#🪲┃bugs-web
1 messages · Page 16 of 1
Stuff is borked today, what happened? Every time I try to reuse a prompt it messes stuff up and doesn’t add the prompt, but then nothing works, not even new songs, had to try to continue from something else and then clear it to get it working again
I clicked Generate and Suno generated 4 versions all at once. All are unusable. 20 credits gone in an instant. Give me back 10 credits at least.
On mobile interface it seems that reuse prompt does not function properly. In addition anything you have in the lyrics and style fields are wiped out when navigating away from the create tab.
I checked on ‘instrumental’, still getting voices.
I am experiencing this too
we just pushed a fix for this issue, let me know if it seems to fix the problem you were seeing
Seems to work
great, thanks for checking
Hello, I would like to ask for help, every time I try to create a song the error appears "Couldn't generate that
Prompt flagged for moderation"
@left citrus not to antagonize, but from a fellow z- whats the business model of deleting my 7k credits when i pause my subscription-
or perhaps it a bug
This is usually an indication that there is something in the lyrics that ChatGPT doesn't like.
its really should be displayed or made transparent
You can usually get a hint by submitting the lyrics in a ChatGPT instance and asking it to point out what might be moderated about those lyrics.
Thanks
don't know offhand, but can look into it
just slightly frustrating and a turn off for the platform for me, just wanted to let you know..
understandable, sorry about that. will investigate
no worries, cheers to v4
HI, I am trying to extend the song that I created, but it gives a system error. Tried seveal times, but it doesnt work. Can you have a look? https://suno.com/song/e8247875-f8b4-4eef-9e49-57154b9e7938
indonesian language strong bass turkish influences handclaps dark male voice melodic progressive house indonesian male singer song. Listen and make your own with Suno.
TIL that starting a continuation with a "[Bridge]" marker has a high chance of the model interpreting that as "just skip this entire stanza and repeat the previous one instead" 😑 After many attempts, I changed it to "[Important Bridge]" and got 1 out of 4 generations to stop doing that... I don't know what the model learned in this context, but that's super frustrating. (Of course, it won't matter as much if v4 or v5 can do 4-5 minute generations in one shot 😁)
this is why i have not created a good song in over 6 weeks. i finished my third album with udio, i love my first and second album done with suno but cannot get anything good from suno which is why i still have 0ver 10 thousand creds left unused. this is because i wont release a distortion sound that sounds like its going through a washing machine. holding out for v4. god i hope it goes back to being good. i prefer the vocal and melodies of suno, but for now it looks like udio is the way forward til things get restored to suno's former glory
Hello! Can anyone help me? I thought that the words in square brackets [ ] were not supposed to be readable, but several songs and credits have already been wasted because the singers in my track sing them. I'm just learning how to work with Suno, so maybe I did something wrong, but it's frustrating to see credits being wasted like this... Here is example: https://suno.com/song/d1ad0fd7-e37c-4a87-9ea0-26b4170d8cea
Another problem is that Suno likes to cut off the lyrics at the beginning of the song, so the song starts and the singer immediately reads the lyrics, skipping half of the first line.
K-pop 2010s, boys band, 5 different male voices, korean, pop, hip-hop, emotional, catchy, emotionally deep, dance song. Listen and make your own with Suno.
they generally don't sing stuff in the brackets, it just happens at random sometimes. one time my 'singer' enthusiastically shouted "key change!" between two choruses, wasted generation but i died laughing. are you having the brackets problem consistently? hopefully it's just a one off. the song starting poorly i get a lot too, a lot of verse 1s just spoken, or sung without music. i found you can generally avoid the first verse getting started weirdly as long as before verse one you put something like this:
"[shouted]
one two three go!
[intro]
[bridge]
[verse 1]
verse 1 goes here"
then 9/10 it should sing verse 1 properly, as long as you put a single line of speech or song in front of it (i do this and then usually clip that line off using audacity)
The mobile UX for playlists seems broken on mobile (Chrome, Android). Duplicate entries, same song added to multiple playlist
Certain widths of windows can cause issues with overlapping lyrics with the rest of the site
Your month might be over? Credits do not roll over. They never did.
Credits in Suno's monthly plans are designed to be used within the specific billing period and do not roll over to the next month. This is because Suno allocates computing resources based on expected monthly usage, ensuring optimal service performance. Rolling over credits could lead to unpredictable demand, potentially causing service disruptions and increased operational costs. The non-rollover policy helps maintain a stable and efficient service for all users.
any clarification on this? I am getting the same problem.
seems to only happen with custom mode
I have a problem. I made a song, and it completely skips verse 2, the most important part of the song 😔 Do you had the same problem?
This is pretty normal (depending on "genre"). If you make an instrumental of classical music, you will not have voices, but if you pick a genre that normally have voices it will have a hard time not to sing. But since you didn't gave any lyrics, it will make them up, and it's called "hallucinations". It's the equivalent of ChatGPT making things up. It's really hard to avoid in some genres.
Try some Jazz, or LowFi, or classical.. things that are instrumental in reallife too.. and you will have really good results.
Thank you for the inside. That is very interesting. I have been getting a lot of luck with no voices too, but I was just curious why that was happening, now I do. I'll definetly give some classical music a try.
Other styles work too.. But yeah, sadly atm certain genres are really hard like rap, country, pop etc
And the worst is that.. i have made a lots of versions of that song, and this was perfect, exactly what i wanted 😂
And most important part of lyrics is missing
It means the AI "lost his way" in your song. What mostly works is continuing 1 or 2 seconds earlier.. Have you tried that?
I will try
One more thing, before verse 2 i have chorus, an chorus was repeated 2 time, and then he skips verse 2
You are in control were you want to continue. If you feel the 2nd chorus doesn't fit, you will continue after the first chorus. Just listen were it stops, and fill in the timestamp. Have you ever seen my video? It will explain a few things: https://youtu.be/lj8utBd_IUM
Hello, I'm 3Daizy, a moderator and guide at Suno. Welcome to my channel, where I share my journey and experiences with AI-powered music creation. Whether you're a budding musician curious about AI or an experienced artist looking to innovate, you'll find useful insights and practical advice right here.
Join me as we navigate the fascinating int...
Thank you so much. I didn't know how extend actually works. Thanks you.
is the thing where it starts out one voice and then like slowly multiplies until it sounds like several voices a big or what is that? Is it avoidable?
This model has a "choir effect" in some genres.. The next version should fix that,
Не работает бот
The main language is English. Please post your questions in English, or post them in a channel with your language here: https://discord.com/channels/1069381916492562582/1154772025940922438
Главный язык - английский. Пожалуйста, задавайте свои вопросы на английском языке или размещайте их в канале с вашим языком здесь.Бот исчез. Теперь вы можете создавать музыку только через веб-сайт: https://suno.com
Not sure if this is a bug or intended but when you extend a song now the "Style of Music" box always reverts back to the styles used in the first part of the song
Good catch!
I get this too, it ‘s useful for making consistent song phases- less for style variation. Are there any tips on strengthening the weight on the extend prompt?
Yeah I totally understand. I already got a response back that it will be fixed soon! @proven socket @gentle mirage
Over the past month, I have been creating music every day using the same method. However, today I encountered an issue where, despite my efforts, the songs I produced turned out to be in Cantonese instead of Mandarin, which was intended.
argh, suno how man credits do I have to waste each time to get an orchestra piece instead of a Wyndham Hill recording from, like. 1993?!?!?
have you tried again recently
we have some bug fixes that reduces the noise issue
this should be marketed, not just a nasty surprise
credis do rollover, unless you unsub
considering i had 7k on the regular pro plan, yeah i think its safe to say theres rollover
witholding the ammount of credits your user already paid for is ridiculous
"cant plan for it"
then dont allocate it?
Only top-up credits roll over. The monthly ones with the subscription don't / never have.
again, my seven thousand credits argue otherwise
It's on the website where you purchase them?
ok
its at the bottom
but that doesnt mean it is a good model
just voicing my distaste, @left citrus already has it handled, no need for parole, officer
In general people have a hard time reading. But I cannot help that. I am just a volunteer telling you what is general. And @left citrus knows I was handling this. He is not support desk. And is more versed in UI
Out of interest what numbers do you have here?
in general people are intentinally dense
same goes for TOS
Well you know how it works now, you know he reason why. If you have concerns about it you can send an email to support@suno.com
I am sure they can help you out.
A few months ago the site didn't have the info at the bottom and the wording on sub info was less clear so I'm glad they at least made changes to the info on the page.
The BIG BUTTON below says "buy more credits"
A lot of plans who work like that. MidJourney, phone plans, webhosting.. 🤷♀️
this is a z to z conversation, not a helpdesk inquiry, thanks
already disscussed, no need for recursive moderation
This is information you were lacking, and that is my job. To give you the information that is known.
I am not only a modator, I am a guide as well.
ok well I appreciate your cadence but it was more dev feedback then a genuine air of confusion
Sometimes extra credits are given e.g. as compensation for downtime or when v3 alpha was available they gave extra credits to help test it.
i cant extend my song anymore longer than like 2 seconds. Its a bug or whats going on? this has been going for a few hours now
I try every day.
I am a paying user of suno from mainland China and would like to seek some help:
Since I paid to become a Premier Plan user, the lyrics of the songs I generated with Chinese lyrics have all changed from Mandarin to Cantonese. How can I restore the pronunciation to Mandarin? This is very important to me.
After I paid to become a Premier Plan user, I did not enjoy the benefit of generating 10 songs at a time. Currently, I still generate 2 songs at a time. How should I solve this problem?
I am also a paying user, today I also had a similar issue where Mandarin turned into Cantonese, making it impossible to fully produce a Mandarin song.
Until today, I have always been able to successfully produce using the same method, but it didn't work today.
OMG
Finally someone understands what I'm saying
That's it, Mandarin all changed to Cantonese, just after I paid on the evening of April 26th
I hope people with similar experiences can share more information.
This has nothing to do with having paid for your account. This is an issue with all languages. And normal in the current state. To show you, here some history to see this is not new and it even works visa verse. Here some examples (and please LOOK at the dates too):
#💡┃suggestions-web message #💬┃general-chat message #💬┃general-chat message #💬┃general-chat message #🪲┃bugs-web message #🪲┃bugs-web message #💡┃suggestions-web message #🪲┃bugs-web message #💬┃general-chat message
I gathered some about Chinese language, but this happens to everyone in every language.
the content moderation seems to be hit or miss sometimes, like it often doesn't like me using the terms "baby boy/baby girl" when making R&B type stuff. I think for obvious reasons, but it seems like I get flagged regularly for using it in its acceptable form
Regarding generating 10 songs at a time, it's not meant to generate all of them with a single press of the Create button, if that's what you would expect. It just means you can press the Create button 5 times in a row, so that a total of 10 songs will be generated in parallel. On the free plan, only 2 songs can be generated in parallel.
"I hate piano, I don't want any piano. I did a prompt viking medieval cinematic score and guess what, a piano... How can I avoid a piano?"'
not got a real solution, but adding "cello string quartet" to prompt is on my list to test eventually - i believe v3 dataset probably has enough qartets this might help - just a best guess
It's fucked up. I tried to make Russian hardbass. I managed to make a good generation with some ‘donk bass’ but the rest is really poor, it sounds more like an electro deriver than hardbass. Pls do an update on this. I've burnt around 600 credit trying to remake similar song
Browser behaviour inconsistency: When generating a song, song length display is different in Firefox compared to other browsers (Brave in my case) - Firefox shows the current length of the song, other browsers show a spinning circle. It's not possible to adjust the playback position until generation has completed.
yo, it wont let me make another song for sum reason and idk why
whenn i create "Instrumental" it still put lyrics. anyone know how to fix
All my songs seem to be gone from my account. Only four pages of songs now, playlists and liked songs are missing. 😦
what happened to suno
Is anyone experiencing some issues right now? was working fine up until now but for the past hour when i'm extending a track, it doesn't matter what i write in the prompt the output is the same
Tried using melodic metatags or even check the "instrumental" option , still has has word and completley unrelated to anything
can't use the word "nerd" in the songs because there's a band called Nerd? 🙂
I too discovered there was an artist called ding dong ..🥲
Just found out "banners" is an artist name.. is this something new? 😮
I click Generate Lyrics, the text is created, but by default there are only 2 verses without a chorus. It creates [Verse 1], then it looks like [Verse 2] but nothing is written and then [Verse 2] without [Chorus].
you can edit the song and add the parts you want to, that would fix it the fastest i guess
Yes, I understand. It just looks like a bug, that's why I reported it
Has anyone else run into the issue of Suno increasingly ignoring the [Verse] and [Chorus] tags lately? I see again that he generates a song and confuses the lines, that is, he sings the verse as a chorus and vice versa? And also begins to repeat the same line again and again.😬
SUNO! Tell us why it starts lyris with 2nd or 3th line of verse ... or fix the thing.
Intrumental intro - OK
Extension VERSE - OK
Extension CHORUS - it skips chorus entirely and start with next verse ...
The best I did in my last time was to DELETE evberything in SUNO and move on with life.
I want to apply for a refund, how to do it
After paying, all the songs generated in Mandarin Chinese will be changed to Cantonese, which will stop all our company's business plans.
mail to billing@suno.com
The official must compensate me for my losses
the paid and the free version work the same... good luck.
I will be fired from my job tomorrow and I can't explain why it's such a mess now
as the paid and the free version work exactly the same i dont see any problem here to be fair.
But no matter what I do, I can't get the work I want and I deserve a refund
Has v3 been changed/updated within the past 2 weeks? Noticed a marked drop in quality of prompt recognition.
