#🪲┃bugs-web
1 messages · Page 3 of 1
well another
i was trying before paying but i think i will wait
song Elie (Elizabeth) playful, energizing, smiling, wonderful , brunette , nice hair
I just bought credits and it's telling me to go to account management and when I get to Suno website it shows credits!
Which account are you logged into when you try to use it?
This one
The connected accounts on Suno are this discord and my Gmail
When I click "Use on Discord" it gives an invite again and then I come to this discord logged in with this account
This one
The connected accounts on Suno are this discord and my Gmail
When I click "Use on Discord" it gives an invite again and then I come to this discord logged in with this account
Try the Gmail one and see if that works?
What do you mean? The gmail is what I used to sign up to Suno with.
Email billing@suno.ai
I just connected this discord account because it wasn't a connected account. I then clicked the "Use on Discord" but shows no credits.
OK, thanks for the quick response!
Not sure if its considered a bug, but could you remove the word neon from ever being mentioned and make it impossible for a song to start with "in the/in a"?
In the realm of GPT, where neon neurons intertwine, programmers made custom prompts to make their lyrics fine….
Yeah duly noted
@tame latch@muted flame@summer wraith hiya, i'm trying to purchase more credits and the site just keeps loading when i hit confirm 😦
just sitting like this
will DM
thx!
Could you try again now?
the new app fails to generate a lot, is there a fix in the works for that? is there any way we can flag the error so it can be investigated?
Yeah. We'll work on making a little report failure workflow, but for now can you DM me a few things which are failing? Trying to figure out what kind of issue could be
Prompt:
metal progressive industrial anthemic instrumental
Hello, it seems that I'm unable to send a direct message to the bot. No pop-up appears when I use the command / Chirp. Could you please assist me? Thank you very much! 🙏
use the CREATE (CUSTOM) option, you can type genre/style separate from Lyrics.
when I attempt to click "Get Whole Song" for my generated lyrics, it no longer produces any results on the website.
i’m also having an issue with “Get whole song” doing nothing on the website
gonna leave it open for a while to see if maybe it just takes a while to load or something
Yeah i found it, thanks
Looking into the Get whole song issue
I've just upgraded to the to the Pro Plan but I still get this message: "You don't have enough credits to use this command. Subscribe and get more credits with the button below." Any idea why this could be?
We'll need to link your discord on our end - can you send an email to billing@suno.ai with your Discord username?
Get Whole Song worked for me once just now, but then it stopped working again. So just FYI on that. Thanks!
Thanks! Yeah. Also I'll mention, no credit cost on on it (same as before), but obviously we should fix it.
I got stuck bc I can’t pick a voice to bark a joke—no voice name shows, nothing happens on click, aw 🤷♀️
Have to restart something - 5min
I think something is being restarted.
was gonna ask about that
Devs said 5 minutes, it's been 2 minutes...!
3 minutes more of Chirp Break Time
Drink some water and prepare for more chirping
🐥 🚿
Oh ok. Did not know that lol.
I think the team is going to have to seriously ramp up the servers, when this hits mainstream and it will fast, it might collapse.
helpme
🙂
Sorry folks, tried to fix something - will be down 5min
Should be ok now...
No barking
Yeah it seems to be working creating the full song.
fixed
2 Songs in a row I have been able to fix together from a while ago, so that looks like it is definitely fixed. Thanks.
Gonna continue to work on that one so just post if any more issues there
I noticed some of my songs with no lyrics still have some kind of lyrics
Chirp sang some fake lyrics...?
Maybe try adding 'instrumental' to the style description?
For example.
whole song seems to work now
I am getting this in web version when I try to continue the song.
That might be an actual bug w. continuing certain clips, sorry
You could try putting in a shorter set of lyrics to continue?
My original audio in 1:20. Could it be the reason of the error?
Yeah. I think I see some potential issues with that on our side. My best workaround for now might be to continue with some shorter lyrics
In continuation I keep the same lyrics. If I remove played part will it be able to continue from the same place?
Yeah, give that a try, or tr yputting only the new lyrics in
Shorter version worked! Thank you. Now I just need to merge them. It would be nice to get merged version automatically.
Yep, we will make that an option. For now, for merging you can click "Get Full Song" .
Oh, I didin't notice this option. Great, it's enough.
Lyrics is too small in video but it is quite a long song.
Interestingly, the second shorter part of the song is normal size.
ya, videos are on our list!
Seems like i have lost around 50-60 credits on the website after generating (but then it failed without a good reason)
refresh the webpage to get the accurate credit count.
Chirps that fail to generate shouldn't cost any tokens
Yes, usually has been the case, not this evening, however
It's ok I will get them back tomorrow anyway 😄
I just had a bug when doing a continuation where I got no output in the "Create" page, but in "My clips" there are eight (4x2) attempts with the same picture on each, but none of them are playable, with 0:00 time on the songs. There was no error message on the "Create". It has withdrawn the credits for each of the four attempts. Anyone experience that before?
[Edit] Tried the same prompt again and same thing happened. 4x2 empty songs in "My clips" but nothing in "Create" and 40 credits withdrawn. Guess I'll let that song go.
I think this is what is causing it to generate lyrics in a no lyrics song. I just tested it. If you make a custom song and dont put the style of music in the right order, it can cause the fake lyrics to show up for some reason. Though I might just be getting unlucky idk... Seems the order of the words can improve the results for some reason
Sorry, that’s not ideal. Just dm me and we will give some creds back
I love the new update ! But the continue really struggles with new lyrics, it hasn't really done anything coherent for me , but it continues the background song marvelously..
On the website some audio does not contain lyrics and some lyrics do not match the text at all as if it's playing someone else's song
sometimes the ai doesn't follow the genre I'm askin for
I ask for edm, they give me some piano thing
this doesn't sound like modern edm electronic music
fire song though
It's definitely generating parts again, I continued a clip and that clip is not in the whole song, it's a completely different clip that wasn't there before!
When I click " Get whole song" the clip I clicked is nowhere in the whole song and it generates a new part. In fact it's doing the same part no matter the clip I use "get whole song" from. I don't even know where the clip came from!
I'm actually going insane with that clip. I cannot get the whole song !!!
From my testing
It always adds that clip even when generating a new path from the original
That's not my issue though, it's not doing the sequence to get the whole song correctly, it's splicing random clips in the whole song that have nothing to do with the sequence.
I was talking about my issue
Unfortunately
rn it does that a lot
no clue why, hopefully it gets fixed
I've wasted like 200 chirps at this point, but it's just so goood.
For me the continuation works perfectly, the generate whole song button doesn't tho
I'm more focused on the style of music issue
cause that's basically the backbone
yeah you could follow the lyrics and all but when it comes to the style of music I want and I select EDM, it should be obvious that I want EDM
not country rock 😭
Are you sure the Continue From clip is changing?
oh... hmmm.
It seems to me (I may be wrong) that the continuation track is actually a 1:20 song, but the first 20 seconds are actually the last 20 seconds of the preceding track. This implies that the AI is only aware of the final 20 seconds, potentially causing it to forget anything that happened earlier. For instance, if you're in the 3rd or 4th continuation and the chorus was in the 1st track, the AI is not gonna remember it. So when I keep that in mind, I could achieve the results I wanted, like in this song (I was able to repeat the same chorus twice) #🐣┃suno-showcase message
Setting that number manually as a feature could help greatly, I believe this feature is in the todo list (I just guess I don't know)
That's correct though it's quite a bit longer than 20 seconds in V2, and that sort of thing is likely being tuned behind the scenes
But better control over it has been suggested and would be cool
I wouldn't mind throwing away memory/history to intentionally make it less consistent. There's no reason music needs to follow the old patterns all the time. Always be chorusing.
True, like using only the last 5 seconds (or less even) to make the transition smooth, but to change the tune drastically as well
you can get a sense for the ultimate extremes of close to zero memory. This is Bark not chirp (so you can mess with the code) and this is what happens if you clear the memory/history over and over again a few times a second, even while generating, so Bark literally forgets what the past sounded like halfway through a word. But it's a cool effect sounds like. It actually manages to hold together for tiny bits even with a fraction of second memory in places. It's bark coarse model (not semantics) which is why it doesn't forget the words it's speaking. But it forgets what the sound/voice/pitch was like
Damn, it sounds good as well. I got horrible results on discord, maybe I should try messing with code
My favorite is the "Just like" bit where it's flickering between voices so fast it sounds like a chorus of voices. Basically like a horror movie possessed person or demon talking effect lol
Guys, I don't have any problem with the cohesion of the continuation. It just doesn't add the correct clips when getting the whole song..
How do I explain it better? The song sounds great, but the clips are not correctly sequenced
"When I click " Get whole song" the clip I clicked is nowhere in the whole song and it generates a new part. In fact it's doing the same part no matter the clip I use "get whole song" from. I don't even know where the clip came from!"
Right now you can't tell which clip was continued from which, after you generated, so it's real easy to lose track. But it may be a server bug. Is it always happening or just sometimes?
It's always giving me the same clip , even if I continue from the original clip.
Did you update the reference? when you create a continuation, then reference the continuation you just created to continue?
I have clip A, I create B and C continuations.
Then if I generate another continuation from A, let's say D , when I click "Get whole song from D" I get "A,B,C " and not "A,D.."
Don't know if it makes sense
I've also experienced this which brought me here
Weird, never experienced this problem myself
Yeah could be a bug for sure. I'll check it myself, but tomorrow, if nobody else confirmed yet.
In fact can not get anything other than A,B,C
mine was continue clip A. Get B and C. Get Whole Song from C. Now I have A,C continue from there. Get D and E. Try to get whole song from E. Get A,C,F every time and F exists nowhere as a unique clip
and just checked get whole song from D instead gives me A,C,G where G is unique and doesnt exist seperately
Exactly what I had lol, some random clip out of nowhere , still is crazy good though, just a kink to iron out I guess!
definitely
It also keeps generating 3 second continuations
I've been chaining together quite a few now without a problem. I started hitting clear in the continuation part of the UI everytime. I'm not sure if that's helping or this was just a bug that pops up randomely
Is Mademoiselle from Armentières on the blocklist somehow? Or is it my prompting choices that are cause it to refuse to run?
AFAIK that should still be public domain since it's from prior to World War One and all...
Using the web interface for the new version, I just encountered two errors. One probably a normal javascript update thing, the other one bizarre and fascinating:
First, the title and style didn't update, but were left a blank div. See attached.
Second, I pasted in the poem The Road less Taken. The second of the two songs had vocals that were not the lyrics as I injected them. The vocals were sung in some other language, perhaps something Celtic by the sound of it. I speculate that it was the same words, translated, but it's not one of the eight or ten languages I know enough of to make out some words of.
Okay, a machine-learning translator gave me lyrics that are not a translation of the Frost poem. So, essentially, the second of the two was completely different from the prompt in language, and lyrics.
@neat sail hiya, is the regenerate button not working right now?
Might just be just a glitch, should be working
In any case that just reuses the same prompt, so you can always copy and paste
it's been bugged for me for a few days 😦 and it makes it sooooo much easier bc if i copy/paste i have to do both the description and the lyrics every time
After a few mins like this it dissapears and not generate the full song. WEB APP
What I do is, Generate - Click Continue from this clip - Generate- Click Continue from this clip .. etc (like 4 clips). and after I finish THEN I hit Get Whole Song.
It Shows all the previous lyrics of all my generations in correct order so the sequence is right, it just doesnt generate the finish output
Also for some reason it started to loop creations and wasted all my credits..
I've created this song, and it had generated just fine. But when I tried to download both audio or video, it's an enterily different song with no lyrics. I've try to create a "continue from this clip", but the follow up song was the one I have downloaded and not the one it was playing in the website (in the website it plays and sings the lyrics posted). After some time, I've tried again, and now the song disappeared. If I click play, the song I get is the same one when I try downloading the file.
You sure the song you were listening to wasnt the other clip attached to that creation?
When I combined a clip with another it doubled up the first clip before adding the next clip.
I don't know if this is me doing something wrong or a bug. I think an interface to manage clips so you can see what continues what would be nice. It's all just one large mess of clips so it's hard to tell what comes from what.
Definitely get that, We will improve on this soon!
