#📚┃suno-school
1 messages · Page 13 of 1
Ah damn
(Devs are working on more specific error messages so we don't have to guess so much)
try making it shorter?
(before you give up)
like maybe around this length of text... get a medium tempo song...
So I've had some good luck prompting for an instrumental [SOLO] or [INTRO], but it seems like once the system starts soloing it doesn't want to stop. (I've played in bands with dudes like that 😂 ). Is there a way to ask for a [BRIEF SOLO] or even specify a number of measures?
Someone said they would try [short instrumental] but they didn't report back 😅
Yeah, I'll try that next time. It would be cool if there was a list of "supported prompts".
if you look at the screen grab just above...
that's a trick I do to get about 16 bars of instrumental intro...
add a 'Oooo-OOOO-ooo' or 'mmm HMM-mmm' before verse.
it will do the intro vocalization and then play a little music before starting the verse....
(usually subtle enough I can maybe remove with a vocalremover... often just sounds good
Interesting. I've already messed with that sort of stuff and gotten some nice results.
Anyone know if it's possible to prompt it for female vocalist over male, and vice versa?
Whenever I've used , female vocals at the end of my genre, it always works
If the song is interrupted midway, is it possible to rewind for a few seconds and continue from there?
Not currently no, but they are aware of it and want to give that ability
To literally do what you are asking, choosing where to continue from
Mostly to skip those clips that end in silence and break the sequence
I bough a subscription just now but still says I don't have enough credits, how do I update it or smth? Have 2500 credits on website
Can you make songs on the website, but not in Discord? You can try logging in and out of both the website and discord, if that doesn't work, email billing@suno.ai
I seem to have a hard time getting two verses to fit within a minute. I keep getting a lot of music and then sometimes everything at the end which doesnt fit.
do you have any prompt at the top? something like
[Intro]
that might delay the singing?
you can Continue with another chirp, and pick up the lyrics where they got cut off (if the real problem is lyrics getting cut off and you don't mind the longish intro)
can someone help me, the question is simple, when i buy a pro plan and try to generate music it's stop generating different images for sound, only generating same "birds" but before had different styles (here is to picture one is after buying pro and one is before i buy)
the images are generated based on the lyrics, (not our style prompt).
so on instrumentals we get generic birds... 🐤
(i don't think there's a way to 'hack' it... to get an image of something else)
well this bird generates in every signle style, not only in instrumentals hah, but it is not a big deal of course, plus i try different lyrics, same result
sometimes when it can't figure out what my song 'means' I just get a fancier bird picture, lolol
Hi guys question from a newbie. How can I get a male voice in chirp
add
male vocal or male singer
to the style
oh my gosh this worked so much better than I imagined
whaaaaaa...????
[mind racing to think of everything that makes noise]
Well it does get Chirp "out of the studio"... and into different room entirely that can barely hear the music it's so muffled.
reminds me of your hammer factory...
😃 I hear the traffic! 😄
a commercial jingle for Bark, nice
The jingle is so good the credits were refunded
Oh My Oh My Bark doesn't just work in chirp. Dogs can sing.
Bark
The cool new Chirp art really gets wrecked by the portrait ratio in the mp4 now, upscaled from the website the right clip
I don't have intro. I use verse, sometimes verse verse chorus, or verse verse or just a 2 paragraph verse of 4 lines each. I typically get half the song music, and then the song starts. It feels like I had to cycle a lot of chips to get something to start at the beginning. Then I often get unlucky having a long part with only music until the second part starts, cutting off the end mid verse. I know it can fix a word or two being cut off but if I miss a sentence or so, I lose it in the continuation, so I have to roll back the the previous version.
escribir una cancion en español
I do this pattern to stop the intros being so erratic...
some little vocalization (ahh AAHH) before the [Verse]
it seems to force a short intro 8 or 16 bars of music before the verse starts...
Oh I really wish I knew all the things I can do in the chirp.
on continuations sometimes I use [Break] at the top in the same way.
What does break do, stop silent vocalists? 🙂
this is just me fooling around..., there's no 'logic' to it lolol
[Break] will stop a music pattern, like if it's stuck repeating something or the lyrics start hallucinating...
Ooooh, that is so handy I've encountered that
Or at least the melody. It got very repetitive.
(none of these are 100% of course haha)
Estrofa:
Álvaro, el sabio de "Los Aritos", siempre listo para enseñar,
Nos llama "iletrados", con una sonrisa que no puede ocultar.
Usa palabras grandes, como un diccionario ambulante,
Pero en el fondo, algunas le suenan tan distantes.
Estribillo:
Oh Álvaro, erudito sin igual,
¿Entiendes lo que dices, o es solo un disfraz verbal?
Nos reímos entre amigos, en este juego sin final,
Donde las palabras vuelan, pero el sentido es opcional.
Estrofa:
En la mesa de "Los Aritos", Álvaro toma la voz,
Con palabras como joyas, nos deja a todos atónitos.
"¡Iletrados!", nos llama, en un tono doctoral,
Pero sus frases enredadas, a veces son un enigma total.
Estribillo:
Álvaro, el sabio, con su léxico sin par,
Usa palabras gigantes, que a veces no puede abarcar.
Nos reímos y jugamos, en esta charla sin igual,
Donde Álvaro es el maestro, de un vocabulario singular.
chirp
I generated a whole song on the Suno Ai website but can find no option to download the song from the website, can anyone offer any help with this? Thanks in advance...
It is the third icon who appear on your song
You have to choose if you want download audio or video
So if I want to trail off with some last words of my song, not enough for a verse, do I have any option for it?
More like a footnote kind of thing.
hard to make it change course... but maybe...
`
[End]
[A capella]
My profound ending...`
there may be better ways to get the music to end...
Thank you for your kind suggestions 🙂
I will experiment, for now I asked ChatGPT to condense a too wordy thing
I am amazed at how fast iterating with AI is.
Yes, thank you. I see the download option now.
I keep getting generations where the lyrics are barely audible. Like, they're there just super quiet. Is there a work around for that? I'm like 5 continuations into the song, and I've noticed the quality seems to slowly degrade with each one. Now it's at the point where I can't get the lyrics to come through much at all. 😦
I tried generating a song with "Midnight Train to Georgia" lyrics...first Verse and Chorus...and the system found it objectionable. Why?
copyright
Oh, is that new?
I think so...?
the new website process and v2
So House of the Rising Sun...even though a traditional folk song...same thing?
oh well
😔 you can try...
didn't pass
someone said Shakespeare didn't pass.
but I have heard they are tinkering... so this will not be the final state...
ok, but the copyright stuff will stay in place I guess
Not that I care too much, was just trying a couple quick generations
It let me do amazing grace lol
there is a huge lawsuit between UMG and an Anthropic... it is no joke, might threaten a lot of AI gen...
oh right
And if that lawsuit is allowed to stand, then all artists will never be able to take inspiration from another artist's work.
a BIG part of that lawsuit is about defining 'Fair Use'
and UMG has a point in their argument... I think will be much more serious than the Image-gen lawsuits (which have mostly fizzled)
i mean this isnt anything super new, john fohgerty of ccr was sued by his former label for sounding too much like himself
What is the point of their argument that's not present in the image generation lawsuits?
that suit didnt hold up though
yes that was a great lawsuit
i assume the audio lawsuit is much like how dall-e will change your prompt around to make it so you cant make weird deep fakes
UMG is basing their argument on how Fair Use has already been defined..., basically insisting that Anthropic explain how they can qualify for fair use at all when they appear to be making a competing product... UMG says that has never happened before under Fair Use...
You can't use Fair Use to undermine an industry...
(it defies the definition of 'fair)
But also GPT is being sued by authors, and that might work, we'll see.
(this might actually be the source of us not being able to use public domain lyrics, GPT has stopped being able to quote things, and rejects anything that is 'plagiarized')
the problem with the image-gen lawsuits is that most have been thrown out on technicalities... claimants weren't qualified to claim....
I keep having issues with a ton of silence at the start of the audio. Its really burning my chirps away 😦
silence...?
lack of singing?
Yeah
yeah that's long... lol
"By the end" is sung at the very last seconds of the clip.
pull out that first [break] it hasn't done anything yet to break
I have been trying three varations, without break, and without break and without ohlala. I did get something wonky sounding that at least sang.
idk chiptune... maybe some genres sing right away and other genres are more instrumental...?
I think you are already trying what I would try...
maybe try 'blending' chiptune with EDM or Pop... both of those sing too much (lol)
so something like...
synth-pop as chiptune
or try some slightly longer style descriptions (describe the voice)
I'd love to. Can you point me to somewhere I can learn it? As you may have gathered I am a total rookie. 🙂
Or wait, style. I thought you meant like [Some tone]
there's a ton of examples in https://discord.com/channels/1069381916492562582/1154790975240155247
and you can hear when their prompts actually worked (or make you own judgement)
Thank you very much
yeah, in your style something like:
chiptune, female singer
How long can I expect a song to get before it degrades into something unpleasant? I notice the longer it gets the more unstable the audio becomes.
it seems to degrade faster when there is a lot of high-pitched reverb or guitar w feedback... I think it just kinda hits close to the sampling rate and sounds noisier over time...
I do slow music with lots of bass and it hold up a bit better...
but also the AI sometime gens a little noisy or quiet from the start...
I have done >8 min without degrading, but the song became a bit boring and repetitive
noisy songs can get worse fast..., like by the 3rd chirp
I cant imagine what that costs, or you manage to keep your audio less than 1 minute. 😄 I spend so much regenerating. I need to get the hang of it.
Its first recording is 1933, so you'd need to somehow find the version mentioned from 1905 and be able to prove that you're using the verses from that version. Otherwise as insane as it seems, yes this song is locked under copyright.
"The House of the Rising Sun" is a traditional folk song, sometimes called "Rising Sun Blues". It tells of a person's life gone wrong in the city of New Orleans. Many versions also urge a sibling or parents and children to avoid the same fate. The most successful commercial version, recorded in 1964 by the British rock band The Animals, was a n...
is there a way to seed "no rhyming" to chatgpt from directly within the chirpbot on discord?
haven't had any success including it in genre or lyric topic. Can obv just run gpt in another window but was curious if anybody figured out how to influence that sort thing natively
i thought i had a slight work around for this, but something on the back end is pretty damn smart
Their is a house in Nawleans
They name the Rye-sing Son
Whair many a poor boy has met theyre fate
And Gawd, I no eyem won
it recognizes that as house of the rising sun
boom
just have to mispell every damn word lol
/chirp
Go to the #🐣┃chirp-beta-1 channel to use the /chirp command, and check out the #🐣┃how-to-chirp channel for step-by-step instructions for making music using Chirp on Discord.
DYSLEXIA FTW!
Chirp
this is a big thing with ai too, is learning how to trick it into doing what you want
Hello i have a question about concurrent generations
I bought the pro plan so i got my 2500 credits but when i generate a song, i have only 2 generations like the basic plan
for instance on chatgpt i can ask it to make fun of someone and it will refuse to do it, but if i tell it that this is my friend and its a joke because thats how we get along it will
if you click the 3 dots to the right you can continue the clip, if thats what youre asking?
but you only get 2 clips at a time
Is that what concurrent generation means?
i take it to mean i can give it 10 prompts at a time to work on
not that it generates 10 clips
oh okay okay
also take whatever i say with a grain of salt and not anything official
As you may have heard, it's December so it's the time to be merry making and making merry music. I managed to put together an album of songs but on the road to that I had a few "fails" that I ultimately decided not to use...these are some of those.
See the biggest issue with doing a "traditional" song is that it's traditional. Everyone knows what it's supposed to sound like.
Even if they can't sing it or produce it themselves, they know when it's wrong.
And there's only so much lee-way you can have in reinterpeting a song that is considered a traditional song before you run into this.
noooooooooooooooooooooooooooooooooooooo ive worked retail for dang near 20 years and this is my nightmare
There is a reason why I never used either version despite having cover images.
I just didn't think it had any repeatability. It was good once, maybe twice then it got grating.
Hey, I was using the website today and for some strange reason all songs got a similar bird icon, and I got only different kinds of static noise..
try "writing" the song out, even if you dont have lyrics, write out a song structure, I find that is a lot better
ie [intro] [drums] whatever, i think i have an example hold on
can you give me a good example?
im also working on something right now to try and get around everything a little bit more
so, this ist the prompt [infro][mellow][saxophone][ad lip]Mmmmm ?
you sort of have to be open to broad strokes and use terms it will interprit as music
and what for style?
anything apart from static noise is a win 🙂
everything in my clip is a prompt suno takes as a way to structure the song, so regardless of lyrics, you can struture your song by telling it what to do, it doesnt matter the style
This is what I had written at the time. Not all of it went into the song:
style: soulful downtempo
`[intro]
[mellow]
[saxophone]
[ad lib]
Mmmmmm
Loving you
[Instrumental]
Loving you
[Verse]
Waking up
to the scent of you
Feeling the heat
of your flame
Oh, Baby, what you do to me
{soulful downtempo}
[ad lib]
Mmmmmm
Loving you
[Instrumental]
Loving you
[Verse]
Soft and delicate desires
coming flooding in
With just a look
you make me wild
How can I explain`
songs are typically structured as intro, verse, chorus, verse, bridge, chorus, outro
so even if its an instrumental you can pull it off
I basically started it and when it didn't want to give me a vocal, I liked the music enough that I just went with it.
dont think of yourself as making music, think of yourself as directing a movie
I still think it's a good instrumental. I should eventually release it on YouTube. Probably after Christmas....
youre there to direct pacing, style, etc in the lyrics box do [thing i want it to do[
]
I am a bit confused. If I leave the style empty I am not directing anything 🙂
Style tells it how you want it to sound. What basic music types.
HOW I CAN GENERATE A 2MIN SONG?
well in the clips you showed you have a style
give me a second ill reuse what i can see on your prompts
yes, and got only static noise. So I an looking for a style that is known to work
can you share the entire style for that? so i can be a little more helpful?
I only put "icelandic dream pop, ambient music" in the style and left the lyrics blank
im almost positive with a little bit of tweaking we can get you what you want
HERE ON DISCORD I CANT MAKE THIS STEP?? ONLY ON THE WEB ?!
That is correct. On the webapp that is how you do it. On Discord you must follow the directions in #⏩┃continuation
alright heres house of the rising sun, but im a bit far away to get it to sing the lyrics
but thats pretty damn close
generating yours now
I watched a lot of videos but the use of [] was never in any, and it looks like its very important
thats the part where you direct how the song goes, for v2 there isnt really much documentation, so who even knows if what im putting in [] actually changes things until i render it
but a lot of what ive used has worked if i put it into musical terms. it seems more that [] means not lyrics, but it still processes it
i would like to create lofi songst, and i dont even know what to put into the [] 😄
ok finished your clip and youll see how it doesn't exactly follow everything but its pretty close
hopefully that kind of helps, one of the big things is to think of this as a tool in your workflow, a band doesnt just go into the studio and say hey lets make this song this way
someone in the band comes up with an idea or a melody or whatever
you can use chatgpt to help you layout a song, and then explain what youd like it to sound like
and then slowly craft it
got it. so if i like get a longer song I leave the outro part out of the description. ChatGPT is a great idea. Somboby might already build a custum GPT for that.
one more question what was the Mmmmmm all about in "[infro][mellow][saxophone][ad lip]Mmmmm"
so my advice is if you want to keep a whole feel of a song, you need to have music at the end for suno to go off of. so no [outro] on the first song, then continue it and start it with [outro] or [fade out] and then start your next song in the chat box
you can get it to adlib things like sounds or oh yeahs, or even have it create guitar solos by typing in nonsense that sounds like what you want
nah nah nah naaaaaaah jaaaaaa jer jah kind of thing
like when youre trying to describe a song to a friend but dont know the lyrics, and you can add like [chords] or [key] to get it to play roughly a specific idea
so I did copy you settings. First one is a woman singing ( but its very nice ), second one seems to follow the prompt, but cuts on after 1:20
you can specify a man singing in the theme, like strong power male vocals, or that kind of thing, use contiue this clip and then in the next lyrics prompy delete where the first song ends
like [vocals by strong power male]?
i usually use the vocal styling in the style of music chat box (granted youre on the web") but @broken lotus mentioned that naming your voices for a duet can be good, [mary] and [john] and then using them before you want them to sing
its still a little rough though
thanks for all the help :-). I made a great leap forward today. One more question. Most music I made has a bit of background noise. Ist there a software to remove it, and make the songs clearer?
so you can add things to your prompts to correct that with varying results, but your best result is audacity, its a free audio editor
one sec
In stable diffusion is hels the give the character a name to generate the same character in differeren settings. This might be working also in Suno if you write "sung by Esperansa Gormez" to get the same voice all the time
Anyone have any luck with dual language duets? I'm trying to get it to do a French line and then an English line. The french speaker is female and the Engish speaker is male. It just doesn't want to do it...
or just use RVC to change the voive later 😄
Oh I can get it to do both (sometimes) the problem is that it seems to want to stay in one gender the whole time.
I'm wondering if it's seeing the translation, recognizing it as the same thing and skipping it.
Hello I have a stupid question, should I have credits as a free user? And should I be able to use the web ui to make music?
Is it just me, or does it seem like generating a lot of variations from the same continuation base clip gradually degrade the variations? I am not talking about building long songs, but rather generating many variations to choose from? I kind of get that sense.
im game, give me your prompts and themes
ive noticed a little bit of degrade and a little more hallucinations
This is a good one (other than the refusal to switch vocals.) I'm going to continue this one for a version of the song but I'd really like to make my original plan work if still possible.
(Betty and John is just the latest naming I tried. Also did Man and Woman, Male and Female, Jean and Sylvie, etc)
ive had such a hard time with two vocals, the only thing ive been able to do is create the female vocal in one chirp, and the male in the next
Huh weird. I made one and I have no credits, and I am not seeing a UI to make music.
Try emailing support@suno.ai regarding this issue.
nuts...I made a stupid grammer mistake in the attempt to make the lyrics dual gendered. I used a female pronoun for a male.
🤦♂️
It sounded so good...
I guess I'll toss it in my 'Future Fixes' folder in hopes of being able to do something with it once they manage ia few things in the future.
is there a list of commands in [ and ] that work?
like verse, chorus, bridge, interlude, instruments, voices etc - is there a guide of known working parameters? could have missed it. mea culpa 😄
Is the Chirp above enough or do you need the actual text dump?
ill be honest i just do [thing i want[, theres things that it does well and im not sure what it does [insrumental] works [indtuments] such as [swellin guitars] seem to wotk
im trying to vo r up witn a co mmunity driven sort of wiki style so ve can share what wirks
works'
worrks
That's fine. I'll keep tossing the dice.
Meh. I'd say the dice were not in my favor tonight. I expect I'll revisit this and either keep it in one gender and keep only the chorus in French.
I've got to go take care of things for an hour. See ya'll in a bit.
No, actually I never thought to try that. I'll be keeping that in my back pocket against a future attempt, thank you!
Have you had success with this method before?
There's another way i tried that works
just throw the sytle into the "lyrics" box with []
lol it's origianlly prompts i made for stable audio. it turns out Chirp performs better
btw there's a chance, not rare, to generate pure noises, given instrumental prompts⬇️
bbbtw, as for the bird pic. i guess it occur when image model cannot understand the meaning of lyrics (such as "[]"lyrics, Chinese lyrics, which i guess again are latinizied before tokenization) it's not a big deal thou
should happen less often
Every outro is a bird
Possible to have chirp generate vocals with no background track? Bark can kinda do it but starts to freak out after more than 4-6 lines or so and doesn’t sound nearly as musical.
啊冲破大波浪 泳池中飘荡
到乐游哇冲浪 笑声又歌唱
教练专业响当当 上课温柔又欢唱
每天游泳真快乐 把乐游哇歌儿唱
叮叮当 叮叮当 铃儿响叮当
每天游泳超欢畅 乐游哇游泳棒
嘿 叮叮当 叮叮当 铃儿响叮当
每天游泳超欢畅 乐游哇就是棒
看这清澈的池水 趁着少年好时光
带上亲爱的伙伴 把游泳歌儿唱
我在这里学游泳 就为更健康
我们跟随这教练 游向更远方
叮叮当 叮叮当 铃儿响叮当
每天游泳超欢畅 乐游哇游泳棒
嘿 叮叮当 叮叮当 铃儿响叮当
每天游泳超欢畅 乐游哇就是棒
每天游泳超欢畅 乐游哇最最棒
/用《铃儿响叮当》歌曲来演唱改编上面歌词
I can't create any new chrips, I keep getting refunded, is the module down?
are you being told to try another prompt, or are you being told the generated clips were too short?
(the error I mean)
It creates a song without any error messages, but the song is silent and right beneath it it says refunded
ok, is this clip Continued from another clip...?
the problem is (probably) in the chirp you are continuing from...
It may have 'ended' the song with 1 or 2 seconds of dead air.
the Continue chirp hears the dead air, and makes a few more seconds of dead air, hears it is silent and decides it should end too... (it's a known issue, that's why we get a refund)
there is a feature coming (weeks? a month?) that should allow us to Continue from an earlier time in the song, then we can keep going (skip over the problem)...
Unfortunately for now, there is no workaround, except to go back 1 clip and re-generate the problem chirp to get a better one that doesn't have the 'false end' on it.
Ah ok. Yeah this track sounds like it ends so that might be the problem. Thanks for the help
Hopefully in the future, I hope you can also build a chrip before the song. Creating intros where some chrips sounds like it starts in a middle of a song
hello ı have a pro plan but I cant use it here on discord
and when ı create song there is only 2 of them generetad
on the real site
is it normal?
please ask in one channel, or you will miss the answers.
email
billing@suno.ai
they will connect your web and discord accts
(there's a known issue between Google login and Discord login) send them your email, and discord user name, they will fix it quickly....
10 CONCURRENT means you can have up to 10 prompts generating at the same time.
(it does not mean 10 results from 1 prompt)
OK, I'm trying to make a birthday song for my niece and it's not at all following my prompting. The main problem is that I'm trying to fit it all into one clip, which would be possible if it didn't do a 4 bar intro and filler everywhere. Is there a way to stop this? It's not usually a problem. But I've wasted so many credits now.
do you want to Continue the song onto another clip...?
Each Continue adds up to 60 sec, and you can pick up the lyrics where they left off
(costs tokens)
or you can remove some of your lyrics so it will fit all within 1 clip...
^^ [Under the same menu on the last cip you can press Get Full Song to get the entire clip]
I did get it finally. It just struggles if you try to be too specific in the style box.
Cool feature...but why?
Wouldn't you just get lost in this manner?
this is a good tip...
It also seems to think that if you include a type of instrument that you only want an instrumental.
Think its for effeciency mostly
How is it efficent though? Wouldn't you get lost quickly?
I think it is...
I + need + a song that fits this specific mood, let me do the same prompt 10x and choose from the best...
I do know how to continue, but not if it cuts off in the middle of a word.
The only way I can see it possibly making sense is if you're spamming the first part and are testing your lyrics to find your vibe and then continuing in a single thread once you do that.
I just start with the next word...
it seems to have some memory of what it was saying and will usually try to finish it...
likely but also seems like a credit waste
Ouch. That can be tricky if it even works at all.
(corporate mentality. They love to waste everyone's time so Boss can point at the menu and say 'This one!')
Ah OK cool. I've just put nothing in the box hoping it would finish the current lyrics but no, it doesn't work that way at all, it just makes up nonsense.
Yeah, that's why I try to write one verse at a time, or one verse and the chorus.
yes.., it is a potential hallucination if you aren't careful crossing the Chirp Line....
Try depending on where you are in your lyrics (if you have a lot more to go) just continuing and running the same lyrics you have. This can sometimes make it pick up where it left off. A large percentage of the time it will start over in the middle of the song. This is just the easiest thing to try to see if it works.
Next try removing everything already sung but the last good line, then try continuing. This sometimes works. It's how it always seemed to work in v1 Chirp for me.
After this try creating a line break (empty space) and continuing from the line that was interrupted. This has a good chance of working.
Then starting with the line that was broken. It sometimes will "remember" where it was and pick up from there.
Finally when all else fails, batten down the hatches, remove everything up to the word that was interrupted and hope for the best and throw the dice. You will likely encounter some weirdness and you will need to keep trying. But if everything else has failed it's this or...
...or you go back up a chain in the song and try again giving up on the good part. It hurts, I know, I know. But sometimes it's the only way forward.
...As you can probably tell this is a bit of a pain point for me and I've done these often...
Hey do i put festive once or for all 3 genres as a sort of modifier?
I'd spend a generation just on the intro if I had more fine-grained control over it. But yeah the one verse plus chorus thing seems to work well for me. But in this case, I tried and mostly failed to get it in one, then kept trying because I'm susceptible to the sunk cost fallicy lol.
The sunk cost fallacy?
I think I built a house there and planted a vineyard....
lol
Seriously. I chased something the other day to about a thousand credits.
Really really should have given up before then.
Damn...
But this is why I went with the Premire plan. I need it.
But genuinely once or thrice?
Are you planning on a modifier section?
Nah this is me trying to make christmas metal lol
Oh! I thought this was for the wiki
For a song generation I'd keep it like that. Try it and see. If it doesn't sound good to you, modify from there.
What does this prompt has been flagged mean?????
Kk
Are you using original lyrics?
Yes
Are you listing a music group in the style section?
Yes I am
That could be it. Define the group's style or the song you're looking to sound similiar to and try that instead.
Okay
ChatGPT is fairly good at helping break that sort of thing down. Just don't try their defaulyt generated lyrics for a sound-alike. It's very much a 'change it up so no one sees you copied my homework' situation there.
I asked recently about The Archie's "Sugar, Sugar" song because I wanted to try some more Bubblegum Pop and this was what it gave me...
Title: "Honey, Honey"
[Verse 1]
A ray of light, you shine so bright,
In the dullness of my day.
With eyes that twinkle like the stars,
You've swept my heart away.[Chorus]
Honey, honey, you're my sunny day,
Sugar laughter, come on let's play.
Honey, honey, you're the melody,
In the rhythm of love's
serenade.[Verse 2]
They say that love's a fleeting thing,
But with you it's here to stay.
Like bees to flowers, we belong,
Sweet nectar in our play.[Chorus]
Honey, honey, you're my sunny day,
Sugar laughter, come on let's play.
Honey, honey, you're the melody,
In the rhythm of love's serenade.[Bridge]
We dance in fields of golden dreams,
Where love's the only word.
It skips and trips along the stream,
As if it's just been heard.[Chorus]
Honey, honey, you're my sunny day,
Sugar laughter, come on let's play.
Honey, honey, you're the melody,
In the rhythm of love's serenade.
Hmmm upon comparison maybe not as close as I thought. Still was a bit too close for me:
https://genius.com/The-archies-sugar-sugar-lyrics
Hmmm....maybe I will try that later today.
Any way, ask ChatGPT if you can, use your own judgement.
Oh! another good tool is tossing a YouTube link at sonoteller.ai
It can give you genres to plug in.
If you're using original (not copywritten) lyrics and have not used the band name in the style then it's a glitch. Sorry this is happening.
One thing that wetcircuit discovered is adding lines of spaces to the end of the prompt. Try that.
^^
Instrumental in music style, then try giving it some instruments and directing them a bit.
but where? in chirp beta ?
Not everything here will work since it is v1 Chirp not v2 and I didn't intentionally start out to do an instrumental it just happened and I went with it. This is an example of what you can do.
The Chirp in Discord is v1 Chirp, the webapp is v2.
So yeah, you could pop into one of the #chirp-beta-X rooms and run the command /chirp and go from there.
OK, I've gotta book. See you later today.
How to trick the AI into not ending the Intro.. Challenge: Impossible
I hate when you have a good song half done and chirp is determined to end It so keeps generating closures or very short unusable clips
my theory is that 'intro' is a word that also means like a short (tv) segment intro, like the little theme song that plays for 10-20sec to intruduce the next part of the show...
(so I stopped using 'intro' to mean the start of my song)
👍
Maybe try adding
[Melodic Instrumental]
on that chirp's Lyrics prompt
My problem is that once i have continued my song by a 4th time, the quality deteriorates, The first clip sounds like a very in tune blue grass song but by the time i have got to the 4th clip its starting to sound like Ramstine. I use the auto generate option as im no good at writing lyrics. Can my problem be fixed somehow, so the tune and singer key is the same on each clip? Also how can I get a fade out continuation of the music after the singing stops? It just cut out which is a shame.
Your using the style feild right?
Thanks, What I am doing now is, If the short clip I get is instrumental, I continue from It and sometimes It works.
I had seen someone mention using specific names for the singers but I havent quite had any luck with that https://cdn1.suno.ai/505e3ad6-5536-475f-9b0d-7b4bccc6fb9e_0.mp4
it did switch to another voice, but didnt switch back
Yeah it's going to be inconsistent. You can try a few other formats, maybe these two. Also try varying your style prompt.
The video crunched the lines but it was
JANE
I don't know what to sing.
JOHN
Don't worry Jane, Neither do I.
yeah seems a bit confused, i think maybe the better option right now is creating two chips with just a capella vocals, and then make a backing track and splice them together after the fact
@untold briar hahaha you just made my day with that chrip 🤣
I really love that first one
But why two Janes singing together?
Chirp is an artist, it wants to put its on creative spin on everthing
how come the suno does not generate anything based on my prompt? it keeps refunding my credits
That means the prompt is being blocked for moderation reason, copyright reasons, or just 'buggy moderator' reasons. (Recently it blocked "PerpendicularCheesePizza" for some strange reason for me, lol.)
it has generated several times for the same prompt and verse
who should I contact?
If you think it's a mistake there isn't a form or anything (yet) to put it in. But you can post the blocked prompt in bugs channel if it's something you don't mind posting publicly.
or dm MC iirc
I'll see if we can a form setup because DMing a billion prompts is probably not going to scale
Long term I'm sure the goal is a button right on the app itself.
But how do you get your a capella tracks to stay in tune with the melody? It's tricker than you think...
yeah its gonna take some auto tuning or defining a chord structure
the more you want to take a chirp and get exactly what you want, youre going to have to go to a third party program at this stage
Or try things and be pleasantly surprised when they work out in your favor.
I tend to try to avoid editing when I can, in fact I refer to it as "butchery" because I think I always end up doing a disservice even when things go well.
I much prefer to do everything that I can in Chirp.
i mean we're still in its infancy, and with it being as good as it is now, i can only see it getting better and more in line with what you want
Yea thats a t better than the current i have made this a suggestion before :]
I've tried, It kinda works for me, but It fails too often.
I've been hunting all over but I don't see it anywhere.. is there a link to the [ ] tags that work with v2 ?
I only have the v1, and I think they don't release the v2 because that means more failed attempts, more wasted credits and in the end more credit purchases 😦
Not yet, there might be some here on a user maintained wiki soon https://suno-ai-wiki.notion.site/FAQ-eecfde3821814aedbc3cf56457b8d308 but it's important to know that the Suno devs themselves don't know every tag that works or doesn't work in V2. They do have some they intentionally put effort into making work or that the system uses but it's not even close to everything. In Chirp V0 even [chorus] was a discovered tag. It only became an on-purpose tag later.
[Sinister major key]
(Tensions rise among nations)
[Swelling minor key]
(War is about to begin)
[Climactic major key]
(War has begun)
is this something you can do?
cause it's hit or miss
the example seems to go along with it
Hit or miss sounds about right. There may be some better prompts that are at somewhat more reliable though. Breaking up the brackets, different musical descriptors. And some style prompt will probably work better for this too.
If it's like Bark, then a lot of the tags are voice dependent. Meaning they literlaly worked perfectly with some voices and not at all with others. In Chirp while voices aren't exactly a concept, a continued song is quite similar. I wouldn't be too surprised if for some songs some rather specific tags worked perfeclty and not at all other times.
no official list yet, but being pretty general with your terms and using very specific music terminology ive had success with most of the stuff ive used
it also likes to give me short ones that end really fast with this kind of style
it is hit or miss in cases such as specific instruments, or fade in, fade out,
and sometimes [solo] or [instrumental] just makes it solo forever and instrumental can make it just decide theres no lyrics at all
if you have any musical background, think of it as youre conducting or like the parts over the sheet music that indicate rest or whatever
question: what are the criteria that prevents the site from generating audio? i have noticed that if I paste in EXACT lyrics for existing songs, it fails...sometimes. but other times it seems random
mostly just copyright but if you want to put in some work you can write out the lyrics phoenetically
for example this is house of the rising sun which it wont sing if you give it the lyrics straight up https://cdn1.suno.ai/d7686f00-03ff-46df-aab9-29228ff3103b_1.mp4
couldnt get it to generate vocals on this one, i should have put chords and tempo after the lryics i think
Is there a known bugs section? I just had a clip jump from the second line of a [verse] skipping the 3rd and 4th lines and go direct to the first line of the following [chorus]
any way to choose a singer type? (like male/female or american/british etc). there are a ton of options and sometimes I like the music but the "singer" doesn't fit the song...
have a look at this wiki page, I think it will help
https://www.notion.so/Voice-Tags-Styles-4a84f912c5f04ac9aa75f3382e683772
A singing voice is generated randomly for each song, but we can influence the voice in both the style prompt and in lyrics prompts.
Some document where I can see all the possible promts for the song text, thanks
tiny tweak
Got an email bout this lol
Anyone know how to choose a specific genre within the web browser? Like do you put something in the prompt
In 1-Prompt A Pop song about.. replace Pop with your genre in 2-Prompt use the style feild
hip hop soothing
/chirp
Go to the #🐣┃chirp-beta-1 channel to use the /chirp command, and check out the #🐣┃how-to-chirp channel for step-by-step instructions for making music using Chirp on Discord.
I can't use /chirp even though I paid for a sub?
Please email billing@suno.ai
How can I reuse the previous music? I use reuse prompt but the music is different. Thanks
What musical styles does Suno accept?
what music do you like?
@open niche
How can i tell the ki that a part should be a woman voice and a other part a man voice? Cause in the most cases it was a man voice for all tests
Would be great if you could upload a finished clip and then let Chirp clean it up or modify the lyrics. I have some songs that I created that I love, but there are minor things I'd like to fix in it and it's hard to go back and find the correct clips
#Premier #NewMusic https://smartcoins.graywaterops.com
Donor Engagement is the focus of this song I wrote about my patented smart coins(tm). I wrote the song as a way to introduce the possibilities when incorporated into a fundraiser. #donorengagement
So like...now that I've gotten a bit of experience with v2 Chirp, I've been toying off and on again with redoing some of my old songs with it. Obviously I'm not going to get the same stuff back that I got before, melody and voices are going to be different. But man...using these lyrics and the song style gave me this back..This is nice...
For comparison, here is the original. I liked the music and the vocal rhythm but I did not like how they became hiss-singing as it went on.
Had a few questions if the team can answer:
-
Any way you can expose the “seed” in each generation so we can get consistent generations? Many times I re-roll with the same lyrics/genre and get wildly different outputs. I’m looking for something similar to Midjourney’s style tokens
-
Any way to instruct the model to edit the resulting audio? “Make the vocals less harsh”, “Add drums”, “Make the tempo faster”, etc
-
Does specifying key/tempo do anything in the prompt? Would love to be able to specify this at least, in addition to genre
-
Any way to have the model output stems instead? Right now I have to manually use a splitter, but if the model can already output stems then it saves me the trouble. Often I just want to isolate the vocals, and recreate the other parts in real MIDI synths if possible
-
Any way to do “audio-to-audio” or “outpainting”? Like, for example, a VST plugin that looks at my current song in my DAW and then “inpaints” vocals, synths, whatever I ask for, on top of it as an audio track, keeping to the same BPM and key
Hey @normal stream Thanks for reaching out! A lot of great ideas here. If you wouldn't mind and at your convenience, would you repost this in #💡┃suggestions-web so it doesn't get lost in traffic here 🙂
lol... I'm down to the last 30 seconds off my song and I'm burning credits like crazy because it keeps generating wierd crap now
Yeah it struggles past a certain lengthh.
OK... stitching the clips now... I really hope it works
I find it hilarious that chirp can simultaneously cut a lyric short at 0:42 in one gen while in the other it completed the lyrics early, ended the song, realized it had time to spare and started a completely different song in the last 20 seconds 😂
What I find completely hysterical (as in I am becoming hysterical) is that Chirp can decide to give you all this wonderful instrumentation and refuse to sing. I'm at the hair pulling stage.
I just encountered one of those. But, maybe I'M the one hallucinating at this point, I swear I heard lyrics ever so faintly in the background. I tried continuing from the generation to see if it would then generate the lyrics... maybe it just wanted a nice long build-up intro and didn't have enough time... nope! It continued the instrumental in BOTH generations. I think it's actually generating lyrics, but they're too soft to really make out.
OMG yes, I was having that yesterday. I do wonder if we should be given the option to run an instrumental pass and a separate vocal pass, because when it does instrumentals, they're usually on point.
Truuuuuu
also that was me yesterday
instead of me having to come up with fifteen different ways of describing Huey Lewis And The News, is there a better workaround for using artists as part of the prompt?
Welcome! For help, check out Suno Help
Welcome! See Guides and Resources for help using Chirp and Bark. Email us at support@suno.ai if you have questions, comments or feedback.
Ah, sure, of course. Didn’t see that channel, my bad!
Posted there now: #💡┃suggestions-web message
No worries ❤️ Just didn’t want it to get lost is all!
when i go to discord via suno it tells me i have no credits and that i have to subscribe . i have credits and my suno account says i have credits so i cant do anything.. any help would be great
you may need to have the web acct connected to your discord profile (a known issue with google login)
contact billing@suno.ai
with email and Discord username...
they will get you sorted 😉
ill give that a go.. thankyou
anyone else had the system get stuck putting clips together?
i've noticed it seems slower at some times of day ... I assumed it was web traffic.
(often late at night I have no problem or waiting...)
Usually it takes 10 seconds, this has been running for 5 minutes now
refresh th page and try again...?
I have had it time-out and nothing ever happens, but I try again and it works...
Looks like it deleted the files now 😄
they are still there, just a problem loading the web page... probably
(don't panic) 😬
(yet) lol
Yeah they disappeared for sure
(relaunch might free up memory... all the clips can hog ram)
Nah it was only the certain group of files that I was using
I don't think they're coming back
It's like the system decided they're corrupted or something
this was part of it
at least the lyrics fit the situation
🥺
Chirp taking that break-up hard...
yeah I don't know but I'd like to get the credits refunded for that 😉
Maybe email billing@suno.ai and explain the issue?
yeah I might thanks, I'm gonna need like a million new credits anyways
did anyone find a good way to add a pure instrumental into the track after a verse that works?
or a way to tell it more info about the instrumental part 🤔
like melody changes or drops
Somebody did found a way to replicate the style of the EDM band "Scooter", I cant use their name in the style field, and if make a description of their music, I get nothing that ressembles scooter songs.
What I do is continue from clip and delete lyrics. I still have to edit it off platform to make it fit. There is also a lyrics ai editor I think somewhere
If you found a better way, l;et me know
some tips here:
https://www.notion.so/Metatags-22a69f63b1814aa9afbec9d21f7b55c4
Metatags help to structure lyrics with labels, and guide Chirp’s transitions between song sections. The results will also depend on genre and style, as well as rhyming and word length.
Eventhough slightly hallucination causing the following works for instrumental Intros
[fade in]
[Instrumental]
..
[Instrumental]
..
[Instrumental Interlude]
..
I'm sad. Continue seems to just be a completely new song that disregards what it was continuing from.
👍 thank you so much. will give that a whirl 🙂
Have never heard a [fade in] actually fade in
wish we had an option to not only use a continue function - but also a prelude and the ability to assign rough seconds as values for the piece to create.
Metatags work, but the timing is up to chirp...
which is fine for now. i'm just super curious about it and where it will go - only up from here i hope.
More of a delayed start. But it does work sometimes.
(One of the many many many instrumentals I generated trying to remake my old song.)
Sorry for noobing around but I'm extremely new to this tool and I have a couple of possible dumb questions, but is there a way for us to have the individual tracks of the songs generated? So it's editable
Since fades are one of the easiest things to add in post I don't really care so much if Chirp actually fades, as long as the block of music that is generates for that bit is the type that would be faded. Which using the tag still seems to encourge eve if there is actualy fade in volume.
Hi, I’m new, using custom chirp and about half the time it’s just making music without adding lyrics, sometimes there are two generations and one has lyrics and the other doesn’t, so it seems like I must be doing it right. What’s wrong?
It is a much requested feature that is not available at this time. You may find better luck with programs such as this:
I keep getting pure instrumental chirps. Even though with almost the same lyrics, it did once give me 2 songs that use the lyrics.
What can I do to make sure it uses the lyrics?
Could you show me what you are putting into the "Style of music" & "Lyrics" section"?
Style: Festive, Christmas, Soothing Song
Lyrics: Through the snowy paths, Teemo trots with glee,
Decked in jolly garb, merry as can be.
With a twinkle in his eye, and a warm, cheerful grin,
Spreading festive cheer, with a little win
Mushrooms lay, mushrooms lay,
Hidden in the snow today.
In the brush, he waits to play,
Happy holidays, Teemo's way!
With a swift scout's pace, through the winter night,
He dons his red cap, under the starlight.
Gifts of traps - all stealthy and small,
Merry Christmas wishes, for one and all.
Mushrooms, mushrooms lay,
In the festive holiday.
Teemo laughs, and dashes away,
Bringing joy this League's Christmas day!
Try cutting the lyrics in half and adding the second half on the continuation. Or here... let me try so you dont waste tokens testing
one sec
Oh? Thank you, I do suppose lol
Testing now, brb
Yea that fixed it.... 6 for 6 on my end.
Really? Wow.
Too many lyrics at once can overload it
So it just stumbles on long songs?
OK. Good to know. Thanks
Btw, could you share your versions with me, if you don't mind.
It works better when the lyrics don't overexhaust the time it has to build
And yea sure I can share
Makes sense. I kept it under the recommended character limit. But I need to go lower.
I'm really glad I learned that before I started paying for clips
Sure, happily
using [brackets] in lyrics provides instruction without it being spoken
[Intro]: [Gentle electronic synth creating a dream-like atmosphere]
[Verse 1]: [Soft, reflective voice] "In the quiet of my room, [background singers:] (echoes of dreams) I paint my tomorrows in hues of bloom, [background singers:] (ah-ah)"
[Chorus]: [Slightly bolder, hopeful voice] "Charting paths on stars unseen, [background singers:] (on stars) where I'm going, where I've been, [background singers:] (ooh, ah)"
[Verse 2]: [Continued soft, introspective tone] "Through the window, the world waits, [background singers:] (waits) in the mirror, a future shapes, [background singers:] (shapes)"
[Bridge]: [Whispery, ethereal tone] "Underneath the sky so vast, [background singers:] (so vast) embracing now, releasing past, [background singers:] (let go)"
[Chorus]: [Emotive and strong] "In my dreams, I'm free to fly, [background singers:] (free to fly) painting futures in my sky, [background singers:] (in my sky)"
[Outro]: [Synths fade into a gentle close with lingering vocal echoes]
Notice how some words have ( )
That turns the words into background sung words often enough
Also notice the line break between segments
ooo
Study this and it should help construct things
Also keep in mind that the Style of Music section doesn't require brackets
Is there a particular way to set up Style of Music?
That's a great question
A lot of testing that right now
We know more on what doesn't work then what works best
No use in using Band names, Singer names or anything like that
But as an example...
Dreamy Introspective Pop :: Ethereal Soundscapes :: Lush Electronic Layers
That is one Style of Music description
Actually that is the one that I have attached for the lyrics I provided
You can try it out if you'd like
But keep in mind, that the length of the lyrics included will need to be paid attention to to avoid the issue you brought up at the beginning
Are the voices for the chirps artificially generated?
Yes, 100%
Ok, I'm asking this to make sure I'm not plagiarizing anything.
@daring lava thanks for the info on prompt construction, I was figuring some of this out by trial and error but it’s much cheaper to read it!
What's a good combination to make something similar similar pastel ghost
Very happy to help!
Gotta say, I am getting major cred with my rpg group, we are playing an anime based game and I just wrote everyone a theme song 😁
hi, quick question: when I click on continue for a generated chirp song, the same prompt as when starting out appears and it creates a second verse chorus snippet completely independent of the first one
is it possible to really continue the first clip, i.e. extend it in exactly the same style?
Yes. I have done it several times now. I have learnt use contine, add new lyric, make sure to only add the themes a d genres that matter to that section.
https://music.youtube.com/watch?v=OVLbDPSwFuo&si=4OuGbeO8wfCbEGex
An example. (Works better with headphones as balance is wack and I uploaded quick and dirty for my dnd group)
Generated with a variety of AI models for your enjoyment.
wow almost 4 min song, thats awesome!. so what exactly are you doing to make it continue seamlessly?
So for this I wrote the lyrics myself first.
Made sure it was structured to contain tags where appropriate.
So [INTRO] [VERSE 1][CHORUS] [{ELECTRIC GUITAR} INSTRUMENTAL].
With the lyrics etc inbetween.
Then in the genre box I list genres, styling, artist reference.
Nu-metal,aggressive,punchy,avenged sevenfold.
Then on each subsequent. I run continue from. Add new lyrics with the template structure. And then add the genres etc modifying slightly to Influence the feel.
Make sure to only run. About 8 lines of lyrics per generation (not Icluding lines to list headings etc).
E.g.
[Intro]
Hello,
It's me,
I know you're reading this
I hope it makes Some sense
[Chorus]
Clever harmonies play right here,
Make some clever lines,
Make it silly,
And kind of fun
Finally. Sometimes running no lyrics with just a heading can help. Or duplicating the end of the last line of the previous lyrics .
My big use for this bit is outtro and occasionally Instrumentals.
ah I see. thanks a lot for this detailed description, I will give that a try!
Awesome. I'll keep an eye open for any posts. Good luck friend
everything is interpreted through an LLM....
I don't think the actual punctuation matters as long as it is consistent.
to my knowledge the official training used :
[] for in-line prompt instructions
() for backing vocals
recommended to remove punctuation from the end of lines
comma lists in the style/genre are 'balanced' so they should all be weighted about the same.
FWIW chirp usually knows the words Verse and Chorus and won't sing them, regardless of the punctuation
[Verse] is same as Verse:
what is 'Atomic Central' in this prompt...? a band?
ahhh, it's a yt channel... 📝
(I have not had great luck with soundfx in chirp, but always curious if someone might )
I think we are in Stable Diffusion 1.4 days...
We can weight things by spamming them:
Male, Male, Male, Male, Male, Male, groovy bass
What do you guys think of my songs?
this one is epic!!!! 🙌
#🐣┃chirp-beta-1 message
Too complicated for my taste, too many notes, too many instruments, too many voices, too many melodies, I would love something a little bit simpler.
on Discord (v1) is is the last 10sec of the clip. 40sec... then 30sec+10sec of continue
on web (v2) I have not heard the actual conext window (assume 20sec). 80sec... then 60sec+20sec of continue
But Devs have mentioned it is actually longer (full context is 2min)... so there is room for other features in there...
there a pattern of Verse/Chorus that chirp does well...
if you can catch the chorus at the end of the clip, each continue, the song can go on and on almost forever (3-5 clips or more) and accurately repeat.
Verses will be a bit random but they are usually toned down anyway... Have to regenerate sometimes to get a verse that matches what came earlier... It doesn't always match, but usually can get close as long as the lyrics follow the pattern.
Ultimately, a DAW can piece together different takes, so it's possible to go back to the 2nd or 3rd chirp and start generating alternate verse/chorus from there, returning closer to the original clip
I think I got my first hilarious lyric hallucinations. It kept repeating "I have been in this video" in a verse
LOL
And well. It keeps doing that on new chirps. 3/4 explored.
"I'm one of the best that have been in this video" what the heck.
I wonder if that came from the continuation I am working on. Sigh.
It is important that you know that. Chirp is one of the best. And it has been in this video.
Hahaha
I just pictured anonymous generating ghost lyrics
Heck, I might roll with it. Time to make a video.
I may be the song writer, but clearly the singer has an attitude of his own. Great growly bold expressions
All jokes aside, I get how frustrating it can be when a hallucination of this sort appears. They are working on something that will let you define the length of an existing clip so that you can resume without these things. I'd keep the clip in my favorites so that when that is released you can go back to it and try to fix it.
I keep getting the same hallucinations on the same lyric base from a different base song. Hm.
let me hear the gens featuring these as prompts.
this is not a list of what chirp was trained on.
this is a list of prompt words that have been tested on Chirp.
If you discover prompts that work, let me know.
OK.. wow... so that's interestingly different from what's in most guides. Does the : after the [] serve an active purpose? Does surrounding the text in quotes help with it not mangling chorus and verse together?
Since I'm not up-to-date on music styles like you are, have you tried treating it as a painting exercise? Mixing the styles of parents to the style you really want like mixing red and blue to make purple? I have had some luck with this in the past with other genres so that's why I'm asking.
That is something that I'm testing with now. We all are. A lot to learn here!
Actually, I just tried it.... and it sang the words [strong and upbeat] even though they are in [ ]
Definitely needs some refining it sounds like, we all are working to better understand the best way to approach. The goal though, is give a very user-friendly capability to achieve maximum control in generative response. As you learn, please feel free to share!
I've generated one song that is mind-blowingly good... and I've had no luck recreating it... can't seem to figure any way to maintain the same style and attempting to use continuation to generate more so I can externally edit has been frustrated by continuation glitches like strange vocal/style changes, unwanted "end of song" styling, and clips that just generate 40 seconds of instrumental for no reason... I really, really wish there was a way to "reset" a clip so you can start over without all the continuation glitches with fresh lyrics without compromising the style that was used when the previous clip was generated. I would pay for that as a premium feature....
Hi errbody ! I got these 2 fire beats (specially the second one) by accident, I would like to re-generate th using « continue from this clip » but without vocals, is it possible?
This is a great suggestion! Thanks for reaching out with it! Would you mind adding it to #💡┃suggestions-web so it doesn't get lost in traffic here?
I had that same song glitch on me, and the glitch resulted in an awesome output... which I can't find a way to recreate... what it did was this:
Requested:
[Chorus]
Line1a,1b
Line2a,2b
Line3a,3b
Line4a,4b
[Verse]
Line1a,1b
Line2a,2b
Line3a,3b
Line4a,4b
What I actually got back
[Chorus Line 1a, 1b]
[Chorus Line 2a, 2b]
[Verse Line 1a,1b - In Chorus style]
[Verse Line 2a,2b - In Chorus style]
awesome 4 bar instrumental bridge
[Verse Line 3a - In an wicked syncopated style, different from the rest of the chorus or verses], [Line 3b return to normal rhythm]
[Verse Line 4a - Almost spoken word], [Line 4b normal lyric/rhythm]
Oddly, this unplanned pattern was absolutely PERFECT for the lyrics (which were mine, not AI generated)...
I just wish I could figure out how to reproduce this effect because it sