#šŖ²ābugs-web
1 messages Ā· Page 4 of 1
Having same error, also for no reason
we'll look into that one
I have already paid, but there is still no credit use command
Please contact billing@suno.ai
dont know if this is reported: when generating Chinese songs, the ai generates a vocal half-mandarin and half-cantonese
Posting here since it might be a bug:
Iāve done two months of the $30 tier and so have plenty of credits to use for the next while, and I want to sit with them without accruing more. However, when I canceled my subscription last month to try this, I lost all my purchased credits. Is that supposed to be how it works?
I took the lyrics from the song generated yesterday because I wanted to continue
Sorry, we are similar to Midjourney in this way where plan credits reset each month, but topup/extra credits do not. If you email billing@suno.ai we can try to figure something out
Is there any chance that partial continuations are rendered with a better quality than the resulting merged songs? Might be just the (already discussed) volume issues, but I keep noticing that I find continuations better sounding
(By āvolume issuesā I mean that as continuations are louder, they might be subjectively perceived as better sounding.)
Hello, I subscribed to Suno Pro and when I try to generate sounds, they tell me that I no longer have a subscription on Discord
plz help
A bug like this is possible, sure. If you can download the audio of both clips and compare in an editor it should be pretty conclusive if they are even different.
I'm trying to generate a full song from one of my 1:00 clips but it keeps generating a full version of a different style.
It's also not even generating the full song.
Do you mean "get whole song" or you are trying to continue it?
@neat sail btw thanks for suggesting using accented vowels, it works much better
Get the Whole Song keeps generating the wrong song- it keeps pulling from a different version than the song I select. It uses the correct cover art, but the music is a completely different file.
Additionally, Continuing A Song isn't generating the full lyrics.
Were you trying capital letters in Enlish prompts or another language?
The problem with a continue is likely not a bug exactly, just how the AI is sometimes. But the Get Whole Song could be a bug. If you have multiple tabs open at once on Suno maybe try closing everything down, restart browser, and continue again. Although it is super easy to mix up Chirp clips, sometimes different clips get almost identical album art, so that is also a possibility.
In Russian
Has anyone succeeded doing purely spoken speech, maybe with some ambient background? I need to put ārapā to make it not sing it but thatās a different thing of course
Could it be that previous prompts are āleakingā into new ones? Eg I ask to generate something with say a cello. Then I change the style completely but sometimes the new generation will also feature a cello
My generated tracks disappeared and only the last 10 or so stayed. WHY?? Where are the other tracks?
ah... why? š make it a endless dynamic list rather?
took me a while to grasp this
Sure it's a blast. empty style prompt is usually all you need, and lyrics that look like a script
Nice job with multi voices
gave her trance and deep techno prompt, she created pop and R&B
Anyone having an issue where all their generations are getting refunded?
I can't continue a song. It just keeps generating a new one with the lyric field and new music style.
That doesn't work. It keeps generating a new song instead. It won't continue the previously generated one.
Please email support@suno.ai regarding this issue.
Hehe, I wouldn't mind that the library is in pages if we had a way to jump to a specific page instead of click-click-click-click. But I'm sure they will fix that.
Hello, I'm using Suno AI, and after changing the lyrics with 'Continue from this Clip' and remixing a song that I like, when I press 'Get Whole Song,' the result is the original song before changing the lyrics, not the remixed version. What could be the issue?
I just wish we could access them via link (like ?page=2) instead of having to scroll manually every time you update the page
"Get Whole Song" just joins the continuation with the previous fragment
100% this is happening. I don't know if it's because I remixed or extended before but I had to fully close the app and come back to get it to accept the new style. Even though I had 100% different text in the style input, it kept using almost the exact same music from previous creations, with totally different styles.
i just had a prompt get blocked, but the funny part is that it blocked 1 generation but made the 2nd lol
so i think i just used 5 credits instead of the usual 10
oh im sorry i didnt mean to post this in bugs, this probably isnt a bug i just thought i was in general lmao
I'm currently on the pro-plan and wanted to buy some credits.
What am I supposed to do when I get the following error-message?
An error occurred.
We were unable to charge your saved payment information.
Probably you're doing something repeatedly that is triggering a refund repeatedly. Could be:
- Something in the prompt is being blocked for appropriate, copyright, or bugginess.
- Something in the segment being continued from isn't playing nice, and all your clips end the song.
Canāt purchase pro plan :(
You'll have to email billing@suno.ai
Although maybe wait 5 min and try again, just in case it's a fluke
hello, I am a subscriber and I generated all the chirps in the website, how do I get the whole list ? currently if I click Create I see only the last 20, where are the previous ones found?
Ok, thanks
Click on Library, and there are pages in top right
š„
Thank you! I do believe it's option 2.
ping moderators.get in contact with me
We're online
the english lyrics don't display correctly in the video preview https://app.suno.ai/song/c30eea31-2dd8-47a5-870d-64f8b5092805
Arabic R&B, Alternative/Indie, Dark song. Listen and make your own with Suno.
I am continuing a song, and I keep getting short clips. The system says "Credits Refunded" for both Chirps for each new attempt, yet on the left side I see my "x credits left" decrease with each attempt. It is as if I don't get refunded credits.
It seems it is a presentation issue. Refreshing the web page "restores" credits (showing what I probably should have).
So I understand this might just be a technical limitation, but I tried making a looong prog-rock opera song (maybe over ten minutes?) and it won't stitch together. https://app.suno.ai/song/f4cd0147-0e4b-4159-8db6-2a9a6ed01307
Is there anything you can do if the song randomly cuts off before the lyrics end? Hearting the song don't seem to make it longer and continuing seems to just give you some random words that weren't even in your song. I don't think the continue feature saves any continuity in style or tempo at all.
i would like to create a song for the german Swift (programming language from Apple) community.
It seems that the word "Swift" is banned.
Is it in the style field or the lyrics. That must be somehow triggering Taylor Swift but I am surprised that happens. I'll check
That might have been a fluke, or related to something else in the lyrics.
Unless Swift has another meaning in German, then maybe it would be that word triggering it...
gosh those swift songs came out great, especially the second
"a hip-hop song for the german swift community on discord"
"a song for the german swift community on discord"
"a song for the german swift community"
"a song for the german swift people"
"a song for german swift developers"
"german swift developers"
"swift"
--> Couldn't generate that.
Suno is for generating original music. Try different lyrics, and use genres and vibes rather than specific artist names.
then I tried "ruby" -> got 2 new songs
Ahh, okay it's on the "all in one prompt" side. Basically it's the lyric generation. I generated the lyrics and tweaked them myself, didn't use the site. So you can just ask ChatGPT manually do it, should go through. Definately way too aggressive moderation on that one.
This has to be some kind of bug. This track starts out fine. However, it quickly gets more and more distorted. Theres gotta be a way to squash this bug. Has to be some kind og copying bug in code? Later in this track, I can't remove the distortion. I usually can. Not this time. It's a really good track and also many credits wasted.
https://app.suno.ai/song/c4d26fbc-6bb6-4008-b851-56be45f474d9
Punk Rock and Appalachian bluegrass, female song. Listen and make your own with Suno.
SUPER THANKS!! the Suno web is on fire!!! THANKS!!! All is ok NOW !! Thanks Team!
I may be nitpicking lol. It might be 6 months to a year before we get much clearer audio. My expectations are probably just too high lol. But I posted that here in bugs just in case.
It keeps on telling me to make original music and not use artist names. 100% if the content, including what i wrote in the song name and style is not another artists name or song
Any follow up on this one? Frustratingly frequent
yeah. I wrote the original lyrics myself and now dumped many rewritings because none are working
Did you try this trick already?
You receive a lot of feedback from us on what all doesn't work. Not sure what changed recently. Here's a comment sharing that it does seem to have improved. No idea how it is working for the others, I would say 6 out 10 production ideas are great results. So, in comparison to before 0 out of 10, it has improved quite a bit since the redesign of the beta website. Awesome stuff, whoever you geniuses are out there! š¦¾
More feedback from the field... Ok, so when attempting to create the 2nd part of a song I have noticed the results not being so successful. Out of 10 attempts the majority has been refunded, 1 fairly ok, 4 results not useful.
Yeah, some clips are really tough to continue from⦠we are working on letting you choose where in a clip you can continue from which should help
Why am i getting this all the time on the website
Trying to make danish songs about me and my homies
experiencing something weird when creating "Holiday Song"...
songs with the same Music Style and Lyrics seem to be almost locked into the same Seed#.
.... and so all sound VERY similar, as if no room to variate the sound?
.
.
Sounding is if they are Continuations, not first Generations.
worth noting: These were made using the "Simple Mode".
Links for reference:
https://app.suno.ai/song/4ef579b2-6a60-408d-9728-3734375fe367
https://app.suno.ai/song/befcfdd0-2213-42a2-ae32-20170c984c06
https://app.suno.ai/song/beb565d8-83f7-4f36-8198-b30143a37d08
https://app.suno.ai/song/cbb0d3d3-3177-4bb1-a695-27851dd8f28b
https://app.suno.ai/song/fa0668e8-e5ee-4e46-8f04-45a8afb35623```
I think that's the idea actually
It could probably be clearer lol
I've been drowning in work and haven't had a chance to try it but I think it is a literally that
ok yeah i just tried also in "Custom Mode" and thinking it would NOT do the same there...
but it did. and tbh that's annoying, double now. lol
I can see that being default on the Easy Mode. fine.
But now, since I can't stand that Melody,
I, The User, No longer participate in your (Greatly Intended and Easy To Use), Holiday Event.
š¤·āāļø as it forces my hand.
</feedback>
Proposed Solution to said Problem I claim ^
instead of starting songs as "Generations",
have (behind the scenes) it add for all who
Click thru that Funneling Button;
a "Style Prefix" - constant value added to the Style of Music.
ie (under the hood) all songs have their styles starting with,
"Christmas Jingle Bells Holiday" or something to
ensure they all get the similarly cohesive sounds that you want the themes to have,
without restricting the user as much and giving more creativity back.
Hello, I've been using the web app and I got triggered everytime the outputed song doesn't end correctly? Why is that? My songs are not even that long and some of them are not finished!
I have a little notepad full of ideas like this. More about making party or social games with chirp, but same concept. Hidden information, taking control away from users.
cool, more positively: "Your layer(s) of intelligence" š
...
You could potentially train some default style models, same way like some of the AI Image generator sites have pre-built models trained for specific cases.
then just toggle to those models depending on what you're trying to pre-program for the user to have available. Each of those models basically being a "Class."
(... in rough theory; like creating Classes and assigning function to those Classes, then basically just having those Classes applied where need be to implement.)
v2 used to generate much more beautifully. After the update a week ago updated now the generation is very bad , at the level of v1
man, the copyright-check is way too strict...
Can't even use lyrics that are in the public domain.
okay, this one's weird: suno didn't want to generate a track for me because of "unoriginal lyrics" etc. etc. (there was none)
By exclusion, I was able to find that this was the part that was causing problem (translated to English):
"[Chorus]
St. Petersburg! I don't want to die yet!
You've got my phone numbers.
St. Petersburg! I still have my addresses,
Where I can find dead people's voices.
[Verse 2]
I live on a black staircase, and in my temple.
I'm struck by a bell that's been torn out of my head,
[Pre-chorus]
And all night long I'm waiting for my guests,
And shackle the shackles of the door chains."
Now the weird thing is, after I entered this part into a new generation, it worked fine. So it was only triggered when put in its entirety.
The thing is, this is a poem by Russian poet Mandelstam. Is this why it might have been complaining? To be clear, it is from 1930, so clearly public domain.
seems like they have updated something recently, I'm too getting these weird refusals
It is stated that short clips are refunded in the UI. In reality the user never gets refunded... this is one example of what Im talking about https://app.suno.ai/song/27eb2cc4-8341-4056-a6db-3109340968d0
From the work in this link, continue from this clip
https://app.suno.ai/song/23c58fd3-b75b-4db9-b843-7572f7016041
I made the next piece, but the lyrics weren't made until the end
https://app.suno.ai/song/9c60b98a-862c-41ca-879c-9c41c4b217ae
This is the lyrics of part 2, but the work of part 2 has been completed up to the red part.
There's no work on the blue part
[Verse 2]
What they say, men do
The word that you break up right after receiving a gift
That kind of trashy behavior, is that really a man's behavior
[Pre-Chorus]
I'll say goodbye to my friends
You're so different
You've been a public servant for four to five years
You who put work before me
[Chorus]
I loved you for seven and a half years
The memories remain bitter
But we met a breakup
This moment when it all ends
[Outro]
It's time to say goodbye. I wonder what you've been doing for seven and a half years
I don't think I can wish you happiness
If you ask me to check the above, I would appreciate it if you could reply.
Lots of continuation problems today, rather frustrating. Very good update with the "Part 2" etc display, helps a lotš
I have a problem with generated songs never ending, is this a general problem, or am I missing some kind of feature that suggests "this clip should be the last one"? There always seems to be overlap.
I am not entirely convinced that it is in Public Domain actually. We only just started getting works from 1927 this year. We don't get 1928 until after January 1st 2024. And I occasionally still struggle with getting some songs that were written in 1922 to work in Chirp.
(This is ignoring the possibility that these works may have been deliberately put in the Public Domain by way of lapsed copyright registration renewal such as is the case for the Bobby Day version of Rockin' Robin but since I saw during a search that the poem was converted into a song by an artist named Alla Pugacheva there could be a situation like that where the poem could be Public Domain but the song is not...)
Hey, each clip that says that you should get 5 credits back. Might take a few to minutes or a refresh. If youāre consistently running into this we can look into more
ill try, but even lyrics made by the GPT you have itself are also blocked
didnt work unfortunately
š¦
This has happened to me twice now. I'm using the webapp, I test out a few things and build up to what I'm looking for. I decide that the segment I like is the one to go with, so I click continue from this clip. The page refreshes and I get nothing. I have to manually readd whatever lyrics are left from the first part, add back the style and try to continue from this clip again.
Is this just me or has anyone else encountered this?
If in a lyric with four stanzas (verses) and a chorus, in the first 120 seconds the suno only harmonizes two stanzas, why when I ask it to continue does it go back to the beginning if there are still verses that didn't fit in the first part?
Traduzido com o DeepL (https://www.deepl.com/app/?utm_source=ios&utm_medium=app&utm_campaign=share-translation)
I understand "public domain" to mean that copyright does not have to be paid, not that authorship is suppressed. The author will always remain the author of what he creates. Occasionally, I may even use part of a song by another author, but I will never take possession of it as if it were my own.
No that's the whole thing. It's in the word itself. Copyright is the right to copy.
As long as something is under copyright that means that it cannot be used without permission or payment for use.
Public Domain is something that is no longer under copyright. It belongs to the public. There is nothing to prohibit you from copying and reusing it any way you want. No one can charge you for its use. It belongs to everyone, which is to say the public which is you and I both and Billy, and Janey, and anyone else you can think of. No one owns the rights to copy it.
So long as a song is still under copyright it is legal property of the ones who have the right to copy. If someone else makes copies (reuses it) without authorization there are fines and possible criminal penalties involved.
The Mona Lisa will never cease to be the work of Da Vinci, or Disney's Mickey, even if you use it. This is easy to identify with universal authors. It doesn't happen with lesser-known authors. This is analogous to opensource. You can use it without paying, but you have to cite the creator. This also works with "suno".
Copyright is a legal term.
Violate the law at your own peril.
You most definitely cannot "borrow" the copyright of others in the way you are describing.
Go ask Disney about taking Mickey for a spin. See how far that gets you.
Just please wait until after the new year, I think they may have finally stopped bribingdonating to Congress Critters to extend Steamboat Willy....
Also with Mickey you also get a bonus of trademark law...
Fun fun stuff for lawyers, not so fun for the rest of us.
Yes, laws change. Maybe I'm interpreting the rules the way they are in Brazil. But tomorrow everything could change. Look, if you use suno to Rhyme a melody from your own lyrics and then sing it yourself using the suno's arrangements, do you have to mention that the work was created by the suno? I could argue that the poetry has its own sonority and that the suno was the conductor and not the co-author. It all depends on the interpretation
It may indeed be possible that the law in Brazil make the item in question in the public domain in Brazil. Suno is incorporated in the United States and bound by the laws of that country, which include its handling of copyright matters. Based on my best understanding of that (which I freely admit I am not a lawyer so I could be wrong) the poem in question is still under copyright.
As for Brazil. there is no fair use exemption (at least according to Wikipedia) if the item is under copyright:
https://en.wikipedia.org/wiki/Copyright_law_of_Brazil#Fair_dealing
I just bought 500 more credits but the credits have not registrer after purchase⦠I will recieve 3000 credits in a couple of days but the idea is i want some right nowā¦
This is all very new, but it's great to create something free of copyright charges. Being able to use a playlist in a public space without an inspector bothering you
Try refreshing the page and if that doesn't work, try logging out and back in. Failing all else please try emailing billing@suno.ai
Where's the suggestion channel?
Thanks š
Another extended silence: https://app.suno.ai/song/6c2669d9-aa89-4a75-b266-bb9317da6ca4
blues rock, ballad, glam metal, emotive vocals, catchy hook, extended solos song. Listen and make your own with Suno.
Music doesn't even start until forty seconds in!
I will keep track from now on. Maybe it is like you say.
I cannot get it to create my song, it says im using existing lyrics? ive tried putting it a hundred ways. "Suno is for generating original music. Try different lyrics, and use genres and vibes rather than specific artist names."
last one i tried "daniel running in the rain on his brithday, december 24" (mispelling on purpose to try and get around whatever it is matching)
Actually, the api is just returning error_type: "moderation_failure", it might just be overloaded from hackernews, and theres a default error message?
Running into this issue as well.
Were those made with the Discord version?
<?xml version="1.0" encoding="UTF-8"?>
<Error><Code>AccessDenied</Code><Message>Access Denied</Message><RequestId>(deleted for privacy)</RequestId><HostId>(deleted for privacy)/(deleted for privacy)</HostId></Error>
I have no idea, but they are listed on the site in my library
Can you confirm you are still logged into the website? Try giving it a refresh or reopen the browser.
okay it was made on discord
So it must be ones i created on discord don't play on the site, but can tell me what to search for to find it on discord
Did you update models today? Either Iām having really bad luck or something is different, I keep getting first generations that donāt finish the 120 seconds, and sometimes even cut off in the middle of the word. One was 11 seconds and then just pooped out. I feel like I should get refunded for eleven seconds when I put in a whole song worth of lyrics! If this is a model change itās bad and please fix it.
i'm using suno on the web to reminx it says it can't generate the music - says "System Error"
Any workarounds or reason why this error message throws up?
The music I made from about a week ago was good until a few days ago. It seems like Suno has changed. The quality is not as good as before and sometimes the sound doesn't even come out. Is anyone experiencing the same?
I'm experiencing the same. I hope the team fixes it soon. I became a subscriber for the good quality. Now I cannot even put together a song (genre: lofi).
I've been having an issue where whenever I generate a continuation of another track, it's just a minute of complete silence
I can't get the links to the tracks because I thumbed them down and that seemingly makes them disappear from my library entirely, but this is the song that I tried to continue from: https://app.suno.ai/song/e9743c5c-2789-40eb-9e63-9599582ada11/
here's an example of it happening: https://app.suno.ai/song/8b92c9a2-6090-43c9-8649-382aabe1c551/. This has happened like 10 times and only occasionally do I get a credit refund
Sometimes, especially when the previous part ends in (even short) silence, itās impossible to āforceā the model to continue it with music. This should be track-specific though
with the original track I'd like to keep most of it but with the very last verse, it assumed that that was the end of the song. after deciding to continue from that, I realize that if I just took the original song ending at the end of the second chorus and continue from there, it would probably work
I managed to get a somewhat decent result that was silent for the first 30 seconds but then afterwards was perfectly good, and then I manually spliced the clips together, but I couldn't make it flow naturally and the transition sounded abrupt. I think the issue was that at the very end of the first part the lyrics bleed over into the next bar, which makes continuing from that point impossible really
Yes that happens often and is super annoying. Afaik the team is working on allowing us to continue generation from any point in the clip
I just upgraded and realized I can only see my 19 most recently generated songs. Is that the limit of history on https://app.suno.ai/create/ and I need to like songs to save them?
Make sure you are on the "Libary" not create page, and there are pagination buttons on top right (on desktop) or possibly hard to see on some other device
@neat sail OH THANK GOODNESS thank you I didn't see the pagination buttons
Should I report songs that have no lyrics as bugs?
Would the credits of non-lyrical songs be refunded?
No, and no. It happens.
For some reason, the web version of Suno suddenly started generating songs without a voice and generally has a poor understanding of the prompt and style. Is this happening to everyone? I have an active subscription. Previously, there were occasional problems with the songs, but only rarely. Now every song is voiceless. What could be the reason?
hi
I get this error when trying to continue songs with original lyrics: Suno is for generating original music. Try different lyrics, and use genres and vibes rather than specific artist names.
Any way around this/anyone I can contact to get this waived? everything is 100% orignal
Check https://suno-ai-wiki.notion.site/Error-Couldn-t-generate-that-or-Try-another-prompt-d6e88752dd9144d6b2c14503ceca43aa and submit the prompt to that form if you can!
I keep getting that when I'm creating a song using my own lyrics. I'm over it.
Add them to the list if you can remember exactly what you tried https://docs.google.com/forms/d/1I6B1GyBSMV3apIaxJMvgZ19nHwFUzUFieCmOoL2s274/
Error: "Couldn't generate that. Suno is for generating original music. Try different lyrics, and use genres and vibes rather than specific artist names." This means your song was blocked and credits refunded.
Songs on Suno.ai are moderated for:
Non-original lyrics (copyright & artist names)
Being overly explicit (in general or in the use of s...
Also paying subcription. I get this too. Did the model/settings underneath change in any way? Alot more "blurry" or non-existant voices last 2 days. Aaand I got another one...
By blurry, do you mean kinda like robotic or static voices?
more robotic and/or fades away into the music
Ok, wasn't sure. Seems to be the same for me more recently. Takes 10-15 attempts to get something good. A lot of credits tbh
My first 20-30 generation I did for a few days had really good vocals and ZERO song-only tracks. But now I get like 50% with no vocals. Albeit different lyrics, but kept the same style using only "upbeat pop music"
"wasted" a lot of credits to only get song-only tracks š¦
"A lot of credits tbh" - hehe, let the conspiracy theories begin š
I don't expect it to be perfect, we are kind of early adopters into suno, but do wish there was a way to get a second chance at a gen without costing
And another 3/4 (will update this number)...Please revert whatever you did with the "Christmas" update
I understand they can't have second chances since it cost them GPU for each generation.
But if they did some setting/update that's just plainly performs worse, would be nice with any clarification š
Sorry to ping you on Christmas @granite tartan but can you maybe just confirm if there indeed was an update that did change some underlying settings? Or if I'm just very unlucky right now
The Xmas thing is its own thing. You donāt need to use it. Nothing else has changed
Okay thanks for confirming
@granite tartan any chance I can purchase an additional 10k credits at $30? Currently it's 4,000 credits for $30, which makes no sense given that initially I got 10,000 for the same amount. Technically it makes more sense for me now to just get an additional account (which I of course would like to avoid)
Thanks. I just filled it out. I generated a song about my ordeal in order to deal with my frustrations.
Listen to Suno Buggin' (Couldn't Generate That) on Suno! šµ https://app.suno.ai/song/15f3c849-5b0d-4300-8932-e5a9c858b8c3
Hard Trap song. Listen and make your own with Suno.
Agree with the price.
But it's possible that they disallow multiple accounts in their TOS (not sure though).
I will never get what was their logic behind this
it makes no sense
I have had a weird glitch sound in joined "full songs". After separating, it seems that it's a one drum beat that gets joined just a little out of sync, thus creating a ghost beat. Great if this can be looked at..
small web feedback: how about someway that it highlights the sung parts in the lyrics (great for when the generated part ends before the end of the entered lyrics (especially when doing songs in a language you don't speak).
Allow remixing full songs for maybe 2x the credits?
is anyone having trouble with connecting clips? I click on continue from this clip, which gives me a useless 2-second clip and refunds my credits. I don't have this issue generating completely new songs just when trying to continue the clips. any help would be appreciated.
Does it happen to any song? If not, this can and does happen from time to time if the model thinks thereās no meaningful continuation except for silence.
it happens for every song.
i am making weird soundscapes, do you think it has something to do with it being repetitive?
Never worked with those so canāt really tell. Does adding various [meta tags] as lyrics in the continuation help?
yes, i think I've been leaving the lyrics tab blank, but when I add meta tags it seems to work, still not as varied as I'd like but I'm able to generate longer clips, thank you
Putting [change of key] sometimes works for āusualā music, so maybe you can come up with something similar for soundscapes.
Is anyone else having issues with suno flagging your lyrics as plagiarized? For one thing, not sure which lyrics are flagged, but the song is about something so specific that I can't imagine it's being legitimately flagged?
and also, even if one of my lines is from a song somewhere else, the fact that the rest of the song is completely different should mitigate that issue surely
by process of elimination, here is the line that the song won't generate on if it is included: "And Bard's posed on the bodies of all we have slain" - seriously Suno?
If you get a chance can you submit the prompts like this to https://docs.google.com/forms/d/1I6B1GyBSMV3apIaxJMvgZ19nHwFUzUFieCmOoL2s274/viewform?edit_requested=true
Error: "Couldn't generate that. Suno is for generating original music. Try different lyrics, and use genres and vibes rather than specific artist names." This means your song was blocked and credits refunded.
Songs on Suno.ai are moderated for:
Non-original lyrics (copyright & artist names)
Being overly explicit (in general or in the use of s...
Thank you!
why does every generation result in credits refunded? š
@chilly violet I'm having the same problem
Bro, im done if Suno's going to start censoring lyrics.
Not getting any vocals on the website
I think Iāll give it a try anyway. Wrote an email to billing@ hoping that it would be possible to merge my accounts when the situation is figured out. I'll still have to pay $10/2500 (so still more expensive than the current highest tier), but at least not nonsensically more expensive.
(Tbc, I have no ill feelings toward the team, miscalculations happen, and no one can expect someone to be super-responsive during this holiday season. Merry Xmas twimc btw!)
Greetings to all! Can anyone help? After the site switched to a new design, I can't generate a song in any way. I tried it on different promts, in different modes, in different languages. The same error appears all the time
Hello, the web app is very buggy today. Its the platform i prefer. A way to fix?
In the text given by me, the AI āādoesn't sing everything but it stops at 1.37 minutes
Suno has been generating worse and worse stuff over the last two weeks or so. I don't know what's going on but I got so much better results a little while back.
Trying the new web app and the generations are ignoring lyrics 9 out of 10 attempts. When it does sing, it's half garbled words and the lyrics are out of order. I'd like access to the original model. The "magic" seems missing the new model. I even find the longer generations aren't a good thing because those continuations were one of my primary creative choices.
Wait, is the official word really that there were no model changes? This is 100% a different experience now, including 1:30 long generations which never happened before.
I haven't really gotten anything great out of it in a while. It used to deliver banger after banger.
is it just the web app? i switched for the convenience but happy to go back to discord for usable output
Discord audio quality is too low for me, and mono
oh is the web version true stereo now?
Ah yes, the discord version is still usable! Phew!
Using the web version I've noticed that chorus consistency is pretty spot on, in fact it's so "sticky" that if you try and change the words of the chorus later in the song it tends to ignore it and just returns the original chorus. I actually don't have a problem with that vs. the horrible chorus-to-chorus inconsistency. I usually change the last chorus in my songs and I've taken to putting [Outro] in front of the replacement chorus instead and that's been working. SADLY after about 5 continuations the quality of the output audio starts to very rapidly decrease, the music and vocals first start to sound like there's an echo of the output slightly delayed and at decreased volume. Then the audio starts to become increasingly tinny until it becomes either unrecognizable or become absolutely nails on chalkboard squeely. I've tried all methods of backing up and coming at it again and the degradation simply comes back making later continuations all garbage. This happened to me starting I think with part 5 about 4 mins into the piece overall
heartfelt country ballad, male vocal song. Listen and make your own with Suno.
Everything I generate starts coming out like that... unusable
When wasnāt it? Judging by the double-tracked guitar sound, itās been pure stereo all along
(I mean, all along for like three weeks since I started using it lol)
I'm wondering what causes "Application error: a server-side exception has occurred (see the server logs for more information).
Digest: 2857688099"
I am getting pretty huge amounts of generated songs that dont include the lyrics I added (or any vocals at all). Is there any known way to avoid that issue?
Sometimes just adding "vocals" (probably with a comma like "heartfelt indie folk, vocals") can reduce that. Or better use musical words that are typically used like "sultry female vocals"
I just tried vocals, but only got instrumentals again. Trying "male voice" atm. Thank you
Unfortunately didnt help either. I got some specific (self-made) lyrics. Out of 6 clips, one had 4 lines out 12. The other five clips were pure instrumentals. Something must be cursed there š
Are you continuing from the same segment each time? If so it may not be related to your text prompt, just some bad lucky in that clip. If not, can you send me the prompt? Is it a genre that doesn't typically have sung lyrics maybe?
Fresh start with no segment, the style is "Edm, upbeat, tropical, male voice". I think there are enough examples of tropical edm with vocals
Jonathan, also deeply appreciate the help and sanity checks. The tool is straight up amazing and I just love what I can do with it already ā¤ļø
Maybe it's the lyrics? I just did a quick test, 4/4 sung: https://app.suno.ai/song/3c5aa844-0633-470c-bb0b-7e8b07d5dd10/ https://app.suno.ai/song/fbc50b0a-6f7f-4464-a9d3-225f14ab4947/ https://app.suno.ai/song/af5ac872-1058-44c6-89f5-fc7459009828/ https://app.suno.ai/song/de20f4d9-dc0c-4078-91bf-37229a81112c/
I think "vocals" may be better than voice but not sure. (Can always use both...)
If the lyrics are a factor you could also try using the first segment as a throwaway segment (with more stereotypical lyrics that may sing better in that genre) and then continue it with your real lyrics.
Will try that. I had one that actually sung it, but out of 1 out of 10 it felt like a lucky shot. Thank you
Also, not necessarily a bug, but what does "remix" do? Doesnt seem to trigger anything for me
I asked a friend. Apparently it autofills custom mode with the lyrics and style. As it was already set to the same thing I wanted to remix, it did nothing. My bad
Remix is simply "Re-Prompt" yeah
/saveid
Hello, the text on the web version is all in black so I can not see any of it unless I high;ight it. I can not see how to fix this...any hints?
Hey, is anyone having the same issue? I created a song, then continued from this clip, changed the lyrics and clicked get full song. but everytime i click it i get the lyrics from the original song, but not from the one i updated
often create songs that lack an intro
often create songs without lyrics
Occasionally there is music with a voice like a child learning to sing, or a drunk guy screaming in some funny tiktok clips.
There are enough reports of this that I have not ruled out a Chirp bug. At the very least the Chirp UI is so confusing for some people (or some devices) it is basically a bug. But so far I can't personally reproduce anything like this...
I think you have a browser extension to change the theme.
I created a verse on discord. Afterwards, I went to the web interface, where I saw the created songs in discord.
But the created songs in discord do not have images in the web interface, and separate pages do not open.
So my subscription/ credits will refill 27th of december. But they are not here yet š¦ any ideas why?
Just a guess on my side: It will happen around the same time you got your original subscription. So if it was 8pm, you ll get new credits at 8pm.
I have no clue though and just might be completely wrong
It might be so. I had the same thought. Canāt remember the time tho
You still might have a confirmation email in your inbox containing the time you got it. If it doesnt update during the day, it is not just an UI/UX issue, but probably a bug.
After creating a song, there should be a way to keep the song/melody generation and option to just change the lyrics
Suno is making a fuzz all the time. "Couldn't generate that".
This is a new one for me...
Seems to have been related to my musical styles some how.
Trying it out and it seems to be working now.
Getting 3 sec audio, after continue or remix? is this happening to anyone else right now?
Is it common for the quality of the song to drop when getting the full song?
It's become incredibly painful to generate a full song with the new update. I keep getting continuations with no lyrics. I have cancelled my subscription until a new update comes.
What kind of error is this?
So that was weird⦠got a fade out on the first generation, not a continue. It faded it out right in the middle of the first pass through the chorus.
Despite this indication, the song was generated and the credits were consumed.
millsup server invite spam attack
#š¬āgeneral-chat message
#š£ļøāchill-chat message
and other channels
How did they make it worse...like...what?
Can we go back a month when i started subscribing?
Was WAY better then.
All of these clips have one or more of the following: garbled voice, missing lyrics, hallucinating lyrics, no lyrics at all, skipping lyrics
The last words in the verse are cut off, although the segment itself is still short.
For example, these two options #š£āchirp-beta-3 message, both are 25 seconds long, but the last syllables are cut off.
@delicate iron Something is wrong with Suno. I can't get any songs with vocals. I've tried many different combinations with eight lines of lyrics but it is refuseing to consistenly make songs with lyrics. Mostly, I'm getting 1:20 instrumentals.
Something DEFINITELY happened about 2 weeks ago. A lot of people have experienced this.
they switched back though
Thank you. The crap is cleaned up now. Please continue to inform us of stuff like this that we may miss here or there despite actively looking for it.
This is still happening. Tried again with new song. I got two more Intrumental of 1:20. This is not normal. There is an issue. @solar grove @neat sail
I paid for a month... but the limit works as if it were free
and the free one works like a one-time use
I can't use 2175 credits left
What are your genres\styles?
(Sorry for the lack of response\delayed response, I didn't get pinged for this message at all.)
My style is "Very sparse deep house with strings and drum machine. Simple synth melody." similar to style I've used before with good results.
And you've gotten vocals before with this prompt?
Not with that exact prompt. I do a different prompt each time.
I got good songs with "Sparse pop, whisper singing, vocal harmonies, synth, drum machine, emotional"
What happens if you add vocal to the list above: Very sparse deep house vocal with strings and drum machine. Simple synth melody and are you continuing from anything? Because if a first or second prompt is instrumental only it is very hard to shift into something with vocals in my experience.
Did not work, I still got two intrumentals of 1:20!!!!!!!
Very fustrating
I changed style to "Very sparse deep house vocal with strings and drum machine" still 2 instrumentals of 1:20
I'm going to test something with some lyrics I am working on from something else. Let me see what I can do.
My month doesn't roll over until tomorrow and I'm using my extra credits so I hope you'll forgive me if I don't spend too many of them on this.
sparse deep house, strings, drum machine, simple synth melody, emotive vocals song. Listen and make your own with Suno.
I altered the style to be: sparse deep house, strings, drum machine, simple synth melody, emotive vocals
That rewriting gave me enough space to add a vocal type to the style prompt.
Mind you with that I still caught one with noting but instruments and the second one (linked above) was the one which gave me some vocals. I'm using a lyrics builder so I am testing that out and it sometimes causes issues with the way it works in GPT 3.5 when it was written for GPT4.
OK I notice you have [intro][flute solo] rather than [intro, flute solo], is that important?
Honestly? Not sure. This was a test prompt I did to gauge if I could make the idea work in and of itself. The lyric script I use does that part. For myself I typically build things like say:
`[intro]
[electric piano]
[electric guitar]
(...)
[coronet]
[saxophone]
[trombone]
[melodic instrumental]
(...)
`
That's an example of my style in prompting.
shingomurata. posted some examples of out of tune songs here: #šāsuno-school message
I have a big out of tune issue very often in the past months. for example, past 2 days I used about 2000 credits, most of those 400 results are out of tune in some way or another. sometimes it's noticeable right on the start, sometimes midway . it happens in "waves" every few days, and very noticeable IMO. start of the week I got about 500 credits of out of tune results, then one day later it went ok, got 90% good results, now past 2 days like I said 2000 credits when about 80% of them are out of tune . here are some examples. https://app.suno.ai/song/d87b8496-596c-43d7-a790-a423ef94ff9b https://app.suno.ai/song/24f33d13-6408-463e-ad55-830ffc1f1c65 https://app.suno.ai/song/ab0e6e01-b644-49d2-b1a3-974c2c82bd1b https://app.suno.ai/song/5ee3868b-1efc-4f31-8c59-7587b7099c4b https://app.suno.ai/song/30bef6ac-2483-4373-98f9-f1b5837bf328
I'll just add I'm subscribed for few months, made tens of full songs, so I'm not just picky about everything. I know it's beta, and when results are good it's good, but very often now when it's bad it's bad constantly, after certain updates it happens much more frequently, to the point I just stop using the app for a few days or weeks hoping next time the out of tune problem will disappear again
Again, I want to thank you for posting those here. I have created a direct link to the above in the moderators and team members section so I am optimistic that it will be seen sooner than later but I cannot give anything like a time frame and it is coming up on New Year's soon. But I have done what I can to make sure you are being heard about this.
thank you , I appreciate it
Sometimes the model adds an unnecessary pause, and there's just unnecessary silence or emptiness even though the text is much shorter than a beat.
It looks a bit strange when you want to make a whole recording or complete a task.
Have to agree. I spend too many credits as a paying sub on bad outputs. I enjoyed what it was 2-3 weeks ago, but something has changed where almost every output is bad. Don't really want to spend anymore credits in the chance there is an update soon
I'm glad I didn't find this service now, because I wouldn't have stayed. I've gotten very few really good things in a long time. My first week here was crazy, I got 20 awesome tracks in a few days that I've been listening to like crazy.
As far as I'm aware the model remains unchanged from 2 to 3 weeks ago (it was changed and then reverted back) but will keep an eye on feedback, AI is complicated, things happen. It seems the same to me personally but my prompts are so weird I'd hardly notice without specifically testing for it, honestly
I got these wonderful new wave-ish tunes with amazing violin solos and just the vibes I was hoping for
(for example)
I got the funniest thing just now it sang to me that it it couldn't understand my lang but it can but like he made a song about it lol
I generated too much but i dindt use all because was only instrumental
When I click on Library I'm getting this error
Application error: a server-side exception has occurred (see the server logs for more information).
Digest: 2637306463
Probably just a fluke. My guess is the song failed but the library wsa trying to display it.
Been like this all day, haven't used it in a couple of days. Library inaccessible for now.
That's a more serious bug, can you email support@suno.ai it's a holiday but that's a significant issue, may still get a response quickly/.
Can you see songs on your create page?
And when I generate new songs I just get "A system error occured" in a little toaster.
I'll do that, thanks
No
I pinged them too, not sure if somebody is around, ping me tomorrow if still no response if you want and I'll ping again.
Just a theory, can you try creating a single song? Maybe that would 'clear' the blockage, if the last song is corrupted somehow
I can't create any songs, they all pop-up the red error box
I'll send them an email, appreciate the help
Just to cover all the bases, did you try creating in both simple and custom?
Correct, tried both
Actually you can even try generating a song via /chirp in Discord.
Good tip let me try that out
Also just to double check, you aren't out of credits?
That woudln't explain the library, nevermind.
Chirp is working fine, that's good. Seems like the data on the web app for me got borked somehow.
It's not worth the credits to test this, but possible if you discord generated 9 chirps to make 18 more songs, and the problem is a corrupted song, it'd push that song off to page 2 and start working. But not worth 90 credits to test a long shot lol
@neat sail Yeah I'll just hang tite, Discord will hold me over just fine in the interim š thank you for the help!!
Getting several random "thanks for watching" vocals instead of assigned lyrics.
Sorry about that, we'll look into it
Thank you!! No worries!
Every try to generate lyrics I get "An error occurred generating lyrics.
Please try a different lyrics prompt."
Skip to near the end in particular
Around 50 seconds
nice phonkš„
I begin getting more and more 'credit refunded' and silence today
I still don't know hat phonk is. I'm resisting googling it...
I am struggling continuing from an earlier generation
It's the clip. For now you will have to backburner that clip.
Can you eloborate? What do you mean with backburner?
It's not going to change if you keep retrying, but future features may allow you you to have better control of cliips like this
Phonk has a lot of bass and very high notes. Sometimes it also has lyrics, like someone talking from a microwave.
What is Phonk?
are you serious??
I checked and phonk is in the Genres/Styles list.
Thank god I ran out of credits or I'd be doing some badly advised late night phonk, phonk, phonk, phonk, phonk, phonk [phonk] [phonk instrumental] [phonk solo] phonk prompting right now
Can't be worse than [pizza] https://app.suno.ai/song/ca8f6abd-38e4-4dc5-bad0-529b25ad631f/
I would like to get the BPM for the generated tracks
But it shouldn't sound like that...
why sometims when you want to continue a song " it says it cannot be generates " ?
I did "ninja, karate, kungfu, taekwondo" once, it came out in 3/4 time
ROFL
I have noticed that this happens to me when I have more than two continuations of a song.
It's the first one that time...
When it has happened to me with one or two I get the impression that it is applying some filter because of the lyrics, or because I ran out of credits, or because there are more songs being generated (although in those cases another warning appears).
Well this is my lyrics, but i thinki it because of the lyrics because when i put randoms lyrics it works
It would be good to know what is the list of filter words being applied.
yeah because it's really troublesome. And i don't know if there should be filter words because u can denounce things by using a word
It could killed creativy
Is there any dev right there that could answer us?
Hey, I would like to ask a question: I get the following error when trying to use BARK ->
AttributeError: module 'torch.backends' has no attribute 'mps' -> Which version of Pytorch is recommended ?
Are you on an Apple M1? That's what that error is about. The regular stable version should work. If you're not on M1, you may have accidentally set a flag to use MPS (or it's a bug.)
When you sometimes continue a song you can't make out word and it's a song made out of good beats but not comprehensible vocals.
It also cuts off the ending is it a problem with removing filler words?
I got confused. The [outro] tag is not used to indicate the end of a song. Was it used to start a song?
Why sometimes in the continue, it goes to an other language ??
i hav a same problem T_T
[intro] is a start song, [outro] the end.
Discord invite spam
#š¬āgeneral-chat message
Thanks for letting us know.
how to fix the language problem when we continue ?
The original song filter is very bad. It doesnāt accept any of my custom lyrics even ones randomly generated from normal ChatGPT. Please prioritize this it makes it unusable
š¤¦šæāāļø
Change the lyrics that made that happens, unfortunatly u'll never know which one, but try first with does that can be problematic for ethic
I got a problem, what to do when the text is finished before the song.....They should allowed more character specially for rap.
ā¹ļø šµ v2 used to generate songs of very beautiful and interesting motif, but now after some retraining of the neural network, it started to generate very boring songs.
When did they upgrade from v2?
half a month ago
suno.ai seems to be having a lot of bugs. songs shorter than 1:10, longer than usual wait times (I'm a paid subscriber), silence and noise. It's cool and interesting sounding but is 100% not functioning properly.
nope.. it is still beta.. š
I mean right now, compared to usual functioning. Also, feature request: why does thumbs-down make a song go away if the credits aren't refunded? thumbs down would presumably be a way to help train suno. and why is the "style of music" under "custom mode" so limited? it is a weird presumption that most people would prefer to start songs out from lyrical themes, thus giving that more space.
Your song is still there in the library.
Being a paying subscriber is more about supporting the creators, and you can't expect a full-fledged product at this point because they are entirely new to this technology and subject matter.
I get that, but isn't Suno owned by Microsoft?
Good to know, although I stand by my request that Suno be more musician/sound friendly. Musicians often start from, or get inspired by, a sound they are looking for. That might mean starting from a BPM, or more traditional notation verbiage (say, "alegretto"), or a solo instrumental by a specific instrument, or a vocal style (or even timbre). And being able to dedicate more space to this in the description forms. I'll also put this into the #š”āsuggestions-web
I understand your point about making Suno more musician-friendly. It's a valuable suggestion to consider starting from specific musical elements like BPM, traditional notation, or instrumentals. Your feedback will surely help in improving the platform. Putting it into suggestions is a good idea!
I paid ā¬96 for an annual subscription but credits are not showing on discord. I wrote an email to billing@suno.ai but I just received an automated response. Also, I cannot find any information on a possible refund, if I don't get what I paid for
Anyone?
I am sure they will contact you when they can.. but it is the 31th of December so it might not be right this instant...
But don't worry, I am a subscriber too, and it is all legit, it might just be a snag atm in the software with adding your credits
The money I spent is real and it's been a hour and a half now
I don't expect instant (although with automated systems nowadays I could)
yeah I hear more people who had some problems. Not sure what that is all about, but they always solve it
In my view of things, if you're open to get my money you also need to be available to assist with customers' inquiries, otherwise temporarily close the shop, because holiday season
I understand your frustration, @shut plaza. It can be frustrating when you encounter issues with a subscription. I hope they address your concerns soon. It's true that timely customer support is important. Hopefully, they will resolve this matter promptly. Wishing you a smooth resolution and a happy New Year!
You too and thank you š
I wanted to have some fun with friends creating songs tonight, I guess we will do something else, too bad I paid for apparently nothing
Seems like the website has trouble at times differentiating between artists and original prompts. I've had several prompts get shut down with the "suno only generates original music" even though they never mention an artist and it's literally just something like "cute jazzy coffee shop tune" or "old school country song about a cowboy" š„²
hit enter to makea newline and it decided i wanted to generate rip those credits lol
Is there any information about the Remix feature? It never seems to work for me. I'm working in the browser version and not Discord.
As it stands now remix just means 'reuse this text to prompt a new song' it does not actually try to recreate the music or vocals of the previous version.
Most of my prompts are not generating any lyrics, just instrumentals
Is there a minimum expected track length, below which it's considered appropriate to report and expect to get refunded credits?
If it's below 10 seconds you get an automatic refund
This one was 8 seconds long. No refund. Maybe only if it was silence?
I think it's 5 seconds, I had a few 7-9 seconds long which weren't refunded
Happy New Year! I am getting an unwarranted "Suno is for creating original music only" block on generating my prompt. It uses original lyrics I wrote, with a very general style prompt. It went through last week and generated a song. I cannot generate a fresh song with the same prompt, I thought perhaps because it was blocked by my already created song. But I can also not create a remix or continuation of my song with these same lyrics. The lyrics and style worked once and there is no obvious relation to any commercial songs.
It's pretty often that I don't get the full 1:20 clip. Is there a reason for this?
That definitely sounds like a bug. Have you tried the things listed here?
Sometimes with a moderation mistake you can workaround the issue with several lines of indeterminate spaces at the end.
It looks like this.
Also, I've been experimenting a lot with "custom mode", trying to get instrumentals, leaving lyrics blank, but in many cases have only gotten static/noise, generally less than 1:20. This has happened a lot. I've tried to "Report Bug", should this (eventually, like after the holiday) result in the errors being credited?
I haven't gotten an answer on this myself. Sorry. The AI and randomness both get a say in the results. While the developers can try to account for the AI in code the truth is randomness is chaotic by definition and sometimes it will give shorter clips.
That's a great list, but none of them seem to apply. I tried the spaces trick and I was able to generate! Thank you
Awesome! I'm glad it worked for you.
@smoky turret @solar grove My personal observation is that it does that with me when it runs out of lines of text. But for me that is just fine, I personally hate when it fills up the rest of the time with just instrumental audio. I get the full length of 1.20 mostly if I have too many text in general, and then it cuts of in the middle of the sentences, and then it becomes a pain to continue from that spot.
I get that, but this bug seems reliably consistent.
So would you say itās better not to max out the characters allowed in the prompt?
It depends. For instance, if I make a rap, it always is fast in the lyics so I end up with a clip of 40 or 50 seconds, but it delivered me all the lines in the music, so if I then use the option "continue from this clips" it works perfectly. But if I make a ballad, then the text unwinds slowly in a song, so that I would advise not to fill up the whole space the prompt allows you.
I've not really found the happy medium with respect to lyrics and time. I think I mostly managed to work within it on the older v1 Chirp. I kept my lyrics short and worked piecemeal to make the song in #ā©ācontinuation and stitched it together. I think I made it work for me to my advantage. The biggest issue I ran into was the chorus because getting it to work twice was hard to make happen.
Now in the v2.xx Chirp I can get a full verse and the chorus with an instrumental break in the first prompt and it is usually able to remember that chorus change when it comes back to it in the second part. The problem I have now is getting a decent intro length that is not too long or too short. Also the occasional hallucination seems to be slightly more common than I remember it being in v1 Chirp.
Which is better? It's really up to you guys, it's your songs. But I do wish that it would stop when it ran out of text or music descriptions, rather than being obligated to fill that time with music.
@solar grove @low cypress Iām using the app.suno.ai window, not this discord. Is discord more reliable?
I use the web app.. never used discord tbh
The webapp offers better audio quality.
This alone makes it superior to the Discord version (for now.)
It also has stereo in addition to its higher bitrates.
The context window is longer. I forget how long exactly but with the Discord version you had ten seconds prior in memory. The rest was luck and prompting. In the webapp it is much longer than that.
šš½āØ
It's weird - I decided to pay for a one-month subscription because the results I got from trial credits, while not exactly what I was looking for, seemed to suggest I could get where I wanted to with some rigorous prompting. But I just got a few examples where it came up with reggae (not even close to the prompt) with what sounded like Hindi lyrics (also not part of the prompt) or even some version of English gibberish. All of this has been pretty far off the mark in terms of written description. It's only $10, which is great, but so far I have to say it's not working out like I hoped.
Tbh, I paid for the full plan, twice... And I have that too. But I was fully aware this was/is going to happen. AI and this technique is still so fresh that I take it for granted. Why? Because they are ahead of the curve... There is nothing like Suno out there.
@low cypress I agree, this is powerful. I did the beta for a few of the text-to-image generators and got almost all impressive results right off the bat (I actually preferred some of the beta weirdness of those to the flawless images they get now), and I've been a GPT user since GPT3, so I guess I'm a little spoiled, and like I said, I was initially perhaps a bit (unintentionally) misled by the results from my initial trial credits.
Yeah, I am long time payer from GTP too, so I know what you mean. I also been working with all the AI forms with images too (Midjourney, Dalle, stable diffussion on my own PC), and now music too. And I noticed that for AI it is easier to learn patterns in text, then images and lastly music (so in that order). Music is so much more detailed. Every beat, every note needs to be right to make it sound good, and then the lyrics make it even a whole different layer above the melody. It's magic when it does it right though!
I am sure when we are at the point we can train custom models over it, the results will be stunning... You see that also with Stable Diffusion.
I'm looking forward to that eventually becoming possible. Actual synthetic bands. Virtual bands in the vein of Gorillaz...
I could be wrong about this, but when you listen to a song like The Beatles "Now and Then", it's not actually filtering out the original backing instrumentation of John Lennon's demo vocal, it's not "isolating" it, rather, it's predictively (perhaps adversarially) reconstructing it. By this logic, this means we should soon (if not already) have the ability to recreate poorly recorded demos, very old or degraded recordings, from scratch. In other words, we could "restore" recordings, but it wouldn't be so much a restoration as a recreation. That means you could also create a totally new beatles song simply by, say, inputting your own recorded 12 string guitar, or just a beat.
I agree. I was already looking into that and found another company doing this with spoken word. I've sent them yesterday a question if they also thought about doing that with singing too.
I cannot continue from my old compositions. Is it expected?
Which company?
Are these compositions from the Discord version of Chirp?
I don't remember to be honest, but they were first ones of those that I created, any chance to get continue option on them?
The reason I asked is that as far as I know the compositions made with v1 Chirp are not forward compatible with later versions of Chirp. Which is a disappointment to me as well. I do not know if it will always be this way but the latest that I have heard regarding this is that it is not going to be possible for now.
Where can I find updates and possible future improvements to Suno?
I'm pretty sure v1 chirps are mono and in a lower bit rate which is why it's not forward compatible. but I hope you can continue from v1 or even your own audio one day
I was experimenting. Definitely not a general recommendation to start with outro.
Okay. I liked your experiment. I replicated it in a lyric of mine and I liked the result. I think that when Suno allows us to fine-tune the generated result (I hope that's in the plans), we'll have excellent options. Thanks for sharing.
Is there any known issue with "Continue from this clip" feature?
For about a week every Part II of my clips are somehow broken (some words in different language not coming from the prompt, or just one word from the prompt language, but not coming from it)
I had the same problem repeatedly this weekend. Lyrics were my own and style of music was either "Mystyerious" or "Happy and groovy." Was able to create parts 1 & 2 last week with same lyrics and style but, this weekend I could not get past the "Suno is for creating original music only." Very frustrating!
When using chromium based browser on android mobile the Download audio/video only works if I switch to "desktop site". Tried Chrome as well as Duckduckgo browser. Same on both. Android v13.
System error occurred when trying to remix previous full songs
Why is it every time I choose "remix" I get a system error?
Just seen your message, came here with the ewxact same problem!
System error occured when trying to even tap on old V1 projects on the website. Remix worked for me but I couldn't tap on it.
This is a V1 sample attempting to load on the web app (instant error): https://app.suno.ai/song/f4953f8a-0851-4e05-af67-ba15f8605e9f
Iām really frustrated as my song wouldnāt generate - discarding with the very wage response. The lyric is my own - no AI - itās in Swedish and I tried changing every word that COULD be sensitive but at this point itās almost a fars. The style of music has been just ā1970 popā and - please! Can we have a button that gives you a report (or another AI) so that we can over ride stupidity ?
Same here! Even when the lyric is nothing but my own - and the style is generic I still get busted - without telling me why.
Yeah this is extremely frustrating.
Especially so when you've already made several full versions of the song you're doing and then are trying for another go at a slightly different style and are several minutes in and it happens on a continuation it was in the middle of...
It's crazymaking.
Assuming it's a glitch have you already tried this trick?
What is the trick? To write the Lyric this way and see what happens?
The trick is to add several lines of indeterminate length (vary them) at the end of your lyrics to see if it changes the profile enough that the AI is willing to generate it. Sometimes this works and sometimes it doesn't. All you can do is try.
Also worth looking at is an adlib, like "Whoa-oh-oh" or Oh-- or something like that.
Once the lyrics go through, then you can remove the spaces and continue or add the next lines (if it goes all the way through your lyrics).
Ok. Iāll try to see if that helps!
Nope. Didnāt change a thing.
I really do want the warning to say what itās all about. Why it canāt generate. This is ridiculous right now. I tried everything and I canāt get anywhere. Iām paying for this. #developers
I can't generate a piano and vocal song without it either turning into a whisper or a guitar, here's the examples:
https://app.suno.ai/song/4cb45156-e708-418c-9557-74b9514e9a99
https://app.suno.ai/song/0fbde826-5899-402e-b720-f0e48fa1662b
Slow ambient piano music with slow western American vocals song. Listen and make your own with Suno.
Slow ambient piano music with slow western American vocals song. Listen and make your own with Suno.
Android 14 bug as well
had this quite often too
I was Googling band names to see if there was a band called "Mysterious" or even a band called "Happy and Groovy" ... there are none I could find. Also, the warning was on both my Android phone and Ipad.
You're not googling very hard, there's at least 38 bands called Mysterious
I guess not. Any bands named "Happy and Groovy?"
Nope lol
š
Sorry it didn't help. I know it can be very annoying when things go well and then suddenly for no reason just refuse to work.
I'm dealing with a situation now myself where the song will not end...two pages of attempts to fade it out or just end the blamed thing and it will not do it. Going to have to fade it out manually in a DAW.
Everything generated in the past week is still out of tune. more then 90% of it. it just gets worse and worse. gave a few examples here a week ago, here is another to show how crazy it gets. I'm speaking of first try, not continue. I tried in another language and it was OK. the less music in it the better it is. the more music it has, it's worse the vocals is out of tune. 3000 credits spent almost all out of tune. I listened to the latest songs in showcase, to see if it's just me, well, it's not just me, most of the stuff people upload is out of tune. like, just make something and see, how come this is not top priority when it's so bad I can't understand. everything. is. out, of tune. I can upload 500+ tries from the past week if you want, besides the examples I posted last time and the one on top. let me recap - tried different styles and lyrics, it doesn't matter, even random generated lyrics are out of tune (V2 web. V1 works perfect). also I'll add that it happened about 2 weeks ago for 2 days )everything was out of tune) then stopped, and everything was fine, and then it came back and stayed for the past wee if not more. last report : #šŖ²ābugs-web message
Tried several times with this one to get it to continue and it refused. Did the spaces trick and it didn't work. Added a metatag for [throbbing vein] and somehow that was enough. lulz what is this even, sometimes...
Argh...this is driving me nuts. Somehow the AI thinks it already sang the chorus and refuses to sing it. If I put the next verse in under the chorus it always jumps over the chorus to sing that rather than the chorus. I hate this. It's driving me nuts!
I've done everything I know how to do. I adlibed, I made sure it was continuing by redoing the first prompt as is with the continuation, I even checked the first part to make sure it wasn't sung in a whisper or something...nothing.
This is after something like fifteen different rolls. So thirty tries and it can't do it.
I've got no choice but to dump it and move on.
I hate this. I loved the sound, the vocals, everything but it's just not going to happen.
This one for instance skipped a bunch of lines and went to the verse combining those together:
https://app.suno.ai/song/c49eed7f-f5fb-4808-bbf3-20edd347a01c
Melodic Disco, Soulful Vocal Disco, catchy hook, haunting chorus, male vocals song. Listen and make your own with Suno.
This is 20 secs: https://app.suno.ai/song/b1f22195-18c0-444c-bc11-cee5f1b84168
This is 51 secs: https://app.suno.ai/song/1026da43-280d-4853-9710-3d91be19a15d
Both of them disregarded the previous part's style and tune mostly.
Melodic Disco, Soulful Vocal Disco, catchy hook, haunting chorus, male vocals song. Listen and make your own with Suno.
Melodic Disco, Soulful Vocal Disco, catchy hook, haunting chorus, male vocals song. Listen and make your own with Suno.
I getting error when try to continue part 2, but I use same chorus was generated now didn't accept
This is another 21 second clip:
https://app.suno.ai/song/62757c55-b4f1-4086-a3e1-d473e13e2ef3
Melodic Disco, Soulful Vocal Disco, catchy hook, haunting chorus, male vocals song. Listen and make your own with Suno.
This one is 37 seconds: https://app.suno.ai/song/f4d19b9c-a602-4cf2-825a-85c0fdc80e26
Melodic Disco, Soulful Vocal Disco, catchy hook, haunting chorus, male vocals song. Listen and make your own with Suno.
I should probably note that these are from different branches of my attempts to work the song. The lyrics are the same but the music is not necessarily the same.
Hey baby can u help me
Try adding spaces like this and adding [throbbing vein] to the end.
Okay I'll try
Nope didn't work
Continuations are hiccuping or something:
https://app.suno.ai/song/c9612916-8420-4e3d-b7d4-0b971e172360
https://app.suno.ai/song/726e0cfa-3572-464d-b85e-860c2817217b
These are two renders of the same beginning from 0:00 to 1:20 and somehow these are both supposed to fit? They're totally different.
Melodic Disco, Soulful Vocal Disco, catchy hook, haunting chorus, male vocals song. Listen and make your own with Suno.
Melodic Disco, Soulful Vocal Disco, catchy hook, haunting chorus, male vocals song. Listen and make your own with Suno.
Sorry. I don't know what to tell you. I'm having issues with this myself at the moment.
then we will wait for it to be fixed, thanks anywayš
As for me, part 3 never matches the rest of the song.
I kept the chorus, I used the same style but in fact the result is basically another different song.
Part 2 generally sounds like part 1, with very small differencesā¦but part 3 is always a mess, even if it often understands the [Outro] word.
Is that issue related to the fact that every continuation is inspired by the previous part only and not by the whole song, in a way that the part right after a (letās say) slow break canāt sound like the first one anymore?
I even tried to create a Part 2 that could be a nice Part 3 (if you undestand what I mean: a continuation sounding like part 1 but fading out) and yet the resulting parts always have a different volume or are impossible to mix using Audacity anyway.
I hope I was clear enough, thank you.
is the remix function currently broken ?
All remix does is reuse the prompt, it doesn't actually reuse the music or vocals. Is that the issue?
the discord is stuck on generating, it never generates anythingn
Just curious, how the heck do you get "fade out" and "end" in the right column?? I am tired of asking songs to fade or end in the song instruction, but it never does.
Yes, hasn't ever worked for me
Tried to generate several parts 2 on webapp and none of them fits the original - nor with style nor with text, just starts over.
V2 is actually doing a good job with [outro] [fade out] in the past month. didn't even had to use [end] in awhile. unless it's a stubborn one that refuse to end, but then it's a loop problem.I give those instructions after others in the same continu (hard to tell by your pic what exactly you wrote before.). so I will end with a chorus for example [chorus] bla bla bla then after that add the [outro] [fade out] to start the process. it usually will start the outro process after the chorus at this point. then maybe if it didn't fade out I'll continue with only [fade out]. mostly it will add just 5-20 secs of fading out at this point. V2 usually does this good in my experience . again, hard to tell by your example what happened before. also I try to avoid spamming of commands since V2 is also prone to confusion. https://app.suno.ai/song/085d5a9b-2f3a-4746-a297-989ff50d6390
Ah sorry, you responded to someone else with that message so I didnāt see it.
I just put the metatag in the lyrics field like always. This is the first time in a long while itās been this stubborn about actually ending the song. As Shingo says itās been working really well up to this.
but how do you get the "fade out" and "text" in the column on the right, all I have are three dots which give me options like download and continue etc.
there's mention of "V2", is that referring to a version2? How do you check which version you have?
Can you tell me how you managed to get that guitar solo after the chorus? Was it just an accident?
Excuse me, why I can't finish singing, I also write according to 4-8 lines, and the generated song is accompanied by 40s in front
Hello everyone, tell me when I write [intro], and I want the loss to be 10-20-30-40 seconds? How do I do this? I write an [intro], and he makes music for 1 minute and 20 seconds, but I don't need that much, or if I write a [verse] right away, then he immediately starts performing it
I asked the same question and was informed that there is currently no way to accomplish duration control other than using shorter lyrics.
How do we add [atherosclerosis] to the end?
I signed up for the Pro Plan, but I can only create two songs at a time. Isn't the Pro Plan supposed to be able to create 10 songs at once?Is it a bug?
@spice flower Where did you generate that song? on Discord or the web app? also how did you accomplish such a long song? I just can't figure it out, I tried on the web and have not been able to do more than 120 secs. Thanks
click the three dots and select "continue from this clip"
I have tried but has not worked for me, usually it start the song over again, it does not continue, also I wish there were a way to delete the left overs, that creates lots of pieces that are unusable creating lots of clutter.
BTW, thanks!
it makes a new part, you must like the part you choose with the heart, and you get the full song (first and second part)
I don't see a heart, but I do like it with a thumbs up and it has not worked for me, what is your work flow? you continue part one and then continue from that part for part two? or all from the first original part? Thanks
The filter is a little whack, love that it will let me do Prokofiev whose work is still copyrighted but not Tchaikovsky who died in 1893. It also allows tons of current pop music lyrics but bans traditional medieval songs if they have been recorded by a major label.
yes/no. when you click the button, 2 songs will always be generated. the 10 songs come from you being able to click 5 times at once
I agree, the page is filled with 30 second clips that are of no use to anybody - the interface on this app needs a lot of work.
I continued about ten times and ended up with a fantastic 16 minute song
Why the heck do I get this error: Couldn't generate that.
Suno is for generating original music. Try different lyrics, and use genres and vibes rather than specific artist names.
It's MY lyrics, and I am on the pro plan. This is not good enough.
16 minutes? Wow!! I'll have to keep trying, I'm also in the Pro plan, I do provide my own lyrics, you too? or let Suno do it for you? and yes lots of Junk left behind, a delete option must be implemented, there is a mess.
not by accident. I think V2 takes guitar solo directions very well, specially mid-end song. sometimes it will X2 the chorus by itself and then make the solo the next continue, again, by itself, but usually I just write [guitar solo] or [instrumental, guitar solo] and it follows in his own way, that's my experience at least.
The web app. but the discord have a continue option too. in the web You have to use the "continue from this clip" function. with the next part of the song in the box, then again continue if you like any of the results. clear the box, write next part
I use the web app, and yes I do continue from the clip but it start the song all over again, it does not continue from where the previous clip ended up, what it intrigues me is the part where you say to clear the box, so you enter the lyrics in parts? I do generate my lyrics in Chatgpt and sometimes are a little long, I do enter the whole lyrics, so do you think this may be my problem? also when you clear the box and enter the rest it does continue with the same singer and style of music? Thank you for your input, I really appreciated.
YES> write the lyrics in parts. I usually use 1 verse and chorus. it depend what you write I guess but don't just write everything, cuz the engine limited to 1:20 max, it won't continue from where it stopped in your prompt, but will continue the music but will read the box from teh start
it continue with the same singer and muic yes, although sometimes it losses it if the prompt confuses "him". yet usually it's all continues very well\
I write about 14-16 lines of lyrics in a part, sure you can write only 2, but don't write too much cuz it either can't fit it all in 1:20 or tries to rush the lyrics to fit too much
Got it, thank you so much, now I understand the process, I'll change my workflow as you recommendation, it does makes sense, have a great day! Yepeee! š
So the song you commented looked like this in my prompt : 1st part verse of 8 lines, chorus of 4 lines. 2nd part the same, but chorus I wrote twice, but when it generated, the 2nd chorus he sang only the 1st line of, so the 3rd part I started the chorus from the exact part he stopped, so it started with the last 3 chorus lines, then the next verse. it's just one example.
Wow! even more details, I truly appreciate you taking the time to basically write me a tutorial, so onthe third part you also entered another chung of the last part of the lyrics right? so that Suno then stich the whole song toghether with the other two parts, did I understood correctly?
yes, continue function will usually continue with the exact song. although usually the verses doesn't sound exactly the same, cuz every generation is different, yet it's still the same song clearly. it's a process, getting a good start is hardest, continue you can try many times if some losses it
when you have all the parts , press the "get whole song" and it will stitch the whole in order. just make sure you press continue on the right generation you liked.
Ok, I'll take that in mind, now lets practice and see what can I accomplish, hopefully one day they make it so that it can generate a whole song in one take š wishful thinking, again thank you so much!
Got it! the cherry on top, thank you very much for all the details
Understood. Thank you.
I have been writing since 1966 so I have thousands of lyrics, I sometimes let suno do the words because they come up with some wild stuff
Wow! I can't imagine your age at that time, respects to you, I'm not that talented my self, I'm taking advantage of AI š
Out of tune problem is getting crazier by the day. can we get a formal answer please? some examples of latest result from the past 5 minutes. tried 2 other languages beside English, no problem at all.
Not seen any out of tune issue.
I've been seeing the same happening increasingly over the last couple days. gens started off fairly solid, and now I wouldn't use what's coming out with effectively the same prompts, without attempting to auto tune them myself.
Ok I'm getting sick and tired of this.. when the song gets to the end it says half of the last word then just cuts off! How can I make it so it finishes the word? would be nice to have a few more seconds of music at the end but at the very least finish the word!
Same here, everything is in tune
select continue from this clip
I was 11, i'm now 68
Wow, you were still a kid, you truly are a musician, thats nice that you still are at it, that is tru passion š
songs not repeating also serval tracks with no sound
It's not singing what I requested, what happened? And does anyone know what language this is and what he is singing/translation? Even though it's not what asked for, it's not bad. (on some of my other tracks, the tracks suffer from audio clipping or the singing is distorted and not as clear as others.)
no song oming with lyrics now hope there some sort of credit rfund
it has a seizure when scrolling to the bottom of the list
Have the same problem.
My solution have been to ad so short text that it will be included all and then continue the next and at final complete the song
Same here, I have had several killer tracks but my lyrics don't appear, I just get an instrumental.
What about the out of tune issue? it's getting worse by the day.
Ive created a full song, but the song doesnt sound close to the lyrics. Is it possible to fix this?
its just a few words off and i like the song so much already
Bug
Acknowledged the paradox, the line between.
Verse 3 (God):
A master tactician, with celestial might,
Iām not sure this counts as a bug but itās hilarious. I asked Suno to create lyrics about the phrase pas de deux and it surprised me by giving me lyrics in French. Well, okay, thatās fair. I asked it to make a song about that phrase in English please and this was the result: [Verse]
(Instrumental) (da-doo-doo-doo, da-doo-doo-doo)
(Instrumental) (da-doo-doo-doo, da-doo-doo-doo)
[Chorus]
(Instrumental) (da-doo-doo-doo, da-doo-doo-doo)
(Instrumental) (da-doo-doo-doo, da-doo-doo-doo)
I think itās taking the ****.
š
Use an AI site like lalal.ai to separate the vocals from the music and edit them back in correctly.
as long as they dont have copyright to them when using their service
And always be careful to make sure that even if they tell you that you retain your rights and ownership they're not helping themselves to a royalty free perpetual license to do whatever they want with your music.
Is this something Suno can help me with?
I tried to generate same again, but voice turn into black metal
š
i could maybe download each of the parts of the song and try fix it that way
It might work. Lord knows I've had to do it that way before. Sometimes it's the only way to get something to work.
new version seem good so far, phew phew
but its female voice instead of god and devil
They don't
well, if you subscribe they dont
actually the copywrite itself is entirely country dependent
for ex no one has any copywrite to any pure ai generation in the us
But you do retain the copyrights for any lyrics you write yourself.
correct
did [Outro] stop working? tried [Instrumental Outro] but it still sing jibberish
16 failed fades so far
What if the main source is written , but rewritten with ai?
Shouldnt it count as intelletual propety then?
When I make my videos, it says I can make ten songs a day but I can only can make 3.
Or when I come back from not being on I can only make one song or something.
It's killing me when I'm trying to make mah anime openings, please help.
Btw was web app are y'all talking abt.
At the Moment 50% of my tries come without lyrics. The lyrics are there but when i playback the song it is instrumental only
This happened to me yesterday and today... Tried to make song, one of them failed, one succeeded. So, it refunded 5 credits... But why??? I mean, you can't use 5 credits unless another generation fails and you get another 5 credit refund.
**It makes no sense to give credits back that you can't use. **
Having MAJOR problems with this bug. Gave up trying last night, and this morning my first attempt with lyrics gave me two more instrumentals, it's a real shame as the music is great.
It would be cool if you could reuse a melody but change the lyrics
is the service having issues ... 35+ gens currently and counting only output silence with some getting refunded and most not
not with me.. maybe you shoudl send them an email on support@suno.ai @echo talon
i dont even klikt on the create button but it has made 3 song so i now have 5 so can i have a refund
definitly bug'd yea
yea i get the same
maybe people who have the free credits are differently handled, but for me (I have paid sub) it works like a charm.. just send them an email
okay
seems to be only with songs that are continuations from parts 5+
like when you extend an full output to add an proper end instead of a cutoff
Does the part you are trying to continue end on a soft note in any way? Because Suno sometimes sees that as an ending to a song and it will not continue it because it's "done already" when that happens.
Same thing if there is a fade on the end of a track.
yea .. but it also happens on some that did have lyrics cut off
Still holding out hope that when they allow us to set a resume point from a previous section I can resurrect and complete this one. It does the fade on the end and nothing I could do would make it continue.
worst is i could just generate as a full new but that will lose the instrumental and vocal audio 99% of the time as we don't have control over singers
Yeah. Seeds is something else I am desperate for to happen so I can bring back a few things in both vocals and in sound of music.
I've got this one where the vocals are just damn near perfect and I'd love to use that vocal again in other stuff.
how i install
Any news about the out of tune problem?
I don't think that is a bug. It's just not trained enough to understand what is "out of tune"... š
I have made 2 song but its only the sound not the lyrics
https://app.suno.ai/song/2d1a016f-392a-4556-81a2-a759f793d22a
Hello, for a few days my songs have been full of 'clicks' and distorted noise disturbances, especially in parts with fewer instruments. Thank you. I'm using Sumo from the website, not Discord."
Listen the end of the song
https://app.suno.ai/song/f74e56d4-e656-4c76-98aa-fb3ab48a9ce6
here, in this song Clicks Crack Distrorsion Noise start al 0:13 seconds, i can't work well š¦
Using my own wrote lyricsš¤
There seems to be a bug at the moment, it's ignoring lyrics too often
I get that every day, it's infuriating
can't hear anything on my end are you sure it's not your device or browser that's having a problem
This is probably a bug, but it is also hilarious. I asked for "Solo tin whistle Irish Folk" and got what sounds like a child blowing on a recorder: https://app.suno.ai/song/8a698f77-0827-442b-bc68-dc0ca8950a97
I was able to get better results with "solo fiddle"
I generated about 20 songs today and only got 3 with lyrics. I tried lots of different prompting. Anyone else having this issue?
Anyone else having problems with the lyrics being unintelligable? Or maybe it translates it to antoher language? I've had several times where I listen to it and I can't understand a single word! Sounds like a different language! Those should refund credits!
Just seems so unfair that we get these "credits" to create songs but 90% of them are worthless and don't work.. Idk if this is the goal of the Suno or what knowing AI is far from perfect so they can charge us credits over and over and over to get something that works? I realize it's very cheap for what we can get out of it but still
when things go very wrong we should get credited back
I've had way long instrumentals as well and some times all the lyrics would fit into the time limit but because of instrumentals it wastes and uses up the time is there any way to help control that?
you could probably bring your song into flstudio and add a cowbell as another track. Depending on how badly you want it in there.
and when I finally get someone singing the lyrics, the music is terrible and it ignored [intro] etc making it unusable anyway
same hereā¦
@carmine basin @neat sail Just repeating two critical issues users have been pointing out here. 1. Out of tune issue. This is most noticable and frequent when doing a solo piano with a single voice and 2. Lyrics not being added. This happens (for me) mostly in Custom Mode in the Webapp when using long and complex lyrics with many tags. Can you please let us know if the reports have been received and if you are looking into it? Thanks!
Both issues are likely technically tricky with no easy answers, but on the second, going 3/20 for lyrics is far below average. Is this using either the same prompt, or is it trying to continue from a single segment repeatedly? If not, there may be something about the tag use that is making it worse.
the 3/20 issue, I tried variations on the prompt. "upbeat, cheerful, pop, light orchestral elements, radio jingle" I tried not including the radio jingle, I tried adding, "female singer" or "Male Singer". I was able to force lyrics with "a cappella" added but then no music. I tried adding [Verse] and not adding it. I tried with and without [Intro] those all failed adding vocals.
Here is one that scored well below average. https://app.suno.ai/song/3aa89cbf-bdbd-48ee-b5b5-5e9a3e92c148 We did about 2000 credits or more over the past week and at least 1/3 is without lyrics. We use custom mode and prompt our own with tags. If not fixing it, detecting that there are no lyrics (And then refurnding the credit) like with short clips now is the case would help. Then we wont have to listen through waiting for a lyric to happen yes or no...
1970s, Jazz/Soul, Soul, Retro, Electric guitar, Bass guitar, Drums, Keyboards, B song. Listen and make your own with Suno.
https://app.suno.ai/song/eb120118-eb7e-49fb-976f-a699ac00ba01
my compositions continue to have a lot of CREEKS and Clip noise especially in the vocal parts
the final part of the entry is horrible š¦ ,
anyone have similar problems?
I find them when I go to continue the song with the second part
All continuation attempts after the initial clip for me are now completely broken. Either generates gibberish clips or short credits refunded clips.
At this point itās 100% replicable if someone on the team wants me to DM the link
Sigh. Now continuations are ignoring the new lyrics and generating a new version of the previous lyric like I did a remix instead of a continuation
lyrics are showing up black once typed so i cannot read them
Final part song usually weird singing sound like under water and music sound really weird
would LOVE some weirdness in mine, most are too predictable lol
Cannot access website. Get a "Took too long to respond" error. @delicate iron @unreal radish @tame latch Is this a known error for you?
Pretty weird bug. Asked it to write a song in the simple prompt section and it returned a little speech about Emma Watson lol
I'm sorry but that'd both horrifying and hillarious at the same time.
That's a classic GPT thing, used to happen all the time, kind of warms of my heart to see they haven't quite RLHF'd it out of ChatGPT entirely
Lol suno say suno make orginal.. When used own lyrics and i wrote myself lyrics about 30 years and current wrote own book.
If it's not something that is already published, then it's probably a glitch.
Try to change the profile of the lyrics with several lines of indeterminate spaces...that looks like these examples:
The key is to get over the blockage by making the AI see it as a new lyric.
You can always remove the spaces as needed if you get past that section.
I tryed that allready not works and when make lyrics ai sometimes happens same thingš¤
I have a lot of lyrics that I've written in a desk drawer, so to speak, and I haven't shared them publicly anywhere. I've shown them to friends. Whether it's a bug or I've just unknowingly written the same kind of lyrics that have been published. But not stuck that i wrote new lyrics lol someday i wrote many new lyrics. It's hobby to me it good way relax mind writing
Not saying that any of these apply but this is a list of what we know triggers the message, from the community wiki: https://suno-ai-wiki.notion.site/Error-Couldn-t-generate-that-or-Try-another-prompt-d6e88752dd9144d6b2c14503ceca43aa
Understand suno is not ready allthing and lot of issues to fix but it's would be cool feature if those error messages tell moore info i think many found that helpfull feature
Anyone else 500āing rn?
Yep, came here to see whatās up
Cool. Letās hope itās not for long
Can access site fine and generate but nothing is finishing generating for me
Same same. Maybe an update rolling out?
We can hope š
Happening regularly for me, with MY lyrics
When that happens, if you can save the prompt, it would be helpful to add it here https://docs.google.com/forms/d/1I6B1GyBSMV3apIaxJMvgZ19nHwFUzUFieCmOoL2s274/
Error: "Couldn't generate that. Suno is for generating original music. Try different lyrics, and use genres and vibes rather than specific artist names." This means your song was blocked and credits refunded.
Songs on Suno.ai are moderated for:
Non-original lyrics (copyright & artist names)
Being overly explicit (in general or in the use of s...
Suno has been trying to create a song for ten minutes, just spinning discs
came here looking for info for this same issue. 30+ mins now for my last prompts, still circling and are still not finished
TY Hepcat
It isn't giving me a maintence message but my Chirps are all silent.
c4d07ec1-87d7-4709-bfa1-ffb48ea012b2
31e42493-3b78-4e5b-bcf4-9df9a8500681
Are we still down?
been spinning discs for an hour now.
Outro really need a fix. I've used perhaps 30 on a outro and still just instrumental
prob more
1300 credits today or so in total
its 03 at night for me
really need to sleep :/
im using the webapp
Not worth the hassle in time. They may check my account if they want
Gotta go, ill be back in 9-10 hrs or so
Still no fix?
NSFW server invite spam:
#š£āchirp-beta-2 message
They hit all of the beta channels
and just noticed another one...
steven_00881 spammed a referral link in several language channels a month ago, that's gone unnoticed all this time.
Is there a way to fix a song's intro that isn't complete? Many times the intro comes out beautifully but I have to give up because it isn't complete.
You can't change a song once it's been generated, unfortunately.
But you can use "Remix" to reuse the same style prompt and same lyrics and try generating a new track with the same settings..
It will likely come out sounding very different from the original, though.
Hey it has been a day since i tried to create something in suno, first i waited an hour then a red bar popped up saying something about offline but will get back soon which was fixed after 10 or 30 minutes. However its still in a loop since yesterday and i can not create something new
It does take credits i believe, but it will just stay like this and after a while i refresh the page it will put me back as if nothing happened
Oh looks like the word āaffairā in title caused it but then it worked before and now I canāt generate a part 2š¤·āāļø idk just some info
Since yesterday two sounds have been in process and have not finished. This happened after maintenance. When trying to play, the error appears:
Application error: a server-side exception has occurred (see the server logs for more information).
Digest: 522368647. iPhone iOS 17.3 / Safari
Suno is getting better day by day. The music output is more consistent than it was like, about 1-2 weeks ago
Easier to make smashing hits now
hello
it shows me that I have no credits
Your info - robert_79432
Subscription
Basic Plan (Free)
Daily Credits Usage
0 / 50
Extra Credits
0
Lifetime Usage
28 generations
but I have credits
What do I do?
Lifetime use? Does that mean that once you've generated 28 sounds on the free account, you can't generate any more? That's curious information.
I have subscription
It's just doesn't seem to be working on discord
It only works on website.
Thank you very much. But if possible, I'd like to add the function of inserting an intro. Giving it too would be very good haha āābecause the intro is now short and the song is up.š¤£
Free users will get 50 credits (5 chirps) per day.
There is no limit to how many clips you can create over time.
Please contact billing@suno.ai with your Discord ID and the email you registered with on the site.
Your accounts just need to be synced.
thank you
ŃŃŃ Š½Š° ŃŃŃŃŠŗŠ¾Š¼ Š¾ŃŠ²ŠµŃŃŃ?
ŃŠ°Š¹Ń Š¾ŃŠ¾Š±ŃŠ°Š¶Š°ŠµŃ ŃŠøŃŃŠµŠ¼Š½ŃŃ Š¾ŃŠøŠ±ŠŗŃ, ŠæŠ¾Š“ŃŠŗŠ°Š¶ŠøŃе, в ŃŠµŠ¼ ŠæŃŠ¾Š±Š»ŠµŠ¼Š°?
is there any fix?
I can't figure out what the problem is, it just doesn't work. I don't see a solution.
when I click on the Create button, this sign appears on the screen. I tried to press it, but nothing happened.
I'm sorry if there are inaccuracies or errors in the text, I don't speak English well and I'm talking to you with the help of an online translator
I got a song with absolutely no vocals š
So it seems if you use (Verse) or (Chorus) instead of [Verse] then you dont get vocals
Check your lyrics and your styles fields. If you copy and pasted from something, (like a notepad after writing your lyrics) you may have too many characters in one of those fields. If you used : anywhere check if removing those fixes things.
been like this for a day now, anyone else?
I'm using it now. Been using it most of the day. Not seen that yet.
(Now yesterday I saw it.)
No idea what you are saying, but if you trying to remix, you need to have less text
Ń ŠøŃŠæŠ¾Š»ŃŠ·ŃŃ Google Translate, ŃŃŠ¾Š±Ń ŃŠ“елаŃŃ ŃŃŠ¾Ń Š¾ŃŠ²ŠµŃ. ŠŠ°Š“еŃŃŃ, ŃŃŠ¾ ГоŃŃŠ°ŃŠ¾ŃŠ½Š¾ ŃŃŠ½Š¾ Š“Š»Ń Š²Š°Ń. ŠÆ не знаŃ, ŠæŠ¾ŃŠµŠ¼Ń ŃŃŠ¾ ŠæŃŠ¾ŠøŃŃ Š¾Š“ŠøŃ Ń Š²Š°Š¼Šø, Šø ŠæŃŠ¾ŃŃ ŠæŃŠ¾ŃŠµŠ½ŠøŃ Š·Š° ŃŃŠ¾. ŠÆ ŃŠŗŠ°Š¶Ń, ŃŃŠ¾ ŃŃŠ¾ ŠŗŠ°ŃŠ°ŠµŃŃŃ Š½Šµ Š²ŃŠµŃ . ŠÆ Š²ŃŠµ ŠµŃŠµ Š¼Š¾Š³Ń ŃŠ²Š¾ŃŠøŃŃ ŠæŃŃŠ¼Š¾ ŃŠµŠ¹ŃŠ°Ń Šø ŃŠ¾Š»Ńко ŃŃŠ¾ Š·Š°ŠŗŠ¾Š½ŃŠøŠ» ŠæŠµŃŠ½Ń. ŠŃ Š¼Š¾Š¶ŠµŃŠµ ŠæŠ¾ŠæŃŠ¾Š±Š¾Š²Š°ŃŃ ŠæŃŠ¾Š²ŠµŃŠøŃŃ, не ŃŠ»ŠøŃком ли много ŃŠøŠ¼Š²Š¾Š»Š¾Š² в поле ŃŠµŠŗŃŃŠ° или в поле ŃŃŠøŠ»Ń. ŠÆ ŃŠµŠ“Š°ŠŗŃŠøŃŃŃ Š² Š±Š»Š¾ŠŗŠ½Š¾ŃŠµ Šø копиŃŃŃ ŠµŠ³Š¾, Šø Š·Š°Š¼ŠµŃŠøŠ», ŃŃŠ¾ иногГа Š¼Š¾Š³Ń ŃŠŗŠ¾ŠæŠøŃоваŃŃ Š±Š¾Š»ŃŃŠµ ŃŠøŠ¼Š²Š¾Š»Š¾Š², ŃŠµŠ¼ ŃŠ°Š·ŃŠµŃŠµŠ½Š¾ на ŃŠ°Š¼Š¾Š¼ Геле, Šø ŃŃŠ¾ Š²ŃŠ“Š°ŠµŃ ŃŃŃ Š¾ŃŠøŠ±ŠŗŃ. ŠÆ ŃŠ°ŠŗŠ¶Šµ Š²ŠøŠ¶Ń ŃŃŃ Š¾ŃŠøŠ±ŠŗŃ, ŠµŃŠ»Šø ŠøŃŠæŠ¾Š»ŃŠ·ŃŃ Š·Š°ŠæŃŠµŃŠµŠ½Š½ŃŠ¹ ŃŠøŠ¼Š²Š¾Š». ŠÆ ŃŠµŠ¹ŃŠ°Ń Š·Š°Š±ŃŠ», какие ŃŠøŠ¼Š²Š¾Š»Ń Š·Š°ŠæŃŠµŃенŃ, но ŃŃŠ¾ Š±ŃŠ»Š° ŠæŃŠ½ŠŗŃŃŠ°ŃŠøŃ, ŠŗŠ¾ŃŠ¾ŃŠ°Ń ŠæŠ¾ какой-ŃŠ¾ ŠæŃŠøŃине не ŠæŠ¾Š½ŃŠ°Š²ŠøŠ»Š°ŃŃ ŃŠøŃŃŠµŠ¼Šµ. Š£Š“Š°ŃŠø, Šø Ń Š½Š°Š“ŠµŃŃŃ, ŃŃŠ¾ ŃŃŠ¾ помогло Ń Š¾ŃŃ Š±Ń Š½ŠµŠ¼Š½Š¾Š³Š¾.
I'm using google translate to make this response. Hope it is clear enough for you. I don't know why this is happening to you and I apologize that it is. I will say that it is not effecting everyone. I am still able to create right now and just finished a song. You could try to check if there are too many characters in the lyrics field or if there are too many characters in the style field. I do my editing in notepad and copy it over and I've noticed sometimes I can copy in more characters than actually allowed and it gives this error. I also see this error if I use an forbidden character. I forget what characters are forbidden off the top of my head now but it was punctuation that the system didn't like for some reason. Good luck and I hope this helped at least a little bit.
š
Thank you very much, I'll try to reduce the number of characters and punctuation marks
This isnt a issue with the amount of characters
So the solution for me was:
Log out from my Discord Suno account and Sign Up via Google
I see. Well I'm glad you solved it. I never use the auto song feature and am always in custom lyrics so I have no idea why that would happen and agree this is definitely a bug.
its saying that i have 0 credits (used all 50 yesterday) and has not refilled
Contact billing@suno.ai
Use lyrics as follows:
(Verse1)
Your first verse and so on.
(Outro)```
[Intro]
[Verse1]
Your first verse and so on.
[Outro]
but also that doesnt apply to his case there
There is a method where if you put in remix or continue lyrics (loop) or (beat) or (vocal) you can get like 15 seconds of audio (sometimes it doesnāt) then those 15 seconds you can use in an audio editor to mix and match or remove vocals using audacity and just mix and match.
But getting the exact replication is not possible using just the ai as far as i can tell
Also those 15 seconds will refund your credits
Hi, first I want to greet the community here and say hello from Austria and second I want to say suno creates cool songs.
But unfurtunatly I perhaps found following bug (or I understand something wrong)
I create a song with suno (lets call the song title1)
I used the function "Continue from here" to change the lyrics (lets call the song title2)
Than I used the function "Get whole song".
I expected to get the whole song from title2.
But I get the song from title1
What did I wrong?
Many thanks in advance š
I contanct support (hello@suno.ai and support@suno.ai) with screenshots attached but get no response
its working appropriately. when you select 'get whole song' it joins the clips together, beginning from your initial clip in that specific line of generations
So you did continue from song 1 then you have asked for the full song on the second version you got, that is how it works it connects the parts
If you just want song 2 download it or remix it but songs generated from song 1 will always include that part after getting full song
Thanks @fleet merlin and @dull quartz : Is it possible to get only part 2 (i like the lyrics more than from part 1)
Yes šš» on web browser you should see an option to download it
Thanks so much š
Anytime ā¤ļøšš»
thats kinda how i usually do it. i join up all the tracks to a song ive made once i get to the end, then individually download all the separate clips, split stems as needed and spend some time in a daw
or other ask in another way: can the music be the same but lyrics changed?
Unfortunately, you can't change the lyrics of a song once it's been generated.
that would be a nice feature for the future š
Hopefully some day!
Hey, I think I found a bug today
You can get another generation without credits
If you open up 2 windows at the same time and you have 10 credits, click generate on both of them
It will create 4 songs, not 2
It sounds like its not a big deal, but with some request manipulation it could lead to unlimited tries
@delicate iron @unreal radish, really think you should bring it up with web guys
Free users are allowed to generate two jobs concurrently. š
(even from the same window)
Unless you're saying it generates both songs with same 10 credits š¤
Anyone else encountering an issue where it finishes the verse, the chorus and then with nothing left decides to start over with the verse again?
It has done that all the time, I now try and add a fresh verse when continuing
This is in the first generation.
It never refills at midnight, seems to happen in the middle of the night.
Annoying bug - I enter a song title, something like "flying high", and enter the style etc, it then creates random lyrics but not about the subject I suggested, invariably it is either "Neon Lights" or "Coffee Shop Dreams" or variations on those, and the lyrics are irrelevant to my subject.
If you put "Flying High" in the lyrics field instead of the song title - and then click Generate Lyrics - it should (sort of) work like you expect, in custom create mode.
Interesting - just tried it and it worked, thanks x
It's a cool little feature. You can even sometimes use it like instructions. Use one verse, and write on a new line "Can you translate this song into a goofier version with sillier words" or something.
Yes it happens to me frequently š¦
It just seems to have started for me recently.
If i connect pieces in a song and i have generated more of same part, will i loose the connection if i choose download whole song? I need to figure out what part is best with the other part of the song
Its hard to keep track of the chain as they disapear after i generate more parts
its the (bridge) and i read that it should be somewhere between 4-8 bars, so one of the parts are better suited for the song, but i like the other one more
Just wanted to put this on the radar for the Remix feature.
When I try it with a bit, it doesn't do anything except make the list start at the top and scroll down, but nothing happens. If I am already in "Custom Mode" will it show any different? It seems to appear the same, there's no indicator showing or saying that I am Remixing from a previous bite.
Anyone else only get the lyrics "Im gonna go, get some water" over and over again?
Even though its not in the lyrics
My 3rd or 4th time š
I'm jealous, pretty cool lineš
Alot of singing
Although he's drinking with Pan
so totally understanding he need every second water
Imagine that with rammstien sound š
Also 1 minute of "Thanks for watching"
Oh, if you want to mess up the linking between song, wich i have done alot.. download the whole song of the parts you choosing from, then continue from those so you dont ruin the whole chain
This is hilarious!
^^
Im cursed, now i got "Im gonna get some water, Im gonna get some water, Im gonna get some water, thanks for watching"
A few times i got some kid speaking dutch, i believe it was
Came across a bug where the AI basically tried to continue where it left off on a new beginning of the song:
https://app.suno.ai/song/4e38425e-5ab2-434e-bb10-1bc04118d1be
blues rock, indie folk, americana, catchy hook, pop, soulful warm vocals song. Listen and make your own with Suno.
This was the first part: https://app.suno.ai/song/b052ac06-8abc-4ae0-a5a8-e283342de2a6
blues rock, indie folk, americana, catchy hook, pop, soulful warm vocals song. Listen and make your own with Suno.
If you notice lyrically it is as though it was resuming from this part.
We're you previously doing continuations? Did you remember to click the "Clear" button before starting the new song?
This was the first part of a song. It was a beginning with nothing attached to it.
I had cleared out the previous song from it.
I have to say I haven't encountered that one yet... web or discord? If web, maybe it's time to dump cache and cookies and start fresh
Yeah it was web. Actually I dumped my cache not too long ago. Don't think it's relevant but...I mean it could be...
I got a funny and catchy verse when I made a song about pirates
Purged a gig of cache when I did that.
can upload it later though
I have full respect for those creating and knows how to mix music well
I've noticed that when they've made UI upgrades to the website that failing to dump cache can make it do some pretty weird stuff... never hurts to try anyway
The ai just completely ignores the text i enter and only generates instrumentals. Have any of you guys experienced this?
I have noticed with some reliability (maybe 80% of the time), when the lyrics of a clip end with a word ending in an "s" sound, that sound gets cut off. The clip is shorter than max length, it seems like it "has space" to generate the consonant, but it has a very high rate of just ... not doing so.
I have thrown out hundreds of clips that otherwise sounded great because of this truncation issue.
For some reason a lot of my generations are not even singing and there just tallking over music
Usually happens if I add instrumental tags like [bridge -solo paino] or [Intro - Instrumental Amapaino] or type of music is deep house
Anybody running into an issue with the "Couldn't generate that" message? I understand it's supposed to do that if you put in a specific band or lyrics from an existing song, but I'm getting it seemingly randomly. For instance, I'm taking lyrics I wrote for a song I wrote years ago. I'll set the genre, title, and 8-ish lines of the verse and generate it a few times with no problem. But then I'll try to generate another instance - without changing anything at all - and, suddenly, I start getting the "Couldn't generate that. This is for generating unique songs. Try changing lyrics, titles, don't use real band names, etc." message and it refuses to generate the thing it's already generated versions of several times. It also happens a lot if I change a single word (to get a lyric variation in a repeated chorus, e.g.) and it just happened again when I changed nothing but the genre from "heavy metal" to "glam metal." Apparently, it thinks "glam" is a specific band or something. EDIT: Figured it out. The first half of one line in the song was apparently too close to an Ice-T lyric. Even when I changed it up a little, it still flagged it (even though it had generated it a bunch of times before without complaint). So I just wrote the line phonetically and Suno stopped catching it. Goofy, but there ya go.
Been experimenting with intros (so the song doesn't just start at the first verse lyric). I've tried things like [Intro, catchy guitar riff] and even just [Intro], followed by a [Verse] tag and 4-8 lines of verse lyrics. In the past 10-15 generations, however, I haven't gotten a single lyric. It just ignores the [Verse] lyrics and creates an entirely instrumental song. A few times, it even gave the song a perfect outro at the 1-ish minute mark, and ignored every lyric along the way. I mean, it's made a few nice, short songs that sound like a complete 1 minute instrumental track... but no singing to be found, despite what I've written.
I have 2 processes thats been generating since i went to bed about 10 hours ago
Reported them as a bug now
Yes, it generates both songs with only 10 credits left
So it does 2 jobs with only 10 credits
That means it generates 4 songs for only 10 credits
putting some sort of singer prompt in the style usually forces singing
Meaning if you have pro subscription you can create 20 songs for 10 credits, or you can run 10 jobs for only 10 credits
meaning you can run 10 at the same time. the cost is the same. 10 for 2 gens
@valid jungle somebody should look this up, it can be a big issue
The devs will see this post and look into it
Its still worth the money for it imo
Cool, keep me updated if you can
Didn't really look at it in burp suit, i could probably do some request manipulation to generate more than 2 jobs on free accounts
I'll look it up when I get some free time from my studies
Yeah, same thing happening to me. Depending on what you want to do, most of the time this happens because you use more than 8 lines of text
But you can fix that with [Break], add break after finnishing the verse, or if its end of the song just add [Finish]
Yeah, in custom mode generating lyrics has nothing to do with song title, lyrics field is the same as chat gpt promt, whatever you want suno.ai to generate, write it like you would in chat gpt. Something like write me a song about a guy named Bob, and then click on generate lyrics
Happens all the time. For me it was happening because it was difficult for ai to find a flow for my lyrics. But I usually get a good flow on 3th or 4th try
You prorably need to add [break] after guitar solo
Btw i fount out that Intro tag breaks everything
Better to use [Hook] instead
I tried that (because I've had success with it in the past). Didn't work this time though. I think the issue was that the "catchy guitar riff" reference in the [Intro] tag was making it think I wanted the guitar to be the main melody line and so it skipped the singer no matter what I wrote in the [Verse] tag. I ended up dumping the [Intro, catchy guitar riff] tag and just started the lyrics with "Let's go!" or "1, 2, 3, 4!" before the [Verse] tag and that helped.
Yeah, [Intro] def seems problematic. Too bad too, bc I liked the riffs it was generating. Good tip with the [Hook] tag. I'll try that!
Can you send me the promt u were using
The thing is you want that to be out of beat, you dont want ai singing that part, thats why ai is getting confused, there is no flow to those words
Thats why you only get instrumentals
You can try to add it in bridge, something like this
[Bridge, guitar riff]
Let's go! (1, 2, 3, 4!)
[Build, Verse]
Some text here
That way ai understands that guitar riff is build up for the verse
Im not saying its going to work 100%, but u can try it
Thanks! I'll try all of that. Fwiw, this is what I actually had in the prompt:
[Intro, catchy guitar riff]
Let's go!
[Verse]
When I stop to look around
I see that freedom is losing ground
Too many people deciding for me
what I can hear and what I can see
It's a song a friend of mine wrote back when we were in high school and I thought I'd surprise him with a full version. š
Yeah. I had the genre as "epic heavy metal, fast, hard rebellious"
Nice! Thanks, my friend, for jumping in and experimenting with these. I also am trying out your tag suggestions and they seem to be helping a lot. I've still managed to generate one or two with no lyrics (or lyrics that start REALLY late) but MUCH closer to what I'm trying to achieve. I'm gonna go play around with these tags some more and see what I can do with 'em. My friend's gonna be stoked when he hears his old song come to life!
Cool, glad to somewhat help
Just don't get your hopes too high, because its really luck based
I generated first song I acutally liked after 50 or so generations
Luck = random š
Yeah, completely random
It could be less random if we had a way to influence the promt itself in custom mode
But giving commands to the ai via brackets rarely works perfectly
Maybe if they ad {}
add*
It could be really cool to see that, I had to generate soo many songs to get dual vocals to work
I actually had the opposite experience. When I first started messing with Suno, 8 out of 10 of the first songs I generated were awesome. Several of them went instantly into my main playlist and I still listen to them constantly when I'm working and whatnot. It's only been recently that I've started having more misses than hits. And I'm REALLY struggling with the "Can't generate that" error message. In fact, I generated a version of the song using a version of your meta tags [Bridge, catchy guitar riff] and it worked great. Then I tried a version without "Bridge" in the tag (so it was [catchy guitar riff], but everything else was unchanged) and it flagged it as "Can't generate that." I put "Bridge" back in the tag and now it generates again with no problem.
Like, this was literally one of the first songs I generated:
Yeah, ai does great when you are making a song without custom mode
but custom mode is weird for me, specially so because im making songs in my native language
I found out that rock is really easy to make without many errors
The thing is you are trying to command the ai here, when you try to do that you get random results
If you just say make me a song about that it will make it no problem
But when you say make me a song with cathy guitar riff ai will strugle with that
It takes a lot of generations to fit everything in, thats why I say it luck based
Yeah, I agree with all that. I'm looking forward to when this technology advances further. It's fun now (and sometimes frustrating) but you know it's just going to get better and better. Reminds me of when I started messing around with AI art back when MidJourney was still in its early days and SD was still brand new. The amount the tech advanced in a short time was amazing.
Doesn't work /bark!!!
Hi! I generated a few songs, but I can't download the whole song... When I click "Get whole song" a different song is generated than indicated :/. Is there anything you can do about it?
When you do "Get whole song," are you doing that on a single clip or do you have clips chained together? The way it's supposed to work is you generate a part of a song (up to 1 min 20 seconds for the first, I believe). It generates two possibilities. You click on the ... next to the one you like and choose "Continue from this clip." Then you keep doing that, chaining the ones you like (which will be marked as "Part 1," "Part 2," etc). When you then click "Get whole song" from the last in your chain, it stitches all those together. If you just click it on a single thing... not sure what that will do. I just did what I've described above and it worked as intended and as usual so, if you're doing all that as well, um... well, I don't know what to tell you. Maybe refresh your browser? Also make sure you're changing the "continued from" clip - You should see it under your lyrics or whatever to the left of the screen. I sometimes don't notice that I haven't changed to the correct "continue from" clip and that can mess up what you expected to get.
Sorry if I'm misunderstanding your question. When I first started messing with Suno, I didn't understand how "Get whole song" worked. I thought I generated a clip that I liked, then clicked "Get whole song" and it would generate the rest of the song (filling in the rest of it, so to speak, in one go). It doesn't do that. It's just a way of chaining together the clips that you liked, but you have to generate them all separately first, choosing to "continue from this clip" on each one that you want in the song, in the order you want them. THEN "Get whole song" - clicked on the last one in the sequence - makes them all one song.
Okay, thx, it's workin š
Glad you got it working! So frustrating when you're trying to do stuff and it's just not doing what you expect. Enjoy!
out of tune problem, log #153, most generations still sound awful, instead of improving the AI gets worse
I'll see about using [break] to see if it helps. I'd been using various tricks like having it end on an instrumental or using (...) to indicate it should play on for a bit but clearly that isn't working or I've been lucky all this time and my luck ran out.
I'd write them off then. These are from Wednesday or something...
If a clip doesn't come in after several hours and is not some huge song that is being pulled together that is extremely long, it's probably good practice to just write it off.
if you want it to play a bit longer, you could probably just specify it in brackets
[15 seconds guitar riff] maybe?
never tried this, don't expect it to work
but i've seen problems with using multiple tags in one bracket like [Break, Build, Hook, Female Voice]
So you should probably separate it
I'm not sure that most AIs know what time is. Since this is a musical one, it might though so I'll give this a test.
It seems to have done absolutely nothing. š¦
You can try [short guitar solo] or [brief guitar solo]
I think I've had them work for me before...
You can't specify the exact length though
Hey i have a song that i wrote but it says that is unoriginal but i put it through multiple plagiarism scanners and it says that itās completely unique
Sounds like a glitch. They happen sometimes.
Well it wonāt accept my song it just keeps saying itās unoriginal
Try adding several lines of space. Try adding new metatags.
Add a new line of the song if you can.
Edit the lyrics with adlibs (Oh--, Ah, etc) anything to change the shape of the song.
See if you can make the lines different by breaking the lyrics into a new line where the vocals would naturally sing it.
Note the metatags you use don't have to be real or make sense. It's about changing the texture of the song so the glitch doesn't affect you and you are over the hump.
Okay Iāll try
Do you feel like sharing the lyrics here?
Maybe I can help deduce what's causing the issue
Yea hold on
[verse 1]
Hope into the inkwell isle,(inkwell isle)
Where the magic runs all wild,(all wild)
Run nā gunning on the ground,
Dogfights happening all around.
[Chorus]
Come on mugman, we gotta fight,
To reclaim our souls.(oh our souls)
From the Devil who is waiting for use to male or rolls.(oh our rolls)
Strange, it blocks the lyrics for me too.
Unless I remove the inkwell isle line.
Must be because that's a location in Cuphead
Maybe
You could spell isle "I'll" or "aisle"
It would sound the same and should pass the filter
ChatGPT also recommended the following:
Inkwell I'll Inkwell I'lls Inkwell Ayl Inkwell Ayls Inqwell Isle
I managed to get this to work:
[verse 1] Hope into the Inqwell Ayls,(Inqwell Ayls) Where the magic runs all wild,(all wild) Run nā gunning on the ground, Dogfights happening all around. [Chorus] Come on mugman, we gotta fight, To reclaim our souls.(oh our souls) From the Devil who is waiting for use to male or rolls.(oh our rolls)
https://app.suno.ai/song/040ca2b0-3b41-4e5d-8814-d6059f135e2b
https://app.suno.ai/song/8c614e66-da67-457f-8504-f39f19058d43
blues rock, indie folk, americana, catchy hook, pop, soulful warm vocals song. Listen and make your own with Suno.
blues rock, indie folk, americana, catchy hook, pop, soulful warm vocals song. Listen and make your own with Suno.
@livid raven
I had fixed it i just put a space between mug and man and took out inkwell and just said isle but thank you for your help
When it's anything to do with a copyrighted name and it's fair use to do it, phonetic are the secret. Glad you were able to write around the issue.
When I generated part two of a clip, it has a really sick instrumental, but the vocals for "verse 3" didn't generate. š¦
[Verse 2]
Hey, I'm Knuckles, guardian of the Master Emerald (Emerald)
City noises surround me, a concrete jungle I meld (I meld)
Past the street vendors and the neon signs
In this urban chaos, where destiny entwines[Verse 3]
Lost in the rhythm of the city's heartbeat (heartbeat)
Concrete pathways beneath my two big feet (big feet)
School days behind me, but the lessons remain
As I navigate the streets, shadows in the rain
(Yes, female vocals for a song from Knuckles' perspective is weird, but I really liked the clip with those.)
I just checked /info a moment ago, and something seems a bit off here
Says the subscription is cancelling on yesterday, and also 50 monthly credits? š
The remix function has NEVER worked for me, tried it loads and all I get is credit refunded
Maybe caused by the transition back to the free tier. š¤
Did you generate anything within the past 20 hours?
Ive got another 6 generates spining again
Remixing just re-uses the same lyrics and the same prompt.
If you were able to submit the prompt once, it should go through again.
Does it give any other errors? (Like the "original lyrics" message?)
It's possible the filter was changed since you created the song.
That would be my guess too, that something just hasn't updated properly yet, so probably not a big deal but i felt it worth reporting.
Also, no I haven't used any generations since i finished up i think pretty early on the 11th. (I saved a 100 credits in part bcuz was burned out and in part to confirm they get burned)
Aside from looking forward to a break, I wasn't planning on using the free creds mostly due to lack of commercial use, but I'll go ahead and spend them real quick to test. brb
The bonus credits will not expire. They will stay forever.
Even if you revert to a free user, the bonus credits will still be there to use any time
Well that was interesting..
Was somehow able to still do like 8 simultaneous generations.
Maxed out the free creds and slightly dipped into the bonus creds.
I guess the next step would be to see if they get refreshed on the next day or if it's really 50 creds per month. š
Here's the /info from after the generations completed:
Yea I wasn't expecting the bonus creds to expire.
And I was expecting my remaining 100 premium creds to expire, but wanted to test for myself to confirm anyways that they got burned. (and they did)
How long have you been waiting for them to finish?
Did you report them as bugs?
Btw am not in a rush for a fix or anything.
Am planning on taking a break from suno for at least a week or 2 anyways since I've got 13 raw suno tracks exported that I haven't even started post processing on yet, and don't plan on using the free creds, so I don't mind being a test dummy to explore this weirdness and see if it fixes itself. š«”
my hope is that in the future we can reconnect the parts somehow
i messed up a few times with that
I have two jobs that have been stuck for hours already.
hmm, might be
I got it to work once, but might be luck, didnt really measure the time
it was something like 5 second pause
Yes, it is a bit random still. It seems to help though.
This easter Egg is hilarious
Despite saying monthly credits I did receive the daily refresh.
Seems like just the text that's bugged, so no biggie.
same here
please revert asap
at website it shows me credits but in discords its not work
we're you able to narrow it down to what specifically is causing that?
oh generating lyrics mb
never tried that
still probably a single word triggering it or something
i would sub if it worked 100%
i was only able to make 1% of a song
wont let me continue a song or generate new lyrics keeps saying error.
I signed in with a new email to get free ones again and the same issue.
It wont generate new lyrics i write in when I continue the song on it will just remove the new writing i put in and only do one vurse and courus.
I bookmarked it and in the future if it works I will put you guys on my stream and make a full day out of it
I can't create songs, and I keep getting sent to a broken link. Has anyone else encountered this issue?
why its not creating complete song
its just use my credits but verse is always missing. any one know the issue?
Same here
works for me
I have enough credits but get this message ??
I still have two stuck jobs.
Oh nice,my $10 Pro Plan is $8 for you for some reason š
I paid for one year in one time
My guess is that you are logged into a different account on discord. Check if you can create in the web app.
no I can't
what message does it give you?
Good evening. I have spent several thousand credits by now until I realized that the following problem was not caused by me. I don't know why, but no matter what I enter, whether "Pop", "Ballad", or "Gospel Choir", out of 10 generated songs in these genres, at least 7 of them are songs that rather come from the Punk Rock genre. Am I overlooking something?
I have also updated and restarted the browser.
Could you post some examples? Find the share icon, this will copy the link to your clipboard.
This message will display..
Then paste the link here.
This way the team members can look at the output and check the backend for the generations.
I've got a weird one... interested to know if anyone else has seen this happen... I have a chorus that looks something like this at the end:
makin' everything feel so right (nice flight)
this thing sure is a happy sight
It's the INITIAL clip, so there's no previous clip I can go back to to try a different continuation.
When the clip plays, the last vocal output is the "so right", it doesn't get to the background "(nice flight)" ... usually when this happens I have no issue, I either add "(nice flight)" as the start of the continuation, or if the word "right" got too chopped off then I start the continuation with "right (nice flight)" and it's always worked for me.
This time, when the "full song" is retrieved, the words "so right" have been completely changed, even though if you go back to the first truncated clip it's clearly singing "so right". In the "full song" the line "so right" is sung as something that sounds a lot like "so rain"
I can't figure out anyway out of this, since it's the end of the first clip and the first continuation that's causing the issue. No matter how many times I try to continue from it, it's always broken in the "full song" version 7 parts later.
Hopefully the feature to reuse every setting for a clip EXCEPT for the lyrics is going to come out soon...
Sigh... the glitch is back tonight that clip#2's are coming out instrumental
I would rather not post seriously written texts here, that's why I have written a nonsense text for demonstration.
https://app.suno.ai/song/8c3bb8e5-8baf-4901-b3eb-7274dc102146
https://app.suno.ai/song/669e0303-7a8d-4fdb-9ec4-774a68d1e826
https://app.suno.ai/song/4bbc383d-962a-4ffb-b0e7-f90574c4063d
Hi, I'm experiencing a degradation of quality on every continuation of a track. When I get to part 5 its usually very bad. Anyone experienced this and any ideas? I can reproduce the problem.
I have enough credits but I get the message "You don't have enough credits to use this command. Subscribe and get more credits with the button below."
Wat shoudl I do?
User as.danjel spamming / selling social media accounts
#š¬āgeneral-chat message
hay alguien que entienda mi idioma?
Thanks for reporting it.
Has been taken care of.
"I'm not a fan of the music"... I keep getting this one when continuing a song. Not sure if I should be offended at a program's lack of musical taste, or file a bug lol.
got a screenshot? that sounds like an interesting error message lol
Wow, never seen this one, whatās it even supposed to mean?
I just did a search, apparently it's a popular theme! It did sound perfectly in the vibe of the song though. LOL
Not a message. It was what was generated.
Lolol
I forget it has a lyric generator. never used it
I guess itās the same stuff like āThank you for watchingā and (in Russian) āsubtitles subtitles subtitlesā. Itās some kind of model bug coming from the way the original dataset was preprocessed, afaik
Not lyrics if I got them correctly, itās āsungā in the audio generation itself
Itās funny until it gets annoying lol
Iāve heard āsubtitles subtitles subtitles N. Novikovaā sung to many more tunes than Iād be happy to.
Would you be willing to share a link to the song segment that produces continues with "I'm not a fan of the music?" (Or DM privately). I'm a collector of those.
Yep, not lyrics. Completely ignores your lyrics and sings that over and over. pretty funny
What an interesting choice for collecting š
This is the best one.
https://app.suno.ai/song/47da0313-2fd2-4d2b-963f-51a6e7f777f6
Post-Grunge, Post-Rock, Emo, Experimental, Male alto, 94 BPM song. Listen and make your own with Suno.
I love incorporate AI mistakes and hallucinations into the concept of the song itself. Used some "not a fan" generations starting at 4:00 here:
https://app.suno.ai/song/c8fe5736-663e-43eb-a64b-49655f20ed78/
parody pop catchy song. Listen and make your own with Suno.
This is great!