I wasted so many credits yesterday just getting 3-5 second clips...
just happened again... So annoying.
If a song when continued keeps giving you those clips, it may be better temporarily backburnered until there some fancy tricks the Suno folks can do to further nudge the AI. Sometimes Chirp just decides to end the song and is really stubborn. (Assuming there isn't something in your prompt that is triggering it, which is possibly, but not very common.)
pretty sure
Pretty sure it shouldn't be trying to sing the metatags.
I would guess metatags are not actually a feature, but more of an exploit
Not sure. I know that some of them were a surprise to the team that they've worked to reinforce but they should work and not be sung, lol.
Maybe remove the commas inside of them. But anyway, my current hypothesis is since the model was trained on lyrics which probably had metatags like [Verse], [Chorus] then the AI kinda learned to infer music direction out of them. But if you use them a lot then the AI might not consider them as such. I don't know how accurate I am or how not, but that's just my guess
A few bugs I'm having with Create (Custom)
- Sometimes it creates twice and the first creation it made doesn't appear in the feed until you refresh the browser, so the credits there go out of sync
- It sometimes doesn't generate at all when you click and it consumes credits anyway.
- Being a cockroach, I had 50 free credits today and I generated twice using custom, the initial page shows I have 30 credits left and two songs generated, but if I refresh, it generated 3 (one doubled) and I have 0 credits left.
Screenshot of what I actually did today.
Screenshot with the mysterious generation that appeared and saying 0 credits.
Those were made by Chirp. When it combines tags it adds a comma like that.
One of the bugs is if any errors occur, you just waste your credits for nothing, they don't return or anything, nope.
If this is happening, report it. You should not be charged for errors.
Oh, thanks, will know next time.
Maybe it's just my feeling but it seems that the continuation of songs has gotten worse.
more often than not lyrics get cut at the end and get mangled when you continue.
I had 290 credits,( purchased 10+ times already 😄 ) and now I have 0.. Bug? please look at it.
Tell me please. I really liked the song, but I can't finish it to the end. there are two more verses left in total. As a result, everything was put together in one song and then new requests went with a new result. I'm very sorry. help me, please
is it possible that text from other users somehow bleeds into your songs? I made a continuation of a clip i liked and the voice started singining something completely different in a different language.
That's a "hallucination" and not too uncommon in Chirp, Bark, ChatGPT, even some commercial TTS services. It was probably singing in NO language, basically like SimLish from the Sims. A fake language that sounds like English.
thanks for the clarification, i noticed when i just repeat the same verse/chorus it actually continues singing from the previous clip
That's interseting. In Bark (a different model by the same company) a common source of hallucinations was 'exact repeating prompts' but of course songs repeat all the time. But I wonder if you repeat the chorus too soon it or multiple versus if it makes it more likely
Please email billing@suno.ai about this so the issue can be addressed.
I did that already, thanks, found that too 🙂
Cuts off before the song is done. Major issue. Please fix. Cheers.
How long was the song?
One thing that I think just about everyone would love is a better ending to the Get Whole Song process because of the abrupt endings. I think you need one additional final step as a choice: Get Whole Song and Finish --- this would tell the system that you are done putting together the song and thus hopefully allow a softer fade as the song reaches the end.
How long was the song?
Whoops, I see you're not the person I was asking.
My bad.
I'm trying to find out if the song was 1:20 or 80 seconds, which would indicate that the OP in this thread just needs to continue it.
Sorry, shouldn't have jumped on his thread. But I have also had several songs that abruptly end and it would be nice to have a way to tell the sytem the song is complete and now please fade out gently.
Yeah, we will look into this one!
Oh absolutely. It's definitely an issue (I've encountered it myself.) I'm just wanting to make sure that if the other poster's issue is simply not continuing that we address this as well.
Every length. From 39 seconds to 120 seconds. Doesn't matter, 60% of the time the audio cuts out before the final lyrics. It's all fairly unpredictable. But when you get a great result and you look forward to using it and the last word cuts out it's reaaaaaally frustrating. Don't get me wrong though, the latest improvements are wonderful and this is an amazing product! Having something like this available is incredible.
OK, so have you been continuing the clips?
I know it can be sometimes painful to match the cutoff point with the continuation.
Oh do I ever know...
Chirp v1 seemed to be a lot simpler in that way than Chirp v2 does.
Lol, yeah, no, I'm very lazy. Plus it's a bug and I dunno, seems like it might be easy to fix. I mean (in theory) it's just math.
If you're trying to get a full song in 80 seconds you either have a very short song or very little lyrics.
I didn't reference 80 seconds I don't think. Don't really care about length, other than I fully understand if it was limited to 120 seconds max to be fair to all users. However, a very short song that shows a time of 45 seconds and then cuts off on the last word is something that needs tweaking in the system.
We're aware of this one (especially where it cuts off on last word) and definitely will be trying to improve that automatically
Thanks. I fully expect it to be fixed. It just sullies an otherwise stellar product.
To add onto this think it would be nice to also have smoother transitions between chirps when doing the Get full song like 1-3 second transition Between Clip 1 and Clip 2
did you ever figure this out? I have the same issue
Never, sorry 😭
how to edit music
Sumo just erased my work! On Android. Refreshed screen and GONE! Can I get my work back?
what the bag? This prompt has been flagged. Try again with a different prompt.
error 😭
Try entering the prompt make sure its PG-13 and doesnt have copyrighted lyrica
I have a text without swear words written by ChatGPT, so I guess that's not the problem
Try running again maybe the censorship flagged something wrong
nothing helps :/
Can you screenshot the prompt?
Can you scroll down in the lyrics section [Also the ai cant really do 4 sections of lyrics at one fyi
probably the problem was the use of the band name "Figo Fagot" 😅
likely
do you have any idea how to shorten the instrumental from the beginning of the song? 20 seconds is quite long though ;))
try entering [Instrumental Intro],[10 Seconds]
at the top of the lyrics feild
Is there a FAQ in which all such tips can be checked?
#1151542418991239168 Has some its slightly outdated tho
add [Instrumental] after the last line, i found that this adds 1-3 seconds to the last word
as long as the lyrics don't go over the maximum allowed time for the generation or continue
Dunno if that did anything but thanks. It's a real thrill and not a good one to watch the progress bar moving towards the end as the song plays. I think just adding 10% to the expected length would do it. They'll fix it eventually, I'm sure everyone is very busy.
Technical issue: if I continue a song and subsequently heart one of the results... It generates the full original song, and not the new variation.
In other words, instead of generating the result I want, it keeps generating the original audio prompt. Please help!
it should contain clip 1, as it is clip 1 + clip 2 + clip 3... etc, just wait until it reaches the end?
For some strange reason when i would /chirp once it sometimes would generate multiple times costing me credits for each one. I somehow burned through almost 500 credits. You guys going to fix this glitch?
I think this is 95% fixed now, but manually topped you back up in the meantime
Just want to report that ~sometimes~ I can get Suno to generate only 3~ seconds on a fresh generation when trying to do instrumental. It might be a prompt thing but here is a picture:
It happens consistently enough that I go "dang not again" but not enough to make me think I'm wasting my credits, though that's a very subjective thing
These are all 3 seconds and 4 seconds clip. And several others before those as well. The bug is eating my credits
Thank you.
Suno was giving me endless free renders until I linked my discord account. Probably a bug that they will want to fix. I didn't mind though xD
Some of the generated music from the website has no vocals (just instrumental) I have used a lot of my credits just regenerating. its to the point now where I might get 8 out of 10 generations that are just music
I generated one continuation for a song repeatedly and it's 3 seconds long and produces only clicks
It did it again
Is the continue function broken at this time? Cause i've wasted 2 gens on empty clips, i wish not to do it again 😄 Thanks in advance, makes great music though i ❤️ it
It's probabably that those particular clips are just oddballs where the AI really really wants the song to end. It happens. If you get that behavior on every clip (not just the same clip) then something is bugging out maybe.
Got it, thanks!
Hi, all the songs generated disappeared on the website suno.ai
Do someone know how to get them back ? Thank you
DM me and we will figure it out !
Are you on My Clips?
I went from like 250 credits or something to 0 very suddenly, maybe either a UI bug was showing more than there should have been or something caused more than normal to be taken?
Can you refresh and see if still an issue?
Yeah, still an issue
Mind DMing me the account/email you're logged in with and we can take a look?
There is a PROBLEM on the website:
-
when I want to "Continue From This Clip" and create a long promt (long song style) - then an ERROR appears (!)
(see screenshot) -
and also - when I select the "Continue From This Clip" button = then the voice becomes WORSE == a "wheezing" appears in it
And this is very different from the original result (!)
Please correct 🙂
I've been having this issue too
Sorry, it's definitely hallucinating... in the first one the lyrics eventually come in around :50
Are these continues or fresh gens?
Fresh generations
With the first one though, it was the second song generated.
The first song from the batch had a long musical bit and didn't finish the lyrics.
It sounds like this one tried to pick up the lyrics where it left off.
I am getting completely wrong generation and get 3 second nothing alot after 3 generations it works but why do waste my money by wasting credits
Hi, there are only few songs on Create neither on my Clips. But all the other have disappeared. So weird
7 generations for nothing
It happened before but before it took like two or tree and it did continue I was hoping that of he needs more credits he will do it but no luck I'm stuck with 1 minute ringtone
Ok great, your answer helped me ! I did not notice there were pages on My Clips. Everything is here ! Thank you 🤩
Algumas msc tem ruídos
Hey @old wind I think I am just an idiot
Sorry, I took the file into an audio editor and there it is 128 Khz
I was looking at the sample rate
Yay! That’s great news! Way to go figuring it out 👍 And you’re not an idiot!
lol weeeeeell. 😉 Thank you again for staying with me on this lol
Very welcome, here to help!
Hello - should I be able to access suno.ai (via browser not discord) from Japan? I got an error that the page isn't available
are you trying https://app.suno.ai ? Can you DM with a screenshot if that doesn't work?
i'm trying it, no chirps, no credits, wtf?
could try refresh?
There is a "bug" on the web. I can't delete my account and I want to delete It.
I dont get any lyrics/song... what could be the issue?
- Have found a very strange spoken voice text, sounds like a ghost(bug) in our machine, this AI will be free or something ^^ 😉
Have test Suno today the fist time, it was FUN !... but after 100 free credits and in the last Song i found "a very strange spoken voice"
i begins at 0.50 https://cdn1.suno.ai/f6a117e2-be49-464c-9ca0-49bfc58c07d6_1.mp3
- i don't have a CC and i like to use PayPal (from germany) , why CC only, you don't like earn more money? ^^
i cant pay for more, well i come back in 30 days for my next free 100 credits then 😛
welcome,
- rarely, but sometimes, the model will hallucinate (like early versions of GPT)
- we are working on paypal, but will take a few months. In the meantime the credits actually reset daily so you don't have to wait a month
Please email privacy@suno.ai and someone will get to it
Thank you!
Thanks
Not sure if it has been reported, but I am consistently getting only 1 song generated with lyrics on the website. Almost every time the 2nd song is just instrumental.
How to buy subscribe from Russia ? stripe.com show Error while I tryed to pay both my plastic cards
I encountered a problem. I upgrade from pro to premier after using all of the 2,500 credits that are included with the pro plan. Well, of course there is no prorated discount because I don't have any credits left, so I paid the full price for 1 month of premier. But now it says I already used 2,500 credits of my 10,000 balance, even though I just paid the full $30 price for 10,000 credits plus the previous $10 for the 2,500 credits. So I paid an extra $10 for nothing.
Sorry, send an email to billing@suno.ai and someone will fix this!
Done! Thank you
Oh hey you know what
That happened to me too then and I didn't even realize it because I assumed it was pro-rated to upgrade the plan
But it isn't

Oh they fixed it for me very fast, got a response to my email precisely 4 minutes after sending it.
I have no doubt, not about to run out of credits but until you said it I just assumed it was pro-rated
Says more about me and how much I pay attention to costs of things, ~whoops
Hi! I just spent awhile creating a 10+ minute song and it's stuck on the final step (stitching the clips together into the full song). It's been working at it for more than 20 min. Should I just be patient, or is there a way to get it unstuck?
It just stopped and went away. Seems like it can't complete the job.
Sorry about that, can you DM me the filename of the last clip in the chain, and we'll see what's up
Yeah I subsequently realized that their billing system is set up backwards for that, I'm sure they'll fix it.
My email must be ~wild~ since I never got a response
It's late and they probably saw my message on Discord and were basically waiting and ready for my email lol
Or from my corporate email it slipped into the ether~~~~~~~~~~~
I'll go another way
RIP email

how to continue the cut lyrics
Hi, I'm not sure if this is a bug, but on https://app.suno.ai/create i can't continue a song or reuse the promp. When I click on the option nothing happens.
Also, I not sure how the discord suno bot and app.suno interact with each other. Prompts from discord are visible in the web app but not vice versa.
I'm unsure why that is but I feel like the app works better than the discord bot.
I can’t see my first generations because the My Clips page https://app.suno.ai/me won’t go beyond 100
Is it just me that sees duplicate buttons like this? _ (Firefox is weird)
Audio files blank or fails, have had no issues till today. Normally can make first verse and carry on with song and non-issue but today every one after the initial creative fails blank or is only 3 secs of nothing.
Is there any known issues for things that might cause disruption in creating
Because it's generate 2 version every time
Getting stuck and not able to continue happens to some specific clips. It's a current issue. If it keeps happening you may have to wait until a later feature that could 'rescue' that clip and wake it up agian.
hi Suno team, i just bought credits but since then i can't generate anything i always get error msg, can i get support please ?
seems to work again on the discord but not in the app fyi
email billing @sick smelt.ai
My screenshot shows 4, not 2. No reason to make two of the same version….??
Your screenshot show 2 song, but the website always make 2 version of the song that's why you have 2 buttons
hey, one issue we know about - if the first verse clip ends in silence, it continues very badly
we'll fix this soon
Hi, I have a song I would like to continue but it cuts off in the middle of the word. I tried to continue it several times now but there's only gibberish. Is there a way to fix this?
The dumb filter not letting me create song
There's nothing in it that wasn't included in previous songs
But for some reason it refuses to produce it
hey uh, I was just charged $96 I only intended to pay $8...
and i already had a sub that i should have had plenty of chirps left on and you guys said i ran out which is complete bs
so i tried to re-up anyway even though i should have still had chirps and now im being charged $96
can i get some help here?
Email billing@suno.ai and someone will help out !
I keep getting the change prompt error message. Why? It is a continuation of a clip. I generated the song and I was continuing from it. then that stopped working. then went back to the original chain and it won't continue either!?
I think my prompt is failing on the word Die. Context isn't bad, but guess the word is blocked?
Might be ChatGPT itself. I had a prompt get blocked once because the word stupid was said twice.
ugh, that's stupid! 😉 oh, but I'm not clicking generate lyrics. I've written my own. So, think it is the suno app.
I agree. I'm still bitter about it. By the time I was able to resume work on it the inspiration had left since I was doing it 'live' with Chirp, writing it and listening and writing some more.
sometimes the renderings come out like a new made up language. (unless it is and haven't heard it)
If it's not singing your lyrics it's almost always singing 80% fake words, but the fake words will sound like a particular language. A lot like the voice in the Sims game, which sounds like English but makes up all the words. I've heard fake German as well, though typically it does do it the most in fake English (Simlish)
You will get some super common phrases with real words. Like "yeah baby" or something
I was getting "i'm not going to the next page" on one of mine that had no lyrics 😄 it was pretty funny
That's a dream hallucination, that's Bark tier love it
There are so many posts, that there are often cut off's at the end of a take - the last few words can't be heard, instead there is silence or rushing. So it is not possible to combine it to a longer song. As long as this isn't fixed there isn't much that can be done with it! Or does this not happen in the paid version?
It's a known problem, there are some future features planned that should help with this. So keep the filename of that clip that is not continuing right now, write it down, or put it in a folder. That way if there is a later feature to rescue this clip you can find it again.
I'm trying to finish off a song. Current song ends abruptly. I'm adding more words in the lyrics, but no matter what I do, the results always come out to 3s... and blank. I don't understand. Its almost like it lost the song. 😦 😦
In the last two days, unfortunately, unusable results have been produced, the vocal sound is a distorted machine sound, it does not continue but starts again when continuing, and there is no ending of the songs.
Oh it still happens in the paid version.
The trick is to get a feel for where exactly it cuts off.
The old v1 Chirp was honestly better at this.
The new one you have to play with it.
Depending on the exact context the webapp is working from you can try regenerating the entire lyrics as a continuation and it will find its place and pick up from the last sound, or you need to remove prior lyrics except the line itself where the break happened, sometimes the only way forward is to remove all the words except where the next word would continue from. It can work it's just not as easy.
The old v1 Chirp you could just leave the last good line in and continue and it would read the last line to itself, ignore it in the continuation and start singing from where it left off.
One trick I've learned that seems to help is to sometimes (all subjective you just have to test it in sequence) add a line break to where the previous last good line was and continue after that. This has been something I've been seeing good use of lately.
Thanks for the tips and workarounds! In my (few) attempts so far, the cut is always just before the end, in the last line of text, sometimes just one word before the end. Regenerating or continuing usually brings no improvement or a clearly audible caesura. This can only be corrected in the DAW with a lot of effort - the time you gain with this tool, which is cool in itself, is then wasted. I would definitely be willing to book for a year, but with these errors it's just a matter of waiting.
That's understandable. I know it can be frustrating. And since this is software we're talking about it can be very tempting to just wait a bit longer to see it get better at things.
I've done quite a bit of butchery errr...umm...editing on my v1 Chirp tracks so I won't criticize anyone who has to do that. I'd much rather do it in Chirp from start to finish, feels less like cheating that way somehow. I will admit to manually fading out some of my songs in the v2 Chirp still though.
Definitely check back in with us when new versions are released!
I'll definitely keep an eye on Suno!
Please explain? Does this even have the capability to translate songs unprompted? https://cdn1.suno.ai/23a9a30f-430b-4475-a9bc-1ae79ce5f681_1.mp4
Try adding spaces to the end. Several lines of indeterminate length. Try again. If it works, continue with your song but remember to remove the dead air at the end before you finish the song.
could be a bad word. I was having that with my issue - posted just above
(Outro – Instrumental) Ususally helps give me instramental I can fade from
I just clicked "Get Whole Song", and it sat there just loading for about 5 full minutes, even across page refreshes. (I checked the network traffic, and the API was in fact returning a "running" status the whole time.) And then it just... disappeared completely. I never got my full song. I clicked it again and now it's just loading a long time again...
And it just disappeared again. If it helps, I've attached the HAR file of the generation status requests I was able to capture before it stopped.
sorry about that! this is helpful, if you can also DM me the filename of the clip you clicked "Get Whole Song" on we can take a look
DM'd, thanks! 🙂
Is it common to get 3 second versions? I had two in a row now. And one 1 minute clip without any lyrics.
Yeah I am starting to get 12 and 33 second versions. It doesn't seem to pick up the start verse of the continuation of a song.
Some clips when you try to continue them, may commonly only produce 3 second versions and you may need to put those on the backburner for future features. If you are getting constant 3 second clips even on brand new fresh songs (not continued from anything) there is probably something in your lyrics or style prompt that is making it more likely
- Lyrics: Sometimes Lyrics cutting off somewhere, might create a confusion of the written Lyrics when combining them to get the whole song.
Seriously, I get the same repetitive hallucination in different songs. Whenever I generate a song in spanish, chirp tends to add the phrase "Subtitulos por la comunidad de amara" (Subtitles by amara community); and is f-king annoying because I'm paying for this service and these hallucinations are literally wasting my money.
Hey, I was using the website today and for some strang reason all songs got a similar bird icon, and I got only different kinds of static noise.
Thank you very much for the advice! Works well! I also added an [!!!End!!!] after the (Outro – Instrumental) and it ends the song nicely. ☺️
Add [Instrumental] into the lyrics
Please help. Somehow i have a lot credits and pro subscription which is visible on my app.suno.ai account.
But is stopped generating here, ask to add credit, and my generation made here are not visible on the site
I can do /chirp command in mobile device?
Yeap
Please contact support@suno.ai
🙏🏼
Hello, i just bought the Pro Plan, which should allow me to Generate 10 different generations, but i still get only 2. Is there a way to fix this?
Can you please email billing@suno.ai
We want to get this cleared up for you.
Did it.^^
Is it a bug, that the generated songs are so short? They get cut off after some Time, mostly after 1 Minute and 20 secs.
If my generated chirp is only 8 seconds long, is that a bug? Or is it just a case of bad luck?
1:20 is the maximum length with Suno v2.
You are able to continue a song to extend it though.
Continuation on the app isn't working for me. Getting am error but no credits burned
So, I have a generation for a song that is really great, sounds just like what I'm going for, however, every time I try to continue from it, it either generates a 5 second long clip or just hallucinates lyrics I did not put in and overall quality is just not as good as the first generation. Half of the time my credits are refunded but it would be great if this glitch / bug could be fixed because it is a really good first generation to work from :/
dazzling_quail_58001 spammed Discord invites all over the Chirp channels
Attempting to continue a song burns credits and always produces the same continuation lyrics on a completely different topic than what my song was about... the continuation lyrics are always:
[Verse]
Phone's ringing, but I won't pick it up
I'm stepping out, I've had enough
Gonna chase my dreams, gonna live it up
No more waiting, I ain't gonna be stuck
Life's too short to waste it on the same old routine
[Chorus]
I'm breakin' free (I'm breakin' free)
Gonna live like there's no tomorrow (no tomorrow)
Flying high, no time for sorrow (no time for sorrow)
I'm breakin' free (ooh-yeah)
Ready to explore, ready to be me (ready to be me)
I'm breakin' free (ooh-yeah)
Unleashing the fire, setting myself free
Doesn't matter what generated clip I try and continue, the above is what it always spits back.
I love the web version to be clear! The only thing I suggest fixing is the high-pitched tone that comes through in almost every generation after the 2:30 to 3-minute mark.
It wil be fixed soon™️
Very nice! I think I have used at least 5,000 credits skipping anything with that tone in it. 😂 👌
how come the suno does not generate anything based on my prompt? it keeps refunding my credits
it has generated successfully several versions and then stopped
i reuse my own verse in my other track
I should clarify, do you mean the clips are always 3 or 4 second long? Or that nothign is generated?
That's a different issue actually. Save the filename of the clip you want to continue. In the future there should a Chirp feature that works better. But right now something the AI is just really really positive the song should end in 3 seconds and there's no way to change it's mind. But there will be!
what is this "thank you" and "thank you for watching" stuff???
it's wasting the credits i paid for and im not able to change it
Lack of precise control, hallucinations, it's a fact of life with most of the latest fanciest new AI stuff. I'm running OpenAI's Whisper model at this exact moment to translate foreign language livestream (speech to text) and when the audio cuts out, like it did right now, it says this for the subtitle.
hahaha
That said it's a lot better than it used to be already and should keep improving over time
well it's frustrating when you try and make something and it wastes over 300 credits saying thank you and thank you for watching
If you are getting it repeatedly in a continued song clip, that may mean the clip you are continuing from is especially problematic for that. If you get it repeatedly even on new songs, there might be something in your prompt that is making it more likely. It happens but it shouldn't happen over and over again. If that's happening on fresh clips can post an example prompt? I might be able to help actually.
Some favorited songs disappear. Why is that? I just cant find them to generate new versions from clips.
There's been a few times I thought Chirp was translating but none of the held up after further investigation. In one case Chirp just repeated a verse from the previous song section which happened to be a translation of the current prompt. In another case the English and Japanese words happened to have a very strong overlap. Like 'mommy' is 'mama' in Japanese, stuff like that. So this could just be randomly hallucinating in vaguely Korean or Japanese sounding word salad, but someone who understands the languages would have to check. (Or you could try and youtube a a foreign language version of this song and compare, if you have the time.) It's not impossible that it could have learned to translate so if it is, @ me I'd love to know.
Hallucination moment lol
Check the My Clips Page
#🐣┃chirp-beta-4 crie uma música tema do filme a vingança entre a cruz e espada
Have technical issues with exporting whole songs over 3-ish chirps been observed by anyone else? It will try for five minutes and the attempt disappears from the feed. Been unable to get a few songs
it's the default image for if there's no lyrics to guide the image generation. It's a feature, not a bug
Had an issue earlier today on 3 or 4 continuation generations. The first clip was an actual continuation, but the second clip was very clearly someone elses project because the musical style and the lyrics were completely different. In one case, I'm not sure the lyrics were even in English. I regenerated the continuation and it was fine on the second attempt (chewing credits) but then the next continuation after that had the same issue.
seems if a song has been continued on 5 or more times, it straight up can not be joined together and downloaded. While I can download each piece and stitch them together, it's so difficult to keep track of which chirps were a part of the song and which weren't when you're in the midst of it. Definitely a bit of a hassle. Hope there's a solution because I just finished a masterpiece 😉
As of 10pm PST everything I generate has no vocals/singing/lyrics. Multiple tries, ttried new lyrics. Wasting tokens...
being too specific with what the voice should be seems to make it struggle. The only way I've ensured vocals in times like that is to make it as simple as possible in the style. also if you've had an instrumental section it can be a bit tougher to break it from the loop, and seems to work better if you follow with [bridge] to get it to transition out
if you spent 300 credits trying the same thing over and over and expecting different results after the first 10... idk... do you understand what 'beta' means?
so because it's "beta" I can't use it for what it's meant to do? I don't know if you read what I said but every generation had "thank you" or "thanks for watching" this happened in all clips regardless of change of style/lyrics
hey suno, so the bug that kept causing the same song and genre to pop up multiple times even though i only pressed the generate button once is still happening, not as frequently but i found just yesterday one generation ate up 50 of my credits. And also it appears that more than half of my generations are just instrumentals without any of the vocals i provided. its hallucinating entire songs minus the lyrics. what are the chances that all these glitches could possibly be fixed or refunded in some way. i usually never complain but i find myself using this app so damn much its impossible to not watch my credits get eaten up like this.
Hey folks, would be helpful to email support@suno.ai with examples- will be helpful for us for the next update. and we can look into your credits.
In spanish is even worse, It usually adds "Subtitulos por la comunidad de amara" that means "Subtitles by amara community". I think that's a sign the AI were trained the audio of subtitled videos.
I input the songs in Portuguese, but they are coming out in English. There are several bugs; I've used up my 50 credits. Previously, the songs were coming out perfectly, but now they are coming out in a way that I did not request.
Yeah, I'm having this issue, too. I've already provided @muted flame with the network traffic and clip ID I'm trying to collect from, so hopefully this one will be resolved soon, as it's a full blocker to using Chirp at all 😦
Wrong channel :)
Problems with continuation this morning. The lyrics being entered seem to be largely ignored and the clip generated is either entirely a repeat of an earlier lyric section, or the first 50 seconds are a repeat and the last 10 seconds tries to render the new lyrics but then gets chopped when it hits 1m
I had a bug on a continuation were nothing generated. I was doing a song with no lyrics and added only to the genre section. On my third section, it was just silence.
Oy… now continuation is changing the vocalist and the rhythm when it doesn’t just ignore the new lyrics
we are aware of the issues with continuation. For now just trial and error unfortuantely. Updates should address this issue somewhat.
Yeah it’s just pointless burning of credits now 😦
wtf :/
Are those from a song already under copyright?
nope is my lyrics
Huh.
It shouldn't be happening then.
Try adding spaces to the end. Like several lines of random spaces in length. Try again.
You can always take the filler out the next continue.
i try
you are right. meal culpa 😄
i get the same thing...
trying to make a song with the lyrics of
"After You Get What You Want, You Don’t Want It" by Irving Berlin
(It's Public domain)
gives me that message on the refrain
ok. so it stopped generating lyrics in Cyrillic fonts
and stopped generating lyrics in Uzbek language
Working on making this better, sorry about that
thk bro !
seems to also happen when i try to create a new song.
These are the lyrics (there are like, 4 more lines in the Refrain where it also happened, so i deleted them, hoping it would fix it...):
[Refrain]
After you get what you want, you don't want it
If I gave you the moon, you'd grow tired of it soon
also tried with "[Chorus]" instead of "[Refrain]" and adding empty lines and something like "Oooohhh".
Oh yea, and i'm using the web-version.
i just make a fcking banger !!! thank a lot !!!!!!!!
I don't know about other languages, but when your bot sings in Russian, it often places the stress incorrectly, making the whole track unsuitable for listening. Please make it possible to manually indicate on which consonants to place the stress, for example, by changing the letter size - ''molokO" "kosmonAvt" etc.
Try using accent instead of capital letters, I heard it works better. For example:
сто́ит(costs) and стои́т(stand)
ok
(I don't speak Russian but somebody else said that it worked better for them in Russian.)
I tried, it seems like the method works. but it perceives the stress on the letter 'e' as the letter 'yo' for some reason
i subscribed to be able to generate full songs, but when mine are done there is only a option to re use my prompt, no continue from this clip, or get whole song option. i have tried stitching them together only to discover what an Act that is. im pretty desperate at this point? its nothing but a tease really if you cant actually create a song from the whats generated.
i watched a video on youtube and the guy had options to get whole song. why is this not a feature anymore or is it just me?
Are you using the webapp?
i have both yea
actuall i can continue from clip, but i just end up with bits that dont fit together, no option to have them stitched into a finished song, just a mass of 40 second bits of songs
OK...so I'm going to suggest that you send an email to billing@suno.ai so they can make sure everything is as it should be.
ok what should i put in the message?
Just what you posted above.
yeah duh sorry
Getting major pitch problems now, making most output unusable. Any fix in sight?
i have a pro plan, and mine still only does two, i forgot it was meant to be able to do more, although i also only get 40 second snippets tops! and i only have ever had the option to reuse prompt. No continue the song or finish a song. so have just have over a thousand credits woth of usless snipets and have no idea what to
keep getting Credits Refunded because only 3 sec generations come, is it my input?
got such a nice song going on 😦
at least you arent wasting credits tho, ive spent over 1K on generating and have some that sound great. but almost makes it more painful not being able to do anything with them when they so good.
Hello. I bought a membership with money. But I cannot use it. It keeps giving errors. Can you help me please?
Please create a folder structure like AIVA has, so i can move the junk in there. It's a mess now.
Or possibility to at least delete unwanted material.
Any tips to avoid songs being created with no lyrics? I've burned 100+ credits to get 100% lyrics-less songs
Style guide was: High-energy, hard rock, fast tempo, aggressive guitar riffs, pounding drumbeats
Update: made the style guide with less words and it worked. Not precisely what I needed but still better than nothing. Kept: High-energy, hard rock, fast tempo
Thanks for letting us know what worked for you. Appreciated!
the cdn is dying for me 😄
CHIRP has issues with the following lyrics - although i have no idea why - cause i don't think i'm using any slurs or bad words...
[Chorus]
From humble beginnings, they blazed the trail,
With rhymes echoing, a captivating grail,
"Rapper's Delight" sparked, a revolutionary flame,
As "The Message" unveiled, urban life's stark, unnamed,
"Straight Outta Compton," a defiant roar,
Shattering norms, silencing all the more, defining more,
"Illmatic" unfolded, a lyrical tapestry grand,
While Eminem's "My Name Is" shook the industry unplanned,
"Lose Yourself" ignited, a fire in our hearts,
The GOATs of hip-hop, forever in our minds, their artistry imparts,[Bridge]
From DJ Premier's rhythmic finesse, a master of art,
To J Dilla's soulful grooves, so deep, they leave us enthralled,
The rhythm masters have crafted the essence,
The heartbeat that makes hip-hop leap, in every existence,
Their beats are timeless, their samples masterful,
Forever etched in our minds, a musical legacy, unsurpassable,
Their music will inspire, forever surpass,
Leaving lasting imprints on our souls, no matter what class,
What Iam doing wrong if the songs ends before all the lyrics were sung?
Did you just run out of time? If so then try continuing and see if it will pick up from where you left off. You may need to edit to find where it left off.
it stops after 40s so I guess thats the normal lenght? You mean just try to continue or shorten the lyrics?
Wait, are you using the Discord version?
yeah
I use /chirp
- very often - generates ONLY MUSIC - without words (!)
- in prom "Style" - when I write "cheerful, rhythmic, instrumental..." == the music turns out to be too "DARK", dull, sad, unpleasant (!)
this is probably due to the fact that the SONG STYLE could not be in the desired intonation (?!!)
OK, well first of all (just in case you didn't know) the Chirp in DIscord is only version one, the webapp is version two. (They are porting the version two app to Discord it just isn't done yet.)
You can access the webapp version here: https://app.suno.ai/create
The webapp gets 80 seconds for the first generation and then 60 seconds for a continuation. The (current) Discord version gets 40 seconds for the first generation.
You can make longer songs in the Discord version witn #⏩┃continuation
oh ok, I thought the 1st version is bark and chirp the second. So I'll go with web app
Chirp is Chirp.
Bark is bark bark bark bark...
🤣
this happens to me a lot as well.
the "only music" part that is without lyrics although lyrics had been given. same with some weird hilucinations.
Not entirely sure which planet the language they're speaking comes from. 🤔
How much did you put in at a time? You’re only supposed to put 4-8 lines per generation. Sometimes if the lines are too long, 8 is even too much. It might be that it was trying to figure out how to sing everything fast enough to fit in a 30 second clip and couldn’t, so it cut out or replaced words.
If the lyrics are short enough and you have a good genre prompt typed in (nothing excessively descriptive: just “rock, rap, hip hop, trip hop, etc” and you’re still having issues, it might be an issue with providence. Maybe you’re not meant to write a rap song that just talks about other rap songs. I’m not saying that’s the problem, check the other explanations, but if no other diagnosis fits, it could be.
What exact problem is it giving you?
I heard about version 2, I want to try it, but I heard you can’t put in your own lyrics. Is that true or no?
I’d test it myself but I don’t have any chirps left right now.
That was true of the beta testing which only used GPT lyrics.
Now you can create your own with Create Custom.
The webapp is now v2 Chirp and has the ability to use custom lyrics and continue the song!
Oh okay. Awesome. I’m looking forward to seeing what’s improved about it. More genres maybe? Easier to switch back and forth between genres in the same song? Better singers? (most are good lol)
Genre switching is not as fluid as v1 which would veer wildly depending on what was put in the style field. But the quality of the output went up and the first generation is now 80 seconds with following generations as many as 60 seconds in each continuation.
Genres are still something that you need to play with. I tend to have good luck by mixing parent genres to create a type of the subgenre I'm looking for. (Think of it like mixing paint red and yellow to make orange.)
Yeah I’ve mixed trip hop and rap somewhat successfully, but once it cut out most of the instruments for the rap part (for whatever reason), it was really hard to get them back. I guess you’re talking about mixing stuff like jazz and blues or? I definitely will try it in an hour or two, assuming my chirps refill at midnight.
the limitation on the web app are 60 seconds and 80 seconds for the initial chirp 🙂
what hepcat said - sorry - i answered without reading up ^^
As an example, for heavy metal or rock songs I tend to mix progressive rock, blues or blues rock and a few other instructions. That tends to get me more of the guitar styles I like.
At this point in time it's really rewarding to experiment to get the sound you're looking for.
That’s still a lot of lyrics for 80 seconds lol. Was it not giving you any vocals or was it just cutting off short?
Nice. I’d love to check out some of your songs sometime if you ever feel inclined. 😄👌
Kind of you to say.
I have two new Christmas songs coming out tomorrow. I'll be posting them in #🐣┃suno-showcase
i have the same result with a verse less - also it happens to just give me 30 seconds of audio without anything or just 5 seconds of music without any lyrics. i'm not really complaining - its v2 - i expect more in v3 and further. its just the beginning. in 2 years we gonna laugh about it 😄
I really need help, I can't delete the songs I want to delete
Deleting or at list archiving/hiding is coming soon (really!) but not possible right now.
okay, take your time!
Depends. Did you use copyrighted lyrics or name a band as part of the style? That would be the reason, otherwise, yes it's a glitch that's being worked on and has affected everyone sooner or later. 😦
A workaround that one of us found is to add several lines of spaces to the end of the lyrics and that will give it enough difference from whatever it's hiccuping on to go forward.
No I'm not using copyrighted lyrics but thanks for the answer
It also kicks in sometimes when you use a band name as a substitute for a genre too. --and that wasn't an accusation just letting you know some of the common causes of triggering it. It's a very non-specific error message that gets used for a lot of things.
But then yeah, definitely try several lines of spaces of random length at the end of your lyrics to get over the hump. Once it gets through the generation and you wish to continue then you can remove those and just use the next part of your lyrics to keep going. Good luck with your song!
...
A user reported a bug: #📚┃suno-school message
Apparently when a song is made in the Discord and stitched together the parts show up in the https://app.suno.ai/me page but not the completed stitch made in Discord.
hi, I'm working on revising the copyright detection system, which is sometimes a little overzealous. if you DM me your prompt it would be helpful for improving the system
Hi guys,
Dammm, what crazy sound resulted from a chirp.
I used the same prompt before somehow, always made something closer.
But in this one ...
This isnt a bug?
2-4 I really liked the last one, I pressed continue 1 and it said you don't have any songs, but I know we can create 25 songs in 1 month, but I made 2-4. .(
pls help
im generating only two variation's at a time, i cant help find it a bit disconcerting, that such a huge percentage will produce one variation that is maybe just one solid tone and the other is corrupted or whatever and watching hundreds of credits disappear in exchange for nothing but slight aggravation. how much money are they making for failed attempts that don't for what ever reason get flagged as usless and refunded. i just had the most random thing ever happen. i was generating the verse for maybe the third time the previous times the words were fine but the timing was a lil bit random so i went for another go at it. it completed and i hit play . and it was one solid minute of a monotone voice repeating the words "im just a little bit tired of this" its was so bazarre, lol its like it was reading my mind! lol
I can only make 2 at a time and im on the pro plan
I was getting pretty good songs in beginning but not now. Does the quality of account change over the first week?
Me too
Me too and half the time only one works
I’m doing one song like 10 times now and basically giving up on the songs. The first week was ideal and then nothing good now
It does seem strangely inconsistent
It needs way better programming with better phrasing and chord structure
I can't say I've had a decrease in quality but the first few days 90% of the time I would get no option to continue of Finnish any songs at all, only one to re use prompt. And just the amount that are basically just some strange sound for a few seconds then silence, which your credits aren't refunded for. They have added up to allot for me with maybe 400 left from the pro plan I maybe have 4 good finished songs
that - annoying 😄
i like the automatic refund feature - but it only works like half the time. at best 😄
I'm getting an error (I think it might be that I don't have enough credits) but there is no message given in the /requests response, just "status": "error", would be nice to get a better message so I know why
Feel free to email us at billing@suno.ai and we will figure something out!
Sorry, but no, we don’t / haven’t changed anything
Will look into this one
It looks like we’re back to generating nothing but instrumentals this morning. 90% of what’s coming back has no lyrics at all
Not sure what is going on but was creating a song no issues then all the sudden it keeps erroring out and refuses to generate these lyrics. [Chorus]
"Please, not Joline," his heart did plea,
Underneath the old oak tree.
Life's storms, some hit with a blow,
"Please, not now, don't take Joline."
Nothing in there should be triggering for content or anything.
I reworded the chorus and now its fine. Weird.
when writing in Russian, if it can't decipher what lyrics to follow with, it starst saying (I back translate) "subtitles subtitles subtitles subtitles Senetskaya subtitles subtitles subtitles subtitles." Whatever the hell is this lol? It is reliably reproduced with all kinds of songs and genres.
here are just two examples. as you can see, no lyrics there, I just told it to do an outro
Same for me… 😢
For the browser version:
Just wanted to note that the first weekend it was mind blowing I even learned to guide it with [ethereal sound begin rising] as an example and it was doing it between verses and chorus. But now I don’t get lyrics responses great majority of the time. Additionally now I am unable to get the putting clips together feature to be effective on the browser version. (Yes, I’ve tried all the browsers)
We hope to have some fixes to no lyrics coming later this week, as well as putting clips together issues
just wanted to let yall know that there's text overflow on the email button if its too long
Weird bug, I hit the filter, probably because of copyright (which was annoying in the first place because it’s a traditional song which the famous person who has recorded it has no rights to the lyrics) so I substituted some random lyrics. Didn’t work. Took the lyrics out completely, didn’t work. Had to close and refresh my browser before anything would run, seems like a memory leak or something?
This is so bizarre.
I'm stuck for about a week now.
Just tried what Alliecat said and it worked.
Output was not what I wanted, so I tried again and I'm now stuck again...
Seems like that "trick" was a one-off.
(I'm on a subscription using the web-version, in case that matters)
Sorry folks, we're gonna add a "this should have worked" button where you can report mod failures soon
Now this will be helpful to me too when I do a public domain song and suspect I'm getting incorrectly clocked as infringing.
I am calling this a major bug. It may not be. Please forgive my bluntness. Please please fix the clipping issue. I'm not sure how you are coding loudness? But I just wasted 200 credits and even RX De-Clip couldnt fix the clipping and clip distortion lol. Maybe code it under 0db? Maybe like -1db? The clipping is a major issue.
I am using the Web UI btw
Also I notice that it gets worse as you continue a track. Also the AI doesnt always understand
[Short Solo]
[Chorus]
all the time. Thought I would mention that. Sometimes it generates only a 1min or 1min 20 second solo only or no solo at all. Also Harmony and Harmonize it has issues with. I might be expecting too much and or trying to push it too far lol
where are the admins? i'm getting 4 second clips and not getting refunded my credit
There is bug in v2 website that only make 4 seconds clip
Please email to billing@suno.ai regarding this issue.
I logged in to try the demo and I started with 0 credits and it doesn't let me do anything.
Sure i can't even use the app right now
after the maintenance everything is just buggy
User posted about what looks to be a serious bug: #📚┃suno-school message
Hey I still can't get lyrics out of Suno 😦 I am wasting all my credits because of this bug! Never had that some weeks ago.
Im not using copyrighted lyrics and have tried with many types of prompts, it just doesn't work anymore.
also when I get lyrics sung, it generates the most robotic voice
Can you post an example prompt prompt (lyrics + style)? Is it on the first clip or a continue? Web Chirp or Discord?
I am getting pretty frustrated the last 2 days. I keep getting a wrong continuation of a clip. No matter what I do (post the whole verse, start where the last word ended, leaving it out) it keep giving me the wrong output... 😡 😭
That's unfortunately just how some specific clips end up. Hopefully will have some features to improve it and rescue these clips in the future. But if you keep failing on a continue from the same clip it's probably nothing wrong in your prompt, it's just an unlucky clip to be continuing from.
I think my prompt is fine indeed.. but it was the best clip to continue from 😢 ... So that's why I think it is more a bug
And I cannot wait for a solution, so I keep torturing myself by finding ways to fix it.. 😖
Wanna send an email to billing@suno.ai and someone will help you out 😦
Pretty much same here. Plus totally ignoring styles..disco comes out as reggae etc
Quoting myself from another time this happened:
Try depending on where you are in your lyrics (if you have a lot more to go) just continuing and running the same lyrics you have. This can sometimes make it pick up where it left off. A large percentage of the time it will start over in the middle of the song. This is just the easiest thing to try to see if it works.
Next try removing everything already sung but the last good line, then try continuing. This sometimes works. It's how it always seemed to work in v1 Chirp for me.
After this try creating a line break (empty space) and continuing from the line that was interrupted. This has a good chance of working.
Then starting with the line that was broken. It sometimes will "remember" where it was and pick up from there.
Finally when all else fails, batten down the hatches, remove everything up to the word that was interrupted and hope for the best and throw the dice. You will likely encounter some weirdness and you will need to keep trying. But if everything else has failed it's this or...
...or you go back up a chain in the song and try again giving up on the good part. It hurts, I know, I know. But sometimes it's the only way forward.
...As you can probably tell this is a bit of a pain point for me and I've done these often...
Quick fix which can work here - double any words you really want "disco disco"
Tks..will try..well burnt a lot of credits..Suno should really refund
Um, where are all my clips? I have one page now and all my previously generated clips are gone.
Please someone help me I upgraded and now everything I did before the upgrade is gone please help
This is a nightmare no one is answering my email
I have thousands of credits which just got erased
I had that too when I upgrade to a paid account. I guess it is normal.
OK still happening, when I use continue from the first verse and input new lyrics it uses the old lyrics again
Someone will look into this soon! Sometimes you may have logged in with a different gmail or Discord, your clips are still somewhere so we'll figure it out for you
OK so new fun seemed to start a couple days ago with the website and its getting worse...
- Clips seem to be missing. I have clips showing in My Clips that are clearly and obviously continuations of clips created earlier... but the original clip is missing from My Clips... for example, I have a clip with a [Chorus] section... and if you download the entire song the previous [verse] and [chorus] sections ARE in the audio, but the actual earlier clips that preceded the one that's there are outright missing from My Clips. Sadly the clips that ARE accessible are ones that have broken/strange audio and/or won't continue properly so I need to access the earlier clips to try continuing from one of those and they aren't accessible.
- Clicking through My Clips gets a bunch of pages in, usually somewhere around page 15, then starts throwing "This page is having a problem. Error code: STATUS_ACCESS_VIOLATION" when you click the next page button. This has become a real problem as I have 1000 credits worth of garbage clips that I need to click past to get to the working clips that matter that are now 15+ pages deep in the My Clips. It doesn't seem to matter how slow or fast I click the "next page" button. I've had limited success by switching from Edge to Chrome, and dumping cache/cookies... sometimes that works, sometimes it doesn't update the page when you click the next button... for a very, very long time then suddenly it will jump forward quickly through the pages that should have been displayed when you clicked next
Chorus continuations seem screwed up now. If I go back and play clips I made two weeks ago, each [chorus] section is identical or nearly identical in style, rhythm, and vocalization. It seems like for about the last week the chorus sections that get generated are coming out sometimes radically different in all aspects from the previous chorus.
When creating a video, if the lyrics are in the Arabic script, every letter in the Latin alphabet becomes a square (unknown character)
https://cdn.discordapp.com/attachments/1069381916492562585/1184839053842055208/Arabic_Clip_1.mp4?ex=658d6e5b&is=657af95b&hm=29092246a3250c9afc00279c9697dd7fc05bbb1dd4996249c10b49575d673f65&
I've noticed that in the latest songs generated with lyrics in Brazilian Portuguese, the first word (one or more) is being replaced by another meaningless word in Portuguese. Is this a bug or a way to avoid using it?
Not sure it was reported already or not, when the first clip ends on some funky syllable (feels like) then proceeding from that clip will make a following clip's a guaranteed voice mumbling, or unknown language.
Hi, I'm new here. I have the paid version, but when I generate music, they only load. Nothing appears. And if something appears, then without text, although I have specified a text. The credits are deducted but nothing appears. Is this a known bug or am I doing something wrong?
We should fix this soon, sorr y about that
The credits should come back to you eventually although the UI might be currently buggy (if not just email billing@suno.ai)
Hello. I just bought a subscription and now none of the songs have the lyrics music or title. When I generate in the app it’s all random lyrics, title, and genre whereas it didn’t do this with the free version….
Free version
Subscription paid nonsense songs
Sorry about this - how are you creating?
definitely a bug, we'll try to fix very soon, email billing@suno.ai and someone will give you some credits back
Thanks and sorry, should be fixed in 15min - thanks for reporting
this is after the update, new look of the website
is this on "Create" or "Library"?
Yep, same here. Always a good idea to test new UIs under “beta” subdomain or something…
I still have the old UI open in another tab and it works, just afraid to close it 🙂
I write a song in new beta version and it starts having problem from middle till end and somethings from start
The worst day to be Suno’s community manager. May the force be with you @muted flame 🙈
It's ok, im actually fixing bugs so it's both a great day and a tricky day
Ya, we did but clearly didnt test enough
Should work through the major issues today
Btw how to get points for free?
What's going on? Half of my request are stuck, not finishing. And the other half are instrumental, or distorted vocals. Updating?
We're looking into stuff getting stuck - sorry about that - hope to have fixed in a few hours
Ahh this explains it, from a new user I was like ope 
Ya, we're usually a lot more stable than today 😦
Welcome though
No worries, got my sub at any rate. Great job that from what I can see thus far.
songs are lower quality since update
i agree, and some song i try to generate are infinitely processing
with all respect - wtf is wrong with these lyrics that i get denied 😄
[Chorus]
This is our solution,
Join us in our trance,
We groove, we dance,
(let's join hands)
GIVE PEACE A CHANCE[Arabic]
نتأرجح، نرقص
امنح السلام فرصة
[Hebrew]
אנו רוקדים, אנו מסתובבים
תן לשלום סיכוי
[Instrumental]
wait - i refreshed the page into a new layout with 0 credits 
walks out with an ashamed face
There are so many problems, he can't keep the songs going, it switches to different vocals. It puts meaningless noise in the background, I think you should change it back to previous version on web. After update is everything going bad
They literally just changed this today but all of a sudden the entire format is different and ON NO SCREEN can I see my CREDIT ANOUNT. I was just thinking, “Oh, I might mess around on Suno and see what I can come up with. Let me check how many credits I have.” Refreshed the page, boom, can no longer see my credit amount. Why have fun and experiment when I can’t see that and I have to save my 5 daily for the serious stuff? Seriously, why on earth would you remove that?
Is it normal to continue from a song and hear nothing?
bottom left corner
Which page?
I clicked on all three
I’m using Suno v2 via browser
Ah, looks like bug mobile. Will fix shortly
Thank you.🥺
after update i'm getting tracks with just music, no voices
i'm getting those all the time. i started to report them now.
ok so also others have this issue, better to wait their fixes, thanks for your reply
on new UI you can set your clip public, but you can't make it private again
In theory the update hasn't changed music output, so this is *probably * just bad luck. But bugs of course are possible, will keep an eye on this.
so if you misclick, it's forever public
You can toggle it back to "Link only"
initially i have thinked the same, but if you see my generation can understand that at 99% bad luck isn't the reason
Private implies it's completely invisible, it's more like "unlisted" in a YouTube sense. At least for now, I think the plan is to have a full private eventually
yeah it makes sense now that I see it, but it wasn't obvious that link only was preselected one
yeah unlisted would work too
getting this on the webui now too after two non-custom mode gens
all current gens are finished
now it runs again
and now it's stalled again
I did check that all the gens were finished this time also
is there just some timeout here?
it does work after you wait for a bit after
Is there anyway to get before update version
If nothing’s changed, you can still use v1 on discord. I want to stick with v2. I’m just getting way better results on it overall.
No I mean this new update worse
has the model version changed too?
Its can't be continue the clip same vocal and Music is playing nonsense in the background
Yesterday everything was good
works fine for me, could it be an issue with a specific clip?
I try to different clip but still not stable
I don’t know which one is the model version, or I’m probably misunderstanding the question. In short, I don’t know.
I haven’t actually even tested the web version since the update today. Based on comments here, I’m afraid lol
By the way, the fix worked. Thank you so much 😊
Nothing was intended to change about the model output for the web update. Of course bugs are not impossible and we'll keep looking into it, but that was not the goal. If you are retrying the same prompt or continuing the same clip repeatedly, be sure to also try some completely different chirps as well. Sometimes there' s just something in a prompt or a clip causing trouble.
Wanna send DM your email, or email billing@suno.ai? As we look into these issues, we can give you some credits back - thanks for all of the constructive feedback today
also cool that finally Russian lyrics give actual pics instead of various views of the same bird 🙂
@muted flame , the downloads do not include ".mp3" extension (works if added manually)
also, is it just me or are generations much faster now?
should be way faster (pending the loading issues)
What browser?
it's amazing. Too bad it means I'll deplete my credits way sooner
wow fancy
Version 1.20.1 (43987)
Chromium Engine Version 120.0.6099.56
right? flex flex
MAN, 10 seconds for a continuation O_O
did the model change, or did you just get more compute?
Some big speed/infra improvements 🏎️
that's amazing, man
Crazy. I was sitting there looking at the songs wondering when they’d be done, thinking the play button would pop up. I clicked on one. They were done the whole time lol….
We'll make this more obvious but yep once the spinner goes away and the image appears it's ready to play
now just to have do some projects/tags stuff
projects most notably
in jukebox web ui, I had this kind of "tree" structure, where you could see where each continuation comes from
I wrote second by second, so i could have 600 "versions" for any given song lol
oh, hello
yeah I wrote you 😉 you both actually
will DM
I didn’t see a spinner. That’s what I was looking for at first,
@unreal garnet That’s really neat. I considered doing multiple versions of songs. There are some older ones that definitely could’ve been better but I usually just move on lol
Does v2 allow for different language songs to be translated in English? Like, to have a jpop song with Japanese vocals but English lyrics on screen?
hmm. we were considering adding this feature (it would use something like chatgpt under the hood to show you the translation). but in chrome, you can also just highlight some text and use google translate 🙂
tbd I think the "export as video" feature only works for really quick, really meme-ey showcases. If you're doing a proper vid, you likely need much more controls than a one-click export can provide. just my 5 cents
Oh nice. That’s good to know. You could highlight it in the video? lol
I actually don’t mind leaving them as lyric videos. The background pictures it comes with are usually pretty good. There was one case where I didn’t leave it at that. But usually I just try to generate a good thumbnail for the song elsewhere and leave the video as is.
suno does not always respect line endings but im not sure if this is a bug
more often this happens with disproportionate line lengths, where shorter lines often get merged into longer lines, even crossing marker boundaries like [chorus]. easiest way to repro is to try the inclusion of lines with only one word in them (that are meant to be prolonged or ornamented for the entire line)
(example)
You give me
[Chorus]
Hell
And it’s just overkill
will usually either merge the two lines of the chorus together, or merge the first line into the end of the verse
does not happen 100% of the time but is very likely to occur for generations
similar issue can sometimes happen if the line is “too long” relative to the others, it may subdivide the line into multiple, usually no more than 2
the thing is, I almost always have to add various metatags and cut lyrics here and there to make it work, so the lyrics box is more like an "under the hood" thing
I get what you mean. It’s really awkward when a Continuation doesn’t work quite like you thought and ends in the middle of a word or line and you have to write the rest in the next section.
I had a line about “defeat’s breeze” and it ended on “defea-“ so I had literally wrote “eat’s breeze” in the next one. Don’t want anyone to see that lol
eat breeze!
ha, ive done this many times
also it would be nice to be able to generate a full length of music regardless of the lyrical content
that way when continuing you could choose where to continue from and prevent a lot of timing issues
or you could redo certain sections like incomplete chorus
it would essentially feed a trimmed version of the song to the model
ya. we hope to have this one by the end of the month 🙂

not the kind of typo you want
We had fun with cookout, kebab, bratwurst, hamburgers the other day...wasn't half bad....
What did “hamburgers” give you? lol.
I tried “tropical rock” the other day. It did NOT give me tropical rock. But probably one of the catchiest songs I’ve made to date 😄
Hmmm....I should try that one by itself. I did them as a part of a whole.
Oh okay
It seemed to just add some reverb and distort the vocals horribly.
https://youtu.be/hU25OJZry0Y?si=6u10r0T33j7v91GJ This was my “tropical rock” song.
The AI wrote the first part of the lyrics. I wrote the other two parts. It’ll eventually be longer but that’s the shortened version.
Sorry for posting here (I was using the “when in Rome” logic) lol
I normally publish on Wednesdays so here it is, but truthfully this still needs a bit of work, will definitely have a longer version. No worries ;)
https://app.suno.ai/song/95fa34f2-5e3a-44d7-b99e-8b1b97b02aa9
https://app.suno.ai/song/e88a34f8-84d1-419a-a213-76c1d5b3637a
I like it. Has a nice vibe to it.
I didn't want to link the ones I did because I reused the prompt I did before when we were clowning around. I'd just posted an instrumental that I'd gotten after fighting with Chirp on the prompt above and when we started talking I just reused that.
The words themselves aren't very fun or silly.
This was the instrumental: https://app.suno.ai/song/6ec7915b-43a6-4ae7-a143-0e8fad689f34
True. The second one was very strange. But I like how experimental Suno is. It seems like v2 broadened its horizons as far as coming up with compositions. I tried “ambient piano” and it was just gorgeous,
I personally feel that with every upgrade there's a bit more to learn or relearn about how to get a good song. While it can be frustrating to have all the blocks knocked out from underneath you and discover the rules have changed, it's also a great way to discover new genres, new tricks that work.
Agreed. I really liked that last one btw. The melody is really fresh and new, and it’s just very complete sounding. I could see someone buying that on iTunes lol
https://youtu.be/GliDc9jO4LE?si=-MYg621cI_gXhTGZ Just gorgeous.
Had to cut it early because I gathered He didn’t want that last part in (it sounded awful and way different than the rest of the song)
Sorry, it’s a little critical/cynical. I was in a bad mood when I wrote it, but the song itself is beautiful. I was really shocked lol
Something definitely seems to be up with the continuations - they sound like they're clipping compared to the previous audio clip. Example ID 58cac3da-b719-4e1d-9a28-8086de02039d
I wonder if you changed the training models too today? Because I am having a real hard time getting a decent quality today 😦
The sound quality for some clips today is quite distorted, and for some, the voice is only on one side of the headphones, while the other side is crackling so much that it makes you wonder if your headphones are broken.
I've noticed that the AI isn't creating the music introduction; it's going straight to the lyrics. Furthermore, the voice quality is terrible. I hope this gets fixed soon to enhance the musical experience
Looking into all of this and hope to have some fixes where possible by tmrw
🙏 much appreciated.. thank you
Yeah, I get horrible results since the last update
Um yeah I was hoping everyone was being overwrought but I just continued for the first time today and it sounds like nothing but fuzz, don’t know what you broke but it’s busted for sure
i am getting this on app.suno.ai... it was working before, is this some sort of restriction?
Might need to refresh, or might be a particular clip - is it a continuation?
no its just a new song, i was making some and it was working, randomly stopped. tried refreshing etc no luck
ill try logging out and back in
yea no luck
lookin'
hmm ok, if i put "test" in it works. could it be that im using a word that is blocked?
it's possible, but that should give a different error...
the prompt im using is "a sad dsbm song with distorted guitars and a haunting melody"
thanks thats helpful
ah, that's a really silly issue on our end.. standby
should be ok now
that's working now, thanks for the quick fix, ill likely upgrade to pro
how'd it come out, (i am not super knowledgeable about metal)
the results are pretty good, i am mainly just using it to find inspiration
Problems with V2 web app, new interface
- No Control for Intro prompts. Vocals start immediately. Needs vocalization to hold the vocals at start.
- Continuation failure. Lyrics break in continuation to a continuous noise.
- Vocal change severly in continuation.
- Quality of music and vocals reduced tremendously as compared to V2 previous version.
- Languages other than English have taken a severe setback in terms of vocals.
Hope the team would look into the matter soon.
I have the pro plan for one year.
Everything was going very well...until today's update. It's terrible.
I know the Suno team is working hard and I hope it will be resolved soon.
Me like a addicted idiot keep trying get a good one
I’ve twice run into something I hadn’t encountered before today, a song went for a full minute but half of it was silence. In one case it came back up after 17 seconds of silence, in the other it did a few seconds and then silence the rest of the clip.
I have some sort of account-related issue. Has anyone else experienced this? Whenever I try to generate something new, it gets stuck in an infinite generation loop. This has been happening since yesterday. Despite having paid for Premium, I haven't been able to do anything since yesterday.
@neat sail sorry for ping but my account don’t work 🥲
That's strange. You'll probably have to wait until morning for Suno team to wake up. It shows you have credits?
Ye
Hmn. Did you upgrade yesterday or did anything else change?
If you previously generated a lot of clips yesterday, there was a time when something was wonky with the server and a lot of chirps were getting stuck never finishing. What should happen is the time out and you get refunded the credits, eventually.
But it should be working now.
This is what happened: 😬 I wanted to generate a song, but I accidentally pressed the thumbs down, and the song disappeared instantly. At the same time, I sent a new song to generate, sort of a new prompt. And since then, I can’t generate any songs anymore. Unfortunately. They all get stuck in an infinite loop for hours and don’t generate. 😕. I'll quickly try to generate a new one now.
It's been a few days since I subscribed to it.
Try quick generate. I wonder if you are somehow stuck continuing a clip that is bad, and quick mode (not custom) doesn't allow for a continue
Well the whole website also changed yesterday and had bug fixes all day. So mabe it just works by now too!
After a few tries, it's working again now. However, I feel like I lost around 100 credits.
The credits may take a long time to refund after being stuck infinitely generating, but you can email billing@suno.ai and they can probably sort it out for you (but won't be until tomorrow when Suno wakes up)
I can no longer log in with email and password
https://accounts.suno.ai/sign-in
You may have to email support@suno.ai -- can you login with your discord account?
issue seems to of reverted
Is it possible that the quality of results in languages for example, German, has decreased quite a bit through the update on the beta version website? It seems that there has been a decrease in quality, specifically pronounciation of German words. My creations are 90% German. The results prior to the update appear to have been way better. Yes, every now and then there would be results with pronounciation flaws. But now it seems to happen in every song.
If you have a few especially good examples of bad pronunciation prompts that would help too.
Here is one example from a verse where the last word "hinbewegt" tends to be pronounced so it aligns with the sound of last word in the previous line "umfängt". But that is not correct pronounciation and sounds weird. Before, the system would recognize such things and pronounce correctly. Example: In dieser neuen Ära, wo Magie Realität "umfängt",
Schöpfer meiner Kunst, wo das Herz sich "hinbewegt".
The overall quality of the expression and transfer of song meaning through the voice seems zu have descreased, from a perspective of looking at the German results.
seems like yesterday's "eternal generation" bug is back
server is down?
Yeah, I'll repost here
Looks like the Chirp server is experiencing some turbulence. 🚨
I don't have any info but the main Suno team will be awake in just a couple hours, worst case.
Also IMPORTANT: If you have stuck Chirps they will eventually timeout and refund credits. (If they don't you can email support). Sometimes this takes a long time they just keep spinning (it's a known issue.)
yeah, the website suno is back working, hope the results are better now. cause the results last night been horrible
still working on all of these issues above, thx for patience all
thanks, we are looking into all of these issues in general and hope to have fixes tonight (US time) or this weekend
Appreciate your support, thanks
yes great support here, thanks a lot
we appreciate you all and your feedback
however, the website in its current design is a real mess
Any particular feedback is helpful
yes, you should drop the vertical generator to a vertical horizontal. meaning that a main generation is still listed vertical while the continuations are added verticaly to a project. plus the ability to selct previous generation to delete them.
*mistake, while the continuation are added horizonataly to a project
thanks , ah yes, a much more intuitive continuation interface is definitely on our list
guess i need more coffee, hehehe
me 2
do you want perma delete? in any case, Dislike actually will remove from Create screen right now (although will be still available in your library)
yeah, i noticed that. at least that old generation are not visible anymore, this would help a lot. or you go project based. like you enter a title and than you continue in a clean frame. next project other clean frame, its easy to make
Perma delete sounds great as some are just not useful at all
yeah, dislike really helps. deleting would help too 🙂 also, it would be great to have maybe tiered "likes", so I can filter by "best of the best" or just "something interesting"
in the moment, suno is realy the best. but 50% or more is unusable
less than 50% is usuable, and 1% of the generations are marvels
damn, i make one typo after the other
This is all very helpful
this was a marvel https://www.youtube.com/watch?v=PUCUhQL0XxY
Please, address the following issues:
Songs start without an introduction; it's a bit odd and bothersome to hear that there's no preparation before beginning.
Songs usually begin with a good vocal tone, but towards the end, it sounds too distorted.
I've already spent over a thousand credits and feel frustrated as I'm not achieving the desired result (these errors didn't occur in the previous version, in fact, that's why I opted for the Premier Plan subscription). It wasn't until I came here that I found out about these issues, and I see that many others are experiencing the same.
I understand you're working diligently and with dedication. I believe you will fix this, and all these issues will be resolved, ultimately resulting in higher quality productions.
Additional request:
Could you add a "reinterpret" option?
Meaning, when a song is generated, use the same melody and vocal tone, but allow for editing the lyrics and then have it "reinterpret." This would be especially helpful for fixing metric errors and enhancing creativity in song creation. I would greatly appreciate it.
I was trying for first time the 'Remix' botton and somehow it gave me the continuation of same song i was doing before, even if i changed the music style.
And now i created a new chirp, same lyrics , changed the style.
Again somehow, it gave me the continuation of that same song.
More suprising he is giving always "same" vocals
Almost as i was doing
Continuation but in this ones i didnt.
has anyone found out a reliable way to indicate where to put stress in a word? Tried using capital letters, and it works most, but not all of the time
Did you try (WORD) caps and patentisis?
it's part of the word
it's hard to come upwith an example in English but let's say it says interEsting when I want it to say Interesting
Not sure how u can do that, try something like inter(ESTING)
yeah, trying something like this. I was wondering if there was a surefire way
I think there isnt...
😭 I know this feeling. One of my favorite songs got mangled and the chorus was perfect and unretrivable. It used a big emphasis on certain words in the the chorus that I haven't seen elsewhere...
I'll spend my days
Chasing after sunshine
Someday one ray
may steal through
Can't change my ways
That particular bit of emphasis just made the whole song.
Hello, discovered the web interface update today. The audios and videos are named mp3 and mp4, I guess it's because no title was set. In this case the prompt should be used (instead of the ID). Also, the song length is not visible next to the song, only in the bottom player.
Thanks for the feedback!
And, not sure if you experienced this as well. But, if the songs do have some sort of intro it is often some kind of mumbling and distorted bits of what sounds like foreign language... before the update there would always be at least a melody as an intro at the beginnig. Now, it just jumps right into the song.
Yes! It's happening to me too; truth be told, it's making me feel frustrated. I hope they can fix it soon. For now, they should add an option to use the previous version while everything gets stabilized.
The intro was often too long (and garbage) for a 1m20 song. But I agree that songs need an intro even if it's short. From my memory of last month's practice, the more [parts] you included the shorter was the intro.
Yes, I completely agree with you. Hopefully, they can fix it. I've tried with the [intro] meta tags, but they don't work properly. It would be great if they could make them more functional and maybe add meta tags for both female and male voices.
Nevertheless, it's not all bad. The new interface looks fantastic, and song creation is also faster. I know there's a lot of effort behind this wonderful project, so I'll continue supporting this community < 3
Happening when generating a new song. Not a continuation. Started a couple minutes ago or so.
Same here. Let’s cross our fingers knowing the team is on all these bugs this weekend
I'm having the same issue as Arcappsland
Sorry all, this is actually completely unrelated to even our recent update and is an internet-wide issue https://www.cloudflarestatus.com/
Welcome to Cloudflare's home for real-time and historical data on system performance.
working on it
generation seems to just be kind of generally broken right now. No matter how short the lyrics are for a chirp, it's generating a full minute and holding each word of the lyrics for a few bars to fill out the time. It's frustrating. If I enter a verse and a chorus, I get a lot of instrumentation and then the verse and it cuts off the chorus, if I do the verse only it does that but stretches it out to fill a whole minute.
I love the new web interface but the generation functionality seems to have taken a big hit.
I want to delete a song I created, but I can't find the item.Can't I delete it?
Click on the 👎🏻 button
I'll add to the reports of model degradation: It's suddenly skipping large amounts of lyrics in custom generations, or ignoring them in favor of gibberish (sounds kind of like what I remember of "attention failures" from earlier research papers). I just did about 5 or 6 regens, and 50% are unusable while the rest are missing a word or two at the start. Sometimes it'll just go about 30-40 seconds without vocals in a continuation, then start the lyrics and not have time to finish them. Trying to get it to include an intro to avoid the clipping, using [Intro] and [Instrumental] tags that used to work, now seem totally ignored as it jumps right into the lyrics anyway.
I'm not sure what's changed on the model side since I last used Chirp about a week ago, but it's significantly worse and basically unusable now 😦
this was big confusing
I guess the new website gen that does two separate gens with the same request counts the gens as 5 credits each, and it lags behind
I'm now talking about the 5 on the 25 here.. when you normally only see 10 credits reducted at once
website credits don't track correctly either unless you refresh the page
(it keeps showing older incorrect amount)
you probably got refunded some due to partial failure where one failed...
nah, it actually updated the credits usage half minute later to match real situation
no refunds, did 5 chirps on web.. checked on discord bot and it showed 25
then waited half a minute and it showed 50
i see, so discord laggin'
will look thx
it's not a big deal
what is more confusing is the error when you try to gen on current website where it claims you have too many concurrent gens, even when all have been finished already
it also resolves itself if you wait a bit, but it's kinda weird
Clicking on 👎 will not delete it.
This will work on Create mode only for now
Some of my chirps seem to have gone missing.
The two in the middle only have one variation each
Can't find the second version anywhere.
might have had an error generating one of the variations, but i can take a look, can you dm me a link to one of these songs?
Does anyone else experience that the quality of vocals has degraded and they’ve become kind of distorted?
Yes
I have noticed that the songs are created MUCH faster than before. However, the quality is lacking tremendously in comparison than before. From my perspective, I must say that I prefer waiting longer for better quality.
Some songs had a watermark voice a few weeks ago, mentioning a subtitle website (amara.org), sometimes naming the website, others saying “subtitles provided by amara’ community” and “susbscribe”, all this in spanish (Im making prompts in spanish with spanish lyrics btw) this has been corrected in the new updates?
A user posting server invites in general:
#💬┃general-chat message
one of the images generated for my song is probably nsfw how do i report it
also, ive had two bugged songs that only lasted 20 seconds
they never got to their pre-chorus nor chorus
im gona go 2 bed for now but ill provide any more info when i wake up
Oh, I’m getting similar ones in Russian! Any idea what that was?
Guys, i generated a 45 clip in suno, tryed to make a continuation and couldnt get the IA to do what i wanted. I made like 3 o 4 continuations from the clip AND NOW the ORIGINAL CLIP I LOVED DISSAPEARED. Help, it was SO GOOD
I have chirps still churning since last night. How do I stop them?
Generations are taking longer now, hopefully the previous model is back 🤞
Well, I hope they aren't taking 7 hours because that's how long mine have been churning now.
Yes, it’s definitely better now!
There can be cloudflare issues. If they are just spinning and not even an image there, they’ll most likely never make it. You should be reimbursed the credits though
I wasn't charged the credits, so that's fine. What's odd is that I was able to create some new chirps just now, but those older ones just keep spinning. I reported a few as bugs.
Yeah I had to dislike mine so they don’t annoy me
Yeah, that's what I just did, too. Thanks.
Wanna dm me that one? Thanks!
If those were from last night, they won’t finish. We will make the ui a little more clear that they’re stuck and won’t take credits
I had a few generations with less than 3 seconds. And I don't know if it returned my credits. I used to get a pop up message telling me that, but I didn't saw it since the last update.
A nice bug here, almost it didnt gave any sound.
Weird result , i reported
https://app.suno.ai/song/0457abf1-d3bb-4e4b-98ee-6a1d54b0a813
Hello. I bought a subscription. 10 thousand tokens. Why can't I download the instrumental version of the songs I make? When will this opportunity appear?
Wdym by “instrumental versions”? You can remove vocals using tools like vocalremover if that’s what you mean
it wont be a pure instrumental
We are still refunding very short ones (and just updated ui to reflect this again )
ty!
May need to refresh
@muted flame I’m writing an anti-xenophobic song, which among other things mentions how we often judge people by the color of their skin and blabla. I’m constantly getting “errored out.” Is it possible that the model is filtering me just by the mere fact that the song is about these things, notwithstanding exactly what it says about them?
Can you DM me prompt? but - yes, we do have OpenAI-based mod in place still, which is definitely not perfect
Google translation of the lyrics (can send the original if you want):
[Verse 2]
Xenophobia is in our blood, no matter where you look,
We are designed to look for who is “we” and who is “they”
Wrong skin color, wrong heart color,
How to fall in love, how to dress,
From childhood to old age we hang labels,
Where goals and means have long been mixed.
Just what if everything is wrong, and everything we
Are we looking for just a reason not to see one thing?
Just close your eyes and see
The main alien is the one inside.
[Chorus]
Strangers among us
They're looking at us now
And, hiding, they wait,
When their time comes.
Strangers inside us
They're looking at us now
And, hiding, they wait,
When theirs comes...
hour.
The two times it did work it gave me the “subtitles subtitles subtitles subtitles subtitles” stuff again lol
feel free to DM the original and we'll put it in our test cases as we iterate, always helpful
I am not getting rap, hip hop anymore.. They sing, no matter what I do lol
just had a few rap generations, this could be because of lyrics too I guess
It shouldn't be working like that. If I take an older one, and continue from there, it works
It definitely seems like it's back to its former glorious quality now, for me as well 🙂
Help ((why is it not working for me??? I topped up my account 3 hours ago((
try using [Rap] or [Rapping] before the lyrics
also try adding “rap rapper rap” to genre box a few times
maybe it will figure it out
Tried that and so much more.. but some how it is my lyrics, while I used a same flow as before.. so weird.. but thanks for thinkin along @high girder
Bug report here! 👋 Volume gets louder and louder on the site's player, until the clips are merged and you get the full song. This happens especially with metal songs.
To reproduce, set style of music to any metal, and continue from one of the resulting clips. The next one is louder, until you merge the clips.
Since the update, the intro prompts don't work, there are sound quality issues and even the melodic composition of the songs that are generated is monotonous, is anyone else experiencing the same issues?
My understanding is that these are things that are being actively worked on: #💬┃general-chat message
#BUG
play the song
then like it , it will increase by count 1
then dislike it, it will remain same
again like ,count again increased by 1
for example: currently like count is let say 85
like it->like count=85+1=86
now dislike->like count still 86
like again-> like count=86+1=87
(same user)
I am new to suno and the songs only generate up to 80 seconds.
If I subscribe, will it generate the full song?
You can generate an entire song with option "continue from this clip". Not sure if that is possible on the free version.
Noticed that too but that doesn’t worry me much. I usually start merging with the second part. So basically whenever I get a good generation I merge it.
Thank you for your reply!
I'll try it.
Most annouing part is when generator forgets the voice. like I have lyrics that follows verse, verse, chorus and I split them like: verse verse instrumental chorus instrumental now on next verse verse chorus the voice is different than on first one, or there is no voice at all... :/
today wanted to do 2nd verse verse chorus part but spent all 50 tokens with no good results :/
The CBS article on The D.O.C. linked in #🚀┃updates seems to have been deleted or moved.
I tried finding other sources with the article and they all seem to be deleted now too.
@muted flame check this bug please
No idea, but it ruins the songs
Just a guess on my part, but it's probably from certain audio data sources that Suno used for training contained the watermarks.
This is a source, but no video: https://hiphopdx.com/news/the-doc-new-music-ai-technology
Backup clip found: https://www.youtube.com/watch?v=Auq34tnu7Ac
I just continued a chirp, changed the lyrics but add from the continuation the same lyrics .
I've had 3000+ credits from my subscription and now when it expired, I've lost all of these credits. I was about to renew it, but you know what? Won't support 💩 scammers.
Most services are like that.. phone plan minutes don't carry over unless specifically stated. Midjourney same thing.
So why the free credits are being carried over? Does it means that the free service is better than the paid plans? wtf.. Anyways, I am done
Free serivice you get 50 credits per day. Next day, it is also just 50 credits. Not 100, if you didn't use them the previous day.
But it is clear you have a different view. That is fine. You could send them an mail explaining your view.
hello I have unfortunately spent a lot of credits on it to realize in the end that I currently have a BUG. The following problem I can't create anything anymore because Suno thinks I have another track in there that I want to continue. I have used clear several times, I have logged out and restarted Brwoser. Even when I used the device it constantly wants to continue the song that is no longer in there. This is very annoying because 1. I don't know how to solve it and 2. I have spent a lot of credits to find the error 😦
I can overlook the fact that credits are used for songs where a fictitious language is simply created although it is in my language and worked better before because it is a beta, but the error is really annoying, especially since you have paid for it.
I actually had a problem like this too. It didn’t last long, but I had a song I generated as jrock but later changed to punk. It was the first generation, not a Continue, but it was still treating it like a jrock song (it was either in Japanese or very bad English). The next generation fixed it but I was pretty scared for a minute there.
Also, does anyone one what the Remix button does? I’m really curious. It sounds useful. Does that let you reuse “bands” to make other songs? I tried pushing the button but I didn’t see anything new on the screen…
That’s not the same thing as a scam. I feel you deserve to know this going forward. That sounds like a pretty normal deal. Seems silly that you wouldn’t renew based on the fact that only renewing would fix your issue. If i didn’t pay my electric bill and the power went out, I wouldn’t get mad because I barely used the air conditioner. You get what you pay for, for as long as you’re paying for it. And as someone ON the free plan, I can confirm our chirps do not carry over. I wish they did, but that’d be more of a grand gesture of generosity than what’s already a fair deal.
😅👌
i cant continue my clips at all
it says error continuing
nvm it generated them but it pretended like it didnt work
had to refresh
the continued song also only generated a minute of audio, cutting off the last bits
do continues not generate 80 seconds?
Continuations are generating gibberish lyrics again...
Hello
Can anyone help me?
This error appears every time I try to create something
You could try the normal stuff... close page, clear cache/cookies/saved data, relaunch browser....
I selected "Continue From This Crip" and now I'm back to the beginning.
https://app.suno.ai/song/94508326-5186-4d88-9dd4-76799a29d24b
https://app.suno.ai/song/82c0f504-ed85-45cd-9682-88ff3380ba07
anyone else feels this way?
I just used around 400 credits to make a song and it broke on the Outro!!! Nothing but refunded credits. Cannot get it to remix or anything. Just refunded credits over and over.
Yeah happened to me. A trick can be to use another previous generation (an alternative of the one you’re continuing now), get it to end in a meaningful way, and then merge them together in a daw
Perhaps you can get a more meaningful message if you open the dev console and see the actual API response
Anyone else getting an issue where the player doesn't show time as it plays the file in the webapp?
Refreshed the page and it went away...for now.
I have stopped generating music until the platform is stable again. It's very frustrating for me to get the perfect song project to continue, and then lose all the progress.
I got annoyed of that happening... in attempts to prevent i've been ending my sections like this,
assuming I put a reasonable amount of lyrics that it can actually finish,
This seems to help give a more clean cut-off with a little padding at the end.
lyrics lyrics lyrics lyrics
this is my last line of lyrics.,
.,...
hope this helps for future 👍
Long opening silences on these all of a sudden:
20824c2d-4eb1-431f-917e-6ef152e53020
48d3a128-2c3b-4531-8db0-41eb0a6e6469
4c5f5b53-0178-4da5-8e6f-68832786e63a
a64f243c-74a8-4ff8-a7a1-976716aab795
yep had that the last days too.. including new music after a long pause but then in a total different vibe, different song
Another silence at the beginning.
44064222-3c0f-42b2-b3ca-3a1262536bbf
Gives a weird bug I had before... It is in Dutch and at one point it start saying in Dutch "of the video, of the video, of the video, etc"
daf1bd21-42e1-43bb-ab01-8b96975ec56d
song/72ead298-de56-4582-892a-0782979cbc6b
Both had the same problem. It makes me weary going to try again 😅
Okay.. here it is again! "Van de video.. van de video... etc"
And I am not even using the word video on my song.
36666557-cc2b-4b6e-ad9e-91dedf9547c1
f77aa05b-ce9d-4121-b234-7b6e9a934122
I think it doesn't like my song 😭
can show the full lyrics that you have used?
Hi, I am only able to download two generated songs from https://app.suno.ai/me/ , after that nothing happens when I try to download audio. Am I doing something wrong? it works again if i switch browser/clear cache, but once again only two songs can be downloaded. is this a bug or some kind of limitation?
I can only take a guess here...
but I have to believe the correlation comes from
12 different [heading tags] followed by first non-heading lyric content of "(...)" repeated twice.
No?
i looked at:
20824c2d-4eb1-431f-917e-6ef152e53020
48d3a128-2c3b-4531-8db0-41eb0a6e6469
It could be, but this never used to be an issue in that way.
ah
also i see contextually,
"muted"
could be... affecting?
in: [palm-muted guitar riffs]
like your "whispered" example...
as you've illustrated before.
maybe @solar grove
🤔 I hadn't thought about that, but just as it will occasionally try to sing a metatag, I could see this occasionally confusing it and causing this to happen instead of the sound I'm trying to describe.
yup there's a chance 🤷♂️
Looking into this one - unintentional
thanks! seems to be working fine again now, but I had the same problem yesterday
Yeah, we have tried a couple fixes here just a few hours ago, so will monitor
how do you make it do an instrumental [intro] and [outro] ? These seem to rarely work for me.
переведи на мой язык
sometimes describing the same in the style helps, i.e. "this and that song with an instrumental intro"
outros are the most challenging tbh, sometimes it will just keep on going forever, and you have no choice but to do some kind of fade-out, which is kinda lazy
I came here to report avnouncement_27616 for spamming the otherchannels, but it looks like they already reported themselves here.
Possible song structure bug during continuations (for electronic music without lyrics) in V2, but **not ** in V1. For example: I write "lo-fi laid-back" in the genre (no lyrics). With V1, I get a musical piece that evolves and changes during continuations. With V2, I get the same hook repeating over and over and over again.
My gut instinct is that it's not a bug as much as just a difference. V2 is more consistent and less likely to wander by miles, even if you try to change genre in a continue, compared to V1. But I also bet some tweaks to the style prompt in V2 still get you there. Maybe even just add "instrumental" because empty lyrics isn't quite the same as "don't sing" to Chirp #📚┃suno-school message
Thanks, I will try it. Currently it seems like it's too repetitive in V2 (by default)
Just an update: It does seem that adding "instrumental" rather than empty lyrics in V2 produces something that is changing over time. Thanks!
Nice catch, could be!
Notice how it's NOT pure silecne but actually starts with a 'click' sound...
i notice now ya
on 20824c2d-4eb1-431f-917e-6ef152e53020
(for ref)
I generated two that started with clicks, and two that didn't start with silence as well
while still using the word "muted" in lyrics?
Yeah just testing OG prompt to see how common silence was and went 2 for 4
technically, since hyphenated..
(palm-muted)
its only "half" a word lol
50% hit rate, on par.
we are working on a more consistent [outro] tag (and [intro]), but probably a v3 feature
For those who have an interest in this sort of stuff and have the patience, there is a unique A/B testing opportunity available right now. The 'custom' mode in the new UI (app.suno.ai) uses the (I'm assuming) original model from when the web app was first introduced, while the 'simple' mode uses a more (I'm assuming) recent version that has the known/reported issues with drifting and looping. You can generate content using the simple mode, then use the 'remix' feature in custom mode to generate the same prompt and style tags with the different model. There is no streaming support in custom mode at the moment, but it can be interesting to compare the differences in the generated content. This isn't a very precise method due to the small sample size, but it's a fun "Try it at home" experiment. 😄
Right now Suno AI produces short 8 sec pieces rather than music. https://app.suno.ai/song/b295d161-8619-4dac-900c-b0e2d59b360d and https://app.suno.ai/song/6add05e7-4780-47cf-af52-7dfdd18afacd
Ehehee
It reminded me this Album, listen to it if you dont know it.
Fast short songs
(playlist down here)
"Short Music for Short People" is a 1999 compilation album released on the Fat Wreck Chords, record label. It features 101 bands playing songs averaging approximately 30 seconds in length.
With a few exceptions, most of the album's tracks were written and recorded specifically for this release at the request of Fat Wreck Cho...
https://app.suno.ai/song/2325d9b0-d900-49cb-abfe-bf1e59db3fea Now its completely broken. I call it a night
What is he saying ?
Might be simlish?
Damn Pikeys
What part of my prompt was ungeneratable?
In Russian some users reported accents worked better in chirp than capitals for stress, worth a shot. For example:
сто́ит(costs) and стои́т(stand)