper 😄
Regga,swing rock Rythmical clear voice,Raw,Garage,Radio,Tango medium music volume higher voice song. Listen and make your own with Suno.
Final product
Regga,swing rock Rythmical clear voice,Raw,Garage,Radio,Tango medium music volume higher voice
Question.. I had to do an edit in a verse on the last line. So if I press extend and put at the top [Verse 1 continued] Would it understand? I only have a few credits left and don't want to waste em lol
I think Suno really should have an option to edit the lyrics without having to risk generating a new song for you. Nothing worse then getting the perfect setup. But risk losing it, cause of a mistake in the lyrics.
if it can see some other structure of melody before the continuation time or text it might make the same thing again. but most of the time you need some generations to make it sound similar again.
Ahh that suck :/ Maybe in future versions it can do it.
not a game breaker tho. Thanks!
how do you get Suno to recognize a guitar solo? (guitar solo)?
round brackets are for (adlibs) and square brackets are for [commands] ... the [guitar solo] idea could work. but just if suno feels that way 😄 it may take some regenerations to make that happen, or it will simply be ignored depending on the style.
Suno is 💩 'in itself today. Not following style, not generating the full 2 minutes.
Had the same problem. I put the next line at the top of the prompt (without any brackets) and the extended part flowed seemlessly. I cut it at the last line I could use. Usually, it works. Though not today. Suno is having problems it seems.
Same thing also happened to me. Also, do you guys have a mobile app?
No, I'm using the desktop version.
But is there a mobile version?
not yet. please use only the website suno.com. it also works on mobile (but you probably know that already 😄 )
Well then please let me know if and when you release a mobile app!
if i hear something from the devs i will post about it here in discord. if the app releases it will be a huge deal.
So would I be able to monetize a song I generated on the free plan if I added a rvc voice, remixed it, in the description “put made by suno”, or change the instrumental?
Make an instrumental track first, then extend it from wherever you like best.
I’ve wasted a bunch of credits today because it either skips over lyrics (in the very first portion of the song!) or sings the commands. 🙃 I hope someone is fixing the music robots.
Is it possible to make a duet
Getting duets... that is more luck than anything else. Try to use male and female sounding names in tags, that might help, but it is still hard to achieve.
A few people were able to do it but at an expense to their credits:
3Daizy's duet
MrTree's duet
Wiki entry
Again, getting it working on purpose is hard. Alternative methods:
-
It is easier to make a duet by doing something like generating instrumental part 1 then Extend it twice (one extension where you specify for male vocals in the styles field, the other for females), or something similar for each alternating part, then edit them together in a DAW.
-
Even easier (and cheaper) is to make a full track then swap the voices gender using other AI tools like kits.ai (see #🗣️┃chill-chat message and #💡┃suggestions-web message ).
You can use a DAW to edit which voice track plays when.
Is there a time when the extend function gets blocked?
I was brainstorming this in the shower today actually... my idea will take a lot of credits, but I was thinking I could generate an instrumental, then do lyrics with one voice, then do them with another as often as it takes to get them similar enough to match up,and edit it together myself.
I spent like 1000 credits on one song and it did not even sound good
Not to my knowledge. It only looks back 60 seconds though so the longer you go the greater difference between the beginning and end of your track.
Hahaha, I know how that goes. I've spent 800 so far just trying to fix tiny problems with my last song.
Yea
okay
One thing I do is generate an instrumental first, and then I can go back to it later for subsequent verses. That's credit-expensive though.
I made an amazing song on the free plan on my first try and I bought a subscription thinking that I would be able to monetize that song then I found out I could not and I have been trying to recreate it ever sense and it has cost thousands of credits
right. i think so. 🙂
I also used Ultimate Vocal Remover to split up the instrumental and got real nice drum and guitar stems out of it, so I can sprinkle in the opening riffs later on to create the illusion of cohesiveness.
Just email them. I bet they'll be like, ok that's fine.
i can extend function my playlist yesterday. but i can not.
Playlist?
Really it was that easy
I dunno. Certainly worth a try.
Do you have their email or like where would I get it
Ok
yes. i correct music. and i make song prompt and if i like song , send my playlist.
Hello! Please email billing@suno.com if you are unable to find the answer to your question on this page.
😎
I’m emailed them, I’m hoping they are not going to say no
Yup, pretty much. It’s been ignoring some lyrics but I haven’t ever had it sing “[verse]” for example. Oh well, hope it gets sorted out. Most pressing issue for me would be some songs deteriorating in quality over time.
God, if you explain how many credits you spent trying to recreate it, how could they? Good luck!
It has the same issues of all LLMs... like, i put in actual chord notation for some lyrics just to see what would happen, like
[Intro]
[Evocative Grand Piano and Bass]
[Verse 1]
[Am] Listen as the wind blows, [G]
From [D] across the great divide...
Voices [Am] trapped in yearning [G]
Memories [D] trapped in time
[Verse 2]
[Am] The night is my companion [G]
And [D] solitude my guide
[F] Would I spend forever [Am] here
And [D] not be satisfied
and most of the generations I did they weren't sung. I can't really tell if they made a difference, and a couple of times it did happen. But it's just that random element coded into the LLMs that makes it such an uncertainty.
QQ is there a way to find a song using title, that created in different account?
Not currently
But soon
https://suno.com/search
I've had some success googling song title site:suno.com
Can anyone tell me how we can change the pronunciation of some words so we can fit in different languages?
You can change the lyrics' language even mid-song and most times it will be recognized automatically.
A notable exception is Ukrainian, that may be recognized as Russian until a typical Ukrainian word or letter (e.g. "i" or the name of a month or day of the week) is found in the text.
Portuguese will most probably be recognized as Brazilian Portuguese.
You might have some success putting a [language_name] metatag in front of the text when auto-recognition fails; I've never tried it, but that's what I'd try in such a situation.
If most of your lyrics are in the same language and you want a single word in another language, Suno may get it right or not; you can try to write phonetically in that case.
As most things Suno, it's hit-or-miss. As in "try until you either succeed or run out of credits". Suno doesn't take orders, just non-binding opinions.
Most notable example of different languages in the same song:
https://suno.com/song/bb55a553-969f-44c0-a6e7-cf3b1fb5fd75
electro swing,Eternal Love, sweet vocal, Collaborative Performance, glee club song. Listen and make your own with Suno.
Hello, Anyone knows if theres any WEB UI (fork?) for Suno Bark?
https://github.com/JonathanFly/bark
https://github.com/C0untFloyd/bark-gui
ok nevermind found those, unfortunately not updated 😦
Hi I have a long song, lots of lyrics (5 parts - 4:40 length)... it seems like the voices are laready modified sinc ethebegining... How to avoid this?
Will check that way 🙏
And if is actually possible to make audio longer than 13-14secs?
Bark is a outdated model and has no support anymore. The thing is, if i would know how it works, i would really like to help you 😅 But it was discontinued before i started helping here.
Oh doesn't looks like that. Any alternatives?
depends on your purpose for it, i usually send the people to suno.com with the hint "suno can also make just speech".
Looking for free alternative (can be selfhosted) for longer texts like 2-8minutes
That sounds great. Will it be a search function for your own songs also?
Yeah, I’m trying to stop the drums of playing.
how do u make just speech
thanks for replying
styles like "standup comedy" "narration" "spoken word" "tv commercial" and so on may send suno into speaking mode. i use this all the time.
ty will experiment with that 😄
🌳 Good morning Mr.Tree🌳
Hello. I canot use "extend". Why?
there might be a temporary problem. i also just joined for today and saw people complaining extend and reuse prompt doesnt work.
ok thank you. I am waiting so.
but it looks like extend is working in the create menu... i have to make some tests
I can not extend other users' songs.
if i click extend then nothing happen.
i was using it until today.
i noticed that too. from explore and my library i cant extend or reuse prompts. just from the create menu
ok than k you
This is a critical piece of information. As reported in the bug channel ( #🪲┃bugs-web message ), the devs tweaked something that broke extend from other people song
hi, I need assist with Metatags inside for Chirp (music).
how do I get a "Break" inside the Music at a specific point?
Thank you.
I checked for Metadata tags but nothing ...
Well, you can only guide it by placing a metatag after certain lyrics, not time points.
The AI also takes these as advisement as well, not direct orders, depending on how the rhythm goes and how much time remains (if you have a lot of lyrics for the allocated time, it would not be surprising to see the AI skip sections)
There is Break, Interlude, Instrumental solo, etc. and their variants. If you want a small pause, adding ", Short" to the tag (e.g. [Guitar solo, Short] or perhaps using an ellipsis (...) should provide a small break.
The following excerpt was taken from https://www.suno.wiki/faq/metatags/song-structure/
I think today I'll take a break to do weird experiments since I don't really have any lyrics for another major song worked up yet.
... Oh, thought I was in the other channel oops.
Anyone know how I can add an instrumental build up to a song with lyrics? All the songs Im working on, if I add [Instrumental] or [Instrumental Intro] or anything of the sort, it just bypasses it and the song starts right away with the lyrics. Anyine have that issue and have fixed it?
Is there a way to appeal the censorship in creating a song? I have a song with a very important line "with your hands around my throat, I'll soon asphyxiate"
It won't create a piece of music based on this line
It got censored from ChatGPT (which moderates the lyrics in Suno) not liking something. Ask OpenAI's ChatGPT itself about "what would censor 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.
The lyrics you provided is quite descriptive to GPT
Any clue how songs are organized in the trash ? In alphabetic order? By date of deletion? By creation date? Randomnly? Today i ve tried to recover a song sent accidentely to the trash with no success (i have lot of files in the trash) as there is no search box.
i think ™️ they are added on top of the trash bin. last deleted part is at the end of the bin. but i might be wrong... i always search the 360 pages of my bin too 😏
I dont think so as i cannot find them on top. Iterating through the deleted songs i see some recently deleted songs in the middle of the list but i cannot say that it is organized in alphabetic order.
my bin is not alphabetically sorted... maybe they go where they once were in the library?
a time/date based sorting?
hmmmmm... i dont know... i will ask
that cant be unsolvable 😄
Hi everyone, I'm experimenting with Suno and trying to generate sounds and I'm struggling with that. I was wondering if any of you tried it or if Suno wasn't trained on this at all. Thanks 😄
sounds = telephone ringing, rain, birds twitting, and so on.
It would think that's too dark and violent. Try being more metaphorical.
for a cleaner version of a song i noticed is better to make shorter parts and extend them
[Instrumental intro]
(Verse 1)
Beneath the bridge where shadows play,
Trolls make their home, where waters sway.
In the darkness, they quietly dwell,
Guardians of secrets they never tell.
(Chorus)
Underneath the arches, under the stone,
Trolls hide in silence, unseen, unknown.
Their whispers echo through the night,
As they keep watch 'neath the moon's soft light.
[Instrumental solo] as a example
Hi, Even though I canceled the payment, money was deducted from my bank account ???
was it on day one of the subscription ?
for all things around billing please contact billing@suno.com to get a reliable answer to that problem. the discord crew has no direct internal connection to the suno team.
Okay thank you
lets say my plan ends on may 30 and lets say i pass that mark and its june 1st does that erase all my credits, or does it keep the credits and aswell if it does keep the credits can i still use all of those credits that i haven't used for commercial use
The day-to-day (free plan) or month-to-month (subscription) credits refreshes rather than accumulate because of resources allocation restraints (and is mentioned at the bottom of the subscription plan page).
Extra top up credits and special credit gifts are, well, special and are carried over unlike the former two
https://media.discordapp.net/attachments/1069381916492562585/1238249445192237106/image0.jpg
Canceling the plan retains what you had for the month but resets the next month iirc (as you are cancelling the renewal).
So if you had 2500 coins and it was the month after after you first subscribed, you would reset back to 50 as your plan switches over to the free plan
Thst is what I gathered from past convo. Contacting billing@suno.com can help clarify the matter but may take time to get a response
It hasn’t been singing “verse” or “chorus” for me, but any other directions (even standard ones like “outro” and “fade out”) are being sung more often than not. I tested last night with a free account so I wasn’t wasting more of my paid credits, and 9/10 generations were completely unusable due to skipping lyrics and singing directions.
But I was also experiencing some kind of degradation. I was trying to finish a song and in addition to the other issues the vocals would get more indistinct and quieter and quieter and quieter. Which is hilarious because I tried a million times to include different instructions to make it more intense/louder/instrumentally complex. Even when it wasn’t singing the instructions in brackets it was looking at them and saying, “Intense? Loud? How about I dial this down to a whisper with no music at all in the final verse.”
In the end I just reported a lot of them as bugs. And then it gave me ~15 credits back, which is something. In addition to wasting a lot on buggy gens, it also took 5 credits one time when I generated the full song. 🤷♀️
hey there 🙂 say, when i extend a song, is there a way to listen/download the full song, meaning part1 and the newly extended one? I know you can download both, use some 3rd party to stich them together etc, but i feel that's a immensely complicated way. is there some way to do this in the browser that i am not seeing?
oh, and suno is down currently?
On the part 2+, after it has finished generating, click ... > Get Whole Song.
A new generated track (no credits used) will appear on your library with the parts combined
https://www.suno.wiki/faq/making-music/how-do-i-make-the-song-longer/
didn't see that option, will look again when Suno isback online 🙂
[Tubas] [Snares] [Vocal Old male] [soviet folk] [Marches] [italy folk] [Marches] for old style audio if u want to change the tune just pick another country for the folk part thought i placed these in the style of music
It only appears on all parts EXCLUDING the base/part 1 and only upon when it is finished generating (which takes a few min) so a lot of people miss it initially.
https://suno.com is currently working for me now
it's back for me now too, but my cloudflare screenshot shows that it was a 524-error on Suno's side, maybe just for a minute or so^^
Is there a way to get it to keep generating the lyrics too?
Is there any plan to extend the song gen to allow for longer songs? I've been finding I have a lot of stuff that I have to try to cut into 2 sections
ok. now i have a official answer. there is no specific sorting implemented, so they should autosort roughly after the creation date.
its about an abusive relationship, its the most important line and can't change it... it will just throw the whole song out at that point. i guess censorship for the win
I just use another AI to generate the rest of the lyrics. Or the whole song, usually.
Seems like it should be an easy fix for Suno to make it easier to create complete songs no?
Easy fix, yes. They could just feed the previous lyrics in as a prompt when generating the next ones. However, that would cost more, and you might get less generations because of it. (Although you can apparently generate lyrics without spending credits.)
Also when you ask an AI to generate a full song, usually it requires a bunch of human editing to get it to follow the verse/chorus/bridge structure properly. Sometimes it'll just write a really long verse and forget to do a chorus, or if you stopped in the middle of the chorus it might not finish it and instead repeat the first verse.
Basically, the longer the pattern is, the less reliable it is, and it might deviate or repeat stuff it shouldn't. The reason it works is because it's short.
Hey can any Devs help me with something?
I've written a song about that kind of stuff, you just have to change around the lyrics without changing the meaning of the line.
Or just spell some words out phonetically
That works well too
Did anyone get in touch with you?
No they did not
I'm not a dev, I'm a discord mod, but maybe I can answer your question?
or help?
No it's a Suno technical thing. I just billed the wrong account and I'm hoping they can switch the credits or something
You already emailed billing@suno.com?
I didn't, no. I thought it'd be faster on discord from experience
For billing things, send them an email... They are usually pretty good about adjusting accounts.
Do I get anything bad for my billing not going through in the month or just end the subscription?
Because I'm fine with using my credits on the account I accidentally billed to instead for this month
You could set the acct to 'free' and when the month is up it should automatically switch over.
The current credits will not roll over..., (I'm not sure if that solves the problem)
Hello) Does anyone know if there is such a possibility in Suno to make one part of a song so that the music is from another one, generated earlier, but the lyrics are different?
I just wanna make sure my other account won't get suspended or anything for the billing not going through is all
no it wont, your account will show your new credits but will be red and you cant generate until its paid
Alr thanks
Not currently..., the closest we have now is to extend from early in the song after the verse and chorus have been heard, then edit in a DAW
That's a pity. I want the same music, but with different words. And I don't know how to realize it ... Can someone tell me how it can be done?
Thank you for your reply)
does the song open up with the music first before the lyrics start?
This part of the song starts with a couple or three notes of music, followed immediately by the words.
itll be tough, but you can extend "Continue From" from the very beginning, and mark it instrumental and see how close it can get to the original music, just make sure the timer doesn't overlap when the lyrics begin
Oh, that's a great idea) I will definitely use it) Thanks for the idea)
Though I have no idea how I'm going to use this part, which is my second part, then, even if I can create an instrumental part, glue it with the 12th part to extend the song and then put lyrics on it.... Could you please tell me if it would be possible?
your best friend right now in suno will be the "continue from" feature, I have had to get very creative with it, there is no way to merge seperate songs together (within suno) that dont stem from the same beginning
or consecutively stem from one another I should say
but if they follow the same beat pattern, its easy to merge it within a daw
Yes, and it's a shame that you can't take the music from one part and make a new part with different lyrics, but with the same music. And I wish the singing style was the same.
post it in the suggestions channel, you definitely arent the first to say that, and wont be the last
It's not working. I have taken on a model that is apparently not yet viable in Suno. All parts, 12 of them so far, have different style tags....
Ok, I'll give it a try). Thanks for the recommendations. It was helpful)
so you have made 12 consecutive generations of the same song?
are you able to post your song? might be easier for me to give you better recommendation on where to go with it
But I haven't tried gluing them together yet. Hopefully, when I create all the parts of the song, because I haven't finished yet, since I can't create the piece I wrote about above, I'll be able to glue them all together.... I'm not sure of the success of the venture yet.
You mean send you a link?
yeah, you can probably start your gen from a different point in your song and then cut out the beginning of the lyrics you dont want
But how can I do that - cut out the words? As I understand it, Suno does not split the excerpts into separate tracks: the music track and the lyrics track....
same way you would connect everything together, within a DAW
I dont mean stem out your song, mean actually remove the beginning
You could try to remove lyrics with RipXDaw, its a hit or a miss of how good the quality will be, but you cant feed that back into Suno
I haven't worked in a DAW yet. I don't know where to get one... I'm just starting to write songs) Although I've been writing lyrics and poems for songs for a long time now)
ahh gotcha, might be a bit overwhelming at first but its something worth trying to learn if this is something you enjoy doing
So, you can't do without a DAW) I get it)
i mean you can remove earlier portions of your song without one and even stem out your song without one, but connecting pieces together, youll need one
I don't really get it...
Thank you) I really enjoy writing songs, it's true) But I wish I could easily implement what I described above)
I've tried pasting a link to the passage, and the same style tag - all to no avail. So I'll try the option you suggested and see how it works))) Experiments!))) I love them so much))))
no problem, youll learn as you go, I've been messing with Suno for the last 9 months, alot has changed but you start to learn the little tricks that work best for what you like to create
That's great - you have a lot of experience with Suno))
Is there any way I can seek your advice in the future? Or is that too much?
I do! alot of headaches as well and learning curves, still do sometimes
and yeah I don't mind, if I am around, I dont mind helping out
Thank you! )
i'll try that
Hello, I have two questions, first, is it possible to edit the lyrics once the song is created? I have some songs that, because they were made in parts, the Lyrics at the end don't match, let's say there are some extra lines or missing lines.
and, second, it has happened to me several times, where I start a song just instrumental, then in the second part I add the Lyrics, but, several times that second part continues just as instrumental ignoring the lyrics, what bothers me is that I lost the credits for that part of the song, any fix or ideas< Thanks.
If you mean edit the lyrics in place, like change one word, then no, not yet. But you can use the extend feature to re-do a portion of the song and just keep doing it until you get a new section that will match up to the old one.
Add a voice to the prompt. I use "Solo female vocalist" for instance. Leave it off for the instrumental, or you'll get nonsense singing. But don't forget to add it on the extend.
how long dose this last when u run out of subscription credits but canceled the payment plan but the free points aint rolling in like normal ?
i wish there was better handling option for generated clips, like marking a bunch of clips together for deletion would be nice. RIght now, managing your clips is much like a nightmare.
i usually generated a bunch of clips and in a first step of organizing i like the clips that i like and start deleting everything else, and then i continue, but sometimes it would be nice to be able to do it in a single sweep.
Hi all, Can Suno generate isochronic tones? I tried to mixed it with other music style and it doesn't seems to work
I'd like to ask also about generating "Manyao" genre...
If i want a bridge where children are singing, how do i prompt this in the extension to get the AI to get this?
You can comma inside the tag, for example [Vocal, Bridge]
I'll paly around with that
who's good at the prompting here ?
what are you looking for?
I'm good at the genres I know...
like 99 problems of Jay-z. ^^
something punchy like it, minimal, u know. I don't know if you remember that song
😕 not a genre I know. I wouldn't have any useful advice.
but i appreciate the fact that u tried to help ^^
No, it's true, because most of the time, people just ignore the message
and that one : https://www.youtube.com/watch?v=7nxWP9BhI7w ?
Provided to YouTube by Universal Music Group
Roads · Portishead
Dummy
℗ 1994 Go! Discs Ltd.
Released on: 1994-01-01
Producer: Portishead
Producer: Adrian Utley
Composer Lyricist: Geoff Barrow
Composer Lyricist: Beth Gibbons
Composer Lyricist: Adrian Utley
Auto-generated by YouTube.
Hello, I've been using Suno to create a Historical Folk song. I'm really happy with the song (instrumental) it bought on, but the lyrics are very scrambled, and cannot be understood; this was a bug while generating where I switched on instrumental but it gave me some lyrics anyway, which were gibberish, and I attempted to make it right but it keeps bugging out to the point where some of my credits get refunded even. Is there any way I can improve the lyrics of my song while keeping my instrumental (what has been generated so far) intact?
use Extend from Time, and set the timestamp before the problem vocals.
How did you figure this out? Is there a list somewhere of the tags that you can use or the format that is understood? I've been stumbling across suggestions like this for months that are super powerful and seemingly not documented and would love to see everything that's possible in one place. Been hoping it would turn up in the Help section but it hasn't made it there yet.
there's a lot of starter hints on the wiki..., but after that we are basically inventing new ones all the time
https://www.suno.wiki/faq/metatags/
Is there any way to edit the lyrics shown on a final 'FULL SONG'? It seems to get messy when it's stitching together 6 different prompt parts where some lyrics had to be input a few times to get results.
I asked Microsoft Copilot to help you. This is what it came up with:
- Rap rock, hard rock, male vocals, aggressive, guitar riff, stripped-down beat
- Hip-hop, male voice, pulsating beat, classic 80's samples
- Rap, male artist, Rick Rubin production, guitar-driven
- Hard rock, male rapper, guitar riff, 80's sample staples
- Rap rock, male lead, aggressive beat, guitar riff
Please note that these are not intended as ready-to-roll, just-add-lyrics style prompts. They may be a starting point though.
This is the prompt I submitted to Copilot (just replace title and artist and reuse at will):
I need you to describe the music of a song. You should summarize its genre, voice style, prominent instruments, and every other peculiarity in a maximum of 120 characters. No need for a bulleted list, or even a complete sentence: just a comma-separated list of the most characterizing features of the song's music. Your output will be fed into a text box in a music library management software.
Some guidelines: start with the genre but omit the word "genre"; do not describe lyrics, just music; do mention the gender of the vocalist; do not name singers or players, briefly describe their style instead; do not name specific instrument makes or models, briefly describe the vibe they give to the song instead; if at all possible, no item in the list should be comprised of more than 3 words.
Please try 5 times, so that I can choose the result that fits best.
The song I want you to describe is "99 problems" by Jay-Z.
This is awesome! Way more useful than the Help link on the https://suno.com/create page. Would you ever think of replacing the link on that page with this one? I went back and tried to find a way to this page from there but couldn't. Seems that this would be much more useful to new users.
hi, how can I make the music to a specific length? for example 1 minute.
I think Udio has 320.
It does, but the actual audio quality is closer to 128kbps MP3.
we can't be that precise with time, unfortunately.
I can usually get very short songs by prompting for things like
crime tv show theme
or
video segment lead-in
apparently one of the ideas the ai has about TV is it has very short music. no guarantee how short, but for me it's almost always a short clip
In a DAW you may be able to cut a few bars from the song to get it closer to 1 min.
yes it has, be i prefer the vocal and creativity of suno.
Thank you so much for your message. i just saw it because i was making some generation to find :). Appreciate a lot, i'm gonna test
and thank you for the prompt
can I order it to give me only one variation and not 2 each time?
Hello
@amber maple What's the deal? @chrome jasper has a Science playlist on the Home page. He added my Math Is How We Relate Song and his playlist disappears from the Home page. Why??? What is the problem with this song? Why is it the only way he/I/we know there's an issue is because something doesn't appear on the Home page? This song is completely tame and at it's very worst the narrator contemplates if two dis-similar people should remain together. Thank you!
https://suno.com/song/a9e8d6fb-7df6-4bbe-a08c-d78525dc190a
quirky, sugar pop, glitch song. Listen and make your own with Suno.
seems tame.. will look into why that's happening
Thank you!
oh also make sure the song is set to public, that could have some effect
ok great
Hello! I'm using a Suno (Pro Account) to upload songs to Spotify via Tunecore, today I received this message for many of my songs, and they were blocked:
"Your release has been temporarily blocked in our front end review process. It appears that you are using the mastered instrumentals for these tracks (ie from the original artists' versions). Any time you sample any part of a recording you did not record yourself, you MUST have a license from the owner of that original recording. You cannot distribute any content through TuneCore that you do not hold 100% distribution rights for.
If these recordings do not sample the original master recordings, could you please confirm they are licensed recreations?"
I was searching the Suno website to see if there is a commercial license linked to my songs or something similar so I can send them to Tunecore, but I can't find anything.
Can someone from the Suno team help me with this? I don't know what I have to show to Tunecore so they unblock my songs.
i had 7500 credits remaning and they all vanished becuase i didnt extent the subscription, why?
yes, sorry, this is a known issue with the way our lyrics critic model interacts with the trending playlists feature. in brief, the lyrics critic has a certain false positive rate and the effect is magnified in a trending playlist because the lyrics critic has to be 100% correct in order for the playlist to trend
So what doesn't the "lyric critic" like about my song? How would I know?
Have you heard my song. Do you agree with the "lyric critic?"
The day-to-day (free plan) or month-to-month (subscription) credits refreshes rather than accumulate because of resources allocation restraints (and is mentioned at the bottom of the subscription plan page).
Extra top up credits and special credit gifts are, well, special and are carried over unlike the former two
https://media.discordapp.net/attachments/1069381916492562585/1238249445192237106/image0.jpg
thanks didnt see that
You are not the only one. From what I had observed, most AI services do the same (because the resources used is costly) so people learn the hard way to check the terms when it happens to them once (credits being refreshed rather than rolled over)
please, i need help. i cant for the life of my get the AI to repeat the instrumental intro melody in part 1 at the end of the song (part 3.) ive tried lots of things, what am i missing?
The issue is that the AI does not remember the original style used. What it does, for each extension, is study the style ~1 min prior to the continuation point to learn/mimic the styling you had (which is why extensions are 1 min long as the other min was reserved for studying).
So if you had a loudness or degradation issue within that minute, it would start becoming emphasized as that is what the AI picked up on and all future extends builds up the issue as long as it exists within the 1 min of continuing.
It also explains why the style can change when you vary the starting point by a few seconds
Since your part 1 is likely more than 1 min away, you are better off extending part 1 again and splice the new second part 2 to where you want using a DAW.
Or try adjusting the continue timestamp, a few seconds can change the pattern
Is the "lyric critic" constantly scanning playlists and evaluating or is there a tag somewhere on my song?
this may be a bug on our part, investigating
thanks, should i try entering an earlier time point then?
melody starts at 0;00, ends at 0:29
im extending from 1:28
I personally would extend from part 1 and use a DAW to splice it in manually (with an instrumental transition so it feels more natural).
Since the Melody is so far away, there is no guarantee you will will get it back with extend and would likely cost quite some credit
one thousand, actually
many thanks!! will try
Djent instrumental intro song. Listen and make your own with Suno.
thats the melody, by the way. i freaking need it at the end haha
In Suno docs:
for example, https://www.suno.wiki/faq/metatags/verse-and-chorus/
Use an ending tag like [Fade Out] or [Big Finish], or an instrumental/interlude tag
https://www.suno.wiki/faq/making-music/how-do-i-end-the-song/
I'm trying to create a song, but when I entered "/chirp" suno said: The application did not respond. Please help fix
/chirp has been retired
The website at https://suno.com is where you make music now 🙂
Oh, ty! Question, how come I can't create music when I click Custom Mode on the website
Once you're logged in, it should show how many credits you have (free users get 50 per day, each creation takes 10 credits) on the left. If you fill in the "style of music" and either enable "instrumental" or put some lyrics in, you should be able to press the Create button.
Thank you so much! :)
Is it possible to cut the last seconds or fix the lyrics?
https://suno.com/song/d3fbf1b7-5aa6-4084-ba17-cf1b9fb2c706
Rock, Pop, Jazz, Opera, Orchestral song. Listen and make your own with Suno.
Hi - has anyone had any luck switching styles mid-song? I am trying to make a song that starts with a old west/cowboy intro, but switches to heavy metal for the main part of the song. I have tried stating the first style in the original descriptor and switching to metal style in the extension (doesn't work), and I've tried putting the heavy metal style in the descriptor but adding a [western...] custom callout in the beginning, which doesn't work either. It doesn't seem like the AI knows what I want. It will either stay in western ignoring my changed style, or it will ignore my custom callouts and stay with the stated style. Before I blast through more credits I thought I'd ask you all. Thanks.
check this page about "a sudden style change"
https://www.suno.wiki/faq/making-music/extend-from-time/
its possible. the continue time plays a huge role. and i do style switching by just changing the styles around and continue. some had success introducing new elements via [metatags] but for me the extension stxle change often gives the best results. sometimes it takes some time until the new style emerges in the music, but if you generate enough parts you may get exactly what you need... here is a example i made where i switch styles about 5 times . its more electronic but shows where i extended and how the parts behave https://suno.com/song/b05c294e-39f9-487a-a3ab-115afe7ef141
uplifting tv commercial, 70s, tape, emotional, spoken word, chatter, sound fx, gunshot song. Listen and make your own with Suno.
Ok thanks - I'll try your suggestions.
Using a combination of the [instant switch] tag and the extend from 00:01 seems to be working. Thank you!
Uh. what? It says I got 20 credits but it says I need 10 credits to create. I'm so confused
refresh the page. sometimes a wrong value is cached
I don't know anything about this but I am very interested if anyone else does.
Hello, can you give me your sincere opinion on this: https://suno.com/song/ed476402-5169-432c-8dcc-20aad142b672 would really appreciate it.
love it
how many generation it takes to get to this version?
That one took me 4 or 5 extensions.
not bad
hello, how can i make nu metal music with slapping bass guitar?
If you doing it manually you can try adding the instrument your want in to the genre then make it specifically to the lyrics
i've try use music style, and meta tag, but still not giving a slap bass guitar
one of my nu metal prompt
https://suno.com/song/19784f59-4f88-4e5d-a87f-55a8a681624c
Meta tags is better to use just 3 words and then connect them with other Meta tags
like [turntable, slap bass] ?
Maybe more like [Turntable] [Slap bass]
hhmmm, K i'll try, thnx Tokrif
Just sharing what I know 😄
are there tags for changing tone of voice
lets say for an aggressive song
voice goes from normal to really intense high pitch and scratchy
at least you are willing to share information, thnxs
[High tone voice], maybe
combination of growling screaming and clean vocals, i use thats on music style
^ You can try clear [gender] vocals, that typically works and other prefixes like Sensual, smooth, etc.
it seems that putting things like 200 bpm or 100 bpm in tags dont really change the pace of the song
Nope but using the Italian tempo phrases should like prestissimo (as well as general terms like really fast)
Taken from Page 25 from the PDF on https://www.suno.wiki/faq/style-and-lyrics/prompts-for-ai-dance-music/
https://media.discordapp.net/attachments/1069381916492562585/1241776732663451668/Screenshot_20240519-113615_Drive.jpg?
I have a question I hope can be easily answered, I'm sure it can 🙂
I got a female lead singer, but I want a male singer to join in on the chorus, but it just don't want to do anything else than the female singer, and yes, the chorus is in there 😉
Getting duets... that is more luck than anything else. Try to use male and female sounding names in tags, that might help, but it is still hard to achieve.
A few people were able to do it but at an expense to their credits:
3Daizy's duet
MrTree's duet
Wiki entry
Again, getting it working on purpose is hard. Alternative methods:
-
It is easier to make a duet by doing something like generating instrumental part 1 then Extend it twice (one extension where you specify for male vocals in the styles field, the other for females), or something similar for each alternating part, then edit them together in a DAW.
-
Even easier (and cheaper) is to make a full track then swap the voices gender using other AI tools like kits.ai (see #🗣️┃chill-chat message and #💡┃suggestions-web message ).
You can use a DAW to edit which voice track plays when.
interesting, I'll see what I can come up with, thx @hollow hare
is it possible to have it generate the exact same song again, just with different voices then?
if you split the original song into 1) instrumental and 2) vocals
and then do voice-to-voice transform on 2)
and then merge 1) and 2')
@lament escarp and how would I do such a thing, remember I'm literally just started playing with suno, still playing around with its features 😉
question: why do the images of the songs not show anymore if you paste a songlink here?
Hello, I have created a song and used the Extend mode very often; because of that, the text shown after clicking on my song consists of many repeating parts. I plan to make my song public, and I want to make the text description of the song more clear for other people who will listen to my song. How can I do that?
How can I make Suno play a piano piece with piano sheet music?
Hi guys, I have a question. As we have exclusive rights on the tracks we have made, what do you think on putting your own name as lyricist and composer when you launch those track publicly and commercially?
I always wrote my name on the lyricist and always on the singer by the first name Suno and by the last name Ai.
Thanks for this, good idea!
Hi all was wondering how could i specifiy a certan tempo for the song that i want generated ?
Why does it keep making two identical songs?
they are not exactly identical. the 2 songs will give you a choice to continue. one is more straight the other one more creative.
I realized this after a few tries.
But the continuation does not use the same instruments and tone. Is there a better way to make it more similar?
i influence the tempo by adding uptempo, midtempo or downtempo to the style. others use words to describe the tempo, often with italian classical music terms
And is there a way to just modify the text in a finished song? Keeping the same voice and rythm?
if you use the same styles and use a time to continue where the music is playing it should often go on in the way it was.
if its finished its finished. only a regeneration can save that in suno. or you use external tools like a daw to cut the songs how you want them
Thank you.
I am just making a song about my cats. 😄 Did not want to spend too much time on it. But one line bugs me, where it garbled some lines. I guess I'll look for an editor.
i often just see the mistakes as a artistic choice of the ai and live with that 😄 to get a random generator to do things 100 percent can be quite credit intensive 😏
sometimes it feels like the voice aspect of Suno is broken though. it has so many vocal bugs where it screws up words, starts blabbering jibberish or just sounds like it has gotten lost in the matrix when it becomes all distorted for no reason. its definitely not taking this stuff from real music, its bugs and glitches with the voice element of it. Whatever Udio are doing for voices, Suno need to copy as its near flawless in comparison.
what happened to the website?
if any dev is here,
how's the training coming along for v4?
i guess you have tons of training data for it by now
are there any tips for metatags and things one can put into styles etc? there's not really a faq or good documentation of things that are known to (semi) work anywhere.
@agile veldt thought I might reply here instead of cluttering the bug channel.
Encodec is a neural "compression" codec.
It is basically the representation that bark knows audio as. Bark creates these tokens from a text prompt and these tokens are then "decompressed" by Encodec to spit out the final audio file.
I have a hunch that the suno music model is very similar just higher quality (sample rate and stereo).
But I can not be fully sure though, but it would make sense to reuse the architecture.
There's the user wiki. I know I've always wished it had more, but lots of stuff I get to work once, doesn't work a second time, so I'd feel an idiot documenting it.
But really, it's an AI. Lot's of things work, or at least influence it towards something. I mean it's a bit of an art in itself. And it's highly dependent on exactly what you want, so my ideas might not work for you.
I've put stuff in sometimes that I'm like "No way will this do anything" and it makes something amazing. But nothing is really programmed so there's no list of verbs or anything.
Any follow-up on this?
Hi all, anyone had experience deploying Bark on a AWS Sagemaker endpoint? Been trying for days with no luck, i keep getting a connection error when trying to do the inference.
It’s a bug in one of our vendors’ APIs, still investigating how to remediate
Hello, is there any information on how does the models used by Suno for music generation work? The results are astonishing and I won't be able to sleep till I read more about it!
😆
https://youtu.be/lj8utBd_IUM might be able to help you, and we also have a community wiki here: https://suno.wiki
Hello, I'm 3Daizy, a moderator and guide at Suno. Welcome to my channel, where I share my journey and experiences with AI-powered music creation. Whether you're a budding musician curious about AI or an experienced artist looking to innovate, you'll find useful insights and practical advice right here.
Join me as we navigate the fascinating int...
🤞
@glacial scarab that makes a lot of sense, thanks for the clarification. Even for an audio minded developer, all these AI terms are tricky to read up on – often being a mix of python (which i don't know) and super high level maths. 😅
Great news from suno
i'm excited about the next update. If they fixed the echo and the downgrade of the vocal over the continue. It will be fire
is there a way to make consistently present vocal?
announcment sounds awesome, please fix bad quality at end of songs
...I even tried [Dracula appears] in a song once to get it to do a creepy pipe organ thing. It didn't work though. 😦
But I mean, I once did "haunted washing machine" as a style and got hip-hop. So anything is possible. 😛
[Call and Response Accordion and Fiddle] worked once... [Playful Vocal Improvisation] worked once for scatting... Imagination is the limit! But it's also very difficult to reproduce anything
In one song I found that the AI often mispronounced "cower and obey" so I told it to sing "cow were and obey" lol
And I've got another song where [rock intro] works well, but the intro also has nonsense words in it.
So I think in future if I want it to actually do an intro instead of skipping straight to the lyrics, I'll just put in some ooohhs or la la la or something.
This typically works for me when all metatags fail
Though in the case of this song, the intro bit is just random "zero zero zero one" because it fits thematically. I'd kinda like to make it be a creepy robotic voice but that mostly doesn't work.
It's hard to get the proper mix of rock with weird glitchy synth stuff.
How do i solve the problem when the ai wants to read the [nametags] even when i dont have to many of them in the song
Roll the dice again. It usually only does it sometimes
done it alot happeneds on every part
Clown,Fair, [Dramatic] [Cinematic ] [Grand voice] style of music tags
one example
Clown,Fair, [Dramatic] [Cinematic ] [Grand voice] song. Listen and make your own with Suno.
/chrip
Chirp is retired. Please use https://www.suno.com to make music.
Also have a look at the community Wiki for a guide on how to use SUNO. https://www.suno.wiki/ ✌️
where i can find, words that are flagged?
It got censored from ChatGPT (which moderates the lyrics in Suno) not liking something. Ask OpenAI's ChatGPT itself about "List what censors the following: [content]" and it should tell you. (If the reply gets censored, ask "what changes to make to avoid censorship" after and it should work).
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.
I remember people getting censored for using bones, ribcage, and skeletons in beign situations but they made ChatGPT assume some sort of horrific tale.
Did someone manage to use metatags to successfully indicate to the ai that the verse is in a different language? I've tried some stuff and it always fucks up the first line but then switches over successfully
Like I've had lyrics with alternating German and English verses and it didn't do too badly but the first line of German was always so so very much English pronounced lol
I'm honestly surprised it worked as well as it did, just was wondering if there could be a way to nudge it more somehow
i always get a Random german song in the start of the song and then english :S
Does suno understand instruments and generate them one by one, or is it creating fully composed sections?
I'm thinking in terms of extracting MIDI or something, but it doesnt seem like it would be possible.
There are of course tools that can do that, which would be interesting to try
My understanding is that it's trained on fully mixed audio, and generates fully mixed audio. There's no separate parts (stems) involved (otherwise I expect it'd be a lot slower.) As far as I'm aware, there's no MIDI involved, so the only way to get this would be to either transcribe it by-ear or use a tool to do this. I've found that there's notes that are either barely audible or just not clearly defined so it makes it quite difficult.
I think they'd need to generate separate stems first, making it easier to identify what notes are playing, or somehow generate MIDI alongside the audio.
Bear in mind though that MIDI would only give you the notes and maybe a little bit of performance information, it won't capture how it actually sounds.
where can i get support, i used premium package and wrote my own song words and use tunecore but they ask about licence or so, how can i solve this ?
Suno does not provide a license. If tunecore is making a fuss about your song, I personally advice trying a different one. A lot of people use distrokid and I use ditto. Both seem cool about AI music.
have you had payouts from Ditto yet? They owe me around $2300 accrued since 2019, I use Soundrop now
Yes, I personally been using Ditto since 2019 too. I never had any problems like that.
who did you use before Ditto?
I've been using distrokid too... but I like ditto better
I used CD Baby, then Ditto, then DIstrokid and now Soundrop
I personally like that distrokid offers "distrolock". Which is great to make sure your music is copyright protected while not being released yet.
how long have you been using AI in Music?
Since november 2023
cool, my synthesizers miss me I think, I am addicted to Suno 😄
You can remake the songs? Then you have he best of both worlds 😉
true 
I'm kinda trying to do that. It's a lot of work. I can't believe I'm paying to use an AI to give me work to do. Bah! 😄
i need help
Is this still about the license? That is already answered in 2 channels. 🙂
i spend all of my creadit and i push my song to tunecore and they hold my song because licence or so but i am in my account, where can i find my song licence as i wrote its words my self and i used premium membership
Like I said.. you've been answered this. But if you are not satisfied with this answer, you will have to send an email to support@suno.com. We cannot give you more support than we already did.
ther is over 12500 songs i generated and i guess there is no search option, how i will find this song?
you don't need a license, you need a better distributor!
Yep.. I told him that 😁
😄
I know that pain.. but at the moment you will have to do this manually.. they are working on a search system
what about licence, they hold 3 of my song
You've been told this 3 times now. There is no license
there is no license, you don't need a license, you need a better distributor
if no licence i need to tell something to them
Your release has been temporarily blocked in our front end review process. It appears that you are using the mastered instrumentals for these tracks (ie from the original artists' versions). Any time you sample any part of a recording you did not record yourself, you MUST have a license from the owner of that original recording. You cannot distribute any content through TuneCore that you do not hold 100% distribution rights for.
If these recordings do not sample the original master recordings, could you please confirm they are licensed recreations?
As a reminder, by agreeing to the TuneCore Terms and Conditions, you have: (a) legally represented to TuneCore that you control or have obtained all rights required to exploit the Recordings (including the art/images you associate with such Recordings); (b) agreed that we may, in our sole discretion, disable access to any master recordings or other materials with respect to which we receive a complaint; and (c) agreed that we shall have the right to deduct from your account or charge your credit card a minimum of $300 to offset the costs of associated legal fees, if necessary, and also deduct any and all revenues from your account which are received in connection with Recordings if we believe, in our good faith discretion, such Recordings violate the TuneCore Terms and Conditions.
Please let us know if you have any questions.
thats what they send
We are talking in circles. It's the distributor who is the problem.
you're responsible to learn about the products you're using man
Again: send an email to support@suno.com. We are moderators, not support. But it's seriously a distributor problem. Not a suno problem.
People had less issue with DistroKid and Ditto (but they will likely deny you if the song sounds very AI like according to some that got denied)
they are not much ai like as i spend over 12500 creadit and choosed juts a few
see tunecore: Your release has been temporarily blocked in our front end review process. It appears that you are using the mastered instrumentals for these tracks (ie from the original artists' versions). Any time you sample any part of a recording you did not record yourself, you MUST have a license from the owner of that original recording. You cannot distribute any content through TuneCore that you do not hold 100% distribution rights for.
means its ok i guess
people are like against ai but they can not resist more... world changed, time changed... it is ai world now.
so they have to eccept.
there are no samples to clear, you need to inform Tunecore that the composition is original work or like others have been saying use a different distributor like DistroKid
if not today then tomorow
yes i wrote all of my songs even spends hours of time
i wrote them
i will ask my refund and use distrokid then
thank you
all
Good luck @fathom relic !
thanks
by the way i think in a few years or less, the sound quality will be prefect...
i like suno, great work
fascinatingly, I had a bug again where my credits weren't refilled, and when I checked, my final creation was yesterday at 22:29, which was also the case the last time this occurred
I'm trying to figure out if this is a coincidence
I won't be able to try and test this theory today because i have no credits but
what is the max character length in the prompt box?
I think 120 char?
or wait
do you mean style or
because I think that the style box is roughly 120 chars
prompt
that one I don't know I'm so sorry
May be dumb question, but I made an old song in the beta chanel before tthey had the website going. It's not in my library and at the time I only saved the video. Obviously the beta channels are gone so I can't search for my mention, but I found it my date ias a chirp bot like. https://cdn.discordapp.com/attachments/1132066051962703992/1186274890114928670/chirp.mp4?ex=664ddc95&is=664c8b15&hm=89233c6d172c24cd1636ea1e920528a40a2d8fac2bdfe7f1f53bca2af2ee6edf& #😍┃v1-songs message Is there any part of this URL that I can paste as the song identifier? So as to replace the identifier in their ural format eg : https://suno.com/song/7f3f3fe3-dd69-4968-9fdb-8b170c6ab482
Yeah.. let's not translate that 😉
what did it say?
200 characters for the prompt
anyone been able to make an instrumental intro before singing, can't seem to get it done?
custom mode or no?
Look at the username pinged, you can guess the context
oh that was just for the prompt box
N word
oh
let me check for the "custom" lyrics box
no need
i was asking about the song description box
1250
lyrics box is 1250 and style of music is 200
lol i was sitting here literally checking the limits that I didn't know by pressing "x" until it stopped filling and then using a character counter XD
oh
Where there is a will, there is a way
"I'm the type of person who really likes to test the limits" (not really lol)
tried extracting the guitar on one of my metal songs, did not work out 😄
Anyone has a nice and clear song i can try?
clear on the guitar parts?
you could try generating something accoustic with guitar only and see if something in classic style or something generates less "noise" from different instrumentation?
you could also try like piano solos? idk "accoustic piano recital" sounds like a very boring prompt but it could work in theory? I have no actual idea though >_>
i'll try tomorrow 😄
the other thing is, if you can hum the 'melody' you hear, I think there are some programs to make midi filles or something weird like that
I'm not sure what your exact goal is, I'm sorry
yes, for sure. Oh im just tinkering, seeing what it does. It generates a lot of weird notes in a thick mix
ahhhh I see
to some extent, if I was deeply in love with a song, I'd ask one of my friends who plays guitar to just tell me the chords etc they're using
Can we get an indicator on a song if it's been flagged by the "lyric critic?"
?
lyric critic?
I had an entry say once something like "the song has been submitted for moderation due to content" or something like that
which was actually wild because it was not only not explicit/offensive but it was the same lyrics I'd been using for multiple generations
(I ended up asking chatgpt which part of it probably caused a flag, because I was just like "????")
It got censored from ChatGPT (which moderates the lyrics in Suno) not liking something. Ask OpenAI's ChatGPT itself about "List what censors the following: [content]" and it should tell you. (If the reply gets censored, ask "what changes to make to avoid censorship" after and it should work).
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.
I remember people getting censored for using bones, ribcage, and skeletons in beign situations but they made ChatGPT assume some sort of horrific tale.
I was just confused because even if it WAS censor-worthy, why censor it only the one time, and during none of the other generations? but yes, I know
ironically I had a song that included bones and it was cool with it
(In all fairness, it was actually a song about playing Stardew, so....)
See 2nd paragraph
yeah, but the thing was, I didn't make a change
not even a small one
and later it was fine with it again
but yeah, I ended up asking chatgpt itself what it thought the problem was
and I was like "yeah okay if I squint and look at things sideways I guess....?"
It's AI, it's not statically checking something, it is a dynamic check and likely that RNG assumed something bad 🤷♂️
I actually don't know how the back end of AI works, much to my chagrin
I understand things like RNG but
when I was studying programming/computer science and asked someone about AI, they were like "how is your multivariable calculus" or something like that, and I was like "okay, so.... later...."
My song keeps hitting moderation errors, but not always, what's up with that? Did it generate a song that ended up being an existing song? XD
yeah
Nothing cray: Orchestral pop, jazz, dramatic ballad, timeless, classic, modern influences, strings, brass, guitar riff, piano, percussion, dramatic, suspenseful, mysterious, bold, theatrical, powerful vocals
[verse]
Silent whispers, secret deals,
Hidden in the night, where nothing’s real.
In the shadows, he makes his stand,
A lone figure, the upper hand.
[pre-chorus]
Eyes on the prize, in a world of spies,
Every move counts in his dangerous lies.
Moderation error mean your lyrics are too graphic or too adult mostly.
It got censored from ChatGPT (which moderates the lyrics in Suno) not liking something. Ask OpenAI's ChatGPT itself about "List what censors the following: [content]" and it should tell you. (If the reply gets censored, ask "what changes to make to avoid censorship" after and it should work).
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.
I remember people getting censored for using bones, ribcage, and skeletons in beign situations but they made ChatGPT assume some sort of horrific tale.
yeah I have nothing for this. Ask chatgpt which line it thinks is generating concern, and replace that line with just
words that sound the same
like "say can you see" would be maybe
"sei ken U-C"
obviously don't use this for evil but like
if you know that your line is not actually adult content or super dark
and the computer thinks it is
just emulate the sounds
and hopefully me making this suggestion doesn't make me sound very sketchy XD
Sounds about right. In a general gist, AI are great at learning patterns, typically using complex mathematical formulas.
They do not understand words per se but pick up their own pattern with assigned weights (and good ol' supervision training).
I tried to study it before and was like "yeah, I think I will come back to you when I have more than a week free..."
what's distrolock ?
The website tells it best. But it is like I said before. You can protect your music even if you do not release it (yet).
yes this is definitely future work, it wasn't much of an issue before but trending playlists have kind of forced our hand
I don't like it when people get snagged by it, but OTOH there's a long tail of extremely low quality stuff that you guys never see
LOL, I see lots of low quality stuff. Oh you meant stuff I didn't write...
I keep trying to get my song to have instrumental before the first verse, but have tried so many things right now from intro, to instrumental, style inside the intro or start, pause, and several others, but nothing happens before it instantly starts on the first verse, any suggestions?
Make an instrumental song as the first segment. Then continue from at around 15 seconds in with the words you want.
so [verse] and then [instrumental] just beneath ?
Then when that completes, if you like it, "Extend" it from wherever sounds good with the lyrics.
You can have other instrumental tags as well. Look at the works by RayanAIR like https://suno.com/song/a98da392-ee0d-4fad-ba9b-937f0c9c5a91
neurofunk drum and bass song. Listen and make your own with Suno.
i get moderation errors for my extending songtext. if i ask openai about it, i always get something went wrong
I like using the following g in my lyric test:
[Melodic Instrumental]
[Build]
[Drop]
[Instrumental break]
[Fade out, End]```
Like with https://suno.com/song/46531a71-c9de-45c7-bf6f-52e72f984b34
Fusion of Mozart and Latin Folk Ballad, Epic Boss Theme Villain Theme and Hero Theme song. Listen and make your own with Suno.
Could be a simple browser cache issue. Erasing cache or using private mode or a new browser may work.
Alternatively, you can DM me the lyrics and I will take a look on my end
yeah that might be it, ran out of storage an hour ago
Glad an indicator is on the horizon. Have you guys also considered an "appeal" process once everyone sees them? Guess I would be the guinea pig for said process; including whether the flag can be undone and how.
is it possible to lower the music?, often I can't hear the singers voice
I had some luck with "clear female vocals" in the metadata style field. It seems to give higher vocal to instrument ratio sometimes.
maybe vocals works better than voice, that's the only difference I have
It's never guaranteed. But just something worked for me in the past
Hey for those who have distributed to spotify thru distro, is sound ok if u dont master a song or pick any of the leveling choices ? I have no mastering skills and i feel like mixea is screwing up the sound.
You need to offer it in 320 kbps in WAV, and mastering is pretty much a must tbh. There are a lot of videos on youtube how can teach you. So I would just look at that. And there are some AI who promise to be able to master your songs, but they mostly do a really poor job.
Thanks for this info.
Thanks
Candidly, I think it's safe to say that I'm as dissatisfied with this feature as it currently stands as you are. Rather than designing an appeals process around it, though, I'd like to do away with it entirely and replace it with something a lot smarter. We're still thinking through what will look like.
An appeals process would be...messy
has anyone been experience website issues today? I've had problem with renaming renders that get reverted back to their original name. my "Create" page has been slowly emptying itself of all the latest renders I've been messing with throughout the day to now there's only the latest 2 in there. and on occassion my library has greyed out the paging buttons acting like I dont have more than 1 page of renders
What I'm doing right now is messy. Appeals are the next logical step after you alert someone that their song has been flagged...they may want to challenge. I'm just prepping SUNO for the next thought after an indicator. Without appealing, how else can we prove false positives? Here's the back story #📚┃suno-school message .
Emptying of the create page happens normally after a while, but the rest I have not experienced myself. This would be a great post for #🪲┃bugs-web
If you by any chance have video of it would be a big help as well!
some of it has kind of resolved itself after refreshing the site. I know Create only holds so many samples before older ones drop off, but i have had renders I made within the last 2 hours that have just completely cleared out of Create
It could be that the team is working on that. I have no sight on that, since I am a moderator and not a staff member. But any feedback is welcome. And the best channel for that is #🪲┃bugs-web in your case. And they really appreciate your reports.
Right I get that, but for you it's 1 song, for Suno it would be 15, 30, 100 or more since other users would also have issues. Trying to fix the current system may just be a quagmire-in-waiting so the suggestion that they'd fully replace it with something else, especially since they aren't happy with how it is now makes sense to me
and that's before you get to the social aspects of appealing and getting denied
You're completely right, mind, your song shouldn't be shown by an incomplete 'critic' to be cut from the main page
Are there other fields we can alter for generation output? Or is it just, a click and pray that it does something you like?
Anyone seeing an issue where Suno is skipping the beggining of a song? So I'm making an instrumental, but it generates essentially the middle of the track, not the beginning or an intro.
I have it happen a lot, yeah. Skipping the first line, some times two. Some times only the first word. But it's always infuriating when the rest sounds great.
Yeah, and I feel like tahts fairly recent. Like in the last month or so it started?
Custom mode is as flexible as we can get but the AI can be surprisingly more accommodating than you think (when it wants to be).
Look at the metatags RayanAIR uses in https://suno.com/song/a98da392-ee0d-4fad-ba9b-937f0c9c5a91
And if you were not aware, check out https://suno.wiki
Eh, I have noticed it happen before that. But it was a LOT less common.
I've been having this happen very consistently last 2 days
some times it will drop into a sort of intro section after like 10 - 15 seconds of starting in the middle of a song and then builds back up, sometimes it doesn't and jus continues without any real intro
Huh, and now I'm getting flagged for moderation, on the exact same input I used before.
I've also been experiencing some really formant-y and kind low quality vocals this week as well, like a total 180 from some of the great stuff I was getting out of it last week (and several weeks before)
I get the scale of the issue and that's part of the reason why I raised the question. I agree that better a.i. would make better judgement calls but, once you flag something people will inherently want to question it. SUNO could choose to just let the a.i. do it's thing and ignore any/all appeals but, at the risk of customer service. Whether or not there's an appeal process people will find a way to appeal...like I am. 🙂
Hey anyone able to enlighten me on why my track didn't appear in 'NEW' section, if I share the link?
have you set it to public?
Yeah, and it has extensions. (is a 'whole song').
I watched the new section about 20 minutes after publishing, for about an hour checking every few minutes.
Shadow flagged?
Hmm. I hope not... :/
It's like my 2nd song lol
And I don't think I'm using any questionable subjects.
Could be...you wouldn't know.
Did you listen to mine?
It's shadow flagged.
Confirmed by SUNO.
I like to share a post Pat shared a while ago about this: When you think about what goes in to curating New, you're probably thinking mostly of your own songs, which are no doubt well-deserving of fame. But think for a moment about the long left tail of songs that other users generate, and whether you want normies, investors, or your parents running into that when they check out the front page. Explore has to be a privilege, not a right. (#💬┃general-chat message)
Interesting. So is that a per song basis? or is your whole account shadow banned now?
Seems tame to me too.
Just the song
I doubt you are shadow banned
This is the song https://suno.com/song/b2315a4c-055b-4492-bf97-cdb51640ccab if it matters.
I understand and respect your opinion. But they are shadow flagging and if you heard my song you would understand my incredulity. I'm not trying to make this all about my song but, SUNO should be able to undo a CLEARLY false positive.
We cannot expect to all appear on New. We almost have 400 k people in discord alone.
Oh no no, I figure I'm not. just wondering if it happens on a song by song basis. I'm fine if single tracks get weeded out for reasons. Just don't want to compromise my whole account. Not trying to do anything bad ❤️
Definitely not begging to be seen ❤️ Just curious on the ins and outs of it! Completely understand there are tons of users 🙂
I really wouldn’t worry about that. I am sure if that was a thing we heard about it for other reasons with reports we made about other songs.
Mine wasn't really about being NEW, it was about my casual song knocking another person's playlist off the Explore page....understanding it's a privilege.
Guys is there anyway i can prompt the AI that i want Guitar and Singer play together whole team the verse is going everytime i tell him its not playing the guitar only singer any fix?
Is there a list anywhere of regulations so one can understand lyrics/terms that might flag the filters? I do listen to explicit music in my day to day, so it's possible I may be influenced in that direction. Want to make sure I know where to draw my lines.
And yes I understand the discord server is PG13 ❤️
There isn’t (yet). It’s an algorithm and I assume they are not sharing it to avoid gaming.
This is how it usually works, yeah
ahh yeah. that's fair.
If you lay out the rules directly, unfortunately people will ruin it for everyone else
well final question then, if we do step over a line, do we get warned? or is it just a no questions asked ban?
It isn't like that, no
Just like with Google. 😄
It takes a lot to be banned
It takes a lot to be kicked
And from Suno itself, the devs have stated repeatedly that they don't do it unless you really earn it
Use your imagination on what that would take
Just to be clear, im not trying to test getting banned :). I have good intentions ❤️ I love suno. Been following for a year now. And am a musician myself as well. So I'm all in and happy to use the tool and show my love and support.
I know tony has proclaimed as much re: the stuff he's had to listen to
Cheers! I appreciate everyone's input ❤️
tony = one of the devs
That's part of my issue.
First, your song generation could be stopped with a warning because the a.i. doesn't like the lyrics.
In my scenario, I never ran into any warnings, never saw any problems. Only found out there's a hidden flag on my song from @chrome jasper
At least it's just the song :). Could be worse!
How do I know it's not more of my songs?
Im having a problem everytime i try to tell the AI to play a guitar while a guy is singing its not happening any tips ? Help please
Just playing the RNG game 🙂
roll the dice
Anthemic Rap Rock song. Listen and make your own with Suno.
If I regenerate/recreate the song would it still get flagged?
sounds shadow flagged to me /s
joke?
Same here. Almost every one now. And t's burning up my credits.... 🤮
I also was getting really good consistent instrumental starts for a particular sound with a set of style tags I was using. and now, this week all of a sudden it's completely lost the sound and its gone a completely different direction
- yes we can undo it, and ceteris paribus I'm happy to. send me the song id and I can do this one but I can't make a precedent of fixing these one-offs: I'd rather just redesign the whole thing.
a9e8d6fb-7df6-4bbe-a08c-d78525dc190a 😉
- to be clear, "shadow-banning" doesn't affect creation or sharing, it just affects what shows up on the homepage. And the homepage is a benevolent dictatorship, albeit one that runs on a lot of benign neglect
Thank you!! 🫡
@icy kindle How can i get into suno team ? any interviews or something else ?
Does anyone know when the suno app will be available
i know that its being worked on or at least its on the plan of the devs. sadly i have no closer infos
Question.. How do I get the AI to do a longer guitar solo? It's only giving me like 5 to 8 sec. I'd like to have at least a 15 to 20 second solo for the song
have you tried like [guitar interlude] or [guitar bridge] ... things like that?
no I haven't. I'll give that a try
Yea I hope it comes soon
That worked! Thanks
How do you change the flow of lyrics?
I want to make a song with lyrics in the flow style of "a boy named sue" by johnny cash.
🎲
I've been extending a song and now it is somewhy, in all those specific segment parts, the audio starts off almost mute, practically a whisper, and takes an absurd amount of time like this until finally being audible (when using very high volumes) how can I fix this?
The best you can do for now is experimet with different Extend From times, try to find a time where the song is loudest and doesn't seem to be decreasing in volume, and extend from those locations olyh.
Ima try that tmr since I'm out of credits today
If you have no other option because the silenced segments are intermittent (not just at the end) it's a bit trickier. You have to manipulate the song history and produce a clip start time and extend from time that omits a decent amount of the problematic segments, extend it (which will lose some long term coherency) and then Get Whole Song to put it back together. I don't know what Suno is working on, but a cool feature would make that that workflow easier, since it's not related to the model itself, should just be surface level software stuff.
When I hear a song on the explore page, is there a way to save it to my playlist? I hear so many that I like, but can't save them...
Is there any place where I can find promts ( I thnk it's called ) that Suno understands? example [verse] , [chorus], [bridge]
you can find some in the community wiki: https://suno.wiki
Anyone knows what causing flagging ?
You can ask ChatGPT, what's the problem with your lyrics, since ChatGPT is the gatekeeper.
Usually it is too graphic or too adult
Whats funny id chat gpt made lyrics for this
Ha.. well, I cannot explain that lol..
Its in Polish military song about Cadia defence
Then only the title and style of music is left
I think its to cool
😹
It sounds realy good i have chills lestining to this
Propably becuse of thus
Its giving me vibes of murdering chaos spawns
How did you generate it with Chatgpt? In Suno, or directly in ChatGPT?
I made it in chat gpt
And then coppied it in suno
I asked chat to write me song
About cadia defence
Ah.. is it more than 1250 characters?
I think its near 1250
Are you able to paste the whole lyric, or did you have to chop it into pieces?
I think word holy war can be offensive
it's about context.. not words
I chop it in to pieces i just put link with song
dark Metal Army Choir song. Listen and make your own with Suno.
That is likely the problem. You will have to repost that part back in to chatgpt and ask why it is flagged. The context is different when you have it in pieces
yeah I know staff would like to see it different too.. but atm ChatGPT is the best solution to use for moderation 🤷♀️
Sure but im okay with the song how it is its ending pretty okay so it doesnt need editing or ending
Not sure what you mean? I though your song is stuck on moderation? I mean, if it is, your only option atm is editing the song.
I just wanted add more lyrics becuse its only 2:35
Ah okay... 😐