#๐โsuno-school
1 messages ยท Page 27 of 1
I think you nailed it... It had never occured to me that, that might be the case... but thinking back.. yeah I bet I did just that..
i kinda had the same issues but tried the same solution as bitsalot was saying, but just burned tons of credits in the process
yeah I toasted about 200 creds.. but ya know.. my first experience with Suno was so moving and positive.. I can easily look past my own ignorance... I am still very much a happy customer ๐
i'm like you: i had a great initial experience and managed to get some really cool tracks... i'm totally hooked to this now, i think i need therapy ๐
Wait till you burn half your 10k sub credits in a few days and then need to budget them for the rest of the month ๐
easier said then done.. once i'm in the mood, then i can't wait 'til i got my masterpiece together..... ๐
galaxy brain: can't hear the static if you can't hear the music
https://suno.com/song/6132686b-afba-4468-8b1d-88ddf3817afe
field recording, beatbox, chiptune, Cappella, Beatbox , chiptune, A Cappella, virtuoso, field effects Cappella song. Listen and make your own with Suno.
How can I create educational music and voice-overs that are not too musical, such as riddles or nursery rhymes? What styles can I use with this?
I have a question. Does anyone know any keywords to help with generating a song that can loop back into itself? Is there a keyword I can use or something to tell suno one of the requirements is to loop the end and the start seemlessly?
There isn't. If you try style prompts like "loop, looping" you will get extra repetitive music that sometimes sort of loops by accident, because it's very repetitive. But that's almost certainly not what you want unless you are intending to make elevator music.
The best you can do is basically try and have the song repeat a verse exactly, just like it generally repeats the chorus.
Why this error occured?
I am guessing you used the "reuse prompt" option on a full song. Which results in too many lyrics. Remove some text in that case.
It doesn't stay on the same vocal, and there is no Chorus in every track. Why is he doing this? The piece starts well, but later on the vocal changes a lot. ๐
Wow I am clearly outclassed \ Like a rowboat with a 22 rifle next to a bunch of battleships... I will just slowly row away.... to the kiddy pool ๐
Is there a way to integrate Suno into a website? We are running a digital campaign and wanted to integrate suno on our website. Can anyone help me with any pointers ?
For that, you need an API, but that is not ready yet. For interest, you can send an email to hello@suno.ai
I am a volunteer moderator, so I have no sight on when, where or how.
Thank you Daizy! I just emailed them.
i've extended a track but it produces a separate file. can i combine them in suno? or do i need an external editor?
tracks end with no outro or smooth ending. is there a way to fix that?
Click "..." then "Get Whole Song"
I was making a song in custom mode, but no matter how much I entered the text, I got an error. I'm doing the generation on the site for the first time (It never happened). I used to do it in discord, but I can't find models now
You are new to the website? This might help you understand: https://youtu.be/lj8utBd_IUM might be able to help you
You can turn the subtitles on in your own language for better understanding
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...
I'm trying to ceate a song, building the Intro, Verse and Chorus etc. But I'm trying to create each part as I want an not another familar autogenerated piece.
The Intro was easy enough. I just used an instrumental based on my Style of Music. I would have liked to add my own direction, such as Beginning with a soft piano and ambient strings to set a tender, evocative tone. I added Soft Piano, Ambient Strings to my main Musical Style, which worked ok.
Verse: This drove me crackers.... I entered just the lyrics for the first verse, but as I'm sure you'll already know when you Generate the piece, Suno has a habit of not just singing your words and then continuing with instruments but starts ad-libbing,
I found a way to stop this by simply adding [Break} after the Verse.
I then decided to, Get the Whole Song as this doesn't cost anything, and you can extend from that.
I listened to this and decided where I wanted the chorus to come in.
I clicked on Extend, set the time I want to continue from and tried to make a few changes in the Syle of Music to develop the fullness of the song. This didn't work.
So I carried on, selected the Get the Whole song and started to work on the Bridge.
I tried Instrument, removed Male Vocalist from Style of Music and added Lyric Tags, but nothing worked. Whatever I did. there was a voice!!!
I haven't been able to get past this point. I can't create another piece and add it to the parts already completed; I have to use Extend. I'm stuck. Help!? Please!
I found suno.wiki, which gives a bit of information: Suno.wiki is worth a look!
Here I found and several Lyric Prompts
[Break]
[Instrumental Interlude]
[Melodic Bass]
[Percussion Break]
[Syncopated Bass]
[Fingerstyle Guitar Solo]
[Build]
[Bass Drop]
The [Instrumental Interlude] seemed to work better for the Bridge section. But would like to be able to Change Key, Change the Style of Music options, but doesn't really seem to take any notice. Maybe it's just me! (edited)
Also i cant chose the version of ai, make a playlist >_<
regular reminder to not forget about V2. v2 can go hard as heck
https://suno.com/song/57da5bc0-a32d-4404-8609-40de3661771b
zombie sound effects, zombie sound effects song. Listen and make your own with Suno.
Have to be a bit more specific about the error you got and when. If iyou were trying to generate lyrics, there are two typical errors you can get:
-
Moderation error - The lyrics are generated using ChatGPT and it didn't like something (e.g. abusive phrasings). Even if it worked before, a small change can shift its learning pattern. To figure out its issue, ask ChatGPT itself something like "What issues are there with the following lyrics: [insert]"
-
If you get a system error, it is likely because the prompt is too long. Cut it down so it can make approx 2 verses + chorus
For a list of errors and some solutions, see https://www.suno.wiki/faq/errors/an-error-occurred/
Strange, I can select either between v2 and v3 on the create screen (just above the submit button. Can't generate it atm due to being out of credits)
For making a Playlist:
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. (For pictures of the selection issue, see #๐ชฒโbugs-web message)
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
If you do not have a playlist, do Add to a Playlist > Create Playlist
Discord is the easiest way to communicate over voice, video, and text. Chat, hang out, and stay close with your friends and communities.
I keep hearing from people asking โฆ Is the quality going to come back to suno any time soon? My question is : is the lower quality output in recent times more an issue of automatic training of the ai using the up and down thumbs to filter quality issues ? Are the up and down thumbs for rating a song automatically calculated into the algorithm? If so then This is a problem because many people donโt use the up and down thumbs properly thus distorting suno LLM adaptation when learning. Hopefully they fix this issue. If this is not the case . Does anyone know what the issue of lower quality output in recent times is from ?
There are 2 models at the moment ("v2" and "v3"). They do not change, according to what one of the developers said here: #๐ชฒโbugs-web message
"v3" was released in March, so there won't have been any changes to it since then. "v2" is the older model from before then, and remains available (and has not changed.)
But there was a difference in how songs worked and sounded when v3 was in beta compared to now. the quality was way higher but the structure had a harder time and just jumped into chorus and verse randomly . The structure bit is fixed but the quality is significantly different from v3 beta . Was v3 beta different then the released version of V3 we have now ?
Ah! Yes. As I understand it, "v3 alpha" was a separate model. When that was available, users were encouraged to use the thumbs up/down buttons. It was a while ago now but I'm pretty sure this was used to help guide the training for the final "v3".
anyone know how to generate a metalcore song that's audio doesnt distort and peak constantly. I can't get a clean metalcore/screamo/alt-metal song. It's always just too loud
How do I start afresh? I've tried refreshing the page, but the old lyrics and music style are still there. I'm not sure that just writing over the lyrics and Style of Music details will give me a fresh, new creation. I hope that makes sense, and I hope someone can advise.
Square Brackets [ ] seem to work sometimes and not others. Does anyone else get this?
I usually flick Instrumental on then off. It clears the lyrics prompt but I have to manually clear the rest
I want a song ownership transfer
I used my main to make songs but failed, so I tried using an alt and I made a song I really like
Now its stuck on my alt
Thank you for your prompt reply! You're right about the Instrumental Button trick... Strange.... We could do it with a Refresh Button. ๐
Just Extend from your main, I really doubt the dev team will transfer it for you (and you will probably get banned for asking since making alts are against the ToS IIRC)
I know
๐ Well, the retainment is a new feature (like it popped in within a few days ago), but yeah... wish we had a clear button. The number of times I accidentally extended when trying to make a new track...
And I wish they used checkmark for selection instead of background highlights, took me a while to realize it was selecting...
Refresh: I'm using Chrome Browser, and I cleared the cache and then Refreshed the Page, and I got a completely empty Suno!
Tip: If the details are not clear in Lyrics, Style of Music or Title, Clear your Browser's Cache.
I hope you don't mind me asking, but what is the 'Checkmark for selection..... referring to?
Can they please stop generate 2 songs at the same time
Sure. Currently, when a song is selected on Suno, the background gets highlighted. Issue: this is not clear at first and when you deselect, the highlight just gets a bit dimmer until another track gets highlighted (see #๐ชฒโbugs-web message)
This caused a few problems when adding or removing from a playlist as several unwanted tracks got added/removed until you realized about the background highlight is Suno telling you the track is being selected rather than actively played (which I thought at the time).
It would be easier to understand to the user if a small checkmark appeared in the top right when a song was selected rather than having the background highlighted
Discord is the easiest way to communicate over voice, video, and text. Chat, hang out, and stay close with your friends and communities.
They are both the same and it wasted half my credits
Now I know what you mean! I've just realised several tracks/creations have a brown/yellow hue.
When you click a track, the colour change isn't really that useful. So, I agree with you, a Tick would help!
When I've got several tracks developing, it's hard to keep track of which belongs where, if you know what I mean.
Definitely know what you mean. I started making Finals and Work In Progress playlists because of that... and learned about the selection thing when I kept getting duplicates added to the WIP playlist ๐ซ
That's a good idea. I just found myself trying to change the Title every time, but I still got myself mixed up! I'm sure there must be a better option.
Hm maybe suggest a tagging system at #๐กโsuggestions-web? It will likely be a while before it gets implemented, if ever
I just resorted to making multiple WIP playlists, a few for my main ones and a "Other WIP" for my junks (and one for my bangers lol)
They have to keep improvising and improving, as others are on Suno's heels! If Udio what I mean ๐
I've never used Playlist, I suppose I should have a go.
Am i the only one or does credits just start randomly dropping whenever i create a new song and listen to it on the Suno website on my phone? For example i create a song and i have 50 credits before it, then it drops down to 40 right? But when i listen to one of the songs, the credits just start dropping 40 to almost 0 without me doing anything. I have to refresh the site everytime for the credits to return to their original values. This doesn't happen on PC though.
Just a visual glitch, likely due to one of the async function during generation. Refreshing the page corrects the credit amount
is there a probleme with the generation ?? I got flagged to moderation, something that i was continue several time and it didn't shows up.
That usually means that ChatGPT didn't like something and kicked it back. You can usually submit the lyrics and ask it on its own site what was wrong with the lyrics to be moderated and it can usually tell you.
it's impossible bro, cause i was just remix the results. I still got some results with the same lyrics
but now it doesn't want to to regenerate
Try altering the wording.
This is the type of thing I've done to work around a false positive.
The [throbbing vein] metatag was me just throwing something random in there to make it different.
The spaces were attempts to change the texture of the prompt to the AI.
hum
I've also added stuff like [This is public domain!] and [I'm quoting the Psalms! Leave me alone] and such... ๐
I really think they should remove that. Because when it's buggy this is crazy. I already take the whole song. i'm just doing the remix in an other genre of music. Same lyrics. But it doesn't want
I hear you.
I don't understand why there's a moderation for creating art. We are note playing fornite.
I lost my bleep one time when I was working through some public domain lyrics I was reworking into a different style. I was on the third edition of the song (I had made it twice already) and was rounding the corner on the final chorus when I ran into moderation.
(โฏยฐโกยฐ)โฏ๏ธต โปโโป
So trust me, I get it.
Crazy bro crazy
It was crazy making. NOW you have a problem? NOW???
same same
On the other hand the fourth version was fire. I was glad that I went through it refining my approach on that one.
Like I said, change things up a bit.
Look at places where good adlibs could fit.
(Oh, yeah, baby, mmmmm, etc)
Try to add an instrumental where one should go. See about trying to resume from an earlier time. All these things to change the texture of the lyrics.
i'm just gonna give up, don't wanna waste my time. i don't even know which lyrics have a problem has there wasn't a problem. And not i got to waste credit to know which lyrics NOW have a problem ?
I understand.
That's why I suggested running them by a ChatGPT and asking it what could be causing the moderation to have an issue.
But then I remember when I ran into something like this with v1 Chirp. Killed my song since I was composing it as I went and I never got back into that vibe. And it objected to my lyrics calling a lover who had rejected the song's protagonist as "stupid" twice.
A week goes by and it was fine. But I never did complete the song because it sapped my muse by having the delay.
(I had a few lines I'd written and saved but no new inspiration.)
what's the question, i have to ask to Chatgpt ?
If you were moderating the lyrics to this song, what would stand out as an issue to be moderated for you? Here are my lyrics:
XXXX
XXXX
XXXX
That's how I do it. Others may have better questions they use.
there's is the problem ( that wasn't there 1h ago ) : " The lyrics you provided seem to be based on biblical events, particularly the crucifixion of Jesus Christ. When moderating these lyrics, here are the points that could raise concerns:
-
Religious Sensitivity: The lyrics reference events that are sacred to Christians. It is important to treat these topics with respect to avoid offending religious sensibilities.
-
Use of Contrasting Musical Styles: Using a "DARK ANGRY RAP" style in the context of Jesus' crucifixion may seem incongruous to some listeners. This could be perceived as disrespectful if the tone or musical style contrasts too strongly with the serious religious subject matter.
-
Artistic Interpretation: The lyrics seem to offer an artistic interpretation of Jesus' crucifixion, which is acceptable if done respectfully. However, some aspects may be seen as unorthodox by certain religious individuals.
-
Imagery and Language: Graphic descriptions such as "the accuracy of a spear in the side" or "pure water and carmine flow in a sublime stream" may be considered too violent or graphic by some audiences.
-
Mix of Styles: The use of terms like "JESUS cri" and transitions between different musical styles could pose an issue if it seems to trivialize or misinterpret the events described.
In summary, the content should be approached with caution and respect for religious beliefs. Special attention should be paid to the use of musical styles and language to ensure that the lyrics are handled with respect."
Number 1 . So christian are offended by the Bible then...because it relates that event....that's crazy...WokeGPT. Number 2. I wanted it that way what's the problem ??
๐คฆโโ๏ธ
It's a robot...
It says that but my money is on the actual issue being words like blood or violent imagery. It doesn't know the difference between a religious accounting and space monsters massacring alien bugs.
These bot only want song about the blue sky and blue sea
I understand now...We're not living in a democracy anymore #RICKGRIMES.
I'm gonna make a song about that i feel that i'm a bird. But ChatGPT won't ask me to consult a psy. it gonna told me " Yes you are ".
It's a thing. Remember we're in the middle of a moral panic with regard to AI right now. And prior to recently the AI was running as an app on Discord, so the content moderation was necessary to avoid issues with Discord's TOS. Now that we're on Suno's own website for generations it's a bit different but no one wants to have some horrible songs made that people can point to and claim as a reason why AI shouldn't be allowed.
Please post that one when you make it. It sounds like it could be good.
Yeah but, people can do it anyway by themselve. Not everybody listen to Dark Metal, only those who likes. And if there's a problem, the problem is not the A.I, is the person that wanted those lyrics. The problem was it Mein Kampf or Hitler who wrote that book ?
If someone paid, the music are theirs, so it's theirs responsabilities. But i understand why they do it. But i think that in Art, it's difficult to say : " don't do that, don't say that ".
In the coming years it will become easier for people to have a local model for various open source AI on their PC, and they will be able to customize it to remove any safeguards or restrictions. It's already possible with Stable Diffusion. It'll happen with music too.
True. On the other hand, have you heard of the 'getting crap past the radar' trope?
It is my personal belief that people striving against limits is something that leads to some of the best art.
no, what's that ?
It seems to be best in comedy, but I personally think it works across the board.
Creativity inspires more creativity.
True
Why do we have a several music genre ?
because of creativity
They started has experimental until it mature
Question regarding the thumbs up/thumbs down. I know it helps train the model as a whole with what is considered "good" and "bad", but does it also affect your own future outputs?
In other words, if I thumbs up songs with certain elements, does it make it more likely for my future generations to have similar elements since the model is learning I like that?
Just discovered Suno, Is there a way to improve the song quality? Many of the songs sounds like very poor bit rate. Also some are exaggerating on the bass
The sound is at 192
We got to wait to see if V4 will go to 320
If the bass is exagerated, on the list of the style put it one or two position to the right . exemple : instead of : " bass, percussion, rap , piano " do : " percussion rap bass piano ".
What languages does the AI know how to pronounce? I mean does it understand how to pronounce Bulgarian if I feed it some Bulgarian words?
*shrug* Who knows specifically what languages it was trained with but it sure knows a lot.
Pronunciation is a different story and it depends on luck in some cases. I think Bulgarian would be fine from my assumptions.
I know several Chinese people complained about their Mandarin lyrics being sung like Cantonese.
Worst case scenario, you would need to phonetically spell some words to make it sound a certain way
I attempted to post this question in here once before, but I'm not sure if a moderator removed it or if Discord failed to send it. Since the last site update (probably just coincidence as that looked like mostly frontend changes), I've noticed my generated songs are getting less and less realistic vocals. In addition to sounding more robotic, every Extend seems to add more and more harmony voices until it turns into a choir. There's an excessive amount of reverb, and I'm getting a lot more noise. It's reached a point that I've stopped creating for a while to save my credits in case a fix is being worked on. The reason I'm not posting this in the bugs section is because it's already been reported there by several users, and I've already seen some proposed fixes to help with prompts. I'm posting here because I'd like to understand a little more about why this may be happening, and if Suno has a long-term plan for addressing it. I'm in some groups who discuss creating with Suno, and I've been seeing a ton of reports of this becoming a bigger problem for a ton of users. Does the music side of Suno learn from all of the prompts it's being fed from users, or is it statically pre-trained like ChatGPT? If it is actively learning, is it possible to restore to an earlier state if the training goes off-course and causes a degradation in generation quality?
Recently there was a site update and a database migration that I remember reading about.
The models (currently "v2" and "v3") don't change (I asked one of the developers previously about this.) v3 was released in March - it behaves the same now as it did back then. I have always had mixed results with it, sometimes really good output, sometimes poor output.
The "choir" vocals you describe have been a common experience with v3 since it was released. Hopefully it'll be addressed in the next model (v4) but there's no release date for that yet.
Hi, is it possible to choose, for example, in each verse of the song that I want a woman/man to sing?
I was trying but I couldn't, does anyone have an example that could help me?
I'm trying to do it like this:
[Verse: Voice Men]
....
[Verse: Voice Women]
......
[Chorus: Voice together]
....
( This is for the duet )
it's difficult to avoid the multi voice
The problem is that it always comes with a woman's voice and never with a man's.
So I don't even know what I can do anymore, maybe just leave a voice XD
Say it in the "style of music" like male voice or male vocals
It works better in the Style of Music that way
And I have some examples for duet.. let me find it
https://www.suno.wiki/faq/style-and-lyrics/duets/ has information and you can use mine as example to learn from.. this is a duet I was working on: https://app.suno.ai/song/5045e5f8-d213-4cac-9d87-99b370e89adb/
intro, rockabilly, minimalistic song. Listen and make your own with Suno.
I use a mix of metatags and style of music to let the AI know
but it is not easy
I've seen it, I'll try thank you :)
It seems that we all encountered similar problems. Before reporting such problems, I was told that this version of the AI โโmodel would not solve such bugs. After a lyric is repeatedly generated, the AI starts making various weird mistakes in order to avoid "repeated generation". It seems that we are driving the AI to go crazy. With the current ability of SUNO, when the style label is clear, the song does not need to be repeated too many times. The generated melody and voice can be matched to the lyrics. At this time, the AI should be manually reset by the user to try new music styles, or the AI should be limited to a certain melody and voice for fine-tuning, rather than letting the AI force itself. Constantly creating "new" melodies (sometimes the "new" understood by AI just seems to simply generate messy melodies and bugs). Of course, for users of the CJK language system, the most urgent thing now is to enhance the accuracy of pronunciation. I have lost About 3/4 of the points were contributed to the AI's mispronunciation of the first three words of the first line of the lyrics.๐
Ive had a similar problem with losing points, but for me the AI really seems to want to skip entire verses, or start them halfway through
In addition to becoming a choir overtime
Those things are because the AI sometimes lost it's way in the songs. AI is still very unpredictable. But a lot of people here who are working some time with it do see patterns to get it back on track. One of these things is going back a second (1 or 2) and continuing from there, or sometimes even bringing back the last sentence so it understands where to continue. It all depends on what you are working off. It's still not perfect.
ive noticed some of those, its a fun tool but a bit finicky. Ive had a couple iterations where its entirely shifted melodies based on where I put the continue from part
[Sprechgesang]
ๆ ่ฏด
[Fade out]
[Sultry]
โๅฝๆ ไปๆฒกๆๆฅโ
[Drum Solo Interlude]
If you divide the lyrics into segments, it is easiest to generate one at a time. Trying multiple singing styles in a generated song usually depends on luck. So far, I have asked the AI to try four singing styles within 2 minutes at most. It is safer to refer to SUNO WIKI for the tags that work.
I created whole songs, till this very day.. It's like learning a new language.
rock slow sad with three backup singers song. Listen and make your own with Suno.
for sure. Im happy with the ones I have so far, and dont see a reason to stop
My "technique" is to make the lyrics Part small enuf so it ends on an [Interlude] and the next part starts with [Interlude] , That way at least there is an unnoticeable join between 2 parts.
Nakasi, 1980s, Taiwanese Traditional, Hokkien Song, ๅฐ่ชๆญ, sad, Enka song. Listen and make your own with Suno.
It will get better, but sadly there is still lots to learn. Some people expect perfection and a manual. But there is none. And it is not like the Devs are withholding us from that. they know what they have "learned" it but that doesn't mean they know exactly how to talk with it. We are all reverse engineering. Which is fine for some people but not for others.
@lone oak most of us do not understand Chinese, so it really hard to help you.
Take a look at the tags and lyric annotations of these two lyrics
We cannot understand it, so what we see is not what we hear.
I know๏ผ too sad
๐ฆ
I understand now, thank you all for your help, I'll try to do this to finish :)
But no problem, I could more or less understand what I should do
I tried it with a moderator the day before yesterday, and everything went wrong in hundreds of tests.
I do like what I hear if that counts ๐ฌ
Yes, a language barrier is hard.. especially if you also write in different characters. Really sorry. ๐ข
Been desperately trying to extend this one. Got a working part two but extending past that one isnt working out so well
rock anthemic song. Listen and make your own with Suno.
extend it from before 1:06 ... after that suno will probably try to make a end. and use the same style as in part 1
No double posting please
That part extened ok, i just posted the root song as it came out the best
its the pt2 im trying to extend off
( https://suno.com/song/43cc939b-ba26-4ae6-bbf1-9cbcad74957e )
rock anthemic song. Listen and make your own with Suno.
ive tried at basically every second from 0:48 - 0:52
Try making the full song from pt2
add [instrumental bridge] after the lyrics and it should make some music after the lyrics
Then extending that?
ill give that a shot, thanks!
hopefully it stops skipping the verse I want it to start off lol
also try what alikan suggested... make a whole song from part 1 and 2 and then extend that
is there a way to cut off where I want the whole song to end? the Pt 2 i made starts up vocals near its end
you can set the continue time before the vocals
for the last part use the style "end" or "end, short" and in the textbox [end] either standalone or after the lyrics... that should initialise a end
Thank you for that clarification!
ill try that too! so far [Outro] has worked for me, but maybe its bein wierd and I need to change it
outro is good too or ending.
That makes a lot of sense. I hope as this tech continues to evolve and improve, we'll eventually have a function to tell the system "this generation is good, but try to tweak timestamp X:X inside the clip to do Y instead". I'm sure something like that is pretty far off though, and will depend on a much more reliable model
For me, in this version, if the AI can correctly recognize the language of the CJK region, I will be happy enough.๐
The thing that always kills me is when it has an absolute banger of a generation, but cuts it off mid-sentence or word. Extend never fixes it, and since it's not allowed to repeat a generation, you're often doomed to never get another one as good.
yeaah, theres been a couple times where I love absolutely everything about a generation, except it mispronounces a word or cuts off a bar
I have to give Suno credit on that though, it runs circles around a certain competitor when it comes to pronunciation. Everything I've ever tried on the other site is loaded with wrong pronunciation.
for sure. Even for words that ive made up, it does pretty well
My favorite side effect of it misinterpreting some words is that it messes up turns of phrase in the most hilarious ways. My favorite so far has been "too many cooks in the stew" ๐
XD
๐คฃ
well it took 450 credits, but I finished Thunder in the Skies
This sounds like "in-painting" which has been suggested as something that may be coming with v4.
guys,, how to extend a music? I dont have the "continue from this anymore" in my desktop/web browser,,, the only option is "Extend",, and this options, completely changes de song
It got renamed to Extend and worked on a bit (and Remix renamed to Reuse Prompt)
https://www.suno.wiki/faq/making-music/how-do-i-make-the-song-longer/
Check out the related wiki links for other topics regarding the Extend topic. Like changing the timing slightly can help change the output
what is the different between unlike and move to trash?
Unlike just moves it to another section in your library and hides it from unmarked and liked songs. Move to trash removes it entirely from your library but holds the clip there until you hit delete permanently from the trash.
got it, thank you!
No problem. n_n
TranslateBot
Only too short music is made. Even if a two-minute piece is made, it ends up not being able to be extended
I have the same problem, the song ends and there's nothing I can do
No, no, no. The song is already over. The extension is meaningless
you can't extend songs, you just end them right away
When extending, make sure that the Extend From portion appears.
the thing is that when I create a new song, I don't finish it, I just finish it and there's nothing I can do about it, because when I want to continue, there's silence or a completely different song, I finish the song on my own at the very beginning, it's never been like that
You cannot create a song longer than 2 minutes
The timing for the continue from is important. If the style changes, try going back a second or two more or to the beginning/end of the nearest lyric.
The AI has to relearn a minute portion of the song so making these small changes influences the outcome (I think this was touched on at https://www.suno.wiki/faq/making-music/extend-from-time/)
I have a song I am currently extending (at the 4 min mark now) and did not have an issue of it becoming very different - other than getting more reverb x.x
mutes the recording in the initial phase of creating the song
This has never happened before so something is wrong
It should be important to note that the base song goes upto 2 min. Each Extended part is 1 min and continues from the time you specified.
Note the key word here, part. You need to use Get Whole Song combine the parts together to get the... well, entire song combined together. This was mentioned in the link I gave earlier.
This option appears on the Part 2+ ... option, not from the base/extended song option
I know all this, it's about something else
everything was working fine this morning
Hm, yeah, sorry, I do not get it without an example.
I think that more people will report it and maybe someone will explain it better. I know that it doesn't work as before, so you can't create songs longer than 2 minutes.
- It's the first part of the song.
- The first song created ends too short.
- The song has a period on it.
- The song cannot be extended.
- Even if a two-minute song is produced, it's the same as above
- The two-minute period is already marked, so there's no point in extending it
- More than a few short ones, I need one longer one
-The first song and the music at 1 minute 59 seconds ended at 1 minute 55 seconds, so there is no point in extending it.
You don't need several short pieces of music, you need long pieces of music that can be extended, but it's already got the finishing touches, so there's no point in the extension
I have the same problem now, created maybe 30-40 songs and they all end up shorter than 2 minutes and nothing can be done (tried different prompts etc.)
Let's say song is finished at 1:30 on first gen, even when I extend it and start extending from 1:00, it still finishes in less than a minute (fade out)
there is exactly no way to finish the song
yep, something broke, maybe devs should turn off and turn back on their machine lol
This would be amazing because you can get a song that hits really well but you want to tweak things around!
If the problem is solved, where can I check the notice?
Does anyone have any suggestions for minimizing the weird screeching reverb that always seems to happen in the 2nd half of songs? Could a good audio engineer do much about it?
I wish the like button on generations kinda acted as a guide for next generation with the same prompt to say "I kinda liked that, keep going in that direction" But rn I just use it to see which ones were the best ones so far.
riddle me this batman, how do daily credits on the free plan work? like, do i have to be logged in to recieve them daily? or? ^^''
its all good if nobody has an answer rn <3
I'm a very experienced producer and even using the best software (I use RipX Daw Pro) doesn't help to get rid of that type of noise that Suno produces. I think that the noise builds up with every extension until it's too much for a good recording. It varies with genres, I find.
Is there a way to recreate the same track but just alter the text a bit? If you generate a good track but want to just alter the text a bit without loosing the instruments.
its starting to sound like a tube something has gone off the deepend
I'm getting so many Credits Refunded when trying to do a 2nd extension today. Problem with Suno somewhere? I've wasted so many hours and so many credits today trying to build songs.
trying to get an extension beyond a min is hopeless
Then you are trying to extend from a spot that the AI feels like it ending. Try changing the continue point a little (1 or 2 second) and try again from there.
naww thats not how it works or has not worked i give an ending this way [outro] or [big finish]
I've been with Suno for months and have never had this problem.
Still, try it.
I got a 1st generation of 01:39 and an extension from that of 00:42, 3rd generation from 00:42 is 2 x Credits Refunded
What happened when you changed the time stamp, was that the 0:42?
I tried both time stamps
Trust me, I've never had these problems with Suno before yesterday. Something is not working properly. I've been with Suno for months and months and made 100s of songs
I am not sure what to say here. I am a moderator and guide, I have 2 premium accounts and I burned over 20k credits per month. I know these things happen. I have written a full FAQ for Suno which will go live soon, and I make tutorial videos. If you ask people here who have been on this discord for a long time, they will also tell you this happens. But if you insist it never happened to you and there is something wrong with Suno, then I cannot help you. Feel free to leave a bug report then in #๐ชฒโbugs-web
Thank you. I will leave a report. Devs can see my library, I'm sure, to view the number of CRs.
Does anyone have a better idea for keeping track of the song you're working on?
The Creation interface isn't that intuitive.
I tried Playlist. I was hoping I could create a playlist for a given song and work inside the Playlist as I developed the song. That doesn't work. I had to go back to the Creation Interface, select the track, and add it to the Playlist. Waste of time!
The other option; change the title each time.
Absolutely this. I find for the genre I tend to generate which is more classic synthpop and darker electropop there's a ton of noise build up that builds throughout the track to the point of unlistenable at times. I'm sure that eventually this'll be worked out, but currently it's limiting what I can do with Suno.
yes, me too. Suno is suddenly getting "lazy" since yesterday. Something really usual. it's not about shorter song/music that ended less than 2 min. Importantly, it's about the creativity of the song also shortchanged. Quite annoyed to be honest.
Suno, pls lookinto this issue. Meanwhile, I won't generate any new song, until the problem is fixed. Just waste my credits.
wasted so many credits for short songs since yesterday. Something is definitely not right
Why can't we get a single consistent vocal sound? Why do different sounds come out from start to finish? Is there anyone who has a solution to this?
there is no way to troubleshoot this if you do not provide the prompt(s) you were using.
the problem is not universal. It might be specific to how you are prompting โ and ironically there are many people who want to know how to make shorter songs so you might help them?
20k credits! Wow. Having said that, I'm sure I've burned hundreds on a single track to get the right structure and outro. I think my peak is 2-4k for a single song!
I think a lot of people just need to know there's an art to getting the best out of Suno. It can be frustrating. Even just changing the extend time by a couple of seconds and one or two words in the prompt can make a huge difference.
Hello, sorry for the question which must have been asked several times. Is it possible to indicate that we want a male singer?
sorry but where is the search bar in Suno to explore stuff I am interested in?
Just add 'male vocals' in the prompt. it won't work 100% everytime but usually it'll work.
anyone knows?
Thanks I try
You should just be able to find other people's tracks from the home screen? Do you mean music or other features?
like a search bar
for example if I'm interested in dubstep and want to browse other people's creations for that genre
There is a conspicuous lack of search in the UI so far!
You can't even search your own library
oh I see, then I guess we need to wait
it would be great to see other people's inspirations
and search what you like
like what other people with similar interest of genres are creating
usually you'll get a similar vocal for certain genres. Unfortunately between generations and extensions the vocal style and expression can change and it's just trial and error to hit it.
Vocalist: Man voice
yeah, we'll see. it depends on whether Suno see this as becoming a social platform for generating and sharing music, or just a tool. Or both. Or something else!
@graceful wind atm search is still something that needs to be build in. But people post there songs also in #๐ฃโsuno-showcase
So you could search here like this in discord: in: ๐ฃโsuno-showcase dubstep
great idea! ty!
hello guyz. Please someone could tell which style to avoid, in order to avoid the crashy voice, metalic voice etc ? (that will save me thousand of credit )
modern music with lots of SFX
future bass, edm, dubstep, and guitar riff then ? those kind ?
Exemple
yes, some of those... I have better luck with older genres 80s 70s etc
that one, after 2 continues for sure it gonna crash
Okay thank you a lot daisy
Why can't we get a single consistent vocal sound? Why do different sounds come out from start to finish? Is there anyone who has a solution to this?
Because the way the model works. To make unique music, it doesn't keep the seed...
"Vocalist: Man Voice" This definitely works. But the choir always spoils it. They need to remove the chorus
Make more complex prompts in the Style of Music.. It often leads to better results. Listen to other people, and look at their prompts.
Even if I don't add [Chorus], it does this on its own.
That's called hallucinations.. if you don't let it sing lyrics, but it is a song that feels like a chorus coming up, it will create it for you
It learned from our music
A good way to start is here: https://youtu.be/kB3W5xLza_k?si=lF2sPwlQGIHYWanc
Hello, I'm 3Daizy, a moderator and guide at Suno. Welcome to my channel, where I dive into the new and exciting features of AI-powered music creation. Whether you're just starting out with AI music or you're a seasoned creator looking to explore new tools, this video is packed with essential insights and step-by-step demonstrations.
Join me tod...
In this video I show you how you can go through these examples and learn from it
do credits accumulate for free plan?
No
Thatโs messed up. Thereโs an app that I use where you can earn credits by different methods.
then why did I have 400 credits when logged in after a few weeks
It resets to 50 daily at least, which isn't too bad IMO (I used it like that for the first month or so.)
There were some issues with the service. Everyone got 500 credits as compensation.
ahhhhhhh
I think they give gift loans from time to time.
I'll probably buy it when they add a search bar and it gets a bit better
That makes more sense cause I definitely havenโt gotten that much yet
Iโm not saying Iโm upset, just disappointed with that. But I do like Suno apart from that.
since yesterday, creating a song longer than 2 minutes has become one big torment and burning credits, please do something about it, it was already so good, creating the songs was brilliant and suddenly something broke.
์์ค๋น ๋ง๋ค์ด์ค
suno finishes songs when I don't want him to and there's nothing you can do about it
This is an English channel. It seems you are trying to use the old chirp bot which is no longer available. You can make music on the website https://suno.com
ok, you can make longer songs, you just have to put in a little effort, it's now the same as it was at the beginning of suno, you have to make shorter lyrics and add more commands like [instrumental bridge]
The advantage is that more interesting songs are released
Hey everyone - I'm trying to make a birthday song for "Simone" however it won't generate because there is a similar artist out there? I don't really understand.
Does anyone have any info or a work around?
Resume from an earlier time than the default to shake off whatever it is that the AI thinks is signalling an ending?
Apologies if you have already been trying this.
Unfortunately not the solution for me. I make instrumental and 95% of them ending prematurely in range 1:20-1:50 on first gen. Yesterday the same prompts didn't have this issues.
When I extend from earlier time it still finishes after let's say 30 seconds anyways on 2nd gen. So for now I'm stepping away until this is resolved.
I bet more and more people will start to report it.
So trying to extend it and set the resume time for say... 00:50 doesn't buy any time in extending it further?
(Guessing at a good time since I cannot hear your music.)
and try typing [instrumental] [instrumental bridge] instead of the text
when will the watermark be realized, does the watermark include youtube content ID tiktok etc
@final hazel
I don't know the technical aspects. but the watermark is already in your songs. You can read here what I mean:
#๐ขโannouncements message
(under trust and safety)
how can i describe to suno an instrumental like this https://youtu.be/6AdMoKKYjoQ?si=7KdLd-7haHELNrmx ?
i mean watermark/copyright Youtube content id like this
You can suggest that to the devs in #๐กโsuggestions-web
isnt that the job of the owner of the song to claim the copyrights? or the distributor?
Maybe try to write Siemone.
makes no difference. It extends for maybe 15-30 seconds and finishes song. I extend from -30 seconds of 1st gen
nope
And I'm not talking about couple songs. Since the morning I generated over 50 songs and maybe 2-3 got 2 minutes length on 1st generation, but I could hear already that they are at the stage of fade out
Hmmm...not a completely unwarranted complaint!
I tested a few of my usual lo-fi prompts and got it trying to fade out and some how resume the song in the middle.
I also got one or two that were just naturally shorter for some reason.
(Currently listening to a track, then I will try seeing if I can rescue the songs that faded out or ended early.)
they seriously do something to AI before they are sued and ruin their AI project
did you get an answer? I don't want to send another email if you already got an answer
So this is the one that ended early:
[Removed links]
I resumed from 1:15 (despite it being longer than that) and was able to take it to 3:15.
[Removed links]
(albeit it does fade out a bit earlier than I like.)
In Style of Music, you can add a Male Voice, and sometimes it works, but I want to know if anyone has any tips for what I should put to get a more consistent response.
Can you select the voice type? For example, if you use a Country Style of Music, you're going to get an American Voice. Can you select French, English, Italian... you get the idea?
Many thanks!
Try it and find out. Sometimes you have to try things more than once to get what you're looking for.
I have, and that's why I'm asking the question. Isn't that what this place is about?
Sure.
I have tried it myself. Sometimes it gets close.
(Or at least I think it does what I want.)
So really you can't help! But thanks anyway for taking the time to reply! Have a good day!
It's AI. We discover and rediscover. We cajole and try to influence. But it's AI and it has a "mind" of its own.
yes 500 credit spent and they gave me is all mess REALLY!!!!!!!
So I went back in continuing the previous whole song. I tried resuming from an early time with 2:56. That only got me another thirty seconds. So I went back further and told it to resume from 2:46 and that gave me another minute to work with. So I kept going from there.
[Removed links]
This one is 5:06
They really take this SERIOUSLY
You keep saying that you spent the credits without showing your work. No one is going to just take your word for it. Show us your prompts and your results.
They must stop credit theft
is there a way to change the thumbnail picture?
Unfortunately no, but they are aware of it
aw
yeah
so search bar missing and changing your thumbnail
waiting for these improvements! Great tool
What I meant is that I generated about 50 songs and their first gen was always shorter than 2 minutes with ending/fade out about 1:20-1:45. Only 2-3 tracks out of this 50 reached 2 minutes , but they were also at the stage of fading out.
Right. I noticed that some of my tracks were ending early as well. Which is why I said your complaint is not unwarranted. I see the evidence of what you meant.
Then I tried to demonstrate that it is possible to fix it and continue working nevertheless.
You could influence what thumbnail you want at generation/extension.
After your lyrics, add an [interlude] + [break] or the like of you are planning to extend.
Then add [End] followed by a line of the description of what artwork you want as the final line.
Sometime the AI ignores/sings the [End] and the last line, hence the intermission/break so you can continue off from a good point if it happens.
E.g.
[Melodic Break]
I made the following up, do not know how it will work
[Catchy Harp Solo, Intermission]
[End]
[End, Short]
[A bird with a Sombrero playing the harp]```
ohhh do you have a creation example?
@smoky obsidian please do not spam the channel with gifs, thank you.
If you were talking to me, this is some of the artwork generated when I was asking for a scene with the sun and moon together
If you want the exact wording, I can grab it. But I noticed the picture thing by some trending songs in the past
Well, the top two were related to emergency food prompts, ignore that, heh
Ah, here is one of the examples
I noticed it doesn't care about brackets and the like. As long as the last line is what you want
But it has been like a week since I last did it so I am not sure if this behavior still happens (it should since it has been a week)
This is what I was doing (if I understood you correctly). Let's say I had a song that ended on 1:30 on 1st generation. So I extended from 1:00 for example. That gave me 2 results. 1 was 15 seconds and another one maybe 30 seconds and song was finished again (faded out). After couple of tries with the same results I just gave up, cause it was just a waste of credits ๐คทโโ๏ธ
Hi @final hazel i have tried starting by an instrumental to avoid the issue of suno starting to sing with the music.
In 9 cases other 10 i am getting suno singing with its own lyrics. That is hallucination. I have flagged these clips as buggy but i still get the issue. Well i dont expect a solution till the in paint feature is out. This is a bit frustrating.
I don't know what your plan is but since credits expire at the end of the month any way I only consider a credit that I didn't get to use to be wasted. Then again I'm on the Premiere plan too though.
And I will admit to chasing a song into a dark alley and getting my arse kicked for it before. So I do understand. Sometimes you just have to wait or take a break or start over. It sounds like you've tried everything else and are now taking a break for a bit. I get it.
For an intro you need often 10, max 20 seconds. And hallucinations often start around 10 to 20 seconds just like real songs. So I personally don't have a problem with them? You can just set a time stamp before the hallucinations kick in.
Hallucinations are often unavoidable. Especially with music that normally has lyrics. AI is mimicking our music. And if you for instance take rap, it wants to rap. ๐คทโโ๏ธ
ohh ty
hello I just created an account and subscribed to the monthly pro plan and got billed twice, however I do not see 0 credits on my account, can someone please advise?
You mean you DO see zero? Try refreshing the page. If still nothing appears, contact billing@suno.com
Or was it billings? One moment yep, billing@...
I now see my credits but there was a double charge on my account
on my credit card statement
I contacted the billing team
Ah then just wait for the response. Nothing we can do here, literally just a place to chat. Devs are not around most of the time
The music (interrupts) when connecting
Why Suno does not have remix function? I would like to change part of the song. Is it even possible?
No, you have to choose 'get whole song' in the menue with the 3 dots. Then you get both parts in 1 song. This you can extend again.
anybody else having a problem with random words in instrumentals? and not just random, i keep getting the phrase "start it up" over and over again
i've cleared cookies and cache, changed browsers, used different prompts, still same problem
Any idea how I can change the text of a song and generate exactly the same song again? I want to fix some pronunciation errors of the vocalist. When I remember right, I could do that with the remixe function, but that has vanished. Instead there is a 'reuse prompt' function which generates totally different songs then. ๐ฆ And the quality is quite bad and the songs are short.
extend and then choose the start time to be towards the beginning of the song, though it usually changes the song significantly if you "extend" it from the beginning
It does not want to generate from 0:00 only from 0:01 and then hung up. Unfortunatly I have to correct the first word in the song text right at the beginning.
towards the beginning? How does that work?
here's what i would do: extend from 00:01-00:20, and in the description start with [instrumental bridge] and then your song again.
Think "Extend" as "Continue From". Telling it to extend from 0:00 will not work because you might as well make a new song
Note that Extended parts are generated for 1 min max while base creations can go up to 2 min
For more info, see https://www.suno.wiki/faq/making-music/how-do-i-make-the-song-longer/
you're going to spend a lot of credits and time trying to get close to what you had before, so don't go chasing waterfalls as they say.
I don't feel like I'm chasing waterfalls but I am blowing a lot of credits with "male vocalist" in the style that get me female voices ๐ฆ Seems like results overall have gotten worse since about 3 weeks ago
i keep seeing that sentiment... i don't think it's getting worse necessarily, i think we are just more aware of the issues it does have
but there are definitely plenty of bugs and glitches in it
I'm also seeing a lot of songs skipping the beginning of the song...
[Verse]
Never thought...```
and it abruptly starts with "thought"
i've had the same issue, i've had better luck with
[intro]
[instrumental bridge/hook/chorus/etc]
then
[verse 1]
blah blah blah
i'll give that a try. not sure what to do about these female voices, i'm like 80% female when asking for male
it can change depending on where in the prompt you put it, and depending on where the comma is/if there is a comma
male vocalist, dubstep, epic, heavy bass, deep bass
try shuffling your prompt around, and try describing the female voice, i've had really consistent luck with "soft-spoken, gentle female voice"
the female voice is what i am trying not to have
I've tried adding baritone, tenor, having both "male vocalist" and "male"
but I just keep on getting girls
might just be a bug then, i've been having an issue "follow me" despite changing prompts, clearing cookies, etc
i keep getting the words "start it up" in my songs randomly
I ended up keeping one song and just shrugging it off as being an unintentionally lesbian love song
start it up! ha
so my conspiracy is that there is some algorithm that is definitely tailoring our results slightly and it's going haywire
but suno does not communicate anything at all so we are just left to speculate and gamble
but i'm still having fun so oh well i guess, i'll be part of the problem
yeah I'm somewhere between fun and flustered... I know iteration is the name of the game but if its going to do the opposite of what I ask its hard to justify the spend
like I'm not afraid to adjust words and syllables and extend from different moments to get my result.. but dang
i definitely hear you there. i'd recommend experiment with different prompts for now just to see what i can come up with. not being judgemental, but you can get a lot more intricate and detailed in your prompts!
righht I was reducing in hopes of getting higher strengths
ohhh i see, that makes sense! I've done the same before
Just repeat it a few times. It gives more weight when you do. Or says it in more ways like : male vocalist, male vocals, male singing, dubstep, epic, heavy bass, deep bass
I've got.. 59 pages of generations and like 14 published songs
and if really feels like accuracy has sloped downward i the last 3 weeks
i've made several accounts with annual subscriptions....
that's another good suggestion
sometimes putting the same thing at the beginning and end of your prompt helps too
I reloaded the create page and put the same lyrics and prompt in and now I've gotten 4 male voices in a row... after like 12 female in a row... ahhhhh the atmosphere ๐
Did not work. Unfortunatly the word which I need to correct is exactly at the start of the song at 0:00, so it cut it. And it did generate a totally different song when expanding from 0:01. Look's like I can only try to correct it in Audacity a little bit. ๐ฆ
This song is somehow cursed - sigh!
ahhhh quatrain, male vocalist, dubstep, deep bass, piano, heavy bass, electronic, high energy, male first result is female of course lol
Any idea when I use custom and instrumentals I am all of a sudden getting "star" or "start" spoken or sung into muliple musical styles?
generally hallucinations... are you putting anything in the lyrics box?
yes, I've noticed this it'll skip the first line of lyrics
hows the election going Wayne?
yes, I've noticed this it'll skip the first line of lyrics I force it by putting [male] Will you put the music type?
well there's about 150 days I think if I'm lucky the CIA is gonna Astroturf me in the summertime there's still a grass move roots movement going on with my campaign behind the scenes from what I can tell. I've really been loving this music machine. I've been making some political science. I'm surprised we don't see music like we did during the Vietnam war that was very political, so I'm taking that initiative and trying to apply that to my own political platform. https://suno.com/@lambright
https://twitter.com/DesertRabbi2024
The only Jewishโก๏ธ Independent candidate for President of the United States.
The future is bright #Lambright24
i'm having the exact same issue
same
i havent used it in a week, but my generations from last night had them when my earlier generations did not
I dunno if "boobs not butts" is the most compelling political message but yeah things were different when you could write easy songs about vietnam and the draft.. out of hundreds or thousands of them I'm sure we can remember 3 or 4 today ๐
wow, really moving honestly though ๐
I'm thinking carefully about it and I can think of 5
Yes of course 'boobs not butts' is a song about a gay man going straight. That is not a political message at all. It's a society commentary. I haven't delivered my most political messages yet I'll craft some today. This is in the direction of the political movement which is the working class being abused and being politically frustrated. https://suno.com/song/9dd7ea75-c2b7-4d87-906b-2ffc76e70f5e
two of which are songs I personally like a lot but maybe aren't commonly known
as long as everything stays pg13 its fine to post here... in special cases it can cause a warning if its too political or something... but thats just a reminder most of the times.
i'm just generally reflecting on the staying power of political music, particularly anti-war or anti-establishment ๐
https://suno.com/song/8112df42-e350-41fa-a582-4c8319e0ab21
just to show you an example of "start" being thrown in randomly
rockabilly with an old school hip hop beat, Interstellar sonic drift song. Listen and make your own with Suno.
this was supposed to be an instrumental. i can deal with vocal hallucinations but this "start" or "start it up" has been pestering me
its adding words of encouragement
thank you for confirming what I thought was true that the antiwar movement was really powerful. As a politician, I'm surprised more people haven't tapped into this for political change. I'm a lefty and a liberal and I believe that we can help everybody. I wanna do my very best to make some captivating, toetapping political songs. My inspiration are bob Dylan, and Johnny Cash.
I was honestly thinking more the opposite, that the 60s music revolution was a popular thing at the time with very little staying power or actual influence.... especially long lived influence, considering the people who listened to them are the people in power for the last 20-30 years and you see what we got out of that
this is way off topic for this room ๐
off-topic sorry
to keep it on topic, i'll just add that the anti-war music of yesteryear sounds like the war music of today. generations change! they use ccr in vietnam movies! if people are going to hear the message you gotta switch up the genres, IMO
thats technical discussion right there lol
I actually came in here to see if there's any technical improvements. This is a very good product and has lots of room for improvement I wish each song was not limited to two minutes where you could put in how many minutes you wanted it to be I also would like the download of music separate from voice.
so your opinion about creating antiwar music, would be the popular format of the day, is my guess such as rap and pop music.?
No, don't just use what's popular. I just think that "you either die a hero or live long enough to become the villain" can be applied to music as well, just my opinion though, i'm just some random dude
i guess it's comforting to know that we're not alone with this issue! where the heck is this "start" ghost coming from?
Fortunate Son is almost the theme song of "its about to go down, war style" these days, you've got a solid point
Yeah it's virtually unusable like this it's happening in other lyriced gens now to usually right before or hybridized slur into the words. Someone broke the model. derp
Can you add your own vocals and remix the beat on fl studio and then upload it to spotify?
it must be reading from some sort of "master prompt" on accident?
you could always spleeter and stem then midi or revoice etc
and if I have the pro plan I can upload my remix to spotify right? @magic axle
I guess, it says commercial use allowed
FAQ covers questions about the plan nicely https://suno.com/docs/faqs
Though note you need a distributor to put songs onto Spotify so that involves additional costs:
https://support.spotify.com/us/artists/article/getting-music-on-spotify/
https://artists.spotify.com/providers
Now that's an idea for a sub topic chat... Distribution chain from suno to ...
sometimes even when we select Instrumental, there's voice. I hope it will fix.
stah star star start
What people are failing to realize is that the entire media landscape is about to change soon.
All media will be generative, on-demand, custom-tailored, perfection for the end user/consumer.
Not to be a debbie-downer but if you're hoping to become a successful producer off of Suno, boy are you in for a surprise soon lol
funny though I used to be amazed by the ghost vocals that arouse in my audiocraft gens
I think that in about a year, youtube, spotify, and everything of the like will be obsolete
thank you so much
Eh, I doubt it will go that far
this will age like milk
or the near 300 llms I have never used more than once sitting on my ssds
They will likely embrace and feature sections about the AI generated songs but I do not believe it will be the cause of a shutdown of those services.
Especially since YT has yet to shutdown despite the massive cost it already incurs. But that is more #๐ฃ๏ธโchill-chat
the future generations will have no interest in rewatching old content. everything will always be new and current.
there will be no need to "store" things. it can be an option sure, but an archaic and antiquated practice
But just to certain it appears the start vocal issues is known. and maybe someone is on it?
I think there's a window and we're still relatively early in it but that window won't last forever. i'm sure someone has already scripted a pipeline from gpt3 to suno to youtube and is flooding or preparing to flood 0 effort garbage
haven't heard anything official about the start problem yet
we are definitely in a window, do what you can while you can!
but most importantly, just enjoy the process
i see suno as more of a netflix of music for myself
i'm able to make combinations of music and songs that would have taken me weeks to make before
Ive had 3YT channels nuked by anti-ai bs since nov 22. It's just a fun whack-a-mole game at this point. They change their tos so often nowdays you can go away for a weekend and find content that had been up for 6 months just zapped wo recourse.
the idea of making any money with it leaked out that winter
you can definitely make money with it by enhancing content with it, but suno music itself is not what a lot of the users think it is lol
I've been using various AI in youtube videos for a few years now, marking them as such since they put hat requirement up.. I feel like the wording of their policy both excludes and requires marking ai performed music... but my youtube channel is old and music is only a small piece of it
I believe that as long as some of the content in your youtube channel/video is human made to some degree, you're good to go. But if you suno'd a song and midjourneyd a picture and threw it on youtube, it deserves to get taken down lol
its fine to share your work but somebody's gotta pay for all that storage
i'm using local SDXL to generate the image for my songs, then putting a lyrics scroller over it
as long as you're verifying the lyrics are correct at the bare minimum then i see nothing wrong with that!
I had 2@75 min albums I put up that I made with audiocraft and animated w deforum that were deleted. That was 30*5 minute clips of 4500 frames ea that I put together. One was some chillspace/binaural stuff and one was a nu jazz elevator lofi set, both of which I've used in the professional massage setting to good effect. And this past nov when they switched to you must label everything was deleted wi the day. It's ridiculous to think that vids that Dj's like Griz did w Pytti didn't get removed as well. It's really more a game of who's established..
youtube is a real pain in the butt
the playing field is definitely not level, never has been. i'd suggest uploading your music with some videos from your phone. throw a filter over it. bam. vlog with cool music as the main thing
they actually deleted one of my channels and after 3 days of asking why they said oops and restored it
but it's "Free?" vid server space... and so for sharing.. a necessary evil
the server space is trivial, its the audience that makes youtube valuable
youtube doesn't want people to be able to "one-button" an entire youtube video. the day that people can do that, youtube is dead.
the server space used to be trivial, ai is a big problem because of that now lol
its not just content being uploaded, but bots crawling the web too. like 75 percent of traffic is bots now
anyone else feel like the ai has a problem layering electric guitar with other string instruments
in my experience, no.
it has a problem of you expecting it to do a specific thing, and when you're listening for that thing, it becomes much more apparent
with enough generations, that problem will go away, but its a gamble
it just feels like it plays either one never both at athe same time
yeah I mean for an individual who wants to put up their own videos
oh yeah definitely. it's way too easy nowadays
there are some easy low cost CDN methods and I wouldn't flinch at paying for my own
Wow, Suno is way worse than it was 1.5 months ago. Just came back to continue working on a song and the extend option is garbage now. :\ It just straight up switches genres, won't reuse the chorus, etc. 
you're not maybe extending a v2 song with a v3 extension are you?
No. V3 for both, but I suspect v3 radically changed (for the worse) between the paid early access and now.
I've tried extending with both the genre box empty and with the OG song's style and it is just a mess. :(
its possible for sure. i haven't tried grave digging old stuff but I actually had a song I wanted to revisit because after listening to it over a few weeks I can pick out parts I don't like ๐
Yeah v3 alpha is not the same as the final v3 that we have now. They're working on v4, no word on when that's coming yet.
hopefully the new competition in the space lights a fire under their butts to step things up
If this is what they consider an upgrade, I ahve no faith in v 4. Already unsubbed but still ahve 10,060 creds to burn RIP.
I find it generally better than v2 at least.
i came in right as v3 was introduced so I don't have experience with v2.. but I for sure have a list of things v3 does poorly ๐
I mostly play around with instrumentals. In v2 when I did put lyrics, the vocals would blend in so much with the music that you needed to actually see the lyrics in order to hear them. You could only extend from the end of each part (no "extend from" time) and for the first month or so I used it, there was no "custom" mode - it'd generate the lyrics based on what subject you gave it.
Pretty sure it would always start in the middle of a song as well. I had some instrumentals that were nice but they just started at some random point
one thing I want to try is to get Suno to perform an instrumental track suitable for putting real lyrics over
like, what happens if you click the instrumental button but also have lyrics, would it try to work out reasonable spacing for them or just ignore them or go ahead and sing them or ??
and by 'real lyrics' I mean 'real vocals'
Is asking what genre something is a technical discussion, or no?
The lyrics box disappears when you choose "instrumental". But, you can put [instrumental] in the lyrics. I don't know if it would try to reserve space for them. I've only ever tried this to type some random stuff in to try to spice up instrumentals (in the hope that the AI will see the lyrics and be inspired by them somehow lol)
I wonder if you can put instrumental or "no vocals" in the style... i for sure have had songs that I put fully lyrics and normal style and they ended up being instrumentals anyway
has anyone had any luck with getting opera style vocals?
i have kinda ish
due tell
try Italian Opera
ok ty
https://suno.com/song/ec92f0b0-2650-4577-b55f-2c54b9f4c77e i used rock opera
you can also add [tremelo] @stone jetty
ty all very helpfull
you can also try using [Choral Opera] before your lyrics, i had mid results when trying to do an amadeus type song
ill post what i came up with when im done
gl
@stone jetty https://suno.com/song/7c69677a-d249-4246-b835-a35990d853d2 i didnt want to spend too many credits, but i found some tags that work
that is amazing
i have the full song lyrics if you want it
ive got lyrics alreadt ty its a song about the emperor of man kind from 40k
ooO interesting, do share when you're done
okay, i extended it to the next verse, and its pre cool that it actually overlapped sung lines too https://suno.com/song/45c4603e-698b-4e3e-a5ab-c3ef2f864160
yo this is wrking very wel ty
Hey team. I installed Suno app on my Iphone, but there is not login button? I also bought subscription and generate some cool content, but how to connect it to my real user? I a not able to achive my subscription and content via web
Installed HOW? Because Suno is only via browser not apps. And there are scammers who offer apps. So beware!
Do you want to say that's app is not from Suno?
There is no official app/playstore for Suno. The apps you see often charges you just to redirect you to https://suno.com
https://suno.com is the only official site
But as far as I understand this app uses API of suno?
There is no API
It is a literal redirect to the website
Notice how that app you have doesn't use the Suno logo and instead uses a musical note
This is a scammer yes.
Make sure you get a refund through the appstore and report the app as a scam.
These apps literally try to charge you to use their app for redirecting to the official site.
If you look closely, they often claim that subscription to the app and the sites are distinct (because they are not affiliated with the service)
Nope. Cannot do such a thing and the team likely never will implwmwnt it because people will abuse it to upload copyright music
understand, thank you for help
btw, seems like this file is from Suno https://cdn1.suno.ai/4499a8e8-15d9-4ebc-a119-a04d04d05584.mp3
Suno.ai was the previous domain. They changed to .com like... a month-ish ago? Which is why it redirects to suno.com
The lyric info page of the song you posted:
https://suno.com/song/4499a8e8-15d9-4ebc-a119-a04d04d05584
Deephouse Club music song. Listen and make your own with Suno.
Wow, thank you
Hello everyone, can we play specific chords?
I can't make a chord progression
thanks for your reply @graceful condor
can anyone help me with this:
part 1 - normal lyrics ( verse + pre-chorus + chorus)
part 2 - normal lyrics ( verse 2 + pre-chorus + chorus)
part 3 - guitar solo
part 4 - normal lyrics - ( verse + pre-chorus + chorus) - but suno changes the style?! why!
Allo. May I ask what flagged for moderation means while using the program? I'm assuming because of the word "kill" or "murder" used in the initial prompt, but those were removed and now I have no idea what is stopping the prompt.
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.
(A copy paste response >_>)
Right on, thank you! I don't think I could use this for DND anymore if it's gonna flag concepts like "Find the Killer" and whatnot. Ah well...
Interesting. Please see my DM to you
It's kind of hit or miss. I got it to generate this song.... somehow, but it did actually lock me out of the original prompt about halfway throug the song. Fortunately, when I cut out the chorus from the lyrics in the text box, it still repeated them on it's own anyways.
Today I asked myself, "What if a Fallout Raider wrote hair metal?"
Royalty/Copyright/DMCA FREE music for streamers, vloggers, and creators to use without fear. All "fair use" is allowed. No content with proper attribution will be claimed.
I'm ALWAYS posting new songs, so let me know in the comments if there is something you'd like me to make! ...
The reason I stopped by though, anyone having issue with song generation right now? I have one song it just literally WILL NOT extend, no matter where I move the extend timer to. I've never experienced this before, any tips? I'm about to try extending another song and see if that one gives me 5 sec songs too
Another new song failed to generate usable extensions, but an old extended song still works to generate exts from the same starting point. For some reason, it doesn't want to extend either of these two new songs I'm working on, no matter where I tell it to start ๐ค
How do I unsubscribe to Suno ?
click on your credit count and then "cancel plan" there you can end the subscription
can i ask why the 1 minute limit forsubsequent generations (part 2 onwards) ?
i have lost so many credits / chorus just because part 4 of the song generated a sound track of 0:58 seconds
Iirc, Suno uses the other minute to read the pattern of the music prior to your continue point
though you wouldn't know it to hear how bad extend is now days :P
Ooh good challenge
has there been any rumblings as to how far along Suno is on v4?
The Chirp bot has been retired.
Please use the website version: https://suno.com/create
If we have a paid subscription, can we still get the 50 daily free credits?
No
When will the sound quality improve? The sound quality is really bad
For me it's hit or miss depending on my initially generated song and where I extend from. I have gotten some good returns, although definitely have had some bad
Suno is getting worse and worse. Each new song is worse than before and never follows the structure anymore. Like it has a mind of his own
I wonder if that's true, or people just setting their expectations higher after the initial wow effect has worn out
A miracle stops being a miracle when it becomes the norm
bassdropnimit is true. they don't follow lyrics and structure at at all and make short( some 1min) tracks ,which was not seen before
any idea how to stop this, listen towards end, athough it is through out...
outro echo to fade song. Listen and make your own with Suno.
you need to "extend music" from 30 sec part in this case
tried it still creeps back in even if i try from 4secs
i will do again and send back
it's horrifying....You need to remake it from scratch
it's very sad to say
outro echo to fade song. Listen and make your own with Suno.
this is then form 30 secs
How terrible ๐ Suno officials really think about this mess. this is rip-off
2nd payment period and "songs" are no longer 2 minutes bout about 1:20. Something is fishy.
i can't get normal male voice unless it's a country song...every male voice i get (when it finally give me a male) is a high falsetto i want some real male soulful voices for certain things
Select instrumental, you might get random Russian female singers on any genre. It's like they're trolling us now. Ready to cancel.
"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
I'm ready to cancel, because something has gone very wrong with how it was working last payment period. What a convenient "coincidence". Absolute rubbish now.
I was able to make the song with the same lyrics previously
Yeah, good luck with that. AI is VERY raandom and unpredictable. I just got what was supposed to be an INSTRUMENTAL and all I got was a robot voice continually saying "START! STOP BELIEVIN'! START!" etc. CRAZY, LOL.
the AI is too unstable in very recent days
Have a listen to THIS NONSENSE, got another one that was NOT supposed to have lyrics! https://suno.com/song/9d896af5-3615-47ee-8e78-efa8fd590a78
Dub at 70 bpm with piano delays, skanking guitar, reed organ, clavinet, with remarkable melody. song. Listen and make your own with Suno.
LMAO.
What a mess
Wrong tempo, but that voice "START! START!" LMAO.
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.
https://suno.com/song/345f10ed-f572-45a5-b6d4-18a00500ba0c it is the best time that I put this song I made before
Hair metal ,glam metal, speedy ,hard ,catchy song. Listen and make your own with Suno.
Hey guys, can you tell me what happens to unused credits the day of the refresh?
Yeah, mine just VANISHED. That's life.
As a paid subscriber, it looks like I've been conned. Songs are now only a minute from what WAS 2 minutes, and trashy results unlike the great ones, last month.
19 second "songs" - C'mon man be serious. Wonder who is behind this site? Granted, credits are refunded for 4-second songs, LMAO.
Was gullible enough to subscribe to Cassette AI for a month - and oh man, don't think my ears will ever recover.
Hi all, I'm new to this, and find out how awesome is Suno. But having trouble with how credits works. Any idea about this?
Good luck, I'm about done with it. I think its 10 credits a pair, or something. I got 3,000 again, but am down to 2,900 in a couple hours with nothing but 20 second songs and some a minute. More random than usual. Way too much ambience on the mixes. Not worth the $15.00 a month (Australian dollars) in my opinion - simply too random in the results - a few really good ones, but still hve to be enigneered to sound right.
hello, I'm having a n issue of missing: continue from that clip is missing, anyone else have same problem? I bouth subscription for 8$ per month
It SHOULD show as "EXTEND" when a clip is done rendering, then you can go several clips from there, and eventually even get the "whole song" instead of messing around like before.
greetings! just wondering if something's up, but i'm noticing at the beginning of almost every generated song, im getting something like "let's start" or "start" before the lyrics kick in. happened in both simple and custom modes. anyone else getting this? just wanna know i'm not going crazy, lol
YUP, this is NEW. Got it even when NOT wanting lyrics. ALSO got one in all RUSSIAN (without wanting lyrics). Gone haywire! Only NEW from TODAY. (Never happened before, so something is up - getting worse over time.)
wow, i just discovered it a few days ago and got so damn hooked, lol. this is already killing it for me. wonder if it's a bug or something
What genres are you trying?
Not Industrial or german industral or something similar by chance?
Yeah, only got it today, and I've been using it each day for months. Was thinking of canceling but it's too bad.
it's happened when i tried stoner or sludge metal, as well as thrash. its kinda funny hearing it at the start, but it ruins it when the song turns out so good
Well, you can use iZoptope to remove or lower the vocal. But the genre DOES NOT MATTER, because I do reggae, pychedelic rock, and more.
i did notice a few "stoner/sludge" tracks were also being sung by a female when i stated male, and the music started to sound more nu-metal through the track lol, weird
"pshychedelic rock" my mannnn \m/
Well, you'll find that AI does whatever it "wants" - LOL. It's the weird Russian female vocals, quite realistic but didn't go with reggae, lol. Unpredictable at best, goes for images as well.
what's your handle? i'd love to listen to some of your stuff!
I have to find it. No search function here for the raw version.
Delete or Not Delete, that is the question. I've been trying to tidy up the clutter; I've selected and deleted items:
In Create Mode by clicking the ๐
In Playlist by clicking the Dustbin, top right
In Songs, ... on each track and Move to Trash
I've even selected Delete Permanently
But, when I return to the library, they are still there! ๐
I've never experienced such a convoluted system for managing files, and the best part is that it doesn't work! ๐ฏ
Found one I edited later (so it doesn't cut off) - WEIRD prompts give OK results sometime. https://suno.com/song/8f06c69f-d81a-4163-a852-681b3fc4523b
Hawaiian HULU music integrated into 1960s psychedelic rock 80 BPM with 1984 "Firestarter" soundtrack influence. song. Listen and make your own with Suno.
I'll post same question here as well.
"Hey
Wanted to ask a quick question.
What is the best way to extend tracks?
I've always done it by getting part 1 part 2, getting a whole song and then adding another part 2, but after a few of those processes, the generations were more and more compressed in sound quality.
Will I be able to see better results when making part 1 and part 2 and then adding part 3, 4 and so on, and only then getting a whole song?"
Home Screen... Nice that you can check out different creations, but is there a Search Option?
Hi, dear Team , If I input a description of what the transformer should do, such as: "Speed 80 beats per minute BPM and Acapella Voice, harmony D-Minor, nothing else," it doesn't follow the instructions but adds instruments, etc. How can I improve communication between me as a musician and the transformer so that more complex music can be made? I am testing this subscription and it should be possible, right? Or is this not a professional music tool, but more just for fun?
hey guys
i found out
if the thumbnail is not colorful or birds it will be a good song
I am struggling with getting suno to start with an instrumental into, it starts with the lyrics right away.
Tried
[Intro]
[Instrumental / Instrumental Bridge / Instrumental Intro etc]
[Verse]
blah blah
Still is starts at the lyrics right away. Any tips on how to have a short intro with just instruments?
Understanding and effectively communicating with Suno's AI can indeed be challenging due to its current limitations in processing highly specific musical commands like exact BPMs or chord names. Instead, the AI is more responsive to descriptive terms related to tempo and mood, such as "fast-paced" or "relaxed". To assist with crafting your prompts, using general musical vocabulary that describes actions or emotions associated with the music can be more effective. For refining your commands, websites like Wikipedia's list of musical terms and Every Noise offer a comprehensive look at musical styles and terms that might be more recognizable to Suno. Additionally, exploring Suno's beta Explore page can give you a sense of what others are creating and which terms yield successful results. Remember, experimenting and adapting your approach based on these resources will help you achieve better outcomes with Suno.
Hi, is renaming a song didn't change the title in downloadable video?
when i input a bridge it never works for me Suno should fix that unless I'm doing it wrong.
The best way tends to be to create an instrumental song to start with. This will be up to 2 minutes long of course but you can then extend it and put your lyrics in this new part. You can set the extend time to the point you want the lyrics to start. It may seem wasteful since you'll sacrifice most of the first part, but it gives a bit more control and might save you having to make several songs just to get it to start right.
(Also means if you like the style and the words come out badly you can retry without losing the style)
Too bad, I thought it would be a professional tool for musicians. Of course it is very random and difficult to produce musically high quality songs with different instruments and tempo changes. I think Suno should really think about recording a professional musical language for the prompts, the Transformer should be able to do that without any problems, because that would really be the game changer for musicians. Until then it's a nice tool for hobby musicians. I hope you take this as constructive criticism, because I think the UI is great
This is still a very young company that's still actively developing. Just give it time ๐
Not much is missing, only the prompts need to be differentiated in terms of exclusivity, tempo, rhythm and harmony ... hope it will be added soon
The problem is Suno is so hyped by everybody. the expectations are through the roof. Its still a random generator. Do you remember the times when AI pictures had 3 Hands and 7 fingers on each... I think suno is at that stage now. And it will get even better.
Yeah. I also stick to Udio for a while because no matter what, I can only create male singers, while I want to have female ones. ๐ Hope this bug will removed soon.
I am not a developer myself, I am a volunteer moderator. If you have suggestions for the team, feel free to leave it in #๐กโsuggestions-web ๐ผ
Here's the response i got. Ironically, i can still wait.
udio might interest you. it sounds a little better, but the composition sound a little bit robotic and stiff. you should always try all options. its a free world.
I must admit, Im not really loyal, when I see oportunities. ๐ & so far, its open beta, so completely free (but you can keep all credits & comercial uses, as long you did everything with ai, so no copyrightharms & stuff).
Is there any way to remix using the basically the same song with slightly different / fixed lyrics?
Not yet.. in V4 they hope to offer us "in-paint" which should give us the option to only change certain parts.
That would be amazing; wonder if it will be backwards compatible with 3
I don't know tbh. But they did it with "continue from", so it could be ๐
Thanks for the info. Any eta on 4?
No sorry..
what on earth are you putting in the input, my average initial seed gen is closer to 1:45
I mean Empty Lyrics will kill a song fast (if not instrumental)
not saying your not unlucky, but perhaps we can help?
Well, it's usually for an extend when it becomes only 4 or 11, sometimes nothing at all. I often get 2, even 3 minutes once in awhile for one first part.
Nonsense, I have the last month or so. I'll see if I can find one.
Extends - if you extend from a down beat of silence... it can struggle to keep going, sometimes you need to drop back 2-5 seconds (it will usually repeat the lost second anyway in the last style)
Extends rely heavily on enough content in the extend to keep doing something (or toggle it back to instrumental perhaps - but it may come to a natural end so...)
Nope.
I look forward to seeing 3 minute initial generations
Please share!
Yeah, there's no search, so it will take a bit.
this AI is Bugged out in latest 2or 3 days
Unsure why's there's no search, or time indication either. Would help a LOT!
Yes the UX is basic, but you are searching for something that should be technically impossible by the contraints of the very system.
so do find it.
No, they actually mentioned it when it first happened, deleted some of them, though. May find one.
I found one, but then it went back from deleted. Should find it soon. Noticed in didn't have a title.
ah so a bugged period from a while back, not a common occurence
There's a few. Here, one: https://suno.com/song/2fd32ab3-6da8-4df6-a59d-24a343401f19
Shamanistic Chillwave at 76 bpm song. Listen and make your own with Suno.
It's 2:42
seems to be, but as you said - when they had a bug
not been like that for a while.
Bug? Should be MORE bugs, then, lol. After all 2 minutes isn't much. Cassette AI had 5, but then, that site was TRASH.
why more, its fixed at 2 then +1 for extends.
Better a "bug" that gives more than a lot less, lol.
A little question: is it allowed, when I use the suno Logo for the Cover, so everybody knows: this is produced by Suno AI Music? Would do: Lyrics me, music Suno.
Yeah that's fine
And if you use the free version, Suno wants you to add credits. Like "made with Suno"
No, it ist premium made. But I like to tribute the company for free, too.
hi quick question:
can suno on web version do instrument solos?
hi
Yes, but it's hit and miss on both quality and whether it follows instructions to solo.
Hey guys! I have a weird use case for TTS models that I need some help with. I am looking for a TTS model that can talk as slow as 40 WPM. Many solutions on the market have a slider to reduce WPM, but at slower speeds it just sounds like slow-mo audio. I am looking for a model that inserts pauses between words to match the required WPM. Any thoughts?
Oh I think this is my bad. I got here from the Bark AI model but I now realize this doesn't include that channel lol
Can the length of the song influence quality of the music?
I have tried to make quite a lengthy song (around 6 minutes)
And the last few verses sound compressed as f.
Especially the instruments.
Do you have any experiences like this?
same here
What especially drives me mad rn is how Suno refuses to end longer songs. It's just absurd lol
Howdy everyone. Is anyone else having an issue where vocals start to sound more robotic and auto tune(y) as the song gets longer?
yes i think this is jusat part of the ai dream. eventually all paths lead to a single outcome
also voices go from a solo voice to as choir as the duration increases
There is a choir effect going, but it's very genre-related, which counts also for the robot voice. I made enough without these effects. Is it perfect? No.. but it will become better with the upcoming versions. Here some I made with V3:
https://suno.com/song/ae066b7a-24e7-4506-8edd-0a6dd36ce97e
https://suno.com/song/dbe9bb07-a1fa-4c3d-a5db-0f7f76f42706
https://suno.com/song/40c3dc17-8a77-42e5-b8b3-1fa5b249888c
https://suno.com/song/075b597c-c4e3-419f-8fee-a86b52501637
https://suno.com/song/daae4616-6b33-4a83-87b9-ea839e67d8e7
Renaissance, medieval, folk, a cappella, harmony voices, duet song. Listen and make your own with Suno.
Swamp Rock, Country Rock, Roots Rock, Blues Rock, Southern Rock, sixties, 60s song. Listen and make your own with Suno.
'50s,deep male rock chest voice, vocal Fry, rich baritone, fifties, electric guitars, driving drum beats, Rocknroll song. Listen and make your own with Suno.
bright high-key piano intro Funk, Soul, Jazz-Funk, Psychedelic Soul, Soul Jazz, atmospheric, sci-fi song. Listen and make your own with Suno.
Celtic New Age,passionate,ballad,sentimental,female vocalist,love,melodic,romantic,ethereal,longing,epic,ethereal song. Listen and make your own with Suno.
Hey yall! How are speaker presets created for Bark?
There is no bark here anymore. You can find bark on huggingface here: https://huggingface.co/spaces/suno/bark
Credits haven't refilled yet. Still at 1700 and it's May
how long does it take for credits to refill from each month
You can see when you get new credits by clicking on subscriptions. Also, be aware, that credits do not roll over to next months (as stated on the website) so don't try to save them.. maybe sure you used them!
Ah I see. So I've I purchased the plan on the 25th, then I get new credits by the 25th?
yes it renews on the same day, and on the same time you got the plan
Thank you ๐
So i wanna make the song i wanna write have fade outs like pauses before they say the next verse in a almosy screaming tone and also add chorus to it but i just cant seem to get that to happen any help ๐ฅฒ
You can try a [Fade Out] and extend then with a [Fade In] going a few seconds in.
Not working?
I would also try things like
[Silent Break], [High Pitch Chorus], try [Vibrato] to hold a note at the end.
I don't if Suno will understand falsetto, but you can give it a go.
I have created some songs with a free account. Now I have gotten a premium one so that I can use the music commercially. I generated 300 songs and...none of them are as good as some of the free ones I generated. Is there a way I could use the free songs commercially or are they out of my reach for good?
Is there anyway to get the same voice from a previously generated Suno song for another song with different lyrics? keep the tone and voice consistent?
sadly, i'm pretty sure that's not possible
especially because we don't have access to the random seed

that emote is another one that cracks me up every time
Hey there, just one question: why are the oauth sessions so short? I have to login every few days, that's pretty annoying. I use the login with Gmail. Will there be any change at some point? Or is it even considered to be changed in the future?
Thanks.
This might be to stop automated bots from using the service, but I could be wrong.
I have not thought of bots, could be a reason. It's very annoying since I have always to logout from Google afterwards...
That's kind of odd. I login with Google and just need to press the Google button to log in to Suno, never needed to log out of Google
I do logout of Google manually, since I hate that Google always tries that you keep your account signed in
It's more convenient staying in the Google session, that's why Google does annoy us. They hope people will just let it happen and ignore this. But i strongly dislike such practices
Is it possible to make a section for all commands for suno?
example: [Pre-Chorus] [Chorus] [Verse] [Bridge] etc. I dont know if it is possible to make a command like [Cut Beat Chorus Piano] [Verse another Melody] or so. Hope u know what i mean
[] is for meta data
So [Flute solo] will probably help you having a flute solo
and how to say instrumental the commands Cut before Chorus 2times then verse then bridge
I am no expert, but I have burned around 5000credits so far. You learn best by trying and testing. Sometimes the success rate is like 1/10
i know my credits are done
Any way to see what words are flagged?
What is a "cut"? ๐ I do use [break] for having kind of pause, but also [pause] can/will work
I believe it's not just words
Paste the lyrics into ChatGPT and ask "what's wrong with these lyrics?"
(ChatGPT does the moderation flagging)
yeah cut is the pause thanks so break will work?
It's meta data, you can probably add an extensive description of what you want to achieve in this tag
ill try that, ty
As I told you, the suno model is no static program, such trained networks will yield different results. So you might be successful in 2/10 times or similar
Ah you were talking about censoring xD sorry thought it's also about meta data
Also check other lyrics to see how they did it
No problem, what is meta data tho?
Can u tell me how to make this in instrumental:
Style: German Hip-Hop with Piano and Guitar, Feelings: Love
Beat starts with: Pre-Chorus, Break before Chorus, Verse with another melody or only drums, snares and Claps
I do always use custom mode, since I do provide my lyrics and song structure. I think it's won't in simple mode to control the structure the way you expect it
I want to sing it by myself. if i do ir in custom with my lyrics i cant seperate the stems clean
You can add song meta information in [] elements like [epic bassdrop] or [drummfill] or [Piano solo]
can u tell suno the seconds for the meta ?
Makes sense, I think you the need to be lucky and always pick the right timestamp for continuing a song. Since you want to sing by your own, you should be able to rearrange it in another postprocessing tool
Never tried it
@final hazel can u help here?
Keep in mind, suno is still pretty immature/feature complete
Also is there anyway to kind of save a voice for other songs? For example i want to create multiple songs with the same voice
i know but suno is pretty good
and sorry for my english. iโm german ๐
can you please put advanced as default
because the hallozinations i hope they will fix fast
D]on't expect direct and full control. And Hip Hop is a genre that has a lot of rap normally, so the chance is big that you will get hallucinations. So that being said. I would do it something like this:
Style of music: German, hip hop, piano,guitar hartwarming,
Lyrics:
`[pre-chorus}
----.----.----.----.
----.----.----.----.
[break]
----.----.----.----.
----.----.----.----.
[chorus]
----.----.----.----.
----.----.----.----.
----.----.----.----.
----.----.----.----.
etc`
what did u mean?
It's hard to fix that especially because there is not much hip hop without rapping. And it's trained on that.
and this i can wrote in instrumental with the lyrics?
i know but we i think it will learn ๐
If you use the lyrics, you cannot turn on the instrumental button, but it's the same if you don't write in text in the lyric box.
i mean the part u wrote with lyric. this will work with instrumental?
There is no garantuees
Hallucinations will happen
ok iโll try it.
is there a way to say suno that the verse will got another melody or only drums snares and claps?
You can try it, but you are expecting too much control. It will cost you a lot of credits if you want every element to work.
i know but then i will got a perfect song ๐ง
if you call something a [bridge] instead of [verse] it should be a different melody.
If you call the 2nd chorus [chorus 2] it could sound a tiny bit different
you can add to a verse things like [verse, drum snares] and it might add drum snares.. ๐ค
But sometimes you have to see yourself as a curator too. You decide what sounds good, even though it doesn't have all your elements.
Thanks a lot ๐ง
when i open suno, it is basic mode.... i always use advanced mode
for instrumental too?
noo... dont you get it... im talking about the user interface
what even that normal text box does? it uses chatgpt? right?
are you suno dev?
i have some things to talk with you....
iโm no dev ^^ iโm user
how yall making songs that are longer than 1 minute 30 i assume the extend feature but i dont really understand how to use it
but im also writing lyrix
On the generation you want to continue from you then click 'Extend'
and now when you create your next song it will flow from that one
You can pick the time to continue from as well
how does this work with lyrics, for example i made a song hwere it cut off mid chorus
I would cut before the chorus and then generate the chorus again on your second clip extended from the first
OR
you can just...finish the chorus in the second clip but that's a weird thing that only sometimes works
Yeah if you are doing stuff outside Suno then offload it to editing
But inside Suno that's the way to do it. Cut before, generate new from a point
when you say cut before, how do you mean? and generate the chorus again? as in using the extend feature and then cut at the time before the chorus?
or do you mean, only write up to the chorus, then extend and add the chorus, but either way is there no way to make it one song on suno?
The first one
Since your chorus is cut in the current clip, extend from that clip, set the time to before the chorus started
and then use the next generation to have your full chorus
Doing this will give you 'a' full chorus but it might not be the one you want because lmao LLMs
i just found out about the get whole song button lol, so i assume you do that and then get whole song
Oh man yeah sorry
i was wondering what the point of extended clips was, i assumed for people to mix on their own
No you can slap that button and it will join up everything generated from before
Hence why songs go past 2 minutes
i fucking love suno dude
It's p. good
its worth every penny
can you undo it if you get whole song? or does get whole song just generate a new song file with the combined pieces?
Getting a whole song doesn't erase the previous clips\
bro ive been writing 1-2 minute songs since i started using this lmfao
You can just go back to those if you want
There is no downside to 'get full song' besides more organizational load for you personally
You just have another song
i'd rather do that than mix it myself
You can always mix it later anyway
Suno sometimes does too long of a pause
Easy enough to cut that out when out of Suno
i noticed when you extend it sometimes adds long instrumentals too instead of having a clean ending
but i guess at that point you can just fade in an edit
hi everyone!! i am a new in suno ...and i would like to know if is possible write a correct prompt...or is that luck? kkkk
Hello. Here are some useful links for you. I highly recommend reading through the entire Wiki if you can as there is not many entries but filled with lots of useful info
Website: https://suno.com
Wiki: https://suno.wiki
Tutorial video: https://youtu.be/lj8utBd_IUM
It tends to fill the time. Having more lyrics can eliminate long pauses, although it's not guaranteed. Nothing is guaranteed to work all the time at this point. Experimenting can help you discover what works best for you.
if i cancel my subscription, is it easy as pressing a button to resubscribe at a later date?
yep
hi everyone how i make it commercial , i already paid for pro , but when i submit to aggregator , it still cannot work
By aggregator, you mean distributor?
yes sir
Unfortunately, many distributors do not accept AI generated songs. Out of curiosity, which company was it? We users are trying to keep track of who rejects
nopi sorry
understood thanks
tune core
but i thought in the FAQ said that item will be commerciall-able once we paid for pro
AI is mostly LUCK, including with images. BECAUSE it is RANDOM. I've used the same prompt in text, images, and audio, and sometimes it's a GREAT result, almost "miraculous" - other times utter nonsense or rubbish - with the exact same prompt. I always laugh when people pretend otherwise. Any type of "guide" or claim of "information" or "how to" is usually pure nonsense but people love to pretend . It's human nature.
Unfortunately, even though you do have the sub which lets you have copyright of the music generated so you can monetize, it does not mean every 3rd party service will accept it - especially if they find out it is AI generated.
All it means is that Suno will not come after you if they find out you are monetizing from a free plan creation.
Ah, unfortunately TuneCore did state in other sources they will not accept 100% AI generated content - https://www.hypebot.com/hypebot/2023/05/are-tunecore-cd-baby-distrokid-symphonic-blocking-ai-music.html
I know DistroKid usually allows, trying to find the list of the others that also allowed.
Ah, here we go:
DistroKid and Ditto are known to accept AI
TuneCore and CD Baby are known to reject
Oh, there is a trick to get the thumbnail art to what you would like: #๐โsuno-school message
(The key takeaway is to [End] so the AI does not usually say the final line which would be the prompt you want for the thumbnail generation)
There are several metatags made specifically so that the AI knows what pattern to switch to, such as laughing, Verse, etc.
But yeah, many of them are just luck or like a placebo
You can also burn a generation on getting just the end, clipping it to Instrumental mode, and putting what you want the picture of in the Song Style box
As long as you have instrumental mode on it will generate a picture from the Song Style box instead
That's interesting, because "placebo" is the word that I use. LOL.
"Instrumental" doesn't always work. There are many times when weird Russian lyrics - or sometimes Spanish come in, but they are always gibberish. Yesterday, I got a country singer endlessly repeating "his nerk is never nark" LMAO - but deleted it.
Put 'Instrumental' in the song style box as well
Try adding "instrumental" to the style prompt, in addition to whatver you already have there.
too slow 
But yeah exactly that
Tell Suno in all the ways you can that no, don't sing
Telling it twice (toggle and song style prompt) is enough, so much so that I don't recall a time where it didn't work
Nope. AI is random, especially when you use it more over time.
Putting "instrumental" or virtually anything in a prompt really is never 100 percent.
Another thing I do is add a different instrument to an "extend". Sometimes it works well, other times, not.. Like I said, RANDOM.
The funniest thing I ever saw is when people talk about "negative" prompts in image generation - as if the AI will not do something you tell it NOT to do. Hilarious - and ultimately makes no difference.
I am curious if anyone has gotten this to make animal sounds such as cows or sheep or cats or dogs... Lion Roar would be cool as well I got real close once almost sounded like a bull but when I really listened to it, it sounded more like a car, if you have gotten animal sounds I would love to know how.
is there not a way to change the talking speed/wpm of tts generated with bark
have you tried High Tempo in style?
Thats it this AI wont react to any tags and lyrics anymore Suno is DEAD
proof? Can you make your music public and send link so we can troubleshoot if we can?
I send them directly to suno official dont mind
๐ค weird, there sure are a lot of songs posted every hour that make you sound wrong.
(must be a conspiracy)
Here is one of the example of [end] tag https://suno.com/song/eb9efc33-c19b-4267-81fd-0b8b18eedf8b
lo fi ,hip hop song. Listen and make your own with Suno.
you can see [end] tag doesnt work and music keeps going on
https://suno.com/song/0902e9bc-347a-40e0-a170-69d799adaa07
and hallucination can be seen far more often than before
rock song. Listen and make your own with Suno.
Here is my comemnt on how i was working with Sumo for 5 days>
The support for 'dangdut' doesn't always work accurately. Many songs end up being rendered with pop or slow music instead.
Several songs are too short, falling below the 2-minute mark. These truncated songs require extension and final touches. While finalization doesn't take up any additional credits, the extension process consumes 10 points, meaning that two songs can exhaust all daily credits. Moreover,
It would be preferable for the song generation to start with a natural length of 4-5 minutes instead of just 2 minutes.
The adherence to [meta tags], which describe chorus, verse, etc., is inconsistent.
Sometimes, even after extension and finalization, the length of the songs remains shorter than intended when summed.
Additionally, there have been instances where the audio fidelity is compromised towards the end of the song.
Considering these issues, I would rate this system at 6.5/10. It still requires further refinement before we can confidently say that Suno is complete Needlless , I will try more to work with Sumo and possibly gotto the basic sub , degrading some other AI generation webapps i m ubscribed
Tags and lyrics seem to be working in this one. I'd extend to try to cut off the hallucination.
https://suno.com/song/3678bbd7-b67e-4fcf-983c-9e513adb774e
this is the example of extension problem ,verse 2 and 2nd chorus have been cut off and lasts only12 min
rock ,male vocal song. Listen and make your own with Suno.
More payment options
the Problem is , this hallucination rarely happens before(some 3 or 4 days before). but now it really often happens
I already paid extra points a lot
Explain
the biggest problem is i can easily remake bugged out songs everytime I try
I need more payment options, like PayPal
sorry it's credit card only
Thatโs a mess
if you are pro plan (like me) you can get 2500 credit per month but you can also buy extra credits (4 USD per 500 credits
Yea but I canโt get the pro plan without credit card
They just need to add more payment options
Yes a lot of people wants PayPal option
they need to cope with
this is the same issue as mine
Good job people!
How many credit I can get free of cost every day?
50
Daily credit refreshes at 0300 UTC+0 (<t:1713236400:t> in your local timezone)
hey i cant pay i have updated my billing method but cant pay pls help
a can continue the next month but cant pay 4 tnis ho can i ?
You know what we really need
A way to make lyrics for a music you input yourself
I have so many music I need lyrics for
ฮost typical is the bin
Contact billing@suno.com
But, if you have not paid yet, you are out of luck if you cannot use any of the available payments methods AFAIK
Well i'm keep trying but no luck its DEAD indeed
ask someone who has so many lyrics and no song for them
I couldnt find the bin at first๐ญ๐
Go to library, there must be trash bin icon on top of the right side.
I want to ask a question, does SUNO's right to use songs depend on whether users pay or whether users share the songs publicly?
Are you asking what Suno (the company) is allowed to do with songs, or what users are allowed to do with Suno songs?
I am asking about the copyright ownership of a song for the lyricist. In my opinion, the lyrics copyright should still be in the hands of the user regardless of payment or not.
yeah i know where it is
Sonu will only display a song on their Trending/New page, if the song is set to: Public.
depends on your subscribtion i think
if you dont pay, suno owns the copyrights
not sure tho
your lyrics stay your copyright.
the ToS allows them to use your lyrics to make the song.
In other words, SUNO will not use any songs that users have not published publicly?
Suno will not publish your song on Spotify and make a lot of money off your hit.
Can you explain what 'use' means to you?
I am not sure I am answering the right question
The question comes from Intellectual Property Rights at suno.com/terms
you understand that Suno has to use your lyrics to make the song...
and IF you set to public, or share the embed link, Suno plays your song.
That is technically them using your content.
Some of this is very basic stuff, like how facebook says they can 'display' your content forever...
Some users think that as long as they use SUNO's service, all the copyrights of the songs belong to SUNO. I don't think so, but they asked me to look at suno.com/terms, and I can't understand it now.๐คฃ
If you use a free SUNO account to make songs the ownership belongs to suno. In this case you can share the song (Youtube, Socials, ...), but not monetise it and you have to mention suno โMade with Suno.comโ.
If you have a paid subscription all songs made DURING this paid period fully belong to you. You can monetise them in any application you want. The songs made with a paid account stay your songs even after the paid period ends. โ๏ธ
They don't claim to own your lyrics..., they just create an exception where they are allowed to 'use' them...
If you are concerned about keeping your songs private, leave them 'unlisted' (the default) and do NOT share the url to the song, or a playlist with the song in it...
The long scrambled CDN number is your protection, don't leak it.
I agree with what you said
THX
Is the moderation system down
I was making a undertale sans geno song and I won't let me send it through
Even though my flowey song has worse words talking about killing
Any tips on how to prompt for okarina? Okarina term itself doesn't seem to be recognized by Suno. I have tried ceramic Vessel flute, but it doesn't resemble okarina.
might be a glitch...?
try adding some blank spaces to the front of lyrics, add a [break] at the end (can trim with Extend from time), add a hyphen to the start of a few words...
you want to force Chatgpt (the censor) to re-process the language tokens, but leave the lyrics alone so Suno will sing them the same as before...
i have been having limited luck with using "..." for a few lines at the start to get that blank space... might try that...
Is there a list of the available song structure meta tags?
Also see the other wiki entries, very useful
Thanks!
I was digging around on the site itself but didn't see anything too helpful
Hey all! I haven't started using Suno yet, I have some questions...
if i have a song where i have written lyrics and melody myself and then let ai create the actual "music", then who owns the song? So if I have a subscription account? Thanks ๐
If you have a paid sub you own the entire song
As in there isn't any strangeness with it. If you are paying, you own the entire thing, full stop
Do whatever you want with it
thanx, but if I end my subscription, do I still own the song?
I just heard somewhere that Suno owns the song if you stop the subscription? but thats not true tho?``
aha ok thankyou ๐
That is not true. If you made your song while a paid subscriber it's yours fully, 100%. You do not need a sub to continue to own the song
You already own it, that won't (and legally can't) be taken from you.
even if, let say the song gets very popular?
I just dont find the terms of this on Suno webpage, do you know where to find that? 0)
Hello! Please email billing@suno.com if you are unable to find the answer to your question on this page.
Specifically:
wow thankyou soooo much ๐
You should use english I guess?
OK!My English is not good and Iโm afraid of using wrong grammar, so Iโd better not ask questions! 3Q
So anyone know? Is it possible to use your own song and make the ai "remix" it? I only find how to give the ai my own lyrics but not my own melody...
There is not a way to do this, and most likely won't ever be. Suno doesn't want to be held responsible for whatever you upload
They have no way to check or know if it is yours
Technically it's possible and 'easy'
Legally, no, it's a giant ball of redtape
aha ok , do you know if there is ANY other ai that allows this? ๐
It is fine, ask away. AI or google translation should be good enough to understand what you want to translate.
If you find it too hard to talk here, you can also try asking at https://discord.com/channels/1069381916492562582/1154780167047217233
Stable Audio is capable, but not nearly as nice as Suno
If it makes you feel better though this capability is certainly coming with all sorts of other not-Sunos
The tech is here already
Someone will release it out
and then that will be that
As a corporate entity though, nobody will touch it
Suno, Udio, etc included
so you mean that if you give AI you own song to redo, the risks for them increase in some way?
It isn't if you give them your song