Right now it’s just terrible.
I was working with spoken word (which I’ve used successfully before) and a style prompt (which I’ve also used successfully before) Spoken word results were about 80% singing, and always the same goddamn female popstar autotune singer, I’m not sure who this is based on, but it is inescapable recently. And it was even putting autotune on the spoken word when I did get it! And the style is a musical style which does not use piano at all, but again I got piano in almost every track! Sometimes only piano, nothing from the style prompt registered at all. Last night I got one clip that was just two minutes of silence, no sound whatsoever.
Generally with AI when you get an unexpected result, you can compare the result and the prompt, and figure out where the disconnect is. But with Suno lately, it’s like it doesn’t matter what you prompt at all, it’s just going to do something and there’s no way you can affect it.
Just feels like flushing credits down the toilet. It wasn’t like this a couple weeks ago, what happened?
And while I’m at it, why isn’t there a dedicated support contact? Clearly there are some very heavy users out there who could provide a lot of good feedback on what’s working and what isn’t, but it seems like Suno just doesn’t care. I mean exactly what are the prompts we can use effectively? A lot of the ones that are listed in the FAQ are hit or miss at best lately. There are others which are not listed which work. This is another area where User feedback could be quite valuable.
I don’t mean to be dumping all over the company, because it’s a very intriguing tool, and when it works, it’s amazing, But v3 prompt recog clearly took a step backwards lately.
horrible
No reason, no insult or something, it worked 10 minute ago, and now i have this popup
Credits refunded do not always refund credits
The models don't change after release. This was confirmed previously by a developer in this channel.
The model may not have changed, but I wouldn't be surprised if they dialed it down with a different quantization (i.e. level of compression) to offset the increase in users.
That would affect everyone though so should be a lot more noticeable. The issue described sounds like it could be something to do with how prompts are being interpreted (but that's just me guessing.)
Using Suno, both before and after I became a moderator in November 2023, I've seen a consistent pattern. Users often debate whether changes to the model have affected the quality (even with V2!), yet at the same time, others produce exceptional work. This discrepancy isn't about the model itself—it's about a mix of factors like your lyrics, metatags, music style, and some luck. Even if you've recently had successes, you might mistakenly think the model has changed if your new outputs don't match past hits. It's crucial to recognize that results also depend on adjusting your input and sometimes spending more credits. Expectations naturally rise after creating several masterpieces, increasing the desire for more control and complexity. While AI remains unpredictable, accepting this and the credits involved can help you understand and overcome biases in your expectations.
Why is it flagging for moderation for words like "mor" "sparka" or even lowercase I's?
Completely broken again. Ignore the lyrics, ignore the language, ignore the style. Sometimes just cut parts away. I really think it's time for an honest explanation, what happens here constantly with Suno every 2 weeks.
I also notice some words getting flagged in one song, but works normally in another
This is the reality: #🪲┃bugs-web message (also for lyrics, ignoring style etc)
This is the final warning regarding your baseless accusations. Criticism is acceptable; alleging dishonesty is not.
I've noticed vocals seem to be getting progressively worse in the past couple of weeks. I'm in a couple of Facebook groups for playing with AI music, and I'm seeing a lot of others complaining about it as well. Is the AI learning from all the on-going generations being fed through it, or is the model static (i.e. it uses previous training but isn't actively learning)?
If it is actively learning, it might help everyone if it can be reverted to an earlier state.
I'm nowhere near done with it, this is just a new thing I'm trying out to see if I can make it run the way I want it to... The lyrics are especially terrible at the moment because I am testing the sound and not too focused on the lyrics or vocals right now.
I disagree that the quality is down. I hate to say it but for a lot of people it seems to me the issue is a combination of skill issue and lacking patience.
[Removed links]
Plus not wanting to invest time and credits.
Apparently criticism is not acceptable when I get a response like this every time I write something, no matter what...
Made this last night.
Am I just lucky all the time?
Some of the stuff I've seen posted is either trying for something super complicated (like a narrated voice over in a radio style breaking into the middle of a song verse) or is written superbasic and not even listing the musical genres they say they are trying to get...and they're shocked when it doesn't happen for them.
(Not saying this is you specifically. But I've seen it.)
Criticism is fine, accusations are not.. simple.
Honestly I wish there were a better way to split actual bugs from the technical discussion. Some of these issues could be solved if they were treated as technical issues rather than bugs.
At least I think so. Bugs do happen.
Every time I write criticism, I get a response like this... I can only repeat myself, and when I point out bugs that are repeated regularly, then that shouldn't be a reason to be given a "final warning". Sorry
Some people actually sit there and count the cost per generation for a song. I can't imagine that. SItting there and keeping a tally going every time I tweak something to roll again.
You're encountering similar issues repeatedly because you might not be fully considering our feedback. It appears that you believe your approach is flawless and that any problems are solely due to Suno. Being a free user, you may not have enough credits to thoroughly test everything, which could skew your expectations. AI can indeed be unpredictable and is not without its flaws. However, accusing the developers of dishonesty may not be fair if the issue lies in these skewed expectations.
I didn't get any feedback on what else I can do, then there's nothing to fully consider. I never thought this AI was perfect, I'm aware of that, this can never be case. However, extremely annoying bugs occur all the time, and that can't be my fault, as this happens in different browsers. I have enough credit, that's not the problem. I also know that AI can be unpredictable, but I wonder why certain bugs always have this regularity of the two weeks mentioned. I didn't say anything against the developers, It's just annoying that apparently nothing is being done about the problems mentioned by all other users, and that can't be the case. No more and no less. Has nothing to do with distorted expectations. And I don't want any further answer to this because it doesn't lead to anything.
I agree.. this doesn't lead to anything because you still feel it's Suno and not you. So be it..Just keep refraining of things like "I really think it's time for an honest explanation, what happens here ", and you can write whole chapters. Have a great day!
bug report: library pages can be clicked indefinitely if you click fast enough
The chorus should only be repeated twice, but it juust goes on and on and on, even when the instrumental changes to be a verse, the song is nice so it's a bit of a shame really: https://suno.com/song/6335ec14-2e2c-410b-aa62-0b417b803cd8
Happy Jazz song. Listen and make your own with Suno.
Very interesting theory, please enlighten us on how skill of typing a simple prompt and pressing the generate button affects quality, I'd really do anything to get the results I did months ago
This also happens if you click on the page arrow and then hold down space or enter on the keyboard (discovered this by accident thanks to a book being on the spacebar!)
bug report: create button says "extend" instead, don't know what triggered it (no song is in the extend queue)
I think I need to stay away from this Discord. I can tell it's getting toxic. G-bye!!!!!!!!!!!!!!!!!!
So I went back to one of my old songs and it's a full song, right? But when I hit 'reuse prompt' it's giving me song parts for it instead of a prompt reuse for a new song.
It's reusing the original song I extended which isn't even in the extended part of the song I am creating now. Gonna try to refresh/relog and see if it fixes itself. Just making a note here about it.
It's fixed itself, but that was weird. Could of been a load malfunction. Maybe cause I'm on Opera GX.
That's not exactly what I was saying and I apologize if that's what you heard.
What I am saying is that in my personal efforts I haven't noticed the huge quality drop off people have been talking about. The skills I was speaking of is knowing what genres combine well and what types of instruments will work well over a long term of use. These are things that naturally are subject to individual taste and which are therefore subjective. I posted the bit I was working on as an example of something created recently (last night in fact) that was not doing this.
I don't know if you can see it but in the playback the album art shifts multiple times showing how I am making use of continue from time to get that style of sound, which is what I was referring to. I have been doing this a little bit here and there since the ability to resume from an earlier time was made available but I didn't want to use an older song and have someone say but that was X number of days\weeks ago. So I used what I was working on recently to be my example.
Sorry for any misunderstandings.
As I said, I've seen people create prompts that don't even mention the genre of music they hope to get in the style prompt and then are mystified at the lack of style.
I must admit, that song of yours was pretty good quality, though the vocals seem to be worse than instrumentals and as a sound engineer I think a trained ear will be able to spot it easily
Oh definitely!
Which is why I said that I wasn't even focussed on the lyrics at the time.
I just wanted something to throw in there to see if that was a good point for lyrics and if it would hold up.
Think of the words as 'unskinny bop' if that makes you feel better.
It wasn't a finished song and I don't really have lyrics ready for it yet. I just had the idea for some of those combinations and I wanted to get them out before going to bed. I hope to come up with some better lyrics and focus on the vocal type later today.
Sorry this is not a bug but thot it time for something positive here.
I have 9 songs that I wrote lyrics for and I created over 40 songs with 30 different /styles/genres. Picking from https://suno.com/explore
All in one month and I still have 1350 credits!
Thanks to 3Daisy I adapted to overcome the CONTINUATION glitches. I split the lyrics of a song ( typically 4 verses 4 chorus) to 2/3 for part1 and 1/3 for parts 2 3 4. Often I had to CONTINUE to the final part with only [Big Finish] as the lyrics.
IM me if you want the styles/genres and the TAGS I like.
Are there any news on problem with voices turning into a choir in 80% of the cases or more?
That can only be fixed with a new model. So I guess V4
yeah... or if they find a backup of v3alpha 😅
i would be fully satisfied with a simple roll-back tbh
Soo... when? 😉
Half Life 3 confirmed?! 😮
I have the feeling, I haven't had for a long time, my daily credits. Or if you have a subscription, don't you get any?
If you have a paid subscription you don't get daily credits.
ahaa, okei, sad
The first song I bought was in the 80s. I paid $4 for a single, not converted to todays currency and it was for a song someone else created and not i created. Suno is a bargain. They should drop the business model for people that complain about the $0,03 per song.
I remember those days. I would buy at most three singles then it was time to buy the album those singles cost so much. Still sometimes you would get special alternative remixes or live cuts of earlier songs and theses were the only way to get those.
I still miss the live version of Endless Summer Nights on Richard Marx’s Angelia single.
Guys, this channel is for reporting Suno bugs... 🙂
Found a big
Bug
Where the text in the box doesn't go Away even when you reset the website
And it creates a major issue
Where the audio changes beats and vocals
Which ruins the entire song
And I know this bc I was testing the vocal generation system
Too see how it works
Probably worth posting an example song and explaining how what you're describing has affected it?
The V3 vocals have become too low in volume and quality and don't sit properly for a vocal in the mix. Some have way too much reverb as well and can't understand the words.
This song has an interesting behavior. I was extending a song with new lyrics, and at one point it completely ignores the typed lyrics and nearly 1:1 replays part of the earlier song with lyrics that aren't there
https://suno.com/song/8c5d1a27-75a3-4748-bee1-e76c662ae610
very glitched glitchy stuttered choppy vocals in a minor key song. Listen and make your own with Suno.
at around :04 at n-n-night it seeks a guiding beam, it instead starts to resing an earlier passage in the same tone and composition
I'll tell you what this song just doesn't want to extend right at all. I'm getting ignoring 3/4 of the lyrics, repeating past parts of the song, and completely empty submissions with no credit refunds
Can i at least get my credits refunded for these two blanks lol
https://suno.com/song/3c75fc1b-dbd3-4f90-b8bf-72fda4893dd0
https://suno.com/song/351393db-741b-41d1-a0e9-ef37089af2e4
edit*
There are three now v.v
https://suno.com/song/55824284-69cd-4517-8ece-a4a6dec87b8f
Too many bugs, apply for a refund!!!
Too many bugs, apply for a refund!!!
Too many bugs, apply for a refund!!!
Both my wife and I are getting two identical copies for each creation, burning through the credits at twice the pace. Is this a bug? Or how do I correct this?
It's not a bug, but a feature now. It used to be different and way better, but it's on purpose to force you to spend more credits and more money in the long run. The whole thing isn't as good as it once was, it's actually garbage. I'm surprised on how fast Suno went downhill.
I posted my grievances in general that's a little more specific here: #💬┃general-chat message
I feel like there is a noticeable degradation in quality of Suno output since last month. Can we get to the bottom of that?
I disagree, I've been using Suno in its early stages and it was creating fantastic songs. It was accurate with my prompts and it kept the same sound and key when I would continue the song. It's extremely difficult now to make it accurate on genre sound and takes waaaay more credits. We get 2 songs with the same lyrics instead of 3 with different lyrics and sound while still preserving the genre. C'mon, that's blatant greed.
it's not a skill issue. It's a greed issue.
It's worse. See my previous comments. I've been using it for a long time, didn't change my style of writing or anything. I don't know why consumers keep defending these companies that are taking their money lol it's wild.
We control the money, we can control the product and it's obvious that the developers have watered it down for the sake of greed.
Ham Man? Hmm... Here's something to cheer you up. The world can be a bad place and greed can ruin everything but let me tell you about this hamsteak: https://suno.com/song/b33cc3bd-ccad-4c3e-b96f-9abf78f1b9cd
Stylish Funky Dark Bedroom Pop, Baritone Sax, Mastered song. Listen and make your own with Suno.
The song I just posted isn't a great example but with other generations the most noticeable problem is that there's a lot of 'clashyness'; like some kind of underlying twangy metallic sound. It's noticeable in a lot of the things I've tried to generate lately from dubstep to orchestra. If the model hasn't changed and if it is anything like riffusion then maybe it IS a parameter problem; like reduced steps or a lower cfg. As far as the inability to understand basic directives like [Fadeout] or [Finale]; it seemed to be better previously and getting worse over time. Like in the first month of v3 I was able to get a concluded song in just 2-4 generations but now it basically ignores it and does whatever it wants. It also seems to not take kindly to sudden changes in style so this leads me to believe it's probably a cfg setting issue. Try setting it to 9.5 and a MINIMUM of 20 steps that's what I usually run on my local gens.
Not recognizing the problem, one that multiple people noticed, by a Suno representative is making people even more annoyed. Just saying.
Here's a good example of something I tried to make today that has a lot of clashyness and overall sounds like it's suffering compression problems. It kinda seems to 'lose quality' as it goes along and while I do realize what is in the song is a product of my choosing or not choosing to accept I really struggled with the fact that it really just wanted to be twangy: https://suno.com/song/0c41d7d7-70dc-4b90-be85-89328269377f
Funky Dubstep Pop With a Touch of Dark Bedroom, Piano, Strings, Horns, Orchestra song. Listen and make your own with Suno.
Even if I refresh the site or log back in, it is still set to create a continuation of the previous song, so when I try to generate a song in a new style, it is still connected to the previous song. In fact, only a one-minute song is generated, and the link links to the previous song.
I couldn't figure out why at first and wasted a lot of credits. What can I do?
Try clearing cookies maybe?
Click the Simple/Custom mode toggle to clear out the Extend. The big yellow button should say "Create" and not "Extend" if that works.
A perfect example of the noise in Suno V3. I hope this can be resolved soon.
Ever since you guys handed out 500 credits, I've been getting 65 credits every time I log in.
Ok, seems that ALL the "full songs" slowly from time to time get the thumbnail broken and the first one below even got the intro distorted.
Are the full songs stored in a different server or what?
I really hope all the "full-songs" don't become broken at some point (already downloaded the last 5)
@low cypress sorry to ping you Daizy but is the team aware?
Is this new? Is this old? Is this in a playlist? Or in your library?
Library
and even in the "create" page
So new, and in library? (and workspace)
yeah
yesterday morning the top song one still had the thumbnail
then yesterday night poof, broken thumbnail
I am creating right now, for the last hours... and yesterday too, not have that personally.. But will let the devs know
the first one from the below (old song) has even got the song distorted
Old songs could be, and I know they are still fixing some of them. I have it in my playlist with old ones. But I will add this to their pile 😬
Yeah i mean it was the first i created when subscribed some weeks ago, it was "clean". Then suddendly some days ago i re-listened to it and had the intro glitched
Not too long ago the moved to another colocation, and moved their database. That broke some images. But they will fix it.
Yeah, so even the audio file apparently got damaged in the process
(maybe)
I don't know about that. It's not that Suno creates crisp audio (yet)
I know, but i'm (let's say) 80% sure there wasn't a hearable glitch in the file on that song, it was flawless
Sometimes it really depends if you listen with a headphone of via normal speakers. And I am pretty sure you cannot damage an mp3 in that way. They are not reproccessed.
I use only monitor headphones, I'm pretty sure the song was ok
but i can't be 100% certain
Plus you say they aren't reprocessed
Degradation only can happens if you re-encode it. And that definitely did not happen.
So not sure what happened here tbh
Alright.. maybe just a glitch. 🤷♀️
But thanks for the heads up on the pics.. Will let the devs know so they can take a look at that 🌼
Ok, you're welcome
Curious about the audio glitch. If you don't mind sharing it could you post the song link? Might be useful to the devs, but I'm also wondering if this could be some weird browser behaviour when decoding the MP3 - do you hear it if you download the MP3 and play it with another program or device?
As i wrote days ago, i know it would be better to share the link (logically) but i wanted to keep it private because i was thinking to distribute the song... (and if it's not anymore recoverable well, i will not distribute it like this and i could share the link)
I may give the link to the dev directly if asked.
Anyway i thought of a browser behavior too, but tested it both on Firefox and Chrome and it's the same; downloaded mp3 has the intro glitched too
Yeah that's reasonable - sorry, I didn't spot that message when I had a quick look.
There is a misconception here. I am not a Team member. I am not a PR flack. Or a spokesperson.
I am an enthusiastic user of the AI system. I like to help out where I can with what my personal experiences with the AI have been. I am also a moderator.
Get annoyed all you want. It doesn't magically make me a Team member. I have no say in how the system is run or is developed. I am a normal user like yourself.
While I have encountered issues with the AI occasionally I have also worked around those issues and continue to get good results. I don't feel cheated. I don't get mad at the magic fairy in the box, I vary my prompts. I test things out. I keep trying. The same thing I have done since version one when I first found it.
From my personal perspective nothing has changed. You have to work to make stuff happen, same as it always was.
Greed? Huh...
Tell me more about this conspiracy to feed you worse results to make money and how deliberately choosing to make their AI system perform worse gets more customers. Go on, I'm all ears. I love a good conspiracy theory.
Wait...let me get ready first...
There. Now I'm ready. Hit me.
good one. thanks for sharing. here's another from my clips. it's a single take, no continuations. sound starts clean, keeps degrading and ends up totally broken
Any updates on whether the Suno devs are actually going to make the sound quality better? Perhaps even giving an option to download a HQ wav file / stems?
Suno doesn't use stems, at all, it isn't how it functions. For .wav the reason the quality is how it is is training, not file format
They could give you a .wav option now and it would sound exactly the same, the data just isn't there because of the data the model was trained on
You can use third party tools to break a Suno generation apart into stems after the fact
And, according to the devs, if they do offer stems in the future it would be done the same way. Master first, stems approximated afterwards
Through software
Any idea on how to force suno to use a male or female voice?
When I add it to the style is ignored and I get bot male or female songs
Idk, i wrote "boys band, 5 different male voices" in style and it always works perfect.👌🏻 how do you write it yourself?
Respectfully I love Suno but please fix this bug.
Chat GPT is NOT censoring my words, it's censoring my ideas for creative artistic expression. Please fix this by switching to another AI. This is definitely a bug. Attached below is a text file, which is my convo with Chat GPT on why it is censoring my lyrics.
It's frustrating to feel like AI is imposing its morals on my lyrics, especially when I'm not even using curse words. Creative freedom is vital, and it's disheartening when external influences restrict the expression of ideas. As an artist paying for this service, I expect support in amplifying my vision, not stifling it with moral judgments. AI should not dictate morality, especially in the realm of creative freedom. AI should enhance, not hinder, the creative process, and it's disappointing when it falls short. It's time to find a better AI, CHAT GPT is not cutting it for paid subscribers. Find an AI, one that respects and champions artistic expression without imposing arbitrary constraints.
AI should not impose its morals on the creative freedom of artists because doing so undermines the essence of artistic expression. Artists deserve the autonomy to explore diverse themes, perspectives, and styles without external constraints. Imposing moral judgments from AI inhibits creativity, stifles innovation, and hampers the evolution of art. AI should serve as a tool to assist artists in connecting with their niche audience, offering insights and support tailored to their creative vision. By respecting the autonomy of artists and facilitating their expression, AI can foster a vibrant and inclusive artistic landscape where diverse voices can flourish authentically.
boy band is also a genre so it will probably only have male voices
I tried to use MALE VOICE, man singer, etc but it doesn't seem to care
the prompts are in italian but the meaning is a medieval tavern ballad
have you tried writing the description in english? my native language is Ukrainian, but I write everything in English and then Suno understands me... Maybe he has some problems with recognizing different languages? If even that doesn't help, it's really a shame...
same for english, also AI doesn't have any issue with translations
I think it gives more priority to the style, if the style is mostly singed by female singers you will get mostly female results
Uhhh, so sorry for this. I hope it will be fixed someday. Btw I love the sound of Italian!
the end result was this but I wasted a tons of credits with these female voice results
https://suno.com/song/30ec1305-5693-40bc-8be4-bbead648fdc3
ballata medievale in stile musica da taverna, strumenti tipici e atmosfera intima song. Listen and make your own with Suno.
Is it the same noise in songs that everyone is talking about? I like the melody and energy of the song, but sometimes my singers sing either fake or like they were chewed up by a computer... It's a shame, because if their voices were cleaner, it would be a cool song...😅 https://suno.com/song/b895cf16-bbdf-4b5d-a402-17ce899d1e43
K-pop, boys band, 5 different male voices, korean, pop, hip-hop, emotional, catchy, song. Listen and make your own with Suno.
Is 50 credits only enough for 4 songs (2 doubles) instead of 10?
50 credits is enough for 5 generations, 10 songs.
This is at best. I have enough for 1-2 songs, because often unsuccessful songs come out (or the melody is somehow strange and illogical, or the singer reads the text completely wrong)...
But whole song need at least 4 generations (first 2 minutes (2 versions) and the ending (2 versions)).
You could easily make an entire song with just two minutes.
Can you check what happened to my last 4 generations, it took 10 credits and than more 10 credits dropped during the generation was going on
I know what your role is.
Hey I hv generated 5 songs using 50 credits but one of them is missing from my list. I could only see 4 different songs(8 in total ) generated.. and right now credits left to zero.
So any help find me one invisible song ?
getting instrumental only without ticking the option, the generative song writing feature is down?
Oh? Forgive me for being confused then. You replied to me and mentioned a Suno Representative here: #🪲┃bugs-web message what am I missing?
so i'm trapped into "Extend" instead of "Create" despite not having a clip selected to extend from
i think turning off/on custom & instrumental fixed it.. but it was weird, reloading the tab and logging out/in didnt fix it
Yes, turning Custom off and then on, or, on and then off, should get you back to Create mode.
Looks like you didn't click "clear" ?
i think, my firefox crashed last night after i stopped using it, so when i came back today, it had a "fresh" create screen, but was still trying to extend
so there was no clear or clip being extended from
seems like a firefox issue in retrospect
Would you have any general tips for having a higher than lower chance of generating something worth downloading and possibly publishing? On general people claim they make money off of Suno creations and if this is not trolling then I really wanna know how, hope I'm not received wrong, I am also passionate about AI, just sharing my observations usually, no matter whether they are positive or negative 🙂
I'm not making money from it; yet, but I do have a growing list of chart toppers I'm holding in reserve. The advice I have for you is to experiment with different words as styles or even nonsense words such as Burningblackness, fathomfunk, moorlightwhisper, etc. Also adding keys and chords to the style prompt helps (CEGmD, etc). Sometimes attaching Dolby, Mastered, or Studio Recording to the end improves the quality. Also note some basic styles consistently output decent stuff like Dark Bedroom Pop. Try mixing it up like Dank Bedroom Horrorcore or whatnot. Also the lyrics section can be used to direct the flow of music. Here's an example of some otherworldly brackets:
Artistic Stomp, Broadway Performance, Instruments Made From Junk song. Listen and make your own with Suno.
problem with play back, you can't start a song playing in the middle, it always defaults to playing from the start
Looks like the 'Clear' button doesn't always show up when in 'Extend' mode. For one, this can happen when switching between Create and Library and Trash and then switching back.
This appears to be a bug related to switching between different Suno screens.
If that happens, toggling 'Custom Mode' On and Off or vice versa should get you back to song Create mode.
sound quality is really bad lately, grainy, white noise
I've been having the same issues yesterday and today when creating song extensions. Getting lots of really bad distortion, extension sound not matching the feel and theme of the song, instruments sound like they are passed through a sharp high pass filter or are otherwise quite distorted, and volume levels suddenly going way too loud. Burned through tons of credits in the last couple of days and could not get any decent extensions. Could be partly due to the style of music I have been working on though, which is ambient/atmospheric type music. Not sure of the cause, though...
thanks for this!
getting a very odd bug
making a new song, but the interface is treating it half like it's an extension. the button says 'extend', but there is no timestamp, nor a 'clear' button. when i cleck the extend button the melody of the song has been carried over from my previous song, and the 'continued from' box is treating it like it's a continuation from that song, but the lyrics portion of the page does not contain any of the lyrics for the previous song, just the stuff i put in now trying to make a new song.
i'll take some screenshots
actually it's not even continuing from the last one, it's continuing from a random part of it from ages ago lol
damn porn spambot don't interrupt my bug report lol
screenshot of interface bug, irrelevant stuff removed. problem was not fixed by refreshing the page, fixed it by going to another song and choosing extend on that (which made the extend interface look normal), then going back to create (which gave me the normal unbugged create interface again)
Try toggling the 'Custom Mode' button On and Off or vice versa, and then after doing that click the 'Extend' button on the song you want to extend. This should reload that song's configuration for extending.
i fixed it, took going to another song and choosing extend on that
all back to normal so no huge problem but i thought I'd report the weird bug anyway
apologies for that, will be fixed tomorrow
all good, you guys do great work
[BUG AO UNIR EXTENSÕES EM UMA MÚSICA SÓ - MÚSICAS LONGAS]
Olá, meu nome é Douglas sou do Brasil. Sou usuário do plano Pro e notei um bug criar a música completa das extenções geradas.
Depois de extender diversas vezes e criar a música completa, ao clicar no botão "Get Whole Song", as partes são todas unidas, porém, em todos os pontos de separação acontecem rúidos incovenientes que "sjuam" a música e deixa incomoda de se ouvir. Isso a cada minuito fica muito desconfortável para o ouvinte.
Por exemplo, nessa música completa: https://suno.com/song/3b48b903-c58a-4b90-9f72-d1c357423dc7
Veja os minutos a seguir e note os rúidos causados (eles são justamente nos pontos de conexão entre as partes extendidas): 2:00, 3:00, 4:00, 5:00, 6:00...
Aparentemente esses rúidos somente acontecem em músicas mais longas (acima de 15 minutos ou pouco mais), pois eu tenho a mesma música com 10 minutos (https://suno.com/song/5ef76675-fbda-44b1-a893-ad4c6ec185ec) e não contém os rúidos nas conexões das partes.
Por favor, corrijam esse bug para músicas longas fluirem corretamente sem rúidos desagradaveis. 🙏
Obrigado e parabéns pela excelente trabalho
stoic greece ambience atmosphere deep focus dark brave song. Listen and make your own with Suno.
stoic greece ambience atmosphere deep focus dark brave song. Listen and make your own with Suno.
Enough for 4 songs, but also take note for extending the songs. I think those take credits too.
After the extension, getting the full song is free of charge.
Continuation song, significantly differs in volume from the original one, original one is too quite. https://suno.com/song/aa7654fe-c26b-4268-895b-b91d6479b922
Darkwave opera, Evil, solo deep and resonate old male vocals, energetic song. Listen and make your own with Suno.
when playing older songs the page keeps scrolling to the end all by itself
All I can say is to keep trying. Keep working on refining your methods with it.
I don't actually have anything available for publishing just yet. I mostly post to YouTube and these days in the form of music videos. Seems to be slow but steady growth so I'd like to think that I'm building an audience and when I do try to publish I will have people who seek me out and want to hear what I make enough to opt for it or buy it.
Hi, there's a bug where when you remove a song from a playlist it removes other songs also. Can anyone advise? This is really frustrating.
This is for the web app.
Hmm interesting, do you think adding tags like studio quality or mastered will change the output? I haven't considered that before
Hi @buoyant flume @unreal radish Can you help me?
Please do not ping the Team members. They are working on making the model better and do not need to be distracted from that.
The automod spamming the previous attempt should have been a hint. Please take ten minutes time out.
Well, this is also an improvement... Who can help me?
I have a paid plan and my songs are coming out defective...
Shouldn't unused credits in a yearly plan roll over from month to month?
They aren't meant to (except for top-ups)
?
?
Ah, ok i see
I deleted it, I didn't know, I won't do it again
I deleted the cookies, but to no avail.😩
Who can I talk to for bug fixes? Is there a priority support channel for paid plans?
When I click on the Simple/Custom mode toggle, the big yellow button says "Create", but it does not unlink the songs not referenced in the extended on the right, and it suddenly creates them from part 3.😩
People tried extending from a different song or different browser (mobile even) which fixed the issue for some. It seems you're not the only one, so I hope at least somebody competent is aware of this problem.
привет
Desktop or mobile? Since suno can only be accessed through the site (the apps are scams), That is more helpful to know
What took me a bit to realize initially is that you may have accidentally selected multiple tracks by holding down a few seconds or playing a song.
So when I add/remove to the playlist on my phone, I add it through the ellipsis options on the current playing banner, or via the song description (showing the lyrics, etc.), not directly from the library
I suspected this was the case as I'm using mobile. I find it kind of unclear when multiple songs have been selected, there's nothing obvious to state that multiple tracks are selected. I realise that they become highlighted when selected, but if I scroll down the page I may not have realised I selected a track and then it either removes extra tracks or adds extra tracks to playlists unexpectedly. When I accidentally remove tracks from a playlist I have no idea which ones got removed, which is quite frustrating.
I can empathize with you as I felt the same. Would be nice to have a Select Marker, like a small checkmark on the top right.
It took me a while to notice the slight difference in the background (selected ones are slightly brighter) but still took a while to notice because of the Deselected highlight.
Attached is an example:
[X] Top = Previously selected track, left selected
[-] Middle = Recently deselected track (slightly dimmer)
[ ] Bottom = Not selected
yes this is exactly the problem!! It took me a while to figure out the highlighted difference, gave me headaches after the update 😅
I'm facing a couple more issues. One is if I am listening to a playlist and I view the lyrics to one of the songs, when the current track finishes it doesn't continue playing the playlist. Instead it plays the top Trending song and then I have to go back to the place I was listening from. This also happens from the Create page which can be a problem if I am going back through old generations, making it difficult to find where I was playing from.
The second problem is where I am extending a song. After selecting the song I want to extend, I press Create a few times after inputting the necessary lyrics etc. I then minimise the Custom Mode screen to see the generated songs. Then when I go back to Custom Mode to make adjustments and new generations the song is no longer selected for extension. This means I then have to go back and find the initial song I wanted to extend and relink it.
Thanks for the replies! 🙏
I'm using the Suno AI web app btw.
Reporting bugs and I don't see the point of having the "Other" checkbox without allowing people to enter what the problem is.
Can't generate anything rn all I'm getting is "Prompt flagged for moderation" and can't tell what it's complaining about...
Like I've literally striped down the song lyrics I'm trying to use to just the first 3 lines and still doesn't work.
Hmm maybe the issue is in the first few lines then. Or in the style prompt that can be flagged as well. If you want to know you can paste your lyrics into chat gpt (free version) and ask it why it could be flagged it will tell you (most of the time)
hmm ok . yeah I tried changing the style to a few different totally different one and it was still doing it... i'll try the chatgpt idea!
lyrics are just this and still flagging:
[Country music intro]
[Verse 1]
Woke up this morning, feeling fine
Gazed down at my beauty Madeleine
maybe its the first line is a copyright thing i guess ... really need better feedback on these things
Got up this morning?
ok i think it was the "Title" actually...
changed that to different thing and it worked w/o tweaking the lyrics
So more info on why things are flagged really needed.
i dont really know if this is a bug but i cannot create songs
You need credits
👍
Not sure if it is a bug, but extending songs further and further always automatically makes the voice more and more in the background and mushy. robotic, fading. Usually it starts at the fourth or fifth or so extension. I'm not sure, maybe the generator prioritizes music higher and therefore the voice goes more and more into the background?
Hi, I’ve used a lot of credits to create songs that are really bad. The voice is 95% of the time bad quality or not enough audible. My texts are in French. But in English, Italian or Spanish, it’s a bit better… do you have any clue to have a nice clear voice instead of a robotic one?
Okay so I found a really weird bug with Suno which most people likely won't encounter due to what needs to be done in order for this to happen. Therefore it is not a pressing issue, but really interesting and bizarre.
Basically, you can cause the generated song to malfunction and obsess over a specific word reliably. I believe the word is almost "instrum" or possibly "estrume" which is apparently Spanish for "manure."
In order to do this, have instrumental mode turned on and simply put a very simple word such as "the" in the Style of Music box over and over again, separated with commas. For example: the, the, the, the, the, the, the or me, me, me, me, me, me, me. This works with a bunch of other things such as singular letters or symbols, as long as you have a long comma'd list.
Much of the time the generation will fail and give about 7 seconds of output (understandably). Nearly every time you will hear someone (usually with an accent) repeat the word "instrum" / "estrume" over and over again. At least I think that is the word being uttered. Sometimes there will only be fragments of the word. At first I thought it was saying "instrument" (Maybe it is? Hopefully someone who understands better can confirm). It also tends to have a generic hip hop beat behind it as well. It may also end up reading out some of the repeated word you put in the prompt and ad-lib other words. Regardless, some form of "estrume" or "instrum" will often be part of what is generated. Examples below:
- https://suno.com/song/d0974f3f-8b82-4001-adbf-6613f7ecbdd3
- https://suno.com/song/6827acb0-b956-4c5c-9ec5-b6ee7d7f9e1a
- https://suno.com/song/2932b081-f8eb-44c6-bc6b-976c2cb8f66d
My main theory is that the part of the code which reads that it is supposed to be an INSTRUMEntal is somehow interpreting that as lyrics? Idk, or it could be a remnant from a testing environment.
Repeating the word is a special case of a more general thing, spamming any word in the style prompt tends to cause it to be sung and send it off the rails (male, male, male = https://suno.com/song/ffef0b04-f078-43e0-9b4d-5dd7276bb3e6. But if it different single letters reliably produce a word like instrum/instrumental, your theory could make sense, very interesting.
Yeah I noticed if I do same things with normal lyrics as well, it will sometimes read the repeated word from the prompt in addition to the lyrics (but the "instrum" hallucination does not happen).
That said, I spam single letters or pure puncutation all the time and don't remember hearing that word. But I am a bit 'word blind' with the lyrics maybe didn't notice...
I think it likely has to be a word which would normally have no effect on the output (so genres or adjectives likely won't work). Singular letters or common words tend to be what triggers it. Obviously not a big deal just very a peculiar discovery
we can ask the suno devs next time they are hanging out in chat. i don't know how it works but it seems plausible to me that instrumental mode toggle is doing something like adding special words to your prompt, at least. oh I just had a thought, I rarely ever prompt using the intrumental toggle because I prefer to use some guidance even in instrumentals in the lyrics field. if it is unique to that mode, then I would not have have noticed it actually.
i'm sort of artistically intrigued by the idea of a song singing what is sort of a system prompt in a text model or something though, fun concept
Agreed it is intriguing, I ended up discovering it because I realized having a single word over and over again seems to be a semi-reliable way to find out what that word does to contribute to an output. So then I branched off and ended up trying some more obscure stuff and started picking up on the pattern of that "instrum" popping up again and again. It definitely stresses the AI out because half the time it'll only generate 5-10 seconds and refund lol
we prompt the same way. the very first thing I do when testing a word is intensify it "oatmeal, oatmeal, oatmeal, oatmeal, oatmeal" etc
Haha yup exactly, it is fun to try and push the boundaries on what is typically done and to get a little weird with it
Bruh, it's so annoying when it generates a kick ass song and ends it at a minute long. Can't even extend it longer. Y'all need to fix that BS ASAP.
Hello everyone
Have you ever had the experience that Suno suddenly distorts the sound during the 2nd or 3rd expansion of a song and makes metallic noises? I'm currently working on a song and these noisy noises are ruining the whole project. Is there a solution to this or is it simply a matter of too many people using the program? thanks for your tip.
Please stop double-posting this. You asked this is in #💬┃general-chat in #📚┃suno-school (where I removed it now) and in here.
okie dokie
I entered the wrong payment method. I would like to receive a refund for one year's payment and change to paying the amount every month. How can I get some help?
You can email support@suno.ai for that
thanks!
BUG - Suno account page still points people to the .ai address instead of .com
When you create a song, the credit section at the bottom goes crazy, sometimes it shows as if 30-40 credits are gone for a song, but when you refresh it, it fixes the page, but this is annoying, I hope it will be fixed!
WEB EXTEND DOESN'T CLEAR BUG Example
Check that the yellow button says "Create" and not "Extend". If the button is stuck on extend, toggle Custom mod on and off then refresh to fix.
@low cypress I further confirm that full songs thumbnails always disappear and bug out. Now in only 2 days, new record time.
I made this two days ago in the evening
Still checking if audio get broken/distorted randomly
Literally "always"? Are you by any chance removing "parts" of that song?
Because that would explain it..
No no, literally always.
I don't touch those songs anymore, they're fully complete.
I don't know if it has been discussed before, but rapidly clicking on the small page change arrows in a user library permits the user to visit not existing pages. That should be limited to existing pages. Letting the user input the desired page number would be much better
Moreover a search by title function seems very important to me. The library is growing bigger and bigger
If i hypothetically generate a song now, extend it to make it full and then done, after 1-2 days the thumbnail says byebye
So you are not throwing away "part 2", "part 3" etc?|
I have created 5 songs in the last 2 days, and for me all thumbs are still there, so trying to wrap my head around this.
Is it just the thumbnail? Or if you open the song is the image missing there too? If it is missing on the song page, what web address does it go to if you right-click the image and choose to open it in a new tab?
I wonder when the problem with the incorrect structure of singing verses and choruses will be fixed? I'm running out of patience, I want to generate music, not spend all 50 credits trying to keep the AI from singing half a verse as a chorus, half a chorus as a verse. 😭😭😭
tekno is a genre.
https://en.wikipedia.org/wiki/Free_tekno
Free tekno, also known as tekno, freetekno and hardtek, is the music predominantly played at free parties in Europe. The spelling tekno is deliberately used to differentiate the musical style from techno. The music is fast and it can vary between 150 and 185 bpm and is characterised by a pounding repetitive kick drum. Nevertheless, bass drum dis...
If i open them, all the songs have this default image
Yeah whenever i merge the full final song, i trash all the previous parts and starting songs
Cause i don't need them anymore
And that is why the images dissapear
They are linked to the parts
Image is not saved in the new file as a new image?
Not atm..
Ah
I was wondering why I only heard you about it. 😄
I will suggest to the team to see if that can change.. but atm, they need to stay
@short forge Devs have just let me know that they will look into that 🙂
Cool!
I was gonna say, that didn't look like Suno
Yeah something was off.. just didn't understood what yet 😄
It's fine... we are all humans 😉
ok, seriously, it has been 2 weeks since i have contacted suno ai to get a price for the rights for my songs, and i still have not gotten a reply. I am paying them money every month, this is unbelievable. I want my songs, this is wasting my damn time. I want to use my songs for commercial purpose and they dont reply to me. Seriously, if they do not write back soon i will send my Lawyer after them. I am not playing around, when it is about my damn Songs.
As músicas que deram bug não estou recebendo os créditos de volta, como faço ?
Como faço para resgatar os créditos das músicas com bug?
this is an English channel. Can you translate it please?
I am just a moderator, but if you shoot me your email address on which you mailed them, I will send it to them and ask to look into it.
/d song
Unless the policy has recently changed, it stated previously in the Suno FAQ that you automatically own all rights to your songs for commercial purposes if you are a paying member of Suno.
He is talking about songs he created while he was using the free credits before that
Hey, yes, thats exactly it. I have now a Subscription and pay them monthly, so i do own the songs now i make, but there are 19 Specific songs i have made when still being a free user and i really like them. To be honest i actually love them, they are bloddy amazing. I did also write the Lyrics for those specific songs and i know i own the lyrics still, but i just want the rights to those specific 19 Songs as well to be able to use them with stuff. I am willing to pay for the rights for those 19 Pieces but i cant do anything when the people from Suno ai dont even reply. I just want my songs. Not like there are not millions of songs made out there every day or week, so 19 specific songs should not even be a problem in the least in my oppinion, and again, i am willing to pay for the rights, even 10$ per song if i have to.
Please send me your email address in DM
ok, thank you so much for your help.
this is not exactly a bug and I don't know if it is a known issue, but the faqs say "4. Add the next section of lyrics" when extending a song, which might be enough, but when it is not enough, to make the model gen coherent tracks, adding a few lines of the preceding verses helps (a lot)
Can I find all working meta Tags somewhere? All unknown the ai takes as words. And is there a working bpm Tag?
We don't think anyone has succeeded with a BPM tag. (If you find one, please let us know!)
You can find some metatags here, at the unofficial Community Wiki: https://suno.wiki
I've been wondering, how can people add to this wiki? It doesn't seem like a typical wiki in the sense that the community can contribute to it.
Isn't it just run by one person?
Good question. I'll run it by them.
I think there is a way to apply to edit but I can't remember what it is.
Ah okay, yeah maybe I'm overlooking something but I see no option for editing and contributing. Otherwise the term "community" wiki is pretty misleading 😅
Anyway, I believe all of these metatag and genres listed should have linked examples for each. And it'd be easier to achieve that if the community can actually contribute.
Well the other one we had for a while was getting bombarded by spam so that might be why this one is a bit more restricted...
Oh damn, well hopefully over time they're able to open this one up to a handful of approved community members or something. There's still a lot of potential for it to grow
The E-Mail should be changed to the new adress (on Manage Subscription side)
This is a known issue. The good news is that they are aware of it and are using an alias so email sent will still get where it is supposed to go.
Good, some days ago it seemed not to work
You can't use the word "clues" because apparently it is the name of an artist? So you can't do e.g. "song about a mystery detective searching for clues" without it getting blocked.
Could you please submit this here: #📚┃suno-school message
meanwhile use "clooz" ( or should I have said "get a clooz" ?) 🙂
Here's a real BUG for ya 995 credits and it thinks I have ZERO
Out of interest, what does it show on the "Subscription" page? Is it showing negative credits there for anything?
Cant see my PAYMENT METHOD details
You should contact billing IMO, billing@suno.com
Thanks my silly bank disabled wrong card - I paid up The only bug is showing 995 instead of ZERO when time to renew
What happened to the Remix function button ?
Hello, I would like to know why when I tried to generate a song a message appeared saying "Couldn't generate that - prompt flagged by moderation" and the song was not generated. My song did not have any offensive content, neither in the title nor in the lyrics. Is this some bug?
To extend the song I created, I clicked the extend button and made settings. However, the generate button was not activated, so I refreshed and my original song disappeared.
Renamed to Reuse Prompt. Continue song is now called Extend
https://www.suno.wiki/faq/making-music/how-do-i-make-the-song-longer/
Got censored from ChatGPT not liking something as it is the moderator. Ask OpenAI's ChatGPT itself about "what was not appropriate about the following: [content]" and it should tell you.
Usual causes are using swear words or interpreting harmful situations. Even if it worked before, it could be a small change caused its interpretation to become different.
But the lyrics had no swear words, no racist terms, no sexual connotations, nothing that was inappropriate or offensive, it was just a normal song
Be sure to copy your lyrics before refreshing the page next time 🤷♂️
As for your missing song, do you mean missing in the Extend portion or from your library?
If you mean from extending, re-extend. If you mean library:
- Make sure all filters are disabled (e.g. Like button is grayed out)
- There is pagination, it could be on the next page
It may appear like that to you but remember that an AI is interpreting what is being said (or from what had been generated)
E.g. someone got censored for having the words bones and ribcage in their lyrics, removing them had fixed it (likely because it conveyed serious harmful image to the AI...)
Or they solved it by rewording it to rybcage, been a while, phonetic spelling y'know
Yes, although the AI once generated an image of a woman with bare breasts, even though the lyrics didn't have anything like that, just the word "woman" and I had to throw the song in the trash because of the image, but I think that really must have been the problem, it must really be what is written in the lyrics
Oh, thank goodness. As you explained, I was moved to the next page. It was always the first song in order, so I didn't know it was sent last. Thank you for your explanation.👍
You're welcome, I had the same panic in the past and it still haunts me lol
Majority of songs generated today are shorter than 2 minutes (and no way to extend it)
Hi guys I was just buying more credits and I noticed when I bought 2000 400 of my original was wiped out how is that happening.
I did a refresh and it’s still minus 400
Who do I complain to
billing@suno.com should be able to help you with that @lilac isle
I messaged them
Can I increase my credits to over 4000
I want to make multiple purchases of credit up to 4 X 4000
Yes you should be able to do that.
Suno bot is no longer on discord. You can now use https://suno.com
thanks
Suno bot is no longer on discord. You can now use https://suno.com/
The music stutters (interrupts) when connecting
My last 20 songs keep generating nonsense lyrics even though I have instrumental on. Is this a recuring problem?
This is called a hullucination and all AIs do some variation of it.
guys, if i get a song with 40 sec or 30 or 120 sec is a bugg or?
the song always tries to be as long as needed. sometimes it ends very quickly if there is nothing left to do. at least thats my observation. adding commands like [instrumental] or else might help stabilize this part
hmm okei, good to know
Concluded that suno is incapable of doing 12 bar blues, despite having pin point lyrical cues, it NEVER keeps to the blues format. Used hundreds of credits.
After the addition of two verses and a chorus in version 3, the analysis of the songs was compromised, as the audio often does not reproduce the complete lyrics. Please make it available as in version 2, highlighting only [stanza] and [Chorus].
Suno is hanging on generations, but it seems to be a UI issue since you can open up a new window and the generations will be done
Page shows this, but a new window will be normal
Looks like it might have cleared up? Weird
I'm having a lot of problems in Instrumental with short songs not extending properly. I've tried many times over 2 days and I'm having so many Credits Refunded and not producing complete songs.
I'm not getting any 2 minute first tracks. My last try was 1 at 1:40 and 1 at 1:34 and this has been the pattern for 2 days.
me too
When will the sound quality improve? The sound quality is really bad
i cant cancel my subscription, how can i solve this?
have you tried on a PC browser? Mobile can be a bit buggy at times
It could just be a visual bug and it's actually cancelled, not sure though
yes, i tried before
Contact billing@suno.com in the mean time if you have not done so yet
Very challenging!
Very often getting the wrong genre, tempo, or mood from generations, as well as songs supposed to be instrumentals coming out with random (and terrible) vocals
I'm getting the same 8 bars repeated over and over for full length of the generation, no melody or variations
https://suno.com/song/82ad6e75-b76c-478e-808f-bd268df76bd7
This sound quality is horrific towards end...i have probably had 100 like this in tha past couple of hours @unreal radish
outro echo to fade song. Listen and make your own with Suno.
can i do anything about it?
hmm these aren't continuations are they
could you link a few more examples?
progressive deep house, house, with, break with 2nd build song. Listen and make your own with Suno.
progressive deep house, house, with, EMOTIONAL , heartfelt, delicate VOCALS, , catchy, phat bass line melody song. Listen and make your own with Suno.
these are from scratch not as prominent but its like a wind noice
noise
progressive deep house, house, with, EMOTIONAL , heartfelt, delicate VOCALS, , catchy, phat bass line melody song. Listen and make your own with Suno.
theres a bad one
progressive deep house, house, with, EMOTIONAL , heartfelt, billieeilish VOCALS, , catchy melody song. Listen and make your own with Suno.
that enough @unreal radish
?
this one sounds fine (or at least doesn't the systematic error present in the first)?
Not gotten credits refunded for any errors or 3 second "songs" in some cases.
a few things that might help in the short term:
- if you're making a continuation, try making a whole song out of your history up to that point before continuing, e.g. if you have a continuation chain A -> B -> C, create whole song AB and continue AB -> C
- If you have a lot of lyrics to get through, try breaking them up into 2-3 stanza chunks and stitching them together via continuation. Sometimes the model gets confused if you try to feed it a wall of text 🙂
- I wonder if the tag
echois hurting you in the first example you linked [?]
Why are the songs only about 1 minute now when they were always 2 last payment period?
Something is fishy.
"Hi guys, I have a problem. I was working on a song about a friend who is a raid leader in WoW. No profanity, no sex, nor violence involved. I managed to create several songs, but now I'm encountering an error message: "Couldn't generate that. Prompt flagged for moderation." I haven't changed anything in the lyrics. I wanted to pay for the app since it's entertaining, but if it has these errors, it's better to wait until they're resolved. Does anyone know what I can do to avoid this error? Cheers and thanks
put the lyrics to chat gpt and ask where is inappropriate chat gtp is very strict at profanity so they may can find where is wrong
I was able to make the song with the same lyrics previously
seems like it's worse on some days, better on others
I recently cancelled my subscription with about 5 days left until the end of the billing period. I figured I can use up my remaining credits by then.
Once I did, I instantly lost all remaining credits (with 15 left for whatever reason).
Quite a bummer. On top of that, because the subscription was still active, I assume, I also didn't get the daily credits you'd usually get.
So... I think the subscription system is bugged.
Yeah, it's a con, looks like. No surprise, really. Just about everything is ultimately a scam on the Internet. Want to cancel, but unwise at the beginning of a period, I think. Songs are often only about 20 seconds and 1 minute or so now. At least I can delete permently, all the hundreds of 20 second "songs", lol.
Actually, have had a LOT of 11 second songs. Unsure why that value.
11 seconds seems a bit strange
I will check mine
Credits are usually refunded at 4 seconds or less, but not 11 it seems. Probably random from user to user, maybe. A lot of it is pure randomness.
I've got lots of 3 seconds, lots of 4 seconds, then 28, 34, a lot around that length.
Yes. Same here.
😦
Out of the 1st gen, which should be 2 minutes, I have very few that are anywhere near 2 minutes
A few are 2 minutes, but not many
Actually, a lot of mine were 2 minutes - only for one period, but I extended them, before, had to edit and remix everything, but now the sound is getting to me, because so much of it sounds a certain way regardless of genre. I can now tell by the dynamics where it comes from (not always. but fairly often).
Checking back, this strange generation started about 20 pages ago, in my library
For me it's getting like AI images - after awhile, I can instantly tell where it comes from. Weird, unlike the people who think six-fingered and three-armed people are the norm.
It's not as bad (quite as bad) as "soundraw ai" which is not even AI but a total scam with a limited number of pre-made loops.
hi folks.. is it possible that the sound quallity goes down if we expend the songs,my ears say sometimes yes 😉
Not for me. But there is too much redundancy at times. I checked in iZOTOPE and it's always pretty much the same. Maybe depends on effects whether prompted or not, though.
????
Good luck, hope you get it sorted. Very difficult with any type of Internet subscription, this is only the second I've ever used.
I've not had any problem with cancelling. I often take breaks from using Suno at the end of my subscription month so I unsub and resub when I want to continue.
?
Sorry, I might've confused two separate things being discussed. I thought you were commenting on the picture just above and what that was in reply to. 😅
It's all good.
Is it a bug, that today I only get male singers, no matter how often I use "Female Voice, Female Singer, Female Singer-Songwriter, female soft voice", in the prompts? That problem sucked so many tokens ... 😣
Keep getting generations that say the words "start" in first 10 seconds or so. Even on instrumental mode. Happens like 15-20% of the time......
Same for me. If I had to guess I would assume there is some newly "force inserted" start meta tag to prevent skipping the first lyrics parts.
That's what I think too. Considering Suno likes to sing the meta tags sometimes 😆
Not sure if it's just my prompting or changes behind the scenes but lately it's really not hitting anything like close to what I'm putting in as a prompt. wrong genre, wrong gender vocalist. Doesn't recognise baritone or tenor, or seem to know the difference between rock and darkwave or synthpop. Maybe just how things are ironing out but it's not reflecting what I'm asking for ATM, like it used to do.
Yes. The more I extend the more noise is introduced to the rendering. Particularly with electropop or coldwave.
I'm having trouble extending instrumentals in V3. It has failed 3 times in a row with credits refunded but I believe I still lost like 5 credits. Anyway I wanted to let the team know about this it seems to be glitched.
10 seconds nonsense generations should be refunded, and are not apparently
i put in lyrics for a custom generation, came out with no sound
Likely the lyrics were too long
Or they updated the censorship message? Hmm...
Yep, looks like they updated the censorship message
Your lyrics got censored from ChatGPT not liking something as it is the moderator. Ask OpenAI's ChatGPT itself about "what would censor the following: [content]" and it should tell you.
Usual causes are using swear words or interpreting harmful situations. Even if it worked before, it could be a small change caused its interpretation to become different.
This is more of a suggestion for the Suno team, but in a future model I wish there will be a different model trained just with instrumental. This way we could add instrumental parts where for sure there won't be any lyrics. Moreover, combine this with the possibility to add instrumental parts to the left (read, intro) and this could be what we all need, the chance to get perfect intros 😊
the bug where generations were often ending short of two minutes was identified and fixed, thanks to everyone who called it in
should also address the "start" and "end" hallucination issue, further bug reports welcome
i just generated a bunch. they were mostly 2 minutes long.
and ouf course, 80% of them had the usual choir effect on the voices 😑
Pretty sure that's just part of the model itself (seems to lean towards having more voices unless prompted in certain ways) and won't get fixed until v4. Whatever the above bug was sounds like it was more to do with the inputs into the model (total guess on my part though.)
That said, I wonder if it's possible for them to work around the choir issue by adjusting the prompts automatically 🤔 might put something in suggestions
i'm pretty sure it's possible because it was working just fine before March 21.
v3alpha was so much better
someone said "Out of the 1st gen, which should be 2 minutes, I have very few that are anywhere near 2 minutes", and i just wanted to mention that I cannot reproduce that issue here.
i get proper length, so that's nice at least
I may be wrong but I think v3 alpha and v3 are effectively separate models (which is likely why v3 consistently wants to add more vocal layers and v3 alpha didn't do that.)
ohhh?
interesting
v3 is better at following instructions than v3alpha, but the sound quality isn't anywhere near v3alpha, especially when it comes to voices and simple instruments.
so if they're different models, that would certainly make sense
During v3 alpha we were encouraged to use thumbs-up and thumbs-down on clips to give feedback. Presumably this was used to aid training the final model, or maybe it just refined the existing model. I guess a lot of people liked the choir effect or didn't dislike it. (Happy to be corrected if any of this is wrong, I would rather not spread bad information.)
i seriously doubt that's the reason tbh
the amount of people complaining over the choir effect in here tells its own story
but yeah, basing anything on just giving people a "thumbs up/down" option is reckless
some more fine grained rating would have given them much better input
if that's really what they did, it's quite an own goal IMHO

Somekind of butterfly effect when little change screwed up the whole model.
it's the curse of IT: just a few typos can have a huge effect. 😅
so many of my tracks lately cut off at 6 seconds, 12 seconds, 20 seconds
hope we get credit refunds
feels like 50% of them
It should have been fixed by now unless it was more recent than the linked post?
i'm going to put this to the test
I've done over 100 generations, haven't been able to reproduce the "start" glitch. looks like it's gone
It's been like that for 3 days, at least. I hope it's fixable, I'm losing too many credits not refunded.
Okay, so. Do we know if Chirp is diffusion based? I'm just curious. I don't need to know.
But, if it is... maybe the problems we're seeing with compositional decay aren't related to the model at all?
See, if it's diffusion based, or similar, it could be a problem of scale and utilization.
Other ai's have had problems like this. Where the quality of the output varies dramatically by the available compute.
Right after 3a, there was a pretty big spike in scale. That's testable, right?
Just thought I would put that out there.
It looks like the short generations have stopped. Thank you devs.
It sounds like your question is along the lines of something that was suggested previously, which a developer had responded to. So, in case this helps:
Q: is this true? is the quality of the clips delivered on a per-server-load basis?
A: This is false
oh thank god
I mean, loosely, yeah.
I'm not asking about per load
It's a common scaling issue with this kind of tech
That's the most specific answer I'm aware of. I've hit the "create" button during times where people have complained about poor quality and got good results.
Right, but I'm talking specifically about issues with the high and low pass.
unwanted harmonies and bleed
From the start of songs or when extending them? (or both)
17 to 43 seconds in
generally on the chorus voice switch
rarely at the beginning
but sometimes
yeah voices tend to get more fuzzy once you get into the song
about a minute into the song it's pretty much always a choir
I've got some sneaky ways around that
do tell
but there's still decay
personally i get the best vocals by using v2
just use what you know about audio filtering. This is where suno's insane mountain off things to train on comes in. Name filters and mastering software.
hang on, I have examples
my impression has been that the training data is like what you get from a web shop like spotify, which means genre descriptions work, but "music geek" stuff doesn't
stuff like "phrygian scale" doesn't work at all
Instrumentals are not instrumentals, there are always phantom words
tbh i've given up on getting any kind of usable vocals out of v3 (unless i actually want a choir)
So this song isn't anywhere near done. And getting rid of the harmony doesn't necessarily get you better vocal quality. Hp lp and notch all work though
https://suno.com/song/2271c45e-86b0-44ee-8052-ac82b2f93fd1
high-pass-filter, emotional female, hiphop, klezmer, low-pass-filter, upbeat, fiddle, acoustic, southern rock, country song. Listen and make your own with Suno.
well it's certainly more multi-voice about 1:30 in
It gets worse when the machine runs out of material
but there you're trading harmonies for gravelly
I'm stubbornly determined to get something useful done in three concats
even if I have to recast the whole thing
check the vocal at about 2:20 into this song: https://suno.com/song/1f6ad2c3-907b-40a2-bb01-2497b7f581e3
in fact, i didn't do anything here to get proper vocals... except, this was v3alpha
Ethiopian jazz in the phrygian mode song. Listen and make your own with Suno.
i think the difference is striking
oh yeah, alpha was amazing
at least towards the end. Took a couple of months to get it right that time too
one thought, the instrumentals have a machiney quality to them in this one. Voice sounds great tho
true
mine were like that too
yeah
some of that you can fix with mastering
no, but it's not hard to recast the musical piece these days
meanwhile the competition is gaining on suno. example:
just disregard the weird intro 😅
Yeah, but the competition who shall not be named also has model fry 40% of the time
lol
but you do get to say you were there
i bought the ticket and took the ride! and now it's bedtime
Not really sure where to post this but i was flagged today on youtube while streaming while generating songs with Suno
Just dispute it. It's an automatic process and all so far were false flags.
Wow!
You probably need to turn autoplay on, the ♾️ button in the play bar
No problem. yeah on song pages it will autoplay trending after the song ends in some cases. But that shouldn’t happen on playlists. More improvements to all that coming soon
If you make your page squished horizontally, click on your profile picture/icon, the trending badges (#1, #2, and #3) go above the profile menu rather than behind.
I believe you just need to set the z-index to 3 in the css class (element?) css-r6z5ec
fixed
wow that was quick
All of my playlists disappeared 👀
Temporary issue I hope ?
ok we're back Thank God
hi my songs where gone after creating them a day ago what to do?
Since yesterday most of my public songs have this picture, or there is no picture at all (access denied). It's not really a bug, but it's really bothering me right now. Can it be restored?
Only on the ending generation, right? How are you ending your songs?
Good Morning KakerMix
It only affects the songs that have been extended and put together with "Get whole song". Mostly about two or three Parts.
bro "phrygian scale" barely works on human musicians
The picture that gets applied to the whole song is always the last generation
I see that happen when I am doing an end generation, usually there isn't anything for Suno to grab to make a picture so you get a bird
glacial, ice age, frozen, slow motion, low, slow, chill, vaporwave, dreamy, digital master, flawless remaster song. Listen and make your own with Suno.
I get a bird because there isn't anyhting else
Nothing I use in the boxes can easily summon a picture
and I think it does lyrics box first, then if you do an instrumental it grabs it from song style
Since this one technically has lyrics it has nothing
Anyway! That means that your last generation that gives you a bird
There are a couple ways to get it to have a picture. If you generation an instrumental as your last clip, and you have room, prompt for a picture in the style box on that last generation
Owww, then I understand what happened: Since I have the complete song, I deleted the individual parts. 😕
My fault - thanks a lot
Oh man yeah
I'd suggest not deleting parts if you can help it
I get the idea to try to keep things tidy
But while "wow organization kinda sucks" is all I hear about those of us that never delete
I hear all sorts of problems with people that delete stuff regularly
I just have to keep the last generation right?
Or i have to keep all parts of the song?
I mean I keep everyhting because that's just how I roll, 'Library' is a never ending pile of generations I may never listen to ever again, and I like and save to playlists the most special ones
You don't have to keep the parts if you have a complete song, no
Okay, but then the picture is gone 😉
But besides organization there is no harm in keeping them
👍
When you play your song, it goes through the pictures though, right
The video portion that you get should 'play' them like a slideshow along with whatever lyrics and prompt used for that section
I can't imagine how the answer would be anything other than "Yes" but hey
Only if i download the videofile, or link to the complete song - anyway, the song can be played.
Right yeah
Can you link one of your songs for me? I want to see the end portion specifically
That gives you your sick awesome birdy
guitar, happy beat, trumpet, sax, single voice song. Listen and make your own with Suno.
the pictures are okay, may cause the videofile was created before. But on the website the pic is gone
ahhh, yeah, same here
My suggestion then is to keep all the generations used to make whatever your final song is and shove them on a playlist just for that song
Never have too many playlists
If you try to show the Picture in a new tab, theres no 404 (file not found) its a "Access denied".
Thats why i think it is restorable
Only Suno devs could tell you exactly what is happening
But yeah it seems like the image isn't gone gone, just a connection of some kind is
I also am decently sure that when any of us trash a generation it isn't actually deleted
i'll do so - many thanks 🫡
We're just told it is for our brains to go "ahhh yes it is gone"
giggle
It worked 100% of the times for me in V3a.
I only used phyrgian, in style, didn't mention scale.
Was this in the first clip of the song, or in Extensions (what used to be called Continues) when it worked?
First clip + continuations, with the same style. I'm sorry I wont be able to find it in the library, since I got too much there. It was Instrumental though. I tried something with V3 but didn't like the result so I didn't try more.
@unreal radish LMAO! Good comment on the phrygian scale for sure. I just happen to be an african music fan, and in ethiopia they use it a lot, and actually manage to make it sound surprisingly natural.
is bot online? Using Chirp gives me app do not respond error(
The Chirp bot is gone. The Suno website is its replscement: https://suno.com
i see, thx
Hey On the website I have click extend oon a song and filled in what I want to add but no generate button is visible. This is my 3rd song so I'm sure im doing something wrong 🙄
You need to have credits.
I have 40...
**user error, sorry. I jusr had to scroll down farther to see the damn button 🫢
hey i cant pay i have updatedt my billing method but cant pay 4 this month
Ugh, not sure if this is a bug or what but wasted +100 credits extending a song, every generated extension has this metal grinding sound which is confusing since I don't have that in my prompt.
Style used is simple, Ambient, EDM, nu-disco, female voice
I had to regenerate removing words individually in the hopes that the metal sound gets removed. No success..
There's an issue with extending words in lyrics. It sounds like Autotune. 10 seconds in and you hear it here: https://suno.com/song/6f537084-c77a-44e9-9763-a8e41b7126d2
This is what happens to mine - https://suno.com/song/8e139fb8-92d2-4c53-89de-84a2e32d4f15
nu-disco, female vocals,
song. Listen and make your own with Suno.
Starts at :22 - all the generations from the source of that, has it..
11 second file with 8 seconds of silence https://suno.com/song/21bbe0bd-c94a-4e28-8c78-33d2ee11ca54
whimsical, creative, energetic, catchy song. Listen and make your own with Suno.
I better stop first and just get back to it another time, I guess..
Not sure if its a bug but the songs are now missing timestamps
The sound is still very dirty, lots of white noise. much inferior to V3 until now, sound quality feels like V2 lately
The sound of the iron drums (cymbals) merges into a horrible high-pitched drilling sound, with a long sustained sound. It's not realistic to listen to, it hurts my head. This is definitely a BUG!
https://suno.com/song/631adfa2-031a-4066-ac0f-1e5c38c4f4f1
Metal Punk
Aggressive guitar riffs song. Listen and make your own with Suno.
Probably a similar effect as tagging 'rendered in Maya' to a Stable Diffusion prompt. If nothing else it adds variation. I may be wrong but it is likely that Suno under the hood uses a fine-tuned Riffusion model (which actually runs via Stable Diffusion). If so all or most of the tricks we use to produce better images should work to produce better music too.
Before the last big system upgrade it did fine on 12 bar blues...
I'm going to put in my vote here to add v3alpha to the list of options. Just do it like this: v2, v3alpha, v3. Then we can choose what model we want. For me v3 is going to cause me to unsubscribe.
hello there. um. im not sure if this is a bug per say. but ive noticed that any song generated thats around 5 mins and longer seems to lose audio quality becomming more and more quiater and tinny. if this isnt then please forgive me.
as im new and could eassaly be doing something wrong
A possible work-around for this (as you're new you might not know about this and it's not obvious) - after extending a song, use "get whole song" and extend the new song it creates (it doesn't costs credits to use "get whole song"). This gives it a longer clip to extend which can help.
so create whole song then extend the whole song. would i add content to that final extention?
Yes so you make the first part, extend that, then "get whole song" on the new "part 2", then extend the "full song", use "get whole song" on the new "part 3", and so on
oh yeah i do that. create vers and chorus sections then when done i get whole song. to create well. a whole song not multiple parts.
its the hole song losing quality thats the issue. sorry if that wasnt worded the rite way.
i can send linnk if that helps.?
It might be useful to one of the developers so yes, worth sending a link here.
https://suno.com/song/7d423c6a-3ee2-467d-a6da-953a6cd48d92 full length song
tavern song song. Listen and make your own with Suno.
start of deterioration from what i can tell
tavern song song. Listen and make your own with Suno.
To the devs: I assume you're pulling your hair out trying to understand why audio is wonky since the hardware upgrade. I don't know exactly what lies under the hood but I did see previously someone say that the model hadn't changed. If the software didn't change maybe you forgot to carry over a setting, if the settings didn't change maybe you should do a couple things to verify the source of the problem. First I recommend running checksums on the current and former versions of the model (even if it is just a copy); maybe in the transfer a few bits changed. It could also be possible that - given a hardware upgrade - there is a difference in floating point precision from the previous system. Potentially the former system had higher precision? I would suggest booting up the old hardware and comparing the output quality to the current as this could isolate and identify enough to resolve. An alternative would be to add v3alpha to our list of model options and we'll give feedback on what sounds better using the current upgraded hardware. Barring those suggestions if I am correct that you are using something similar to Riffusion you might want to make sure your spectrogram inputs and outputs aren't being jpeg compressed ^^. I'm well versed in the dark arts if you ever need advice feel free to reach out.
also thank you devs for such an amazing ... program? websight? A.I thingy. 😄
After a bit of testing today I'm fairly certain instrumental versions do not suffer the distortions; it seems limited only to generations with lyrics. Following the earlier suggestions to get a full copy of the song and continue from it did help to complete a song with vocals but I still feel like it 'barely' worked and dozens of generations just had so much metallic, twang or chorus added to the voice that it was not usable.
On a brighter note generations are really fast these days; so kudos for that. If you're making instrumentals it's great 🙂
It will depend on the genre. I get plenty of distortions with instrumentals, making many unusable,
It really does seem that the 'Extend' feature has really degraded since the way it was working two to three weeks ago. The quality difference between creating the first two minute song which is often good or very good quality compared to the very bad quality Sumo has been producing when Extending recently is quite a big difference. Burned through many credits recently and could hardly get anything useable when using Extend. Much distortion and sounding like instruments have been put through a sharp high pass filter, and the volume in extends suddenly going way too loud, etc. Has something significant changed in the Extend functionality in the last couple of weeks or so?
I've had the same problems for the past 2 weeks or so. Burned through a lot of credits to end up with just a few usable instrumentals.
Can we please look at why the time of track development has been removed. it helped with organizing.
You mean this thing?
for me it depends on the browser. chrome doesnt show the time, firefox does show the increasing time while generating
Yes, there was the timestamp as of yesterday
I never knew it had an actual time
Dunno why it's missing
I'm going to guess they didn't mean to
There are a few quibbles that make their way into the UI that certainly feel unintentional
Prob got removed accidentally during a release. Yea the modal hover over the UI when you delete a track seems like it should be placed at the top of the browser and not over the UI also
It's always been over the UI though, so that's just a choice they made early on
I too doubt that's on purpose in so much how anoyying it is
Would be a better experience if it was moved up. Also would help if the track that is playing is more predominately highlighted. There is a slight gray and orange text but it would be more user-friendly to highlight the entire row better. just a quibble
The Ui design on mobile is… questionable 🥸
I’m still getting mostly songs without any sort of structure in the melodies.
I waste lots of money on these. How do I not get charged for these
Who would've thought "folk inspired experimental avant garde psychedelic art dance pop, fast tempo" meant a somber piano piece that sounds like a kid bored in band class
Glad I don't pay, 50 credits a day, 10 "songs", and only like 3-5 are in any way close to what I asked for
Unauthorized withdrawal from my credit card! I had automatic debiting turned off, but for some reason you debited it and left a free plan in addition! I want to remove the card from your system and refund!
We here in Discord have no access to any billing information. Please contact billing@suno.com to get a reliable answer to this question.
Piano seems to be its default go-to if it doesn't understand. If this is in custom mode, try putting commas between the parts of your prompt eg. "folk, experimental, avant-garde, psychedelic art, dance pop, allegro, fast" - it seems to work better as a list like this, though it may focus on only certain items.
I recently tested the free song composition function of Suno. Today I found that some songs I tested before (within 10 days) were gone. I did not delete them myself. What is the reason? Where can I find them? If the generated songs are not downloaded, how long will Suno keep them on the website?
the clips should be in the library. deactivate all filters and look on all pages. also make sure you use the right login.
Thanks for your reply. I did everything correctly, but I only found some in the library, and some are missing. If I don't download it, how long will the music I made be kept on the website?
no clips are deleted automatically. if you accidentally deleted them they go to the trash bin. only if you delete them again from the trash bin they will be gone forever.
Got it, I'll look for it again, thank you. At present, all the songs I need need to be extended. If I keep extending a song, how many minutes of the whole song can I get at most? Is there any limit?
After each command (as well as after pressing any button) an interaction error appears
i have not made a good song in over 2 months, my finger is hovering over the cancel button. then ill hang around for v4 and try with free creds to see if any better, sound is poor distortions wishy washy, vocal chorusy. its like listening to a good track through a washing machine. such a shame i got 2 great albums out of this before it started degrading
Thanks for including more animal songs! :3
Maybe its just for me, but my Rhinoceros Song is 2 times in there? Either it got copied twice or maybe, you tried to put V1 too in there?
I'm guessing it is unintentional but with the latest change to the site, any song that wasn't renamed by the user is now "untitled" where before it would have a name derived from the lyrics
Now everyone's clips, unless they had renamed them, are all 'Untitled'
we’re looking into this, thanks
40 times I tried to start a song with the instruction "guitar riff intro", 40 times the vocals kick in right away. It's maddening.
should be all set now
Workaround, make the first part always as non vocal. and thyn start from this part and Time where the lyrics shouldbegin .
Spent much time and money for nothing. Could you make A.i following the lyric please, and dont repeat anything (User can build the struture) The "extend from" must be exactly and continue to linking the lyrics... ah....
Minor UI issue: When you go into the Library with no song selected and click on any of the buttons, the side panel with the song information will pop open and the action you were trying to perform (e.g. extend) will be ignored so you have to do it again.
Hello, today Suno loads almost completely then throws an error: "Application error: a client-side exception has occurred (see the browser console for more information)." This is for 2 accounts in the household and both of them are paid. Chrome browser.
Typical of results I've been getting. Not useable, sadly, so many tracks.
identified the issue affecting some users. fixing
try again please, when you get a chance
it's working fine, thanks so much!
Does anyone here know if Suno is planning on doing a large update any time soon? I noticed a huge drop off in quality. I thought that maybe it was just me. However, I've come across a bunch of articles recently that have made the same claim (not to mention the abundant comments here stating the same thing). To my ear, Suno has become unusable in its current state. Sad, but hopeful Suno will respond. Big Update coming??? v4?
maybe just RNjesus? i have had good results, but i've never used v3A, so i am not familiar with the rant
it also might just be familiarty, the longer you use it the more you notice the little flaws that add up over time informing your opinion
Model v3 was released on 21st March and won't have changed. If you have noticed quality degradation since then, this will be a combination of the prompts being used along with some randomness. The overall quality has not changed, but it does vary.
Also yes, familiarity with it - plus confirmation bias from others noticing and commenting on the same thing. A lot of people remember v3 alpha and might be comparing to that because they preferred how it sounded.
Have a listen to a few of your older generated songs for comparison. Developers are working on v4. No announcement on when that'll be available yet.
Here's one from about a month ago whch gets quite scratchy in parts: https://suno.com/song/b1a3a1a8-8241-48cc-b9e4-01b9de4c428f
Just clicked on my "view profile" link on the menu and was taken to the profile for a user named "undefined".
Did they have anything good?
That’s not great, let me take a look
Only happens the first time. I was able to replicate it. This is in Firefox. If I stay logged in, close the browser, reopen it and visit https://suno.com sometimes there's no songs showing on that page. At that point if I try to view my profile, it's undefined.
So also if I try to view my profile really quickly before the page has finished loading, I get it.
If the songs do show there, it's fine.
Thanks, that’s helpful to know. Should get patched soon
Subscription Payment Issue
I hope this message finds you well.
I am reaching out with an urgent matter regarding my attempt to subscribe to Suno AI. Unfortunately, upon trying to complete the subscription process, I encountered a persistent issue wherein my card was declined
Hi, you have to email support@suno.com to resolve billing problems. (Unless there is a new bug where payments are being declined, but I am not aware of such a bug.)
Yo whered my old songs go? I had two that i really really liked and now theyre just gone
Not in the library or anything
why i cant press active on free plan
it does not show anything when i click on it
helpppppppppppppppppppppppppp
I had 10,000 credits because i was subscribed to the $50 subscription ( 10,000 credits a month ) and now i somehow have only 850...
The question about "Active on free plan" above was answered in general channel and refers to the "Active" button on the free plan on the subscription page.
On the subscription page, does it show any extra credits and what date does it show for renewal?
I feel the same to be honest. I'm using the same prompts as always, but of late not getting anywhere near same "dark, indie electropop De/Vision and Mesh style vibes" genre results. It's not recognising vocal prompts too well for me either, creating that same generic mid Atlantic vocal with a ton of reverb etc. or that soft spoken Eillsh type vox no matter if I want a British baritone or German accent male etc. It's feeling very generic results of late no matter how I prompt. I tried V2 yesterday and although the results of the prompt was even more way off target as genre, the sound quality was far better IMO. I burned through a ton of credits yesterday trying to get a decent result but trashed them all. Hate to say it achieved what I wanted within two generations on"the other rival platform".
yo, i've found you can combat this problem by adding any kind of vocalisation at the start of a track. it's not perfect as a solution, but better than the current thing where the ai is really inclined to jump straight into singing (often without instrumentals) and ignore all instructions for an instrumental intro. i format it like this:
"[spoken] lets go! (or whatever other spoken/shouted thing)
[instrumental intro] [guitar]
[bridge]
[verse 1]
blah blah blah"
and have decent success rates in getting it to actually do an instrumental intro like you want
the spoken part is often clean enough that i can just clip it off of the finished track in audacity too
You can also start out by creating an instrumental in custom mode, then extend that with a vocal part. So far, that has worked very well for me. Right now I'm working on a long-form slow heavy djenty song that has a 2 min. intro which is basically one long big buildup, then instant drop when vocals enter. From there it builds again into a chorus and now I'm working on just breakdowns and buildups for a few minutes by extending the song with more instrumentals.
yeah that works well too, though i am loathe to sacrifice the initial 2 minute generation and only use 20 seconds of it (especially with the current problem of tracks degenerating more the more generations you stitch together for it). still a good way to guarantee a great intro!
Does it? I have noticed that sometimes it generates something weird but usually I can get a good section in 4 or less generations. Like the song I'm working on here is on part 8. One of the parts I used litterally just the first 10 seconds because it had a cool breakdown, but I didn't like where it went afterwards, however with another generations it pretty much nailed where I was going.
i'm lucky if i get one complete in 4 generations haha. most of mine are easily 6-10 to get the vocals and melodies consistent and the instrumental parts in the right places, and even with generations of 4 or five the vocals often degrade to being basically just noise by the end (unless i got a great roll and they were extremely crisp to start with)
Yea I can see that when you are generating clean. The genre of my song is sometimes very "noisy" so it kinda fits with a little bit of background noise. It has nailed the drums and guitar pretty much perfectly, but the synth and vocals sometimes gets close to over doing it, but I'm still very impressed by the fact that it actually is enjoyable to listen to 🤘
i'm lucky to be messing with genres where scratchy vocals aren't unnatural sounding too! yeah agreed, even with the current problems i'm happy to put suno songs on my playlists sometimes
Don't know if you are into long-form metal songs, but this is what I have so far. But you can hear how the intro goes on for 2-minutes which is common for the feel I'm going for, and also you can hear how there is usually a little bit of noise in the background which fits the genre but it gets close to too much. I'm up to 5:10, the rest I will probably replace. Also, the synth/piano peaks a little sometimes, might be able to fix with an equalizer 🙂
what time do credits refill? Today is my renewal date and credits havent refilled
One of the mods/guides could answer it more directly but it tends to be late at night EST time
It's a specific time
Something like 03:00 UTC (there's a way to display it in your local timezone apparently in Discord but I don't know it)
Got you covered fam (you can copy the text [via Discord context menu, not ctrl C] to get the timestamp tag included):
Daily credit refreshes at 0300 UTC+0 (<t:1713236400:t> in your local timezone)
What do I need to do to make songs longer than 2 minutes?
Click the 'Extend' button on the song.
Further details are in the FAQ:
https://suno-ai.notion.site/FAQs-b72601b96de44e5cacd2cd6baa985448
Hello! Please email billing@suno.com if you are unable to find the answer to your question on this page.
dont change Das
When I produce the song, it automatically gives it as 2 minutes. I want to do this in 2:30 minutes. how will I do?
This has been driving me crazy today. Im trying to generate instrumentals, but almost all the generations has vocals inserted into them. And I absolutely did check on the instrumental switch. (The prompts are weird because i use this to generate guitar backing tracks for myself)
The 'Instrumental' button is not a guarantee there will be no vocals. Not sure if that button does anything more than just hide the Lyrics box. Based on your 'Style of Music' text prompt of 'indie folk rock', which typically does have vocals, it appears Suno is inclined to throw vocals in there even though you did not provide any lyrics. I have run into the same issue on several occasions. Maybe try including the word 'instrumental' in the Style of Music text prompt, but that may not make much difference.
"The 'Instrumental' button is not a guarantee there will be no vocals" Is this stated somewhere, or is this just from experience?
Just from personal experience using Suno. 🙂
So its a bug, the button should do something
everything you tell suno are more or less suggestions. some genres and styles love to sing.
disable instrumental mode and write [instrumental] in the textbox to give suno something
When I produce the song, it automatically gives it as 2 minutes. I want to do this in 2:30 minutes. how will I do?
When you shoot something for more than 2 minutes, the button turns red and gives an error.
theres an extend button after the first 2 min version is finished. You cant generate one longer than 2min but you can hit extend and then stitch them together. This documentation is outdated, but the process is generally the same
https://suno-ai.notion.site/FAQs-b72601b96de44e5cacd2cd6baa985448
Hello! Please email billing@suno.com if you are unable to find the answer to your question on this page.
Just a FYI, the more up-to-date version is at suno.wiki:
https://www.suno.wiki/faq/making-music/how-do-i-make-the-song-longer/
Thanks, i see they even answer my question there
For now, I think the best you can do to try to address this is to include the word 'instrumental' as the first word in your 'Style of Music' text prompt. As you have found out, there currently is no guarantee Suno still won't still try to generate vocals, but it should increase your chances of not getting vocals, at least a little. Suno devs will have to answer on what exactly the 'Instrumental' button does, but it seems it may not do much more than hide the 'Lyrics' text entry box. 😯 🙂
Which raises the question, why the heck does the 'Help' link right on the Suno webpage not take you to the latest help/FAQ info??? 😄
if you mean suno.wiki.? That's because its created by us moderators
Right, but the question remains why is Suno not keeping their help/FAQ up to date? 😄 With the influx of new users recently, it shouldn't be a cash flow shortage that is stopping them from hiring a technical writer. 😉
They are working on it, but it's small team.
Yeah, but it is something that is hopefully a priority, but I understand the world is changing. What would once have been seen as a given in the software development world, is now apparently seen by some these days as a whatever... 😄
It is.. but development and FAQ cannot keep up. Not sure what to tell you more? Many other things that have priority too. 🤷♀️
I wasn't looking for more, I was just making an observation on the matter. What once would have been a given is now a maybe later... 🙂 😄
And I agree, but as a volunteer there is not much more for me to do in that regard 😅
Sure, but I wasn't expecting you to provide answers on that. Again, just making the point that it needs attention. 🙂
I was answering your questions 🤷♀️
I do know the answers since I've seen the new faq and I know why it's not there etc.
Sure, thanks for providing your input on that. Much appreciated.
Playlist remove and add songs don't work right way at mobile version
yes the multi select UX is confusing on mobile, we’re working on a fix
how does the multi select work on desktop? i didnt know it was a thing?
aaaah, shift+click
noice
I discovered you can also use Shift-left click to multi-select items on your Trash screen, to delete a full page of items in the Trash screen with one click.
All of my stuff on the website app like all my full songs that were older are gone. the clips are there for some but with no way to relink them or even try to extend them to rebuild them
s'funny, I just posted an hour ago making the comment that the quality in recent days is WAY up on before, I am making some bangers at the moment.
Conversely, someone earlier today was commenting in the technical-discussion channel about the quality seeming to have got worse recently. It really varies - the service itself can't have "good" and "bad" days, in terms of audio quality at least, since the models aren't changing.
HI I SUBSCRIBED SUNO FOR A MONTH. BUT BUYED YEARED SUBSCRIBED AND I CAN`T CHANGE IT.
You can email billing@suno.com
thx
Why is it only doing one song at a time now
The models need to get better though. Are they improving them?
They don't change the existing models. v4 is being developed.
Okay that makes sense
Anyone know why it’s so slow and only doing one song at a time?
https://suno.com/song/1c86d3de-c916-4780-8898-5e6f5c1f2a96 It doesn't follow the lyrics I add
warlike drums dembow song. Listen and make your own with Suno.
Report it as a bug
I've just done it
You will need to contact billing@suno.com
thx
The display name box should be limited so it doesnt overlap with the song list if a user has an extremely long name.
For reasons unknown, the following is not made clear on the Suno Create page 🙂 :
Important Note on Using the Suno Create screen:
Currently, at least, the use of meta-tags in the 'Lyrics' text entry box and all keywords entered in the 'Style of Music' text entry box are taken as "suggestions" only by the Suno model.
There is no guarantee that Suno will closely or even partially follow what is entered as meta-tags in the 'Lyrics' box and in keywords which are entered in the 'Style of Music' box.
Nor is turning on the 'Instrumental' button a guarantee that Suno won't try to generate some vocals in the song part it generates.
Likewise, Suno can sometimes also ignore some or all of the lyrics a person enters in the 'Lyrics' box, or otherwise mangle them in some way.
I have emailed but no response from them
I have a problem paying with a Master Card. I stupidly can't pay for the purchase of loans in Suno. Why?
I've already called the bank and asked what's with the card, they say everything is fine.
I wrote to those in support, so far nothing has been answered.
I do not violate the rules of the community, I work for my channel, pay for subscriptions and so on. But why can't I pay the loans? My creative potential is being wasted.
Please help me, I can't pay for the fifth day already
Sorry about the long delay, but email is still the only way to contact people who work at Suno who can actually fix or resolve billing problems.
Some users with payment problems have found alternative solutions, though it usually is related to a specific country. For example, I know some German users use the website https://n26.com/en-us and apparently that is free and works with Suno. There may be something like that you could try.
I was emailed and told to wait for a response. I hope they will solve the problem. Thanks for the answer
Heh)
Thank you for pointing this out, we will take care of this
i am pretty proud of my alphabet playlist that is so broken that it now has 4 scrollbars and is barely usable at 100 percent scaling 😄 i have to zoom to 75% to make it all visible 😄 https://suno.com/playlist/b41f7bc1-df29-44b1-a553-be5c418c7457
Listening to song... remove the song you are listening to from playlist.... try to play new song on playlist... doesnt work
I've also noticed this towards the end of a bunch of my songs..you get the "tinny" sound. Hopefully they are working on it
The ability to download your songs is also not showing up in Playlists.
Sure you are trying to download your own song?
For me it shows up in playlists
Yeah I'm borwsing my own playlists.
I did a Sign Out and Signed back in and I can download.
Ah yea then the authentication somehow broke partly I guess
Yeah I signed out and signed back in and that fixed it.
SUNO is having problems again. The sound quality deteriorates as it extends to the later parts. By PART3, the decline is quite noticeable, and by PART5, it becomes practically unusable. What's going on?
https://suno.com/song/a880865c-ab91-4263-821f-57412b1f3f44 Listen to this link, it's a example.
Pan flute, harp, Egyptian, a prophet walking in the desert song. Listen and make your own with Suno.
Why's this in the bugs channel? And the suggestions channel? And the technical-discussion channel?
Because it's a kick ass song and will burn the hairs off of your furry beans.
To be fair, I didn't put it in the memes category. So there's that. lol
Please move it to either #🐣┃suno-showcase or the #😜┃suno-memes channel (if it is actually a meme song.)
When did meme's have music? I'm 43, not up to date with the latest foolishness of you youngins.
When there's a song called "I Glued My Balls To My Butthole Again" you know that we've entered the era of meme songs...
Thank you for moving it.
This is the bugs channel and we need it to be clear for the Team to browse through when looking to see what issues are being reported.
Isnt that a bug in the lyrics moderation prompt? 😂
Any latin here? Chilean?
so a couple of things, and not sure if they are bugs. First, I know you probably get this a lot because it happens on tracks all the time for me, but the quality of the audio just gets worse and worse as the song goes on. Second, I made sure to do an instrumental portion to start and then only used a bit of it because I wanted a great guitar riff (congrats, it totally rips), and the riff pops up throughout the song, but then at some point, it won't comeback. I wanted to finish the song with the riff and tried all kinds of different ways and wasted a load of credits trying, but it was just gone. Here's the track. https://suno.com/song/07890138-247b-4912-ba51-f554b80511ed
rebellious gritty rock song. Listen and make your own with Suno.
You'll note the riff appears pretty far into the song, but then it just... isn't used again.
I have the same problem as you, the sound quality keeps getting worse, which is really sad😩
Is it me or the song generated is some put together sets of different loops ?
Also why am I getting pop tracks when I asked for something waay more elaborated?
Why isn't there way to negative prompt and why does it feel like the tracks are filled with noise as if the generation stopped at 95% of the diffusion steps?
Why are my credits not refunded when it provides 1/5 times boring pop tracks not even following anything from the prompt ?
Why does the prompt written gets deleted when you click on simple buttons ?
Why are a lot of synths sounding like it has been arpeggiated on default synths from FL v2?
Why does the drums get significantly worse the more you extend an idea ?
It's like the drummer got some arguments with the sound engineer midway of recording session and the sound tech decided to throw some bitcrusher on the drum mixing channel
Why is it so limited in terms of instruments ? A lot of exotic instruments are sounding like an insult to the real thing, completely robotic, badly sampled and mapped on a keyboard type of output
Also, why does it tries to take cash from the account at 27 days of the subscription ? I've never seen anything reroll before a full month
Wtf is this
Rerolling 12 times for 1 song part feels like it gets quite expensive and takes the fun out of the thing
Especially if the end choice is between : not interesting or worse
https://suno.com/song/be3de8a6-befb-4fa3-8459-74c9370d7486 I really wanted to extend this, but just keep getting a little box appear with (! An error has occurred. A system error occurred),
I reloaded page and, tried from a different point in the music, still getting the same system error 😭 (and Yes I do have credits left).
mixolydian 432hz with universe fx clear upbeat catchy fantasy wave natures drum beats electronic and space music clean prog-rock song. Listen and make your own with Suno.
Ok so I just tried to extend it as well and also got the error
however
If I clear out the 'style' box and attempt to extend it, it works
Which means it's something in the Style prompt box
Thank you for your help, will try removing the style prompts.
I am doing more testing
So, if you clip out a section, from the front, or back, it goes through
So then I deleted one letter at a time, right, from the end
Delete one letter, try to generate. Delete another, try again.
I had to delete, one at a time, rog-rock from the end of that whole prompt before it would generate.
(this is more for the devs rather than you @faint jungle )
Here are all the gens I got from your song:
https://suno.com/playlist/eb88ead9-0419-47f4-be6a-1c7b9883aaa3
You don't have to use them but I spent the credits so if you want them, they are yours
"trap k-pop" cannot be clicked in Explore because it is covered by "tuareg motown"
Thank you for your help, will listen and see which I like and work from there.
"baritone male vocal" apparently means soprano female vocal
Just so you are aware, you can be quite dramatic with your style prompts after the first generation
Suno already has the song it's building off of so it doesn't need the exact prompt every time
I didn't re-add the list, Suno auto did when I selected the option to extend, also normally it just selects a few prompts, not the whole list I originally put in before getting Suno to create from those prompts.
Right it does that yeah. Just you can change the prompt to whatever you want. If you are willing to experiment you can get lots of neat things
I like to put a few genre styles together and see what happens,. Although wasted a few generations, as Suno doesn't yet know what a Wurlitzer organ is.
It might but you may have to try to ask it for things a Wurlitzer organ would be in
But yeah, experimentation burns credits
Sometimes the results are not bad, when mix things up a bit, look at the combo I used here 👀 https://suno.com/song/35a65a9e-1741-4c42-88c5-ef3ca79aee5c prog-rock cinematic gothic folk
prog-rock cinematic gothic folk song. Listen and make your own with Suno.
Tried to create a Playlist and got this error - Application error: a client-side exception has occurred (see the browser console for more information). Any ideas? thx (That was going into 'playlists' and clicking directly on the create a playlist square)
investigating
I can add songs to an untitled playlist and then rename it/add details, but had the error using the way mentioned above.
Should be fixed shortly
it's fixed now. Thanks for reporting
I tried to extend a song with lyrics but the results are all fade out outro even i did not type any [Fade Out] or [Outro] prompt in it, please help me and tell me how to fix it, thanks
my suno is not playing sound for some reason
Sumo is not consistent. Only thing you can really do is try again.
@left citrusHi,he sound quality deteriorates as it extends to the later parts. is there taking care of it?
in my case some noise adds up, like disconnection noise
is your case the same?
This is known and hopefully will improve in future models. One thing I noticed is that sometimes it helps to continue from the last individual part instead of getting the whole song and then continuing from it. Of course, if this part is short you lose some context, so it's not a perfect fix.
I post a link, you can listen to it https://suno.com/song/a880865c-ab91-4263-821f-57412b1f3f44
Pan flute, harp, Egyptian, a prophet walking in the desert song. Listen and make your own with Suno.
almost all my song-images are corrupt and shows just "song image" and no picture. How can I fix that?
You don't have to do a thing. The team is working on this. It will be fixed soon.
hello, i know u get this question alot but do u guys have plan to release API access anytime soon. if yes whats the estimated time?
I do not know when. I am a volunteer, not staff.
aha ok tnks
Even though I made my new song public It did not appear in the "new" section
@left citrus um my suno isnt playing any sound idk why
check the volume slider, is it muted?
nope i tried adjusting it aswell
Do other websites play music? If you're on PC, is the tab muted?
its not muted, it isn't even showing the audio icon
but i might try closing spotify
nvm it was cus spotify was open
mb
nvm it isn't working again
@humble cobalt i tried closing my spotify and it worked for a second but that was it
ok i have to download the songs for some reason
I keep getting a "credits refunded" error when trying to extend a song. Says that it’s too short. When in fact it’s a full song with 4 verses and 3 choruses. ???
I have made countless attempts to extend songs beyond two minutes (and even hundreds of attempts to reach those two minutes, with little or no satisfaction at all), but each time a separate segment is created, sometimes lasting 10 seconds and sometimes even 1 minute . This way it is completely useless, furthermore even if you wanted to use a DAW to try to join the two parts, there would be the problem of timing and structural coherence, but even this is not possible because the new segment does not take into account the further instructions given and is only a stale repetition of the previous part. Is there something that can be done to overcome this inconvenience, because among other things I have made a subscription and I am practically wasting time and purchased credits. Thank you.☹️
oml i had to wait for the songs to load im so dumb lmao
After you've created the next part you can use "Get Whole Song" to join it to the previous parts: