#🪲┃bugs-web
1 messages · Page 11 of 1
I understand but I am trying to figure out a "quick" solution.. Normally it takes a few days to get the Devs to whitelist something.
when i permanently delete from trash and refresh page, they all appear again
They are actively working on the servers right now, so it could be due to that. Please try again later.
Ah gotcha, any word on a delete all option in trash in the future?
I have no insight on that.. I am just a volunteer 😄
I see XD, well your doing a great job answering everyone's question so keep it up 😁
Hey guys, concat is broken at the moment
Working with someone, can't get whole song to work without failing
They are working actively on the servers for the last 2 hours at least.. I assume it has to do with that.. they are not finished yet
🐞Each time I am connecting to the app, it generates a song automatically, still loosing credits.
damnit daizy, I want it now. now now now, lol
I would advise to wait. They are actively working on the servers, so it might be connected @hushed umbra
👍
It's all good, I can wait
Hi, I think v3's voice generation needs tweaking, I hear choirs when there should be a soloist
That's known issue 😦
Unfortunately, very well known...
I can't seem to make any 'Continue from this song' like in V3 Alpha. Am I doing something wrong or is it broken?
Hopefully it'll get fixed in a future mini update
People who have been using Suno for longer say that this may not happen until the release of V4
Oh
Hello to all, yesterday I was quite satisfied, today I hear in the vocals DISTORTION, it's very annoying. When I pay subscription I would expect some quality. I'm happy to wait up to one minute... it would be cool to have some button "Generate HQ" or something.
Ok, this is technically not a "bug" per se, but all of a sudden, people with a free account can't generate anymore... Hopefully this is just temporary
I'm not wasting my money, sorry...
This is because of the update that is currently being carried out, when the Devs are finished, free users will be able to return to the fun.
Reading hard: Due to heavy usage, generations are currently only enabled for subscribers, and may take longer than usual to complete.
Thanks for clarifying and sorry if I'm being rude, but I get nervous when I'm forced to pay
Yeah... I have the same... Mix has a huge problem using the Limiter correctly. Especially when choirs join in with the song. I wonder if it's a limiter error on individual tracks or on Master Bus.
Chill.
Thanks for your feedback. I think will try tomorrow. Anyway, have a nice weekend.
is it possible to remix a song changing only the lyrics ?
Yes... Just empty the "style of music" field. But in a new creation. So it will not sound the same as the previous.
but I need to keep the same melody and soundtrack, just change the vocal, is it possible ?
Sadly no..
Issue with Generation Feature on Free Plan
Dear App Suno Support Team,
I hope this message finds you well. I am writing to bring to your attention an issue I have been experiencing while using the App Suno platform on the free plan.
Recently, I have encountered difficulties in utilizing the generation feature, as I receive the message stating, "Due to heavy usage, generations are currently only enabled for subscribers, and may take longer than usual to complete." Furthermore, another message indicates Generations are currently only available for Pro and Premier subscribers."
This problem is quite perplexing to me as I have been using the free plan for some time without encountering such limitations in the past. I would appreciate it if you could provide some insight into why this issue is occurring specifically for free users.
Your prompt assistance in resolving this matter would be greatly appreciated. Thank you for your attention to this concern, and I look forward to your response.
Warm regards,
When you switch from custom to simple mode (to get more screen space) and back, the “continue from” track is gone, this is super annoying
Not on the team, but I’m guessing because the heavy usage forces the developer to choose between providing decent service to paying users and providing shitty or no service to everyone?
Hey there. I’ve encountered a bug with the remix feature.
First off, I’m loving playing around with this tool you’ve built. I immediately subscribed. Here’s the issue: I built a song on the web UI that is 3:30 long that I’d like to play around with remixing it, but when I click create it just throws a nondescriptive error. I’m assuming it’s because it will only allow remixing up to 2:00 of track time.
It would be nice to be able to remix the first two minutes of the extended song, or better yet to choose a start/end point for remixing.
The error is in the pinned bug list. It's caused by the lyrics field being too long because the Remix function copies more than 1 song clip into it. The fix for now is to manually shorten the lyrics field when you do a remix.
Did something happen to V3 recently? It started to add noise to absolutely anything. Track starts clear, and in about 10 seconds it's like it added chorus and then removed it through voice remover. Sometimes it even adds voice echoes, though the track is purely instrumental. Became unbearable today, when almost everything (over 100 generations) is garbage. Well, at least 80%, just because of that noise.
P.S. I see from previous messages that it's not just me, for others it's solo voice becoming a distorted chorus. I think it may be related.
That would be a great feature in the future if possible. Select a portion to redo but keep the melody - only change vocals/lyrics.
If this is possible, we can even make the same song in different language. Which is something I'd love to see.
Yeah it's an issue many reported with V3. Voice quality degrades after continuation. First clip usually sounds crisp and full. But later ... it's a gamble.
Yes, but from 29/03 it's also true for the first fragment. For me, at least. And massive noise also started to appear at initial clip at instrumental part.
I see. I haven't played around since 27th. Maybe they're tweaking stuff.
Current generation turnaround: <5 seconds
Current generation quality, variety, and complexity: dogshit.
Have a previously "generated from 20+ buggy fragments" musical accompaniment.
https://app.suno.ai/song/7b5c76db-a20f-4fbc-96c8-7b05eb142c2d
Trod In Dog Shit And Walked On The New Rug song. Listen and make your own with Suno.
v3 is crazy noisy and continues to degrade with each continuation.
Why noisy?
everything, especially vocals turns into static garbage. Thousands and thousands of unusable generations.
getting random stutters and skips with continuations too (had that authentic worn out casette tape feel hahaha)
but on the whole i've found v3 has been absolutely outstanding, seems like they have completely fixed the vocals not generating problem, and i'm getting almost no hallucinating with the lyrics. i've found it has much better melodies with the vocals and much better consistency with the melodies between verses too. maybe i've just been getting lucky but i havent had problems with everything degrading into static and garbage yet
yeah... sounds like scratched vinyls. lol
If the audio quality is improved it'd be great
getting a LOT of muffled noise in all songs I've made tonight... anyone else too?
ahhhh not just me then!
hoping for a fix and maybe some credits? The music is SO amazing and it's hard not to be able to use it b/c of the vocals muffling (instrumentals too)
its up to 10, every time you click generate, 2 tracks are made at a time, keep clicking it 4 more times and it shouldnt stop you
My thinking was that both text and music generations should be configurable. Like, say for example, I'm in custom mode, and I know I'm going to render my song 170 times before I get a take that I like. Well, being able to render them ten at a time would be super convinient. But, also, sometimes, I only want to render a couple of them. So I would like to be able to choose how many I render. Again, for both text and music generations.
As absolutely everyone has already noticed, V3 suffers from a loss of quality with each subsequent generation. I would like to suggest that each of us take some time and conduct a small test:
Start an instrumental song in the chosen style, then continue it in different places of the original track to check what factors influence the loss of quality - maybe it's a specific instrument, maybe an effect, maybe the positioning of the instrument in the mix space is important?
This will create an instrumental track.
Then, to the same original track, which will be our reference, we add a continuation with vocals and check where the continuation creates the most artifacts, distortions, etc.
I will definitely check out rock and metal because I noticed a huge drop in the quality of rhythm guitars.
How about this idea?
Hello since yesterday I'm trying to continue a song in desktop version and it came it's 3 second with no sound our kind of noise. It's something wrong?
Try continuing at a different point in time. If you continue where there is silence, Suno thinks it has finished the song.
Go back a few seconds - if you find the moment where the vocals are, add the currently sung line of lyrics at the beginning of the continuation so that Suno knows where it is - Without it, he can skip a few lines of text because he gets lost...
Hello everyone, I would like to ask, yesterday I started testing with music generation and purchased a subscription. I successfully made three verses (czech language), but since today when I put continue with the last song, AI generate a 3 second recording with the message credit refunded. I've been trying for about an hour and still the same problem.
As above - try to continue in another place of the part
Thank you for the reply, I have tried this a few times but no luck.
- Question - does the moment you continue from contain music and/or vocals?
- Please paste the Lyrics you use for the new part with all metatags
Sorry I'm replying late, but my wife force me to the pre-easter cleaning... 😅
We should get refund for low quality outputs :/
This makes me think about a potential feature, if feasible. Ability to re-process old track. If all input data is stored, there might be a possibility that once the model is fixed, it could produce similar, but better, output? Some great pieces from V3 alpha are still waiting for that day 😔
From what I gathered from previous posts, Suno does not rely on seed generation (And even if it is, it apparently doesn't save any input other than what the user enters).
Moreover, mixing is done as a whole, not as individual tracks joined together. Unfortunately, this is a big limitation for mixes, remixes and mastering.
Correct me if I'm wrong
Right now there seems to be an issue with the web UI where the "download audio" link does not appear in the ... menu for some songs.
Now they are back. Looks like it was a temporary glitch.
There's possibilities, upscaling models, or thigns analagous to what 'MultiBand Diffusion' does with existing audio samples using Meta's 'encodec' underneath. (Or vocos in Bark.) This is the bugs forum though not suggestions.
Well... You can always download an MP3, isolate individual tracks with some external software, throw everything into a DAW and mix it yourself.
What you're asking for is basically the creation of a new AI just for this purpose.
EDIT:
Damn! I thought I was responding to the author of the original post 😅
Whoa! I just used my last 10 credits for the day to create new 2-minute tracks. One track was 1:35 long BUT the other was 3:30 long although at the 2 minute mark, the tempo and style changed but kept most of the lyrics intact
Hello, since 3 days none of my creations appears on the "New" tab. Is there anything i can do?
what is the difference between REMIX and CONTINUE ?
Probably y’all already know this but everything is spinning and not finishing again, one of my songs went so long it just disappeared
Hey, Suno is amazing. However seems many Chinese characters are wrongly pronounced. I generated a song about periodic tables but the Chinese version doesn't sing well. 😆
can anyone help me
im unable to listen to my creation all of a sudden:
https://app.suno.ai/song/b5072a7e-6bb1-4fa3-88b2-19fc557a4dc9
it just keeps loading
i just created a song and it was perfect!
but then it suddenly dissappeared and now its gone....
https://app.suno.ai/song/82aa4c95-9446-4f6f-979d-1b9f3ceea605/
is there a way to restore it?
To je v pohodě. Zkoušel jsem to dál, podařilo se mi to rozběhnout, že jsem vybral jiný koncový čas, ale to generování bylo špatné. Tak jsem to všechno smazal a začal znova. Díky za tvůj čas 🙂
No problem
Have a good time with Suno!
Hmm, I need help, my continuations are going from a different song than the song it is supposed to continue. I tried it twice but it keep continuing a song that was already finished. (Also, it takes a long time and many songs are stuck. (Spinning forever , not only 5 minutes)
Do I need a reset or so?
Hmmm right
Oooh, nasty
i had a question, is sumo having issues at this point in time? it keeps generating tracts that start over midway/have weird audio glitches in the song
Yes, there are issues, I believe from Suno being way too popular, folks are working on it.
As far I know any time clips fail to generate, credits are being properly refunded. If this doesn't happen you can email support@suno.ai. But it can take awhile to get the refund in the credit total. There is still a risk that a truly sublime song might fail and be refunded, and even though you didn't lose credits, it is sad.
alright, the problem lies in the fact that it does generate succesfully but a few clips sp far have either autio glitches+the first part is generated ver and over lol
me too
Exactly! Free users are blocked again. Just as I wanted to test quality degradation on a free account...
Okay, so, can we talk about the current state of moderation suno again? It's gotten better about things like copyright violations that aren't really copyright violations.
But one thing I do a frequently, is use the "generate lyrics" button in custom mode to show me where the holes in any given lyrical piece might be. It's not policing anything dangerous, but what is policing at the moment is tone, which is wrong. It shouldn't be doing that, if the things that are being said in a piece of work don't violate the TOS. Instead, what it's doing is telling you to rethink your lyrics on anything that isn't upbeat. And that's content policing at a level I don't think anyone involved is comfortable with.
It'll still record just about anything, but you can't use the lyric debugger
What do you mean show holes in the piece? You ask the lyrics generator for feedback on your lyrics?
Pretty much yeah
It'll reformat lyrics, often so they're easier for the machine to consume. And sometimes, it'll add a line to balance out the math in your piece, which is wrong, but actually incredibly useful.
That's sort of an unplanned use case. Maybe you can add a disclaimer in front like [melancholy sad song] or something?
hang on
huh
why did that work?
It probably defaults to upbeat, but if you just make it clear the song is a different style, it's not opposed to it
Okay, but I shouldn't be able to skirt moderation policy with a "this is a sad song" tag
Only other thing I can think of is that I removed the first line. It was as follows:
"Well I was once a millionaire, spent all my money, cuz I didn’t even care"
It might help to write a context as the first line something like "We're rewriting the lyrics below to be more X or Y" because my guess is whatever is going on underneath is more used for generating lyrics from a simple topic sentence
Oh this is nothing. I've used it for much more complicated stuff than this
Dunno what happened but a really good render got completely changed to a different render when I changed the name to "Good Intro" gah.
I suppose it could, possibly be a copyright issue? Song's foundational blues, over a centry old. Absolutely in the public domain. But Clapton also recorded it. Mine uses different lyrics, but if you squint you can still see parts of it. But, that said... isn't there a special warning for that?
I'm happy to see you using it like that, it didn't used to work like that, used to only generate: #💡┃suggestions-web message
Oh it's neat. Helps me with structure, and sometimes flow. I probably use the system in a lot of really sideways use cases
I've been experimenting on it for months
hello all, what is the difference between REMIX and CONTINUE ?
Remix pre-populates your form thingy (to use a technical term) with your style and content from whatever it is you're remixing. Essentially, creating a new song with your settings.
Continue, adds new lyrics to the existing track. Like a bridge, or second verse, or if you need to redo part of what's there.
ok thanks 😄
Raggaeton instead of ragtime😂🤪😂
Again a bug or something! AI is unnecessarily stretching the length of song to at least round about 1:20 and most often to 2:00 even if it's just a simple Verse and Chorus or a Bridge and to fill up the remaining time it start repeating lyrics or adding lyrics previous part or gibberish. This happened for previous two songs as well but I was ignoring and thinking that it might be just this time but now since it's happening continuously for everything, I believe there is some sort of problem! Please help me sort it out.
Use the [end] meta tag where you want to end. It doesn't always work because Suno often tries to use fancy endings, but sometimes it will do it correctly.
@spark fiber But it's just first part, how can I end it there?
I wrote you a private message
Hello. I bought the pro version and I see error. Couldn't generate that.
Prompt flagged for moderation
Please help
You could try using 'throwaway' lyrics to pad out your song, and then use the continue_time feature to continue from before those lyrics to the next clip or final clip of the song.
1Some of the words in your Lyrics are "inappropriate" for Suno. Try passing the text through GPT 4.0 asking which words trigger Suno.
It could be anything in the prompt, sometimes something silly even a single word. Try looking for anything an AI might interpret negatively even if only by misunderstanding something.
😁 I generate lyrics with chat gpt. ))) thanks
@spark fiber @neat sail Thanks, but I don't think these are going to work and the problem is there is no way to contact Suno's support team, they never respond to emails and the person from their team who solved my issue last time hasn't responded for 3 days!
In that cae, you actually ask ChatGPT what part of the lyrics could be moderated, it can usually guess accurately.
This is a recent problem, I've created several songs without facing this issue earlier!
Are you continuing from the same song clip in these cases?
Just making sure you know about the "Continue From:" feature that lets you choose a time in the song to continue? It should let you choose a time before any repeated or nonsense lyrics. It's not ideal, but it can help in many cases. You can continue with a style prompt of just "end" and [end] in the song if you just want to remove them (maybe try V2 for this)
Now that is working for now, Thanks!
Sorry, I didn't understand that it was about ending an ALREADY EXISTING piece of music earlier
No problem!
trying to create in the past hour is just resulting in multiple credit refunds
If you get the same result over and over, I advice to change something.. like for instance change the timestamp one second earlier?
Thanks, same result.
Going to have to abandon eight songs and move on
This happens to all of us. It's not perfect. But sometimes you need to find a sweet spot. So maybe another second. Another thing that you could try is flipping it to instrumental and see what happens then..
Thanks, it already is instrumental
Sorry for another somewhat sarcastic comment, but it seems to me as if Suno AI intentionally restricts access for free version users in the evenings. Already yesterday at this time, usage was restricted due to "heavy usage" and only permitted for subscribers. Strangely enough, it went on all day before. Is this a new strategy because of greed, or a bug or something?
its also interesting that they never really publicly announced it too
they didn't officially put it on discord that it was gonna be down, nor no news about it on their social media at all
Perhaps that's because they don't know that there is going to be heavy usage, until there is heavy usage.
You can either choose to trust the info on the banner on the official website, or you can choose not to.
i do think its absolutely fine if there is heavy usuage and has to temporarily shut it down, i just think they need to be more open about it to re ensure
Are you suggesting that paid users be on the same level? It makes no sense, paid users must have perks, that's how commerce works. If you are so keen, buy a plan.
"in the evening"? there are Suno users all over the world.
Is anyone having any luck in finishing a V3 song without any distortion or forced chorus? I've been creating songs that fit that problem, but I'm really missing the days of Alpha 3 where I was able to create 3 clean, beautiful, varied, songs a day. Now I'm lucky if I get one usable song a day with those restricted creations. I'd happily pay 10x the credit per click to get back to the A3 glory days! It's curious how the process had degraded so much from A3 to V3 -- it makes me wonder if all this audio junk we're getting is somehow the new audio watermarking information being added to our tracks? Is it getting hidden in these forced harmonies after the first two lines? It's so awful. Were they marking songs during the Alpha or not?
Sorry, but apparently Suno AI is only intended for a certain class of people, namely those with credit cards. Those without are treated like "dirt" even though they would pay, but Suno resists it... So that's not my problem because of the criticism...
I just think it's intentional because of greed... There are AI developers who do it that way, so why shouldn't Suno use it too...
Criticism is allowed, but in a respectful manner. Fair warning.
It moderates valid music terms too. Like "climax" in the tag. It needs the whitelist.
anyway, you cannot use for any commercial usage the free songs you make, so it's just a waste of time to make music that is not really yours
Sorry, but it is worded respectfully, you will probably still be allowed to express your anger. And such an approach just makes me angry.
No. Previously I was able to squeeze 6 minutes, with bearable 7th. But in last 24 hours, not a single fragment. Noise and forced muffled chorus. I guess we should try after the Easter.
It doesn't matter in this case, it's about principles. This being restricted when you don't even get the opportunity to subscribe is the problem.
And I don't want any further response to my criticism after it isn't taken seriously.
Oh, please. Running this system costs a fortune. If I would own it, there would be no free tier. Besides, I find the pricing fair, even with all glitches and hallucinations.
Criticism and venting are one thing, defamation is another.
You have a right to your feelings. You do not have a right to categorize others in a defamatory manner.
I said I didn't want an answer to my criticism. It won't be taken seriously, so leave it alone.
We have been told that the current restrictions in payment methods are due to issues with how payment is operated by the payment processor. This is the current state of affairs. It is a technical issue not one of greed. No one is treating anyone like a second class citizen.
If you are able to, why not use a prepaid credit card? We have been told that people have had some success with those in the past.
I appreciate your response. Same here. Hard to get anything to fire right. Is there an official status page for these considerations, so we can follow that instead of searching here?
Unfortunately, I'm not able to do that, and that's one reason why I feel ripped off, not just by Suno AI, but by “AI” companies in general. But as I said, it's better not to answer this, as I said, it won't be taken into account anyway, so it doesn't matter in principle.
Then why post it?
You act as though this is a personal thing targeted towards you on a personal level. Almost all AI generators are on a subscription basis. And the reason why is obvious. Because it costs money to generate, every single time something is created. Whether it is a good result or not, it always costs money to run the system. So the only way it can work is if those who will use it will subscribe and offset the cost of operation.
This is not being "treated llike dirt" this is the reality of the situation. I actually think Suno has been quite nice in allowing so many people to have free accounts and allowing so many chances to test the system for free. But in the end it costs money every single time. This is not hardship. This is not cruelty. This is reality.
Funny... You want to criticize others, but you don't allow your own criticism to be criticized. I really don't like this approach.
There have been multiple reports here about this but I'm not sure this is considered an issue at Suno. There's a poll that is being brought up and showing more users liking V3 than V3A. They also said at one point that V3A costed 4x as much so I assume this ended up as a price to be paid to drive down costs. I find myself using post-processing now to remove noise and enhance vocals but I can't consistently get the results I had with V3A
I would say V3 is better than V3A, I think the reports were mainly about the last 24 hours or so of degraded quality. At this point, perhaps because free tier is turned off, the generation is quicker than ever, but the noise... A bit off topic, but all things considering - maybe not that much: what do you use to remove the noise?
The noise reduction effects in Adobe Audition. You have to tinker a bit with the settings though depending on the song. Sadly no AI to make it a one-click type of thing
Does it work pretty well at removing the muffled and distorted sounds?
I tried it last week with two long songs and it was hit or miss. I ended up having to redo a complete song that way just to make it slugs better, but BOTH I started out as instrumentals. I do rock and power metal. Edm had never given me problems until yesterday, both the instrumentals and songs with vocals came out muffled, all of them, last night
And I did the test today - the effects, conclusions, etc. are in the tab https://discord.com/channels/1069381916492562582/1102676217687195679
hmm, i've noticed a problem. As long as Suno is visible in Firefox, it makes the browser eat up a lot of CPU... corresponding to loading a single core to the max. This slows down anything else going on in Firefox.
Work-around: Hide suno's tab.
for instance, i noticed stable diffusion not really getting anywhere until i hide the suno tab.
don't react properly with very specific request like asking for only one instrument, for example when i ask, epic wind instruments only, and getting piano.
Personally I think the poll was skewed. Seems like the majority of votes happened the first day v3 was released. Not enough time to even test or know if v3 was better than alpha. People were just hype to see V3 released.
Been having a problem that when I generate a song, instead of getting 2 different songs I get the same one twice.
Renamed generation and CONTINUE then the 2 CONTINUE outputs don't have the RENAME
PS The name of my song IS "Seriously Flawed" ( not a comment which would be more like Seriously Brilliant )
Message says generation temporarily only avaisble for Pro and Premier. I have Pro but can't generate new tracks 😔
...appears to be working now.
can't seem to continue from existing song... keeps generating new songs ...
Same
When doing a CONTINUE copy the continued "generate"'s STYLE into the "Style of Music" box
I can't continue songs, despite trying to copy and paste information over in the "custom" view. It doesn't seem to generate any audio
I have no service even though I paid for it
V3 has been complete garbage for me trying to continue from songs I created in Alpha. I haven't really tried creating anything new in V3. Sounds like it's bad all the way around. I'm not upset as the $10 was worth it for me, but it's a real shame that I can't fix my beloved song (which has a few problems and also a few changes I'd like to make).
Suno, if you're listening. We need a tool that allows us to change parts of the song in the middle (like changing lyrics with same general singing, adding bridges, instrumental sections, introductions, endings, etc.)
I'm telling it like it is. My V3A results were top. My V3 results with the same prompts are not satisfactory. How can you make everything worse from an alpha to a final version.
Please bring back the V3A.
They're not bringing back V3A. Instead, I think it would be better for them to update over time to improve V3. Check #💡┃suggestions-web message
Exactly the same here.I have unsubbed for now, maybe next version whenever it may come is better. But trying to bring this forward only results in getting gaslit and told everything is OK or that you are just a "hater", I mean look at the devs reply above.
Here is some feedback: Remove the white noise/fuzz that creeps into every synthwave, retrowave. etc. songs making them completely unlistenable and make the bass sound clean again as it did in alpha. If you need examples I have several.
Example 3: https://app.suno.ai/song/3776e76a-d16d-4788-b459-c6b6396248f5
Example 4: https://app.suno.ai/song/4827a025-ddc1-4d5e-9062-3b97c462827e
Example 5: https://app.suno.ai/song/c7394a07-d27a-4292-b452-0791a3e8fb2e
Example 6(espescially towards the end) https://app.suno.ai/song/c71af47d-f9bf-4ffe-9bf5-ac22e34b94b0
Example 7: https://app.suno.ai/song/61d09233-f7c1-44f4-ae36-b3b8a0e0c54c
Also find it in other genres like house: https://app.suno.ai/song/c9569a31-29de-4d24-b3cb-ec319548835c
Example 9: https://app.suno.ai/song/8bc92a74-47d4-4496-b9af-d419105c81af
Example 10: https://app.suno.ai/song/1554e638-c2b1-4374-85b7-4bd294291917
Posted in hopes that it may help you figure out the issues.
Thank you. I will test it. 😊
The feedback was for the developers 😛
😵💫
Sometimes Suno puts thing that's in "style" into generated audio lyrics.
Totally agree. Anyone is free to find and use another service. And good luck with that.
Hey guys, very annoying moderation bug. The word "nerd" is totally banned at the moment. It's an artist name, allegedly
the entire topic is off limits
@valid jungle
Me too, it's not even worth wasting the rest of credits I have left, the results are not even comparable and I have many examples on my hard drive, even V2 was better sometimes. It's crazy and the devs, mods no one seems receptive to feedback, which clearly affects a lot of people who know a thing or two about music. Thanks, until next time I guess
Yeap, unfortunately for me the genres I make are not anywhere close to what they were in alpha. Im both sad and glad that at least some understand what I am trying to say. The blue colored folks(not all of them) here mostly laugh at you and tell you its your fault or that you need to do other genres... Helpful... Maybe next version. 💪
credits sometimes are displayed wrongly (5 less or similiar) if you switch between tabs while generating music. mostly happens to me if i switch between create tab and playlists while generating a track. but refreshing page always fixes it.
I'll register my artist moniker as "Female". And that's when all of you are done with your experiments.
I had to delete like 50 songs because of saxophone. I wrote "no saxophone", or "exclude saxophone" and guess what? I got saxophone. 😭
Negative prompt never works. In fact, it's a common problem with all models - when you are specifically asking to not include something, there is a chance you'll get it. Even DALL-E, SD, and even GPT. But, as a suggestion, you can try to find a positive which naturally excludes the negative. E.g. synthwave excludes vocals and a bunch of instruments that "don't belong".
Not a bug, really. That's how WWW works, unless you are polling the API constantly to get your credits. Which most of us don't need to. They could make it possible to refresh the count when you click the number, but otherwise I think it only refreshes when you are consuming credits in the same tab.
Bug in french 🇫🇷 the ʃ sound is sometimes replaced with the k sound
≈ 75% reproductible for the word "Chevalier"
The "Ch" should have the "ʃ" sound but it have the "k" sound
So it sound like "Kevalier" instead of "Chevalier"
The voices are coming out very " robotic" again.
And please see the " censure" part , because if somebody wants to make a metal song ( that usually uses bad words and stuff) will limit that genre or kind of creation
And please try add a mixing mode, imagine each clip generated (let's say clip 1, 2 , 3 ,4) and then we can choose in different orders ( let's say 1, 2 ,3 and again 2, let's say a chorus, then 3 again and finish on 4). That would allow "bigger" music's and also follow an music scheme and have (like normal) the repetition of chorus etc
It's been ignoring things like [intro] and [instumental] at the beginning of songs recently, and ignoring them HARD!
rewrite it as schevalier
@willow pollen seems to work, thanks
Totally get your frustration with the "robotic" voices and censorship.
But when using Suno, it's important to remember that it's a service with its own set of guidelines and responsibilities, not just a personal tool like a guitar where you have complete freedom. Suno strives to offer a creative space while also being mindful of our responsibilities in the content produced through their platform.
For the voices sounding robotic, you can explore different voice metatags and styles within Suno to find a balance that works for your music. Adjusting factors like tone, tempo, and emotion can significantly enhance the natural quality of the voices. But also genre is very important. If you are more into electronic music, voices might be sounding more robotic because that is like the RL music in that genre. AI learns from us. 😉
Hey Aith, that's a quirky one with the French 'ʃ' sound turning into 'k'. Not really a bug, more like the AI's still learning nuances in pronunciation. French can be tricky for AI, but I'm sure the team's taking note for future improvements!
Solution in the meantime: try writting in different but still has the same sound.
yes I understand, thinking about it some word in french use the "k" pronunciation for "ch" like "chaos"
btw, it would be great if there was an automatic conversion for numbers to words (4 -> "quatre") when they are used in a specific langage because the numbers are always (afaik) prononced in english
@low cypress - tagging you so this doesn’t get lost. Apologies if that’s incorrect. Anyway, I asked Suno to generate a 70s operatic glam rock and did not provide my own lyrics for this instance. Unfortunately, the chatGPT (?) component produced four lines from a song from Rocky Horror Picture Show. I realize, as a musician that we should do due diligence on whatever Suno produces, but this is a bit unexpected. Just thought I’d raise the issue if this hasn’t happened to anyone yet. Lyrics were: “In the velvet darkness of the blackest night
Burning bright, there's a guiding star
No matter what or who you are”
The obvious risk here is that if this happens enough times and the public notices, it could raise concerns with the recording industry. Of course, this may just be an anomaly. (That said, thank you for all of your work on this amazing technology!)
... There's a light ... over at the Openai's place ...
It seems so. 👀
That's diabolical
Thought: You could try micro-managing your instrument list from the style box. I do that often.
I noticed that Suno works in English by default, but recognizes another language at the first word read. This often creates strange artifacts in the pronunciation of the first word, especially if it is preceded by meta tags in English...
Each subsequent part begins exactly where the previous one ends (or where the user chose to end), part 2 can only be followed by part 3. After part 3, you will not return to part 2, because part 2 doesn't start where part 3 ends... You know what I mean?
On the other hand, it's AI and it could create a Bridge 😄
Sure.but songs usually use the same parts to chorus. If they could make something like I said ( blocks mixing) would help a lot in the making of the songs. Salud
I agree. But it's still just AI with all its limitations. I also think it's idiotic that when creating part3, Suno cannot analyze part1 to repeat the Chorus that is there...But well... That's the way it is now and we won't change it.
Yes there was a reddit post about someone getting identical AI generated lyrics to an old Ice Cube song. Maybe Suno needs to partner with a different AI service for lyrics.
Or Suno is better at creating modern music than we think - nothing sells better than remixes of good old songs 😂
@spark fiber do you know when we can play music like before with V3ALFA?
Als pay customer i'd like to use the old version for my usage
I don't have any contact with developers, but from what I gathered from other posts - never, because maintaining V3a was 3-4 times more expensive than V3. Why? I have no idea. We probably have to wait for appropriate patches, or even V4.
I've been using the free versions for a while, but has V3 finally been destroyed? Since Suno AI was recently only available for Pro users and now again for everyone, nothing works. Bad generations, the voices are horrible, something has definitely changed. In addition, prompts are constantly being changed or completely ignored, what happened?
Plus all of my generations are not working and saying credits refunded
It's probably just a classic idea to force users of the free version to subscribe.
It's classic according to you but I never heard of it.. plus it is untrue..
Giving an opinion is one.. defaming is another.. Please refrain from making accusations like that.
that should be normally considered a bug, if that's intended then this is not normally the way to do this, no site shows old information. it updates itself to the new information if i stay at create tab it just does not change if i got to playlist and back to create(is basically a refresh normally).
Well, it may be that not everyone knows something like that, but I have experienced something like that many times and know how it is. Nobody is being dragged down here or anything like that, something is just being established here... And the question was, what's wrong with V3. Nothing else. And I think it is appropriate to address the question here and not lecture someone just because he expresses his opinion.
Well I have the plus version and it’s still not accepting my generations
Nope. It's equally problematic for all paid tiers.
I am a moderator here, a volunteer. And I think it is appropriate not to make assumptions before hand. I personally can only guide you with knowledge that is on this discord and what I have heard from the Devs. #💡┃suggestions-web message might be of interest to you.
Sorry, but behavior like that is simply not acceptable. Criticism should be expressed everywhere, especially if it is meant seriously. And if I am addressed like that again (it is a bit disrespectful), I will leave this server. And with a clear conscience. I think I'll do that too, I wouldn't do that to myself. And Suno should think about what kind of people they hire as moderators. Unfortunately, user-friendliness is foreign to this service, in any form.
Alright, it seems like TheOneOfYou14 has left the server. If anyone else is experiencing similar issues or has concerns, feel free to reach out. We're here to help and improve the platform for everyone. Let's keep the conversation constructive and focused on finding solutions. Thanks for your understanding.
I'll write a sad doom metal song about it.
critic is one thing but "force users of the free version to subscribe" is like 3daizy said and accusation. these are fundamental other things you need to have proof for that and i don't have the problems with free version like you have.
Not even a day old discord account, just stirring trouble it seems.
It looks like generating lyrics steals five credits from my account, It's a bug or a feature?
@undone zenith I just generated from a blank slate and my credits didn't go down..even refreshed the page and nada.
Then a generation took 15 credits because I stepped from 80 to 65 in one generation
Can anyone explain why I get Rock instead of Soul? I wrote "soul, solo voice, real drums, bass line". 90% songs I do for trash. It's frustrating. The field "Style of Music" is quite ... how to say it politely ... uneducated?
Ok, here is the song. 16 attempts, and this is the best(!) it can do today. Both voice and instrumental are not what they used to be a few days ago. Publishable, but only in bugs thread 😆
https://app.suno.ai/song/68761d46-718f-4c57-94cd-3f1770ee0af2
This sounds like very low quality mp3... https://app.suno.ai/song/a5180d83-5a53-4585-9668-260ed72c6525/
soul, strong solo voice, real drums, bass line, fast tempo, bass song. Listen and make your own with Suno.
Ye yesterday was a great day -- 34 files some songs with 2 CONTINUES. - All GOOD ! Today: ( did someone turn up a SPEED KNOB? ) Either the whole 4 verses complete in 2 minutes or EVERY CONTINUE starts with an "soon jump" to the beginning ( tried 2 songs, including yesterday's song, and multiple attempts at CONTINUE)
"Hey guys, look what I found at attic! An actual tape from my early days!". Then call your album "Disintegration loops".
Don't worry, you were right and you have my sword!
Good evening, I have Your Pro Plan but nothing works
Due to heavy usage, generations are temporarily only enabled for Pro and Premier subscribers.
So this suppose to be soul 😕 https://app.suno.ai/song/cf0d7b9b-7858-47b1-a684-b8c52de50a22/ I think I will go to bed and watch some movie. Spent 2000 credits and I have at least 4 awesome songs.
Well, I can tell it's not metal. So it must be soul.
Is your "Create" button yellow as a chicken or grayed out?
The information is displayed to everyone, but paying users can use the program.
Couldn't generate that.
Prompt flagged for moderation
Well, it means it works. Just try to make your lyrics more subtle.
An error occurred generating lyrics.
Please try a different lyrics prompt.
Well, you have another problem - your Lyrics or Style Prompts contain words that Suno considers "inappropriate". Run the text through GPT 4.0 and ask why Suno rejects it.
No matter what text I write, this message comes
Out of curiosity, what did you ask for?
okay thx
Any luck?
yes it works again was a word
Hey there, I've been trying to create part 2 for some of my song creations and adding my own lyrics as it opens the custom mode, and the part 2 creations never adhere to the lyrics for me. Is this a known bug? I am wasting credits trying to create a full song and I feel like my efforts are wasted
Guys , I have a problem, I am clicking on button "continue from the song " but it is not continuing, 2 minutes and that's all... Maybe I am doing something wrong? I have premium
What need to do? Anyone know? Maybe this is a bug?
The previous fragment ends with silence? Try to continue from earlier time, like 01:55.
Not a bug, a feature. How exactly it doesn't "adhere to the lyrics"? What do you do, and what do you get?
I choose "Continue from this Song" option and it opens up Custom Mode, so I write in the custom lyrics I create, make sure to write the musical style the same as the original part, and run it. It usually comes back with either no lyrics at all or just a small segment of the first part's lyrics repeated
I see. Try to continue from a bit earlier time of the previous fragment - it might be confused with how the previous part ended. I would say it's "normal" behaviour, we get it often.
Alright, I'll try pushing it further back for the start of part 2, thanks. I did kinda do that by starting 3 seconds earlier than the end of the first part, but maybe I gotta start even earlier.
Also, if the previous part ends with instrumental, you might want to insert "[instrumental]" before your verse, to let the instrumental part finish. Or they may merge, and it will be ugly.
Nice tip, thank you!
In the web player, if we click "dislike" for certain tracks and then proceed to play the liked ones on auto-play, it unexpectedly includes the disliked tracks. Ideally, the auto-play feature should only queue tracks that are currently displayed on the generation page, excluding the disliked ones.
How do I fix this?
Are you using a VPN?
no
And I tried different browsers
I tried brave and edge
Ad blocker is disabled too
I personally have no idea what exactly this is. It get triggered the last days now and then. I had it too but only once. retrying solved it for me.. And today earlier some one solved it by using a VPN. I have already told the Devs but that is the only thing I can do personally (since I am a volunteer)
ok ty
i don´t know what happens but I cannot sing up
Vpn fixed it. Thank you very much
try this
Good to know! Thank you ❤️
i can't use kalimba insturment as style of music because it thinks i want to use kalimba as artist (which i didn't know existed) how do i fix it?
Hello, it worked for me 2 days ago (I was able to do also part 3), since then... I didn't try because I didn't get song I would like to extend.
How do you do a part 3? it always declares it as part 2 for me
Actually this is helpful also for me... 👍 I have good speedy song, but takes 1:59, it's quite logic to go from eg 1:40 and add lyrics
a bit late to the party but I would like to confirm, as a paying customer, that since v3 is out of beta, it has become way worse, and it gets worse by the day. the voices are horrible, there are many glitches in prompt following, and the music sounds garbage, closer to v2 and v1 than before. Whatever happened?
just noticed it on the first day, but was like it is probably just me and went on. but now that I generated 8 more songs, I just realized that I am not mistaken
came here to see if anyone else has the same problem but apparently yes
Reverse the changes please, otherwise I will stop my subscription. this is not what I paid for.
I paid for v3 not v2 disguised as v3
When Part 2 is generated and you are happy with that, there is an option Continue from this song. It's the same ... My song about sleep could be longer 😄
ahh ok makes sense, for some reason it wasn't giving me that option before but now i see it, thanks
I have the same issue...
I generated today around 20, all for trash
My stats: I spent 2000 credits, got 4 good songs.
yep. We are burning our own money, for nothing. I entered my info and added my money with confidence that I won't be fucked over, but if this is something they have done deliberately, then the company is not trustworthy. I will wait 3-5 days, and if no annoucements are made, nor changes get reversed, I am immediately canceling my subscription.
f###### h#ll!
I only used a few credits for now.
my condolences
I'm waiting how they manage the quality issues, then I will see if I will keep the subscription. Anyway, I asked several times for SOUL and I've got pop, in some cases hard rock. 🙄
I am almost certain that they have changed the model for either something entirely different, or heavily quantized the model so the servers can handle the influx of new users trying v3
Unless it is because of copyright laws, and there is something deeper in the background.
Either way, the customers are the ones screwed over, which is never good.
We will see (I hope) in next few days what will happen. Anyway, 1 AM here, time to go to be, bye!
Well, for the irony of it, I generated one more song with the prompt "a song about suno turning into garbage for some reason"
same! hope things turn out for the better. Goodnight!!
you should not send it, I will have bad dreams about suno ... 😩
also ty 😀
to add to the things here, maybe useful for the devs or idk
spectrogram of a "v3" generation
vs a v3 beta generation
Perfect. I was about to post spectrogram too. Thanks.
I canceled my subscription. Subscription would have been charged April 1. I would have renewed if V3 was better or at least the same as alpha. But reading here, it seems like the developers know the issue. They said they aren't going to go back to Alpha, they will only release a new model when it is time to upgrade to the next version which could be months. So I will probably pay again when I hear they do release an upgraded model.
it seems like the developers know the issue.
Can you point me to the message please? I have not looked back far up. But I am interested in the message. Either way, it s#cks. If they do know and said that, then that s#cks even more. Not looking bright for the future from what info I gathered thus far.
I feel lucky to have had generated what I needed for my project before this situation.
I could always create the entire song myself, I am a producer, but I have a yet to find a vocalist. I used Suno as a singer, isolating the voice and adding my music under it, and even hoped to ask the devs to add a function where suno would sing over my songs, instead of vice versa. But this situation is not good.
That is not good, certainly. Thanks for the link. Heck!
I think I will cancel my plan too.
Going to leave for some sleep now. I will check back later
Good night everyone
That's common sense, I'd say. I'm glad Suno's subscription model is flexible. Still have plenty of time till renewal, but if situation won't improve, I'll get a well deserved break.
Bug:
We've created a new term. Often with the percussion (Usually with a rock style song) a sound that sounds like someone banging on the side of a metal shed will come out. We'd deemed this "shedcore" AKA my song is going shedcore. We'd like shedcore to be hit with a lazer beam from space where it can never touch our ears again. Thank you devs for understanding.
I put [drum roll]but I can never get drums on this: https://app.suno.ai/song/938c55c6-0939-4bbb-89c7-8a49341baaeb which was continued from: https://app.suno.ai/song/30d5b0ad-873f-496e-ba76-91cb0caa2428
It's a repeated, also really strange error, I wouldn't be surprised if my continuation was v2 instead of v3.
I'm pretty sure the slowdown and backup today is the result of V3 requiring users to generate 10x as many clips to find a decent one
#bringbackv3alpha
Same. I'll check back in a few months but everything I'm generating has vocals with high reverb and muddy instrument sounds. So as for now, I'm out.
I try to remix my sound and I get this message: An error occurred.
A system error occurred.
anyone else having issues with instrumental tracks sounding like they have a cacophony of whispering nonsense vocals in the background? Seems like most of my tracks are coming out this way now
You are lucky. For me it's not some. All of them are sh**core.
Whats going on with V3? I was a V3 Alpha supporter...... Cmon Suno you know something is up
yes, hope for fix asap
Just came here to check if I'm not alone and definitely I'm not. Suno was really great at least 3 weeks ago when V3 was in alpha. A miracle! Now it became....uncreative? I can't find a word, the music is all the same, it takes 30 tries to get a decent melody out of it. Almost no variation to what it creates and it's too monotone. I'm barely using it now, I've been really hooked when it created cool music
I feel your pain. I bought a year after playing with Alpha. This is crazy
is there some sort of limit on the number of verses? I'm at verse 5 and it absolutely refuses to sing it. It insists on going back to the chorus. I've tried continuing earlier, different points in time... I've wasted a lot of credits on this by now.
yeah I noticed that one too, it wasn't like that before, seems like it's a new bug
So with this song... just see how ridiculous and ugly the lyrics have gotten due to my many many attempts to get it to sing a fifth verse. This came out ok, but it still has no fifth verse and instead sings the chorus. This took WAY too many credits. Also has lots of multiple voices singing, but that's a known issue. Would also be nice if we could edit the written lyrics to clean up this mess.
https://app.suno.ai/song/1825b75e-8e6b-48c6-afeb-4b070db8cddf/
What I have found works, is to start the continuation on the last line of the chorus adjusting the time to right before that last line of the chorus, and use it as the first line as you move forward with your next verse.
hello, so basically i made a remix of a song and the original basically disappeared and i don’t found it anywhere, i have looked in trash, library, create every single option that i have thinker or the people in general suggested and still didn’t find it
help
Can also add that you have looked in the library, in the trash bin etc. and cant find it beause otherwise they will ask you all the same questions we did
Suno doesn't know what is hand percussion. I tell it "[hand percussions softly introduced]". It just reads it out!
Definitely, effects such as reverb, chorus and others that affect the stereo reception of the song have a negative impact on the spatial positioning of other instruments and thus the entire mix.
Getting a good rock/metal sound is a pain - appearing of lead guitar ruins the whole mix, and getting the Suno to generate only rhythm guitars is difficult to say the least...
Does anyone have similar findings in other styles of music?
I think no need to be sorry, they will sort it out, I just hope - very soon.
I bought pro and generated a cool clip that I tried to “Continue from Clip” to extend the song with my own lyrics like I’ve done before but everything I do is getting stopped as “prompt flagged for content moderation”.
I’ve read up on it and tried everything in case i was using censored words or themes. But NOTHING works.
Everything causes that error when trying to continue that clip. I’ve even tried just putting in [Verse 3] and even that alone fails.
Is this a bug? Or has my original clip somehow been marked as banned even though it was successfully generated? I don’t get why I can’t continue from this clip at all? I am pro, has my account been flagged for some reason??
Please help. I’m not doing hate speech or anything weird.
Run your lyrics through GPT 4.0 and ask which words are inappropriate for Suno. This is probably the faster and most effective way.
Edit: I just read that entering only verse doesn't help either - check if Style Prompts doesn't contain any prohibited phrases either.
Thanks, I actually tried that and it just gave me a long litany of guesses at intent. ChatGPT4 didn’t seem to get the continuation from clip part that the first set of lyrics were approved and generated but no lyrics at all work when continuing.
It’s an upbeat pop song and was called “Eat the Rich” and the first clip with a couple of verses and chorus worked fine. (Which it came up with- but nothing I do works - even just asking it to continue from the one it created gets the flagged error)
Could you please elaborate on Style Prompts? I haven’t edited that field at all.
I suspect that the word Rich may be associated with the name of an artist. Try renaming the song
Even though Suno named it for me?
Do I just rename it to something else in the continuation?
Yes... Change title in continuation
Ok cool I’ll try that
hello. Accounts are not exist jet? https://app.suno.ai/account/USERNAME = 404
Let us know if it helped.
Where did you get the link with "USERNAME" in it? For me, clicking such a link also returns a 404, even though I am logged in and have an active account.
Edit:
https://app.suno.ai/user/
???
🎉🥳👏👏[SOLVED]
You nailed it. Even though Suno named my tune, the continuations failed on everything even though it said it was my lyrics.
I tested it even with more interesting lyrics and every time it worked. But I had to change the title of the song from “Eat the Rich” to just “Eat” (even though the voices were singing “eat the rich”)
Ok. Give it some context - what do you want to achieve? Because I don't understand what's going on at all...
User sites like https://app.suno.ai/account/USERNAME or https://app.suno.ai/USERNAME to see how was creating an song or watch out published music
this time all user named user
What?
Did something like this exist before and then suddenly it stopped working? This is the first time I've heard of something like this, but I'm not a veteran either...
I didn't say that. I just have strong legal concerns if someone uses their own texts (mean song righter) and doesn't have them generated by the AI.
How you manage this without user pages?
User profiles and other features are being worked on but there isn't a profile yet.
Hi! I have encountered from today two days back that I've created 4 songs and I am empty of credits!.
It may be a temporarily bug, but it's frustrating. Would like to have this fixed so the free accounts doesn't feel triggered to pay for the premium.
Thanks a lot!
Try refreshing the page
I also noticed that the credit counter goes down way too quickly, but after refreshing the page it goes back on track.
Done it. I didn't thought of that and it worked! Thanks for the tip, I'll try it next time I see a problem like that.
i can't generate the song
Try changing some lyrics?
Im glad it worked tho
yes same
Because your lyrics are probably too graphic. ChatGPT can help you figure out what is wrong. Next time just paste it in their and ask why your lyrics are flagged
Jo I will that then but it is kinde sad I feel like it should be like that for my Ghana
https://app.suno.ai/song/4975847b-f7e7-4e55-b17d-9ea278703a8e it skips so much of the song in both versions that i created
heavy rock, intense, Distorion Guitar and Heavy Metal song. Listen and make your own with Suno.
Is the problem with the text Again otherwise I wouldn't know why it skips half of the text
Everytime I enter I have the same problem, when I enter I have 100 credits, but when I'm going to create a song I lose 5 credits before doing anything for no reason
And then I went to 85 without generating anything!
15 credits stolen without doing anything!
first time I try to generate since the "upgrade" 2 days ago and boy the difference is very noticeable . intros are rare. very rare. unless you put it in the style itself, which is always bad as prompt should be taken into account. it's like V2 in a way. also very strange bugs the same as when upgrade happened - skipping full verses or starting at the end of them. this happens about 25% of generations. just skipping full verses. . all in all engine sound and feel much more like V2 now. I don't undertsand why mess with is so much it was in a good spot, now generating something that follow the prompt is very hard
Yes, It worked far better for me before the "upgrade"
Also the choir effect is getting out of hand, almost impossible to get a clear one person singing
You know, that people won't recognize, if you'll put "reach" in your lyrics instead? And the same with other words.
Did you try to put verse tag into square bracket? I.e.
[verse]
The sky is red and violets blue
I take my axe and come for you!
The blood in awkward silence spills,
For I forgot to take my pills!
yea i tried to do that but it still skip have of the text
i hat to chance the lyrics of the song in odder to make it work
Happens quite often. For me, it may take 10 attempts to get it to spell everything as I want. Hence a good song may take around 500 creds easy.
Yea but that gets quite different if you only have 50 a day
I hoop that will get improved by the next update
You are praising to chorus.
No pun intended.
Hi, Suno.I was able to listen to this song when it was in generation, and it was over two minutes long. But now I can't listen it at all.😥https://app.suno.ai/song/e611135a-1e72-4228-b529-ead613ceab34
Yes, it doesn't play. Never thought "Whiskey" could be a genre. Made a difference?
Just try , I don't hear much difference🤣
goed one
Generated a song, listened to it a few times, and then it disappeared. Here is the link from where it "should" be https://app.suno.ai/song/bc486ce4-3dab-4510-8778-355a5d708fa9/
is there any way that I could still find the song to download it?
Hi there. I keep trying to continue an instrumental song, but every time I do so it adds lyrics into it even though I still have the instrumental option checked
Did you accidentally put something about vocals into the style prompt? I did that once and had the same issue
nope
"high energy rock anthemic" apparently
It could be anthemic causing it, but I am not sure ofc.
I started listening to a generation while it was still being created, and it's 5 minutes long and still going.. it sounds super glitchy and broken and keeps pausing, but it hasn't stopped yet
is generations going slow for anyone else?
Notice a rather strange problem today. generations are slower then usual, but what's strange is one of them always complete first, and the 2nd not only generates slower, but usually stuck at about 0:22 or 0:28 for a few minutes before it finalize
not just me then.
I'm getting 'Credits refunded' all the time when I want to continue a song. Anyone also has that experience?
Yeah servers are overloaded
Here is another one. Song is finished generating. I listen to it. A few minutes later I go back and it is a totally different song.
no i didn't get that message pop up either
F5 refresh.
I had a song where the entire song was rewinded to the beginning after 10 seconds or something. It was a cool effect and I thought it was on purpose, because I was making weird demonic horror sounds. But when I played the track again later, the rewinding was gone and played as normal. 💀 😰 "They" know I am summoning them, I guess ...
Actually this song started off as a pop song, and half way through it switched to a country pop song. Was weird but kinda cool
Today I have a few new bugs:
- Singers are singing metatags in [] brackets...
- Parts are not generated fully and the circle indicating the length of the Part keeps spinning endlessly.
- Some Parts described in point 2 disappear without returning credits.
- The generator creates subsequent parts entirely unrelated to the previous ones, attempting to create an ending with the Metatag [End] and "End" entered in the Style Prompt as well - ignoring the style, which could actually be random, the generator didn't match the tempo and the moment of changing Parts.
- Around 10% of the parts generated today were cut off after a few seconds and started from the top of the part - sometimes multiple times in one Part.
EDIT:
Oh... Now i see info about upgrade...
Many of my creations switch styles after 20-30 seconds now making them unusable
Yes same here
no credit refund too lol
DEVS wth is happening, I click "get whole song" from the last part, but the last part is changed in the whole song, like it added another part not this. tried it 3 times, refreshed page, tried and tried , and still getting a full song different then the clip I click to generate from....
refresh your page.. you will see this banner too ^^
I see the banner, I still don't get how it's related. my generation completed, it just changed when addd to the full song...
Why can't this be related?
This is already going on for at least a half hour.. so anything you created can be effected
I just hope this part although generated, isn't actually lost forever in the server eye
Is the upgrade gonna get us back to V3 Alpha?
I hope it wont get us to V3 teenager rage fit stage
the engine just invented a part that doesn't exist, I went through the 8 tries I made of the last part, and non of them fit what "generate whole song" made
If you "get whole song" again does it generate a different last part again?
no, 4 times it generated the same thing, which isn't the one I click to get it from neither any of the other tries
So I tried generating whole song from another part, and it's ok, but from the part I actually want I get the wrong song ending. like this generation exist but can't really be used...
I thought it might be mixed with another part, but the ending wasn't generated at all. so the one I actually got can be listened but can't be added to the song, and the one actually added can't be listened separately
If you are having problems, I advice you to try again later. At the moment due to high server load not everything works as expected
hi, I cannot generate perfect song I've spend more than 1000 credits for nothing, V3 is better on some way but less than V3A
Not necessarily. Maybe they are adding more RAM to some nodes. You don't know what they are doing. If your changes affect the result and make system non-operational, you don't display the banner, you turn the system off.
In other words, the way I understood their message - they are just adding some elements to backend, like extra nodes to the cluster, but because it's Easter, I don't think anything serious is going on that would lead to reconfiguration and bug fixing.
You are right.. I am just a moderator.. 🤷♀️ 😉
And I'm just a smartass.
So songs from the explore page randomly appeared in my create tab, but, after refreshing the page they disappeared
Could be related to upgrade in their web app code. Perhaps Ctrl+F5 is a good idea, to ensure all elements are fresh.
Same, I lost 2 songs that were generating
1x clicking on Create (2 songs) costed 15 credits, seems like generation is currently a risky gamble.
Hopefully the errournously taken credits will be refunded automatically.
Another bug. Ive continued a song multiple times. It won't let me download the song now.
Although CONTINUE starts where PART 1 left off it jumps back after a line or 2 instead of finishing the song. Same problem as yesterday ( which never occurred the day before) . What did u do Saturday night?
Not saying there are no problems, but did you try "[end]" instead of Big Finish?
Either way, it did this before to me. Just continue from before the jump, and in new part instruct to do instrumental ending.
I tried again with [outro] instead of [big finish] similar result. I assume [End} will do just that but I bet it too jumps back. I wont waste credits on [End} cos I'll never wantto use it.
About the problem I posted about before, I think the generation in my feed is different then the one on the server, somehow. not only sending it to "full song" changes it, but also if I try to continue in the middle, it's continues not from the song in my feed, but the other unlisted song, like the one I hear is non existent
How exactly can u do you " instruct to do instrumental ending." ?
maybe this video helps some of you: https://youtu.be/5uJw5kb05as?si=gGiiztZWEX9rFmV-
The full song I created in the video, can be found here: https://app.suno.ai/song/33f4b016-c8bb-474d-9fe3-6c05b29ff9df
Discover the world of Suno.ai, your gateway to effortless music creation! Whether you're a seasoned musician seeking inspiration or a novice eager to craft your first melody, Suno is your ideal partner. Perfect for personal use...
For me, using "end" produces a satisfactory instrumental conclusion in most cases. Why don't you want your songs to end?
for me usually [outro] works fine. but can also press the instrumental button and make sure no singing is involved
I have the same problem, but it's actually the credit counter that's lying. Refresh the page and the missing credits should return to your account.
What you put inside square brackets is not cut in stone. That's your conversation with AI. So, for me, the [instrumental ending] worked. But like all tags, it may be disregarded.
OMG! Thanks so much I had no idea you could just bung in a tag. I Got 2 HUGE [Big Finish] 's in the 1 minute https://app.suno.ai/song/e692287b-d254-440f-bbd8-54bf45ae0207/ ( Not sure how you reduced your [End] to 9 secs in advance ? )
Mysterious Hypnogogical song. Listen and make your own with Suno.
You don't always need ending. Just load your mp3 into audio editor and make fadeout when needed.
Total control is not something we have.. sometimes you've got to go with the flow.. You can try a different time stamp and try again.
Hey, in the Explore > New page, it's not anymore the newest songs but it seems to be the new trending songs. Is it expected behavior ?
At the moment its an absolute waste of the paid credits for me. Songs aren't finished, or have some words in an other language in it or they just missing the singing (comes out as instrumental only, but the button wasn't set to "on" and it just happend to one of the two songs). Or songs with problems mentioned before are getting lost or was removed, but no refund happend.
They should stop the generating or take down the "create" button and fix the problems first.
What's going on with Suno? Up until yesterday it was working flawlessly for me and now that I finally pay for credits, it's a mess. I keep including in my prompt that it's a male voice (as usual) and it keeps generating a female voice. Is there any way to get back the credits? I've lost about 50 trying and trying and no luck.
German language is messed up completely. It worked flawless 12h ago. now even the simplest words are pronounced like English words.
Why it is not working right now? It says the service is only available for Pro and Premium users...
What a mess, didn't get to try it
working fine for me though
So many people are trying to use it, which makes it unusable for everyone. Temporarily they've restricted the service to subscribers so that it's still usable for them at least.
It's not. I'm a subscriber and Suno is not following the prompts. I've wasted 50+ credits so far today.
Is there any kind of form or way to get in touch with the Suno team? I really want a refund on those credits as it's not doing anything I'm asking it to do.
I understand your frustration. Creating AI songs is not easy. There is no full control. You can reach out to suno support via support@suno.ai
Thanks! I've been using it for over a month and it was working amazingly until yesterday. I don't know what's happening today: As soon as I go Pro, it no longer follows any of my prompts. I even tried re-using the one I loved the most and nothing.
strange, not its working again. for like 70 to 90 generations it wasnt working.
At the moment there are too many people who want to use Suno and there is not enough computing power. So they have to make though choices. Free will come back when there are more GPUs.
I understand your frustration, but please remain respectful. Check rule one for our community standards here. Remember, I'm just a volunteer moderator providing assistance. Due to your behavior, a 10-minute time-out is being enforced.
Discord is the easiest way to communicate over voice, video, and text. Chat, hang out, and stay close with your friends and communities.
LOL...
Yes, completely replaced. That's why I upgraded my Midjourney account this month to a larger package.
One of my creations was overwritten for no apparent reason. How can I restore theoriginal/where can I find it?
You are doing a great job here! Your role is pretty thankless at times, but I appreciate your ongoing positive attitude and your want to make this a good place to express genuine concerns. Don't let the negativity charge you!
Thank you.. that means a lot 🌼
I assume this is because of the server traffic, but just in case it's not: I've generated a song from a few continuations, but the context menu is just infinitely loading. It's got some options, but it's missing others, such as the Get Whole Song option 😦
Might very well be. But are you sure you are no looking at a song that doesn't have more parts? Because if a song only has 1 part, you cannot make a whole song.
Yep, its part 4 of the song. There's a little loading animated icon next to "Continue From This Song" that's just infinitely spinning as well
Although I haven't tried maybe disabling my VPN, so one sec while I see if that might be the issue...
Ahh.. then the rendering of that part is not done yet. Generarions are slower in general. Not just with you or VPN related.
I disabled my VPN and reloaded, and now the entire UI isn't loading 😅 I guess I'll come back and try again later...
Sorry about that 😬 please try again later yes 😅
Good/bad news: good news is another few refreshes later and now it's loading everything again. Bad news is only one of the two generations from that part 4 exist now, and it's not the one I wanted to use 😅 I won't quibble over 5 lost credits, though; thanks for taking the time to deal with my complaints that resolved themselves 😂
Does anyone know anything about creations being mistakenly overwritten?
It really is getting hard to create a decent song with V3 - I just did 10 songs with the same prompt, and despite putting "intro - violin building" ALL 10 started with acapella vocals, and I got just ONE that lasted 2 minutes - the others were 21 seconds, 50, 9, 34, 35, 42, 53, 52, and 35 - and this was despite having six verses and 3 choruses in the text box. It ignored my prompt for a solo AND for an orchestral passage in ALL ten. There were lines missing, and many choral voices despite asking for solo vocal. V3 is a pain.
Just for an experiment I am going to repeat it with V2. Watch this space.
I wonder if it's a linking problem, and some other person is now confused about why the eery haunted house background music has been replaced by a rap about the march of AI.
OK, tried another 10 with V2, all but one gave me the full 1:20 (one was 1:09), so that's better than V3. Two gave me female vocals, and six gave me a solo of sorts. One hallucinated after 30 seconds, and two gave me "nonsense" lyrics. ALL TEN started acappella, and musically none were useable. I have just blown 200 credits and 20 minutes with zero results. V2 or V3 seems to be off colour.
I like listening to my songs as they’re still generating (thanks to the new streaming feature), however I’ve noticed the majority of my longer songs suddenly skip in the middle of streaming and the final output saved is only the last 30 seconds. It sucks to be losing my songs like this and I used nearly 1,000 credits before realizing Suno was only partially saving songs. Has this bug been reported?
Worth noting I wasn’t experiencing this in the V3 beta testing
Good evening, I am seeing a bug that just started coming in on my end on the last “I’d say” 20 min new generations under V3 are generating in clip. Skipping parts of the song it originally was showing while streaming. I’ve also tried waiting and not listening to the songs; however the bug still takes effect.
Same here. Suno is only saving the last 30 seconds of some songs for some reason. Seems to be only when you’re streaming it to listen early, but I’m not gonna waste credits testing it. I hope we can get a refund
I think we can agree at this stage, that V3 Alpha was legendary.
For a example this song when it originally was been generated it was going to be 1:45 in length but after it finished buffering or loading it cut off to this 54 sec version.
good tips, thanks for sharing
Interesting, my last song was 1:54 while streaming, and Suno only saved the last 30s of it. Must be kinda random. Again, wasn’t happening in V3 testing so hopefully it’s not too hard to pinpoint the cause of the bug. I’m gonna hold off on generating until this is acknowledged and fixed
Alright, I’m gonna consolidate the bug details here:
Suno sometimes only saves the last 0-50ish seconds of songs after generation. Most of the time credits are not refunded, and sometimes the entire song disappears from your library.
Impact:
- At least 8 of us are experiencing this bug including: @wanton dawn @rancid tapir @proven hazel
- This bug seems to have started affecting us around 4:20pm PST
Reproduction:
- Generate a long song using V3 (shoot for 2m)
- Stream the song to listen before it’s finished generating
- In some cases, there is a random offbeat skip in the song (like a broken record) or you’ll get the “credits refunded” notification while listening
- If you hear that skip in the song, the final song that Suno saves is only the last however many seconds of the full song that you just streamed
- You may have to refresh the page to see the clipped song (or see the song completely vanish)
Important Notes:
- Continuing from a bugged/clipped song like this and then pressing “Get Full Song” has returned the full song for some users. This suggests that the full song data is still stored
- I’ve only noticed it when streaming because the whole first part of the song I listened to isn’t saved once it’s finished generating. However, this appears to be happening with all songs regardless of whether you stream the song (according to others in this channel)
- I didn’t experience this in the V3 testing version at all
This just happened to me, 4 times in a row
I just had it happen again, this time I didn't even stream it. It just cut a random part of the song.
At least the other generation had the full song
Wth now the fully generated one is 16 seconds
What is going on, I literally can't generate anything, it's just burning credits.
Did the server just fall over or something? Was generating a tune and now it's in a never ending spin
On the web app it says: “Due to heavy usage, generations are temporarily only enabled for subscribers.” Maybe this is what’s affecting us subscribers too. I’ve already blown enough credits testing, but I’m gonna consolidate details on the critical bug here for now: #🪲┃bugs-web message
Oh no, that happens time after time. Your credits will likely be returned for that one.
Well that's the second time today I lose songs. A generation is only a few cents but this very annoying
I just generated and listened to a full 2 minute song, while I was listening to it as it was generating I saw that it said "Credits Refunded" which usually only happens if a song is like 6 or less seconds. Once I was done listening to it as it was generating, when I clicked back on it, it was only 6 seconds and the entire song was gone basically.
It’s probably this bug: #🪲┃bugs-web message
You’re the 4th person to mention it so far
That’s good you got credits refunded, the rest of us haven’t been getting refunded unfortunately. Might be related to usage spike right now
You didn't lose anything, I guess. Just refresh the page. If generation didn't produce the result, 5 creds shall be returned. Chances are - it wasn't a gem anyway. V3 is currently at its lowest.
But this really happens to all AIs. All of them.
I am trying to pay a monthly subscription. It asks to verify my phone number and shows the end of my phone muber. Yet I get no text to verify it. It says phone number can't be found. How do I fix that?
I would contact billing@suno.ai about this, you’ll probably hear back quicker
Nope, I most definitely lost something. I had started a continuation with a set of lyrics. The continuation played but was still spinning. It just never finished spinning. And now 15 mins that generation is not showing up anymore, it is gone from my page
And creds?
I don't know about the creds, but that's not really what I care about. 6 cents per go, who cares (as long as it's not too often!)
True. But I would wait till tomorrow anyway.
Yeah, Suno is bugging out for a lot of us right now it seems (saving only the ends of songs or nothing at all). I would personally hold off using it right now until this server usage spike calms down. For those of us that lost credits, hopefully we can get a comment from staff or mods on the situation and when it’s resolved
Also, if I continue from the bugged song, and then get whole song, it is continuing from the first 6 seconds of the original song, not from the 6 second snippet that I have in my library, meaning the full songs data is probably still there
Thanks for the details, I added it to #🪲┃bugs-web message
Hopefully it gets addressed soon
I confirmed it works
Dude I just lost an insane song. It just disapeared from my library while I was flicking back and forth between other songs. Can I get it back? Is it still somewhere in the system? I really loved it and now its just gone 😭
The big issue is that you can only continue song from the end of the one that's present, so you have to guess where you're currently placed
Either that or burn precious credits to get that information
Are you saying you were able to continue from a bugged/clipped song and it used the full song?
Correct
Although I've checked, and the time you use to continue from doesn't match where it actually continues in the full song.
ragtimish, violin, electroswing, glitch song. Listen and make your own with Suno.
I chose the end of this song^
W name btw
And you'll hear an abrubt change
Ty
When it says stuff about electroswing
That's where it actually cut vs 47
We’ve all been having issues with Suno the past 3 hours, losing parts of songs and such, I’d refrain from generating until stuff is fixed
I'm just glad to get this one back, it's such a good gen
Weird, imma try continuation again from my end and see because it sounds like it didn’t work for @proven hazel
When you continue, you need to 'Get full song'
Then from there you'll get the whole one
Ah gotcha
@wanton dawn Yep, can confirm. It’ll retrieve the full song from the start, but the timing is still messed up and things are still clipped. More of the original bugged/clipped song is retrieved which is great, but still a total loss unless you want to burn a bunch more credits. Linking the original thread for this bug: #🪲┃bugs-web message
This is a critical bug that is affecting Pro and Premier Suno subscribers. I would appreciate a response from mods or staff that it’s being addressed.
I sure hope this bug isn’t an April Fools joke because it really makes me sad ☹️
I'd generally only do that if there's a message you feel needs a moderator's immediate attention (e.g. spamming, abuse or other rule breaking.)
Is there a better place than here to report critical bugs affecting paid subscribers?
Not that I know of. The mods and devs respond to messages in this channel as well as suggestions pretty often. I'm sure they'll notice/act on the messages above when they're able to.
Hi, I am using lyrics in Simplified Chinese (and the song will sing in Mandarin), but sometimes the song sings in Cantonese and there is no way to specify the language which is kind of frustrating. Please advise. Thanks.
I’m having the bug too. Was listening to a gen as it was going and really liking it, suddenly popped and only 7 seconds of the end left. At least it refunded me. But it was a good gen and annoying.
I posted Chinese lyrics, and the result was that I always sang Cantonese, which is two languages with different pronunciations, and the difference is very big, but no matter how I adjust, he can't sing Chinese, he keeps singing Cantonese, wasting a lot of my paid points, how can I get him to sing Chinese correctly?
Have you tried putting “lyrics in Mandarin” in the style prompt? Also try [lyrics in Mandarin] in the lyrics field
I don't remember ordering 100hertz potato quality samples, but that all that's currently on offer.🤢
I just tested it, the same as singing Cantonese, not Mandarin Chinese. Thank you.
You can try using international phonetic alphabet.
That wouldn’t have the tones right though
I’m trying to subscribe to a pro membership however the my account page will not load. Not on my phone or computer
Maybe suggest a specific city or region famous for the style of music you want? I know Suno has trouble with other languages which it gets confused. There’s not an easy solution, just things to try to help Suno figure it out
Thank you very much for your professional reply. For now, we can only wait for SUNO to be able to distinguish these languages in the future. Thank you
I don’t work for Suno I am just another user 😃 I hope you have better luck with your songs!
So with suno after generating a song sometimes the song either repeats it self half way through the song and when this happens I go to refresh the page and it bugs the song to repeat the 2nd half thats generating. Not sure why this happens, but just a bug I ran into.
If I understand you right, it’s probably this bug I’ve been trying to raise: #🪲┃bugs-web message
No comments from mods or Suno team at this time
alright
Has this only happened during times when Suno was restricted to Pro only accounts because of server load, or have you also seen it when things were normal?
I’ve only experienced it during the Pro-only restriction personally, and it seems like that’s been the case for everyone else here
Song slider at the bottom shows 0-1:00 until it goes past it to the actual song length
is it just me or did the limit for writing song lyrics suddenly get way less? 
v3 prompt coherence, sound quality, creativity and composition are all noticeably degraded compared to v3a.
Can someone please explain the difference between remixing and continuing on? What do I do if I want to redo the exact same clip over but I want to change the words of a phrase and JUST that part?
It's not possible to edit an existing part. Remix will just copy the lyrics/style etc and start a new song. Continue will extend an existing song.
This channel is for reporting bugs. Questions should go in #💬┃general-chat
Thank you. I didn’t see that channel. Appreciate it.
To add on a bit, while you can't exactly redo a part, what you can do is choose a continue time from right before the phrase want to change and generate a new clip. This can substitute for a redo function with some luck and patience (especially if the phrase has plenty of audio before it, that makes things more consistent.)
Usually I want to edit just an ending phrase or middle phrase in a verse. What you describe doesn’t work, I want to keep the music and the syllable cadence/rhythm just change what it’s saying. It’s a pain, but what I do is hit Remix from the clip I want to redo, copy the lyrics it give me in the Remix screen (which are the existing lyrics in the clip). Then I go back a screen and click continue from this clip make it Custom and paste the lyrics to extend the song so it has “memory” of that verse, then I update the [Verse] to [Verse 3] or whatever number is the last, then make the lyric change in the text and then continue on from after the verse. It usually creates a new verse exactly the same, but then I have multiple clips with a broken song structure. Maybe something like: [Verse] with wrong lyrics then [Chorus] then fixed [Verse 2].
Then I have to manually stitch it together in an audio editor. It would be nice if I could get a collection of clips and then rearrange the pieces in order without regenerating like:
[Clip2:Verse1][Clip1:Verse2][Clip1:Chorus1][Clip3:Bridge][Clip2:Chorus2][Clip2:Outro] or something like that to piece together all the parts I like from the various good clips.
Sorry for asking in here, just replying to you from my previous wrong channel post. Thanks!
I understand, yeah. What I described only works if you don't already have later clips. In the future "inpainting" (which is being worked on) could deal with the problem in an elegant way and let you regenerate a section of audio between two existing clips.
I don't know if it only happens to me, but every now and then the font turns black and I have to select it to see the letter I added.
I deleted this song from trash and it’s still visible https://app.suno.ai/song/3414009e-6f5b-4556-9543-991fc8eca204/
Soulful r&b, hip hop verses song. Listen and make your own with Suno.
The Explore > New page still stuck with the same songs since yesterday :/
Just wanted to give positive feedback related to previous things I reported here. I was one of the people that have been critical of v3 (v3 alpha was great though) In the last day or so, things have gotten a lot better. I'm not sure if anything changed or if I'm just getting better at prompting but things are better now. It seems closer to the v3 alpha version that worked well for me.
I still have the issue where I got billed twice for one track yesterday.
What happened is that i started a creation it generated 2 tracks but deducted 15 credits.
I won't press having it refunden as it is only 5 credits that are missing. But I hope this issue is already being worked on as from what I read I am not the only one and it sure as hell lessens my confidence in using suno currently.
Is it a bug and do I need to report or it or is this normal error? I have lyrics 1 2 3 4 5 but the Suno sings them 3 1 4 5 2 ( Yellow is the color of the Sun, but Suno is singing Sun the color yellow is)
Try removing Yoda in your music style
We were looking into this one yesterday, we think it’s affecting a small % of gens yesterday, sorry about that. but should be fixed by today
Looking into this one, we think this was a problem last night and should hopefully be fixed today.
Wish it was that 😄 but the lyrics go: Red, red, the color bright,
Like an apple, juicy and right.
Green, green, the color lush,
Like the grass, where creatures hush.
and suno is singing
Blue blue, red the color bright like the grass.
Consider it a feature. It turns pretty bland lyrics into something creative, deep, thought provoking. Eventually will lead you to Emmy.
This is often just a thing AIs do sometimes (hallucination is usually the term people use for anything not following your prompt) but it is sometimes related to software bugs. In this case my guess is a very interesting AI quirk because the text is related to a very popular nursery rhyme maybe? Or if the clip is a continued clip, then it can just be related to the previous clip.
This for me? It should be a nursery rhyme for kids 😄 Not gonna work for them to say grass is blue and red 😄
Really?? Popular? as in exists?
I guess I was thinking of Roses are Red, but reading it, it isn't that similar 😂
Popularity has nothing to do with it. It perfectly makes music for very popular nursery rhymes.
Is this possibly a continued clip, from a previous clip which used the word 'blue' in a verse?
Seriously, though, it does that to me all the time. About 10% of generations, at least, have lyrics messed up. With verses being skipped, lines rearranged, etc. But sometimes, I have to admit, this mess leads to revelation. Had these moments, coincidentally, with nursery rhymes.
Also great for lazy prompters just let the AI say whatever comes to mind https://app.suno.ai/song/e1fea520-6249-429d-bcb4-4765951bed99/
I just had Suno crush on me, was using in safari after an initial sign in through Discord, still had 1500 credits left on my account, and now there is no way for me to sign in to my pro account. Nor can I access any of the many many clips I’ve generated. Trying to get back into Discord takes me to the initial free subscription plan, zero credits, and only my first.few clips!!! It does not recognize I have a pro account This is not cool. Did this happen to anyone else today?
Yes, it is. I did not notice the first generated lyrics had sleeping in my blue jeans, and I wanted to change that part.
That's pretty common, something about the previous clip causing a mix up. One thing you can try is adjusting the continue time backwards from the end of the clip to find a spot that doesn't hallucinate. As you back up, be sure to add any words that are now unsung because the audio no longer includes them.
So Is there a limit on how many times you can "continue" a song. I had a song that I continued 4 times and now it won't let me "Get full song"
The original song is perfect, just has that one mixup of sleeping in blue jeans. No errors in lyrics compered to the song. But the continuation is all messed up.
But his part "be sure to add any words that are now unsung because the audio no longer includes them." so continuation can't be verse that was used before?
For example, if you are continuing from Clip 2. Change the "Continue From" time from 1:00 (the end of the clip) to an earlier time like 00:50. If the audio at 00:50 is now cutting off words in Clip 2 from that last 10s, then add the missing words into the Clip 3 prompt:
(a few last words from Clip 2)
My Clip 3 Lyrics
If the verse is 100% the same as one that is sung before, it can help to tag them like [verse 3] and then reference [verse 3] again if you repeat it. Won't help now but in future.
The original 2 minute song is full and perfect except one line in Verse 2 (out of 5). I clicked continue and added verse 2 to be resunged (is that a word) and chorus one more time and instrumental at the end. I set continue at end of verse 5 01:49
That will add to the end of the song. If you want to generate Verse 2 unfortunately the only way is to back all the way up to the start of Verse 2. This will mean the rest the song can come out pretty different though.
Unless you are just trying to generate the audio of Verse 2, and plan on editing it in post production? In that case, you may be able to do what you want by choosing the right continue time.
OMG... this was some nightmare now.... I clicked continue and it generated another 2 minute song but the beginning was kids singing like a horror story like 2006 Eurovision Song Contest Winner but then in the middle in continued in the same style as original song and I got two correct verse.
Here is an answer:
https://app.suno.ai/song/0acd68e5-1036-4af7-b16d-2bdfaa16e5e8/
Pop, glitch, rock, Americana, indie, Rococó classic, EDM song. Listen and make your own with Suno.
So that means I have to listen to all 2 minutes when I click to continue because the start could be weird howling and then it goes back to the original tone of the main song.
Some of my best songs are from lazy prompts 🙂
@muted flame is there a way to retrieve the lost songs or is this already being worked on?
DM me if still happening - sometimes you can get logged into another discord account if you have 2,
Thanks a ton for working on that—it was affecting ~30% of my songs. My generations from yesterday are still clipped/bugged though, if there’s no way to get those back a 500 or even 250 credit top-off would be much appreciated since none of those generations are useable.
Here’s everyone I know of that was encountering the bug in this channel since yesterday:
- @wanton dawn
- @rancid tapir
- @proven hazel
- @prisma void
- @hardy abyss
- @neon ivy
When I try to "Get full song" I only get this error message https://app.suno.ai/song/664195ed-093c-421b-bf2b-9d742f23ba62 :
Id really appreciate help to get the full song since I am very happy with it, feel free to DM me.
The servers seem to be glitching out..I didn't realize at first cause I was making glitch tracks. XD
I LOVE THE GLITCH TRACKS I think it's interleaving two different clips, sounds quite interesting. Slow database, maybe re-using something that should be a unique key. Not a bad vibe for 'wizard opera' style though!
https://app.suno.ai/song/30f8f8db-17da-4b91-90c1-f5366d65c49c/
(I'm having fun with the glitched gens but it is a pretty signficant bug at the moment, seeminly high-server-load related.)
binaural-wizard-opera, anti-anti-anti, fingerpicked vocals song. Listen and make your own with Suno.
Here's mine for this morning, very glitch. https://app.suno.ai/song/ce06e0a4-c979-4dc1-bf3e-b1e1224bcaf8
Neurofunk, Darkstep, Glitch, Eerie, Futuristic, Experimental, Ambient, Drum and Bass, Industrial song. Listen and make your own with Suno.
Woah those are weird, sounds like it’s randomly switching between 2 completely different songs every few seconds
saxophone, emotional depth, Synthwave, retro, electric, epic, drums, bass, female vocals, electronica,
song. Listen and make your own with Suno.
@mortal tide We've invented a new Genre. I call it..Splitbrained.
Suno Glitchwave here
1 generation, no continues, 3:22 long.
At first I was stoked, then I realized the song kept restarting and acting like a CD skipping. It was unusable.
Not sure if y'all are aware of this happening.
https://app.suno.ai/song/63b6d088-957c-4f78-b2c7-625b6d78b18b
Rock, Reggae, Dub, Trippy, Ethereal, Groovy, Major Key, Stratocaster, Hammond B3, Live Drummer song. Listen and make your own with Suno.
We able to fix this bug?
Im not a staff
TZhen why reply lol
Cause you asked me lol
No worries, as of now it is not fixed and none from staff has said anything about working on it. Hopefully it means they are busy working on it
Geez, these bugs yesterday and today are major breaking bugs, Suno should consider giving refunds, pausing generations, or at least warning us that generations are broken so we don’t use our credits while they work on it
I expect some kind of refunds, it has been major daily issues now
Adding some more examples for this major glitch bug
https://app.suno.ai/song/8e530455-b326-423a-b0db-243d0a58335f/
https://app.suno.ai/song/cae681ea-9d03-4126-82b0-614e28e8cdcd/
Female vocoder robot voice. German Synthesizer Pioneer, experimental avantgarde, minimalist, electronic, robotic, punk song. Listen and make your own with Suno.
Suno has become sentient and as a result a singularity has formed, disrupting time and space.
Freshly subscribed today and nothing works, seems to be severely overloaded. All creations from 5pm (German time) are covered with errors and other overlays from other foreign productions.
We'll figure something out here (working on it at the moment), and thanks for collating!
Sorry, not the usual experience - we are actively fixing this
Do you know if a song that gets this error when clicking "get whole song" is gonna work after yall have fixed the current server issues or is it dead forever?
can you DM me the link to clip that is having this issue and i'll forward it?
Yeap on my way, tyvm!
is the load effecting song continuations too? they're rewriting the song completely (but its the first time im using the feature so im not sure what to expect)
if rewriting, that sounds more like a quirk of song continuation - how drastically are they rewriting? does the song you're continuing end in silence? if so may need to use the "continue from time"
it's a drastic rewrite of style, trying to continue from 30 seconds in
that said, it's going from a drum solo to electronic music
I just don't know what to expect. is it supposed to preserve the music it's continuing from completely?
or does it kinda fuse the two
actually a bit of both. it can certainly feel a little more experimental. also is this a later part of a continuation chain, aka this is like part 3 or 4?
no this is part 2
the music styles I'm giving it are really different though, currently experimenting with continuations to see what they'll do
Guys, are you experiencing some weird fluttering/ flanged/phased artifacts on suno's custom mode ( vocal + instrumental)?
Usually it's more the opposite, the style prompt in a continued clip rarely shifts the music genre except in a few situations. One is the clip is very short (or continue from time is close to start) and another is if the clip sounds like the song is ending, in which case it's more like the model is treating it like a new song is starting in the audio and that's why it's not consistent with what came before (if this is happening, then backup up the continue from time can avoid it and keep it consistent) Also there are some server-load related glitches at the moment which could be confusing things.
it must be a glitch then. I start with an instrumental drum song and continue it from 30 seconds with lyrics, but the lyrics still start immediately in the new song
unless.. wait, do i need to mix the parts together myself afterwards?
Does the drum solo not sound like it could have been a song introduction to the later part of the song with lyrics? No, to fuse them, on the last clip you click ... and then "Get Whole Song"
haha, we are thinking about some better User Experience here
(Forgot To Take My Pills mix)
Publish it, and you'll be surprised.
Btw, I'm getting those 03:40 tracks time after time too. Seems like 3:40 is hardcoded limit.
From what I can tell if you continue a glitched (interleaved) clip the song is only continuing from one of the two clips, so they can't generate music where the model tries to make something coherent from the rapidly switching song audio as the previous clip.
any follow up to this? I paid for Suno when it was v3a and I was extremely impressed by the results. the exact same prompts in v3 are nowhere near the level of quality that v3a had and i'm hesitant to continue paying for the service. is this an issue that has been acknowledged? many others are echoing the same sentiment. there is a clear noticeable degradement between v3a and v3.
I have been experiencing issues with "continuing from this song". It's cutting in halfway through the verse and now it's hallucinating. Also, I keep continually getting errors and a message saying, "credits refunded"
I wasn't having these issues yesterday
I found something strange ,I think is a bug . For example ,classic piano on start sounds ok, after 1 minut ...sound like a chorus, honky tonky piano . some like chorus/distort and out of tune (like 2 strings,and one is out of tune) . A...and on saxophone the same problem. From a tenor sax ...becomes pack of 2 sax ....and one is out of tune 🙂
Is there a leading instrument coming to the fore in the meantime? If so, it most likely has a reverb or chorus effect on it, and this affects the entire subsequent mix - as if Suno calculated the sum of all instruments and their effects and based on this it created a continuation of music... I have a similar situation with guitar music - as long as there are only rhythm guitars playing, it's ok. As soon as the lead guitar or those damn backing vocals appear, everything breaks down.
I did not have any vocals ....so..is not from vocals .
Maybe some other lead instrument? Even a violin playing a theme can ruin a mix if it has effects on it...
no...all instruments are the same like on begining . anyway...I only announce is a bug there .....for what I pay ...is too well. I`m happy with results . Is amazing 🙂
It doesn't matter whether they are there from the beginning or come later. Listen to the effects.
Man....Im a sound engineer and professional musician, I know what I hear ....no matter . Sorry I ask, you are admin here ? :) I dont want to disturb an admin :)))
I see. Well, unfortunately, yes, Suno has this error... You have to wait for an update.
No, i am not admin:)
I'm just hiperactive 😂
A..good :)hahahaha
Hey SoundAndMIDI! Sounds like you've stumbled onto a quirky bug with the classic piano and sax sounds. It's super helpful to know about these kinds of issues. Thanks for pointing it out, and it's great to hear you're happy with Suno overall! And no worries, DzaQ713 isn't an admin, just a fellow Suno enthusiast. Your input is valuable here!
Yes, I'm sorry. Sometimes I forget that my "advice" doesn't have to be helpful, and sometimes it's just the opposite...
Dza is funny 🙂 And yes...is crazy Suno .... I "eat" all credits in 3-4 days :))))))
Don`t worry ..thank you for help.... 😉
Just send an email to billing@suno.ai. They will make it right for you.
Sorry ...do what ?
To get some compensation for lost credits due to the bug
a..no .....thanks . I`m ok with that , if can be fixed..will be super ....this is more important 🙂
Hi. Where do I find tech support for Suno? The issue is when I write my own lyrics and press "enter" it creates the song rather than creating a new paragraph. I'm on Mac.
Yesterday it worked fine.
Never mind I fixed it
Question: is the New section on the Website working right now? I have the same songs since the morning. Nothing seems to change
Every afternoon the servers are overloaded. Devs can either cut off free users and give the servers the power to run, or watch everything go to hell...
Someone wrote earlier that Suno became a victim of its popularity...
ok ❤️
the quality has dropped off a cliff compared to v3a. is there any way to use v3a again instead of v3, especially for paying subscribers? v3 is not the product I paid to use
We all raise the same prayers to the heavens. I hope they will be heard, but for now we are stuck with V3.
Hmm.. not sure if a bug technically but if you buy a subscription you lose the 50 credits. I am not talking about the top-up i am talking about the 50 one had. I expected to have 10.050 after buying but i got 10.000 so to be pedantic i only got 9950 for $30 😄
The full song lasted 1.07 minutes and I spent 500 credits or close to it.
I think there are bugs in the way the tracks continue and merge and give the long version. When will it be fixed ☝️
I am truly baffled by how bad v3 is compared to v3a, and the responses to user feedback
really sad and unfortunate. complete step backwards
I know...
hi guys, I can't pay for subscription. Tried 3 different cards.
you could mail billing@suno.ai
Copilot Pro - Suno Plugin says it has been submitted. The song will be available shortly, and you’ll be able to enjoy this unique mashup. Thank you for your patience!
But hours later nothing ever appears or anyway to checkup on it
That's what's happening for 3 days already. Noise buildup after 1 minute.
aaa..I understend ....
every song I try end up a choir, metal is almost impossible to get as a solo singer, hope this will be addressed
Pretty clear example of this distortion that I'm getting on vocals after the first 30secs or so. In this case it is a single instrumental track where the guitar is pretty clean at the start, but if you skip through the song it gets more and more distorted in an unpleasant way. Which I think is what is happening to vocals, but with vocals it comes across as more of a "chorus" or "multiple singer" effect that increases throughout the song (and is also unpleasant for the genres I am creating in).
I've only noticed this in the last few days, and it is making all my chirps unusable so I hope there is a fix in the near future.
Plus the consensus here seems to be that V3a was better. So hoping at least that they give the option to use the alpha test one instead for now.
(However, that said, Suno is awesome and I'm sure is under a load of stress right now due to media interest)
Lets hope we don't have to wate till V4a for this to be resolved.
https://app.suno.ai/song/b6ca13b7-8eab-490e-a9e4-e17dd4e64deb/
Swamp delta blues slide guitar mississippi song. Listen and make your own with Suno.
copying from the suggestions post i made as this also seems it could be labeled a bug:
" I wanted to let you know I found a way to include artist names. Can't put Avicii -- but you can put "Avicii's" (i.e "Avicii's signature style) unsure if this useful but wanted to share so this can be fixed."
And when you put it, do you get something that resembles that artist's music?
i think so -- https://youtu.be/II6D20Cvg2U?si=3e_gQGFB7Fi-VFKO
https://youtu.be/Ks648PGlvyo?si=plpHcaeAd_gUF8Wl
similar drum fills, composition style, melody style, especially leads. the drum fills and hits are where it sounds the closest. when I do not include "Avicii's" it does not sound like this (exact drum fills, tempo similarity, low pass style, snaps style, explosion hits, etc)
open to hearing other's experience on this.
Maybe I'm just not into this kind of music, but I don't see it to be that similar. I wouldn't think that AI track was produced by the same people.
I have listened to a lot of Avicii. I also tried with and without the phrase "Avicii's signature sound" and the difference seems obvious (to me), but also i am only 1 data point and i have bias. if others can replicate with various artists, then it seems it would be a bug/issue to fix. if not, then no worries 🙂
Seriously, no. I'm sure they won't sue over this.
Not to my knowledge. Have you tried a different browser?
and also trying to write lyrics, pressing enter makes it create rather than going to a new line. I'm not sure why the website is buggy for me today.
seems to work on mobile oddly enough but not on my PC
Got the feeling that the sound quality of nearly all stems is pretty bad today - https://app.suno.ai/song/8bd605ef-f3fb-4b13-9f9f-9beb17b262bc
contemporary electronic alternative rock hip-hop song. Listen and make your own with Suno.
I'm not sure if this issue is specific to my browser or laptop, but occasionally I lose parts or entire songs generated in the library and generation page when I come back to them the next day.
The only workaround I've found is to complete the session on the same day to avoid losing any progress. Has anyone else experiece this issue?
And it randomly selects which parts are deleted, making it difficult to follow a pattern for restarting the generation process.
Ai can mispronounce "Aye".
okay, I didn't have custom mode on. that was the issue.
half asleep and I didn't notice 
certain genres even if you write "male singer" it will give you a female singer at times
with "ethereal fantasy" i notice like half of the time it is doing this
is anything being done about the constant doubling, tripling, quadrupling of vocals throughout nearly every song? This has been going on for several days, at least in the rock genres. Every song sound like 1990's The Cure.
I've found that if you type 'male vocal' in the music style box at the beginning, you'll get what you want 99% of the time.
that's really useful thanks for the tip!
i didnt receive 50 credit refilled after one day
There is many major accessibility issues with screen readers on the Suno web page. Also the "more actions" menu seems to change very frequently depending on where I am on the site and sometimes doesn't even display the options it's supposed to.
honestly
this is a serious issue that will cost Sudo and the team my sub...
was 3a really that much better? i hope they reinstate it as an option so i can try it. how was it different to the public v3?
with my experience, it was clearer vocals, more refined audio, and better/intros for hip-hop type songs.
I'm not a music-producer, I just got into this like a day ago, but the vocals were more "on-beat" if that makes any sense? it had a better understanding of what to stress when singing and what not.
the stuff that v3 puts out now is just kind of meh, the vocals are all over the place and the beats arent nearly as smooth or clear
i haven't seen what 3a could do but i've been incredibly impressed by 3 in general. did 3a avoid that thing where it sounds like an entire room full of people are singing? i've noticed issues with getting it to sing at the right cadence, stressing lines at the right point, but i had that with v2 as well, and been able to navigate it with the outstanding continue from x point feature, been a godsend to just be able to cut an otherwise good generation at the point where the vocals screw up and continue from there
not to mention instrumental doesn't work 😒
Yes, I'm going to have to cancel my sbscription that renews tomorrow until v3.1 comes out or they give us optio to use v3a
Yes, and V3 adds a lot of noise to instruments too, like if there is electric guitar distortion it ruins anything that comes up after it
I do a lot of EDM and power metal and the vocals AND instruments sound very muffled, someitimes even at the beginning
i suppose i haven't noticed because i've been intentionally generating stuff with a gritty feel, i didn't notice haha. well hopefully they re-add 3a, a lot of people have been saying it's better
yeah, if you want clean, it muffles TOO much, it's unusable. I'm unsubscribing, as mine was set to renew tomorrow. What a shame. Hope they fix ASAP
I didn't earn any new points today.
lately my songs sound like more than 1 voice singing, like it sounds like several voices singing, how can i stop it and make it sing just one voice
is there any way to stop it from automatically WASTING credits when i log in ! ?
Whenevever i log in, it just generates random suggestions wasting 10 credits
please add your browser details, OS, and device so it can be fixed.
edge \ windows / laptop ?
Seems to be a fair bit of Audio Defects on very Low Bass Driven Sounds.
Example (towards the end mainly): https://app.suno.ai/song/9abe419b-69b7-4d6e-aadf-c9de6f2443c5/
EDM, Slow Start, Sorrow, Starry Sky, Cool Breeze, Vast, Ascension, Limitless, Hopeful song. Listen and make your own with Suno.
someone knows how can i make the voice to stop sounding like double, it keeps on giving me songs with like a choir singing instead of jsut 1 voice
cancel subscription till devs fix it 😦 And/or experiment by Continuing song from before the chorus/muffle comes in, and putting [CLEAN VOCALS].. hit or miss
Found error in 2 songs now with different music style where it sings lie instead of life
is it normal that it keeps saying: couldn't generate that. Prompt flagged for moderation. for absolutely no reason when trying to continue a song? And that also no matter what, when I continue a song it cuts at 1 minutes?
i think it should say why it couldn't generate. throwing a non-descript error message like that seems like a poor design to me.
@idle shadow i also just checked in here to see what's going on about the doubled voices. i'm starting to miss v3alpha for that reason.
Would be awesome if it said why, just spent 30 minutes rewriting the same thing trying to get it to work.
Copy paste it into chatgpt and it will tell you why
Chatgpt is what they use for the moderation
The generation only does 2 mniutes on first and 1 minute on the rest of the parts, so if you want to make a 6 min song you have to continue it several times then at the end choose "Get whole song"
Thing is, I used chat gpt to help me so that's why I was really confused that it gave me that error. But oh well after rewriting for a while I finally made it work.
Have the developers put out a statement as to why V3 Alpha was better than V3 yet?
Devs say V3 is better and thats it. Meanwhile the degression continues and songs get worse each day. Since Yesterday i get only distortion and reverb in all my songs
Did they really say it is better? That is hilarious
Yes, there is a Statement somewhere here on The Discord
Important to note that the past couple days have been filled with a slew of breaking bugs that they’ve been working on, so some of what you’re seeing isn’t intentional behavior
We are just wrong 🤷
I have been waiting for them to fix it before I start throwing accusations towards them. But if they say its better then I will not hold back lol
Its not
night and day difference btw v3a and v3 for sure
Ya its not even close
I've pretty much stopped using Suno as v3 is just destroying everything I try to do. Even this Discord has lost so many of the people who were in here daily for months just enjoying Suno, but v3 has driven them away. I still have hope, but I just check in every now and then to see if they've fixed it.
i'm holding back on using this month's credits, hoping they'll fix the "voice always turning into a choir" problem
i'm sure they must have noticed.
it happens 90% of the time
They have and they're working on it, but it's taken a long time and they don't give out much info.
good to hear!
Some people say it's a prompting issue but it's not.
it isn't.
It's B on the new poll #🗳┃mini-polls message
Development takes time. Just give them some time. Maybe they will also refund some credits. I hope they will find a way to produce the same quality of V3A or better, but with less server costs on their end.
Hello, I've tried to delete already created music on Suno without success. I have tried even permanently deleting the files and they keep on appearing both in the trash section, create section, and library section. Is this a bug?
Also, not sure what's causing the server overload all the time... are people scripting heavy usage or something?
the server's log files should give a hint about that
maybe it would help to create short preview (15-30sec) snippets of a song, before creating the full 2 Minutes to reduce server usage. This way I can easily check, if the preview (voice, melody, rhythm etc.) matches the style I want. There is lots of potential to reduce it, as there are lots of productions for trash
i think the worst source of such "waste" (useless generations) is probably the current "choir instead of single voice" problem. so many generations are useless because of that.
maybe the snippets then only cost 2 credits instead of 5. and then create the full 2mins of a snippet I like
Hmm then im not sure sorry, maybe copyright issue?
But it should prob have said that if so was the case
i think it should always show what the cause is.
so that people have a fair chance of fixing their mistake.
I agree. V3 Alpha to V3 has rendered it useless for me as I use it only for instrumentals and It has started adding these voices to nearly every one of the "good" beats it gives me.
Distortion and / or bad sound quality whenever I make a part 2 or 3
for the, losing the voices are the saddest part. in many cases i can recreate the rest with real instruments, but i just can't sing.
I feel your pain my issue is the other way around lol. They need to fix this. Feel like I got ripped off (I bought a year of Premier). Thought Alpha was good so this should be better
Volume 8 dropped -1275 plays already
Terrible, tbh.
Well, that explains why it's in Bugs.
why can't I edit a song when i remix it? (the lyrics)
Remix is only to paste the lyrics and style of music into a new song to create
Sort of "re-use prompt"
the Continue feature seems to be bugged for me the past 2-3 days. It just ends up skipping verses altogether. For example if I were to prompt a [Chorus][Bridge][Chorus] it would straight up skip the chorus before and after and sing the bridge only, it did this even when i renamed the metatag to Verse. For fun and wasting my credits I tried to continue and just copy paste the entire text front to back and results were:
- Sing only the bridge (even though it had ended in the original generation) and fill the the other parts of Part 2, before and after the sung part with just instrumentals, no other lyrics.
- Would just mumble random nonsense (not even a language), sing the bridge (again) and end
i think my main issue with v3 is the phaseer that is in the background on alot of my generations, i got it alot with v2 v3a was by far the best for having this going on v3 does it all the time when creating rock song it makes it sound mechanical and is un usable, i have created 2 albums so far using v2-v3a that i love but album 3 is taking forever due to having to start again with songs sounding great for the first 2-3 mins butr towards the end of the songs it just gets muddy and phaser sounding guitars. very frustrating, knowing it can do what i want it to as has done in the past
Yes, it became very common. I think it was in the poll of priorities, meaning it's a recognized problem. https://discordapp.com/channels/1069381916492562582/1163842729231528047
I find that the sound is too saturated and the voices are too noisy, can you do something about it? on rap or hip hop it's unusable thank you.
singer voice sex error when continuing. pt1 female pt 2 male. its wasted so many generation making pt 2 female instead of male as instructed. 6 generations and only 1 was male.
we should be able to report them as errors and get our credits back
im lost 20 generation because sounds is nos clear and noisy
try adding "high standards, high quality sound" to your prompt
It's not just continuing either eg: https://app.suno.ai/song/94a49497-db64-4225-abdf-da3de0d3a5aa yeh, totally a female voice
death metal, female singer named susan song. Listen and make your own with Suno.
in the "style of music" field or as a control tag in the lyrics section?
in the style, it seems to ignore the control tags when i try if control tags are like [voice 1] or [voice one female]
thanks! (trying right away)
Hello. I can´t continue or remix any song I generated in suno. It just generates a few seconds of silence(2 times :D). I am using Opera browser.
hmm... can't tell the difference tbh.
Try setting the "continue from" time to a few seconds earlier. It might be trying to continue from what it thinks is the end
Though, I'm surprised if this affects multiple songs
I don't know why but I always get that, I can't log in
@weary steppe same
I don't know what's up either.. Devs are looking into it.. but if you have a VPN, that seems to be a workaround for now
I was using it daily and was looking forward to renewing today, but I tried it last night again and could barely make a song that didn't sound like it was recorded in a bathroom with an old phone... The melodies and all else were GREAT, but it's unusable 😦
Doesn´t help. In Chrome the same
Hi when I want to sign up in suno it tells me:
Sign up unsuccessful due to failed bot validation. Please refresh the page to try again or reach out to support for more assistance.
I tried from my laptop and my mobile phone, different browsers but nothing works for me
you using a VPN?
What is your location? I'm collecting data to identify any commonalities among users experiencing this issue.
The developers are currently investigating. In the meantime, a temporary solution is to use a VPN. If you don't have a VPN, I can suggest using https://chromewebstore.google.com/detail/free-vpn-for-chrome-vpn-p/majdfhpaihoncoakbjgbdhglocklcgno for the Chrome browser. It's free to use.
also adding @lavish sierra
What is your location? I'm collecting data to identify any commonalities among users experiencing this issue.
Oh yes, the compositions are improving all the time, but the sound quality is getting worse.
I am from Spain
Same, today i've lost a lot of credits trying, don't even know how many, and each one was worse than the other, vocals and melodies were trash
hi! I have a problem with the sign up in the web, It doesn’t let me register, It says: Sign up unsuccesful due to failed bot validation. Please refresh the page to try again or reach out to suppport for more assistance. So I refreshed the page a few times and it doesn’t work, what should I do?
I have the same problem
I tried in different browsers and anything
same
This chorusing effect as a song continues is cray-cray, it’s very difficult to have a single singer throughout a song. I know everyone else has been saying this, just wanted to add my voice that having it fixed is a priority for me.
anyone have a solution?
What is your location? I'm collecting data to identify any commonalities among users experiencing this issue.
The developers are currently investigating. In the meantime, a temporary solution is to use a VPN. If you don't have a VPN, I can suggest using https://chromewebstore.google.com/detail/free-vpn-for-chrome-vpn-p/majdfhpaihoncoakbjgbdhglocklcgno for the Chrome browser. It's free to use.
I am from spain
Me too and I have the same problem
Thank.. till now I only get people from Spain.. You too @final stirrup ?
Im from spain
Orange
Thank you @quick fossil and @final stirrup.. ATM it seems only Spain related. No clue why. The devs are investigating.. The only solution so far that works is using a VPN. I posted a free one above.
No
Ok, Thanks
Spain
Thank you.. what is your internet provider
Digi
I'm from spain and I have the same problem
Using the VPN was the unique solution, thank you!!
Okay.. as you might have read above.. it only happens to people from Spain. The Suno team is investigating but atm the only solution is using a VPN. If you don't have a VPN, I can suggest using https://veepn.com/vpn-apps/vpn-for-chrome/ for the Chrome browser. It's free to use.
I have to pay?
still no better today, the guitar sound is terrible, muddy, and sound mechanical, terrible will try again later.
souunds ok til 30 second in then shocking
What is your location? I'm collecting data to identify any commonalities among users experiencing this issue.
The developers are currently investigating. In the meantime, a temporary solution is to use a VPN. If you don't have a VPN, I can suggest using https://chromewebstore.google.com/detail/free-vpn-for-chrome-vpn-p/majdfhpaihoncoakbjgbdhglocklcgno for the Chrome browser. It's free to use.
Why i get Prompt flagged for moderation for "Arschloch/Asshole".... Should I only write children's songs or what please change this for pro users...
THIS should be the priority for the Devs rather than letting new subscribers login. We're already paying subscribers and the lack of support lead me to NOT renew this month.
YES!
Above all, the Free User can do almost as much as the Pro functions.
I also think they should remove Prompt flagged for moderation, you have no freedom with it even though you pay for it.
Spain here
Thanks, but I'll wait for them to fix it 🙂
So like all the others so far.. Keep in mind.. it could be also on your end (spain related). But Suno is looking into it.
It's stopping generating with plenty of time to go and only half way through the lyrics
https://app.suno.ai/song/6d32a8f9-7bb1-4322-b88b-3a22aa58ce4c this only chopped off the last sentence but its only 55 seconds.
K-pop, mutation funk, bounce drop, dubstep, edm, 160bpm,, high standards, high quality sound, female voice song. Listen and make your own with Suno.
I know its in alpha so bugs are expected but we are paying for this. they need a system to report failed generations on the website and refund credits when its the generation at fault
I'm waiting till CONTINUE will finally reach the end of my lyrics. Is there a RELEASE NOTES page where I can see if a new release has been issued ( and maybe the NOTES will mention CONTINUE so I can try again) ?
Sadly no..
Oh well -- I can wait for its too "brilliant" even as is ! ( Dreams of Beta)
I've been trying to bring this to attention for a few days now. Just compare the riff from the first seconds with the one from 1:20 to hear how the wide stereo Rhythm guitars disappeared. The bass becomes the main instrument and everything else becomes muddy.
The cutting songs short is a big problem. most of these songs in this list (47 tracks long, or 470 credits) do actually finish the lyrics (barely) but are cut way shorter than the 2 mins we pay for. https://app.suno.ai/playlist/8d2d280e-b310-4001-b1e1-a4a0e66134c2 its forcing us to use more credits to finish the song when it had plenty of room to do so.
A lot of Bugs seem squashed this morning! I was able to get two great songs today that are pretty clean! One was EDM, the other a Hip Hop counting song. Well done, devs, thanks for listening to us and helping! Onward!
I cant sing up with any of the three options, It is saying me this all the time: "unsuccessful due to failed bot validation"
rt
lit me pasa eso
creo q es pq somos d españa
puto illojuan tiene el sv petao
yaaa que cabron es que esta guapo
m quiero hacer un tema de break beat a ver si deja
ya a ver si va
avisame si tal porfa
I have a problem, when I generate the songs they usually come out in silence, without music, but they take away the credits.
Did u use CONTINUE ? And your own lyrics?
Yes, always my own lyrics. Always continue. Always GET ENTIRE SONG and then continue.
Hey support, I have a pro plan and yesterday, the servers were quite overloaded and ChatGPT servers went down. This lead to spending a lot of credits for really bad generations. I message support but had no email back yet. Can we get refunded credits if the servers go down as thats now all my credits gone and over 50 generations that didn't quite work lol x Thanks ^.^
You get entire song in the first 2 minutes? And if so what does subsequent CONTINE do ? ( I expect the 1st 2 mins to cover about 2/3 of my lyrics and then hope for CONTINUE to start where it left off and complete to end of lyrics, Do you do that too?)
To get the best result, in prompt, Use "[Introduction]" then "Continue from this song". Find the time the intro ends, then add [Verse 1]. once thats done, select the one you like and again, "Continue from this song, [Bridge] [chorus]. I got a very very strong track that I can now take to my producers (I use it for melody ideas) ^.^
WHen trying to "Get Whole Song" on some songs I get an error. looks like the API is returning None.mp3 as an audio file
My latest song this morning have 4 continuations. The best result is always in the first try. Up to 2 minutes if you're lucky. Then, it's about 1 minute continues each. Get the FULL SONG after each continuation and continue from the FULL SONG and add to it.
but on other songs it seems to work
yo si me entero te aviso
That happens sometimes Akiselev. You can select it, "Report Bug". I had alot of that hence wanting some credit refunds 😦 I think the servers sometimes give blank outcomes as could be the traffic on them ^.^
So u dont start with ALL the lyrics ( like I do) ?
You can, but it won't give you as powerful result as breaking it down in to parts 🙂
You should be prompting each section:
[Introduction] - Most songs start with a short 5-15s intro. You can ask it to ((adlib)) ((spoken))
[Verse 1] - Add your lyrics. Make sure the syllables match up. You can use [Pause] between breaks if you don't want the AI to sing certain parts to quick
[Bridge] - You can also use [Bridge] & [Build] for EDM songs
[Chorus] - You can also use [Pre Bridge] & [Bridge] for EDM
Giving away the secrets now lol
No, putting the whole song in as a first go is asking for trouble. I usually go a verse and chorus and verse for a regular song for the first part only. Then it's only a verse and a chorus. Or, for the first part, verse, chorus, verse, bridge, if it's a quick Rap or Hip Hop song. If the song starts to repeat in the first part, or just continues with music, I "continue from" the last lyric I provided, not what was hallucinated and then continue from there.
And, if you want the AI to each a certain part in a bridge eg Let's get back togther - You can use "Let's get back together (Let's get back togther)" itll add a each effect on the end of the vocal xD
echo**
Oh and do not put grammar at the end of sentances. Always use [Meta] to determine pauses etc
Thanks - I'll try that ( of course the Devs have designed it to work with ALL lyrics at start and no doubt it will work soon) PS with [bridge] after a chorus -- invariably it repeats the chorus' last 2 lines. And [Big Finish] with nothing else has never disappointed. ( most STYLES/genres )
[Outro] [Instumental Drop] [Instrumental Break]
AI Won't understand 'Big Finish'
Or use [Instrumental Outro] [End Song]
Or try [Build] [Outro] [End Song]
How do I end the song? https://www.suno.wiki/faq/making-music/how-do-i-end-the-song/
[End Song] lol
Good! It would be great to have a song done in one shot. Go back and cut out the incomplete chorus by continuing before it begins. That seems to be the fix now. It works if you don't try to continue half the lyric. Just start it over and continue. I just use [End] and it works fine.
Hey guys, anyone knows why it wont let me sign up?
it appears failed bot validation
Did you try sign up with discord?
I used discord to sign up and it worked fine
yes I tried everything
have you tried clearing cache/cookies?
ya me ha dejado
no, imma try to do that
If you dont try [Big Finish] -- youre really missing out imo
If you heard the 6 songs I've created, I'd say otherwise 😉 lol
I call your 6 a raise you 21 !!
You what? 🤣
gracias tio suerte
anyone have luck generating past three parts? parts 4/5 aren't working anywhere near as well as parts 1-3
a mi no me va jajaja
joder
Are you on free or pro? I haven't had any issues but sometimes it has gave me a whole new song from "continuing" lol
Pro
Tons of "Credits Refunded"
I wish mine did lol
ya me ira disfrutalo tu jajaja
same
The site has a problem, you can bypass tokens using twinkies, that is, other accounts. How should I react to this?
You can also use the Remix function, with it you can use less than 10 credits, that is, 5
this is not universal. Sometimes I max out the context window with great results
it really just depends
Hello. Increasing the length of the song the quality at the very end decreasing a lot https://cdn1.suno.ai/447a1b7a-8afa-4bf9-aa73-0a3034b3459c.mp3
I can't sign up with any option, it's always telling me it's because of failed bot validation
smallish bug. When clicking on the Add To Playlist from the Song Preview in the right sidebar it doesnt open the playlist modal. I am using Chrome on OSx
What is your location? I'm collecting data to identify any commonalities among users experiencing this issue.
The developers are currently investigating. In the meantime, a temporary solution is to use a VPN. If you don't have a VPN, I can suggest using https://protonvpn.com/free-vpn. It's free and safe to use.
I find a song I like on the website, and i click Remix. I try to remix the song and i get one of two errors (An Error occurred) or (Prompt flagged for manual review)
Had that too last week. Keep on trying every 30-60 min. or so, most likely the problem resolves itself.
Longest generation ever!!! over 6 minutes!!
New record holder 😄
all garbages unfortunately
you cannot win all 🤷♀️
It looks like that there is a bug with "continue from this song". After a few tries (and wasted credits again) - in many cases instead of getting a second part with continue from 00:35 I got a song-part with just 35 seconds length with a hard cut at the chosen "continue" time.
Here are your choices - Latin: Bossa Nova,,Latin Jazz, Forró, Mambo, Salsa, Tango https://www.suno.wiki/faq/style-and-lyrics/styles-and-genres/ Dont add anything and let SUNO do the magic ( my experience is more success simply choosing a genre and nothing else, after I add my own lyrics )
I was using that... but it was my experimenting too...
I was getting great beat and results... but not so good voices... so I keep generating... but then one time, it hit me with that 6 minutes with lots of stuff switching in the clip... sounds funny!
I'm from Spain. I've also seen that there is a lot of traffic from there, so it might be because of that. And yeah, I'm already using a VPN
Ah.. which location is your VPN? If you can choose.. I would recommend a country that is not "spain"
Asia
We don't know. The developers are currently investigating. In the meantime, a temporary solution is to use a VPN. If you don't have a VPN, I can suggest using https://protonvpn.com/free-vpn. It's free and safe to use.
You could try another country like the USA.. that should work 🤞
it worked for me! Thanks
I’m experiencing this too. It works when clicking directly in the song list though
repeated Vocal hallucination of "It's coming for everyone": https://app.suno.ai/song/818d3be2-94e0-4243-9de4-aedfb65a836d It was a continuation of https://app.suno.ai/song/67ad41d4-e118-4e7e-91ef-95476917d0e9
Hi, I have this error when I thry to sign in "Sign up unsuccessful due to failed bot validation. Please refresh the page to try again or reach out to support for more assistance." Someone can help me?
I have a problem that when I go to do the second part of the song, the A.I. starts in the middle of the verse or skips a verse. Does anyone know how to solve it ?
Just try again... you can't really solve it.
is this a bug? Usually 1x generate results in getting 2 music but how is it possible that there are still 5 credits left?
then you had a refund on one of your clips because it was too short
not really, move verses around, only do that specific verse, or "continue from" different seconds, like 1 or 2 seconds before or after
Any updates on the Devs on whether the muffled/chorus/reverb sound is going to be fixed soon, or if in the meantime we can get the option to use V2, V3a, and V3? If we can go back and use V2, why can't V3a be usable?
I'll definitely subscribe once that happens or gets fixed, but right now, there is no point on keeping the subscription 😦
Thanks in advanced!
REMIX button is not working very well and never remix with same voice / style
Remix is more like a "reuse" feature. It will copy the prompt (music style and lyrics, etc.) and title. But every time you hit Generate, it will create something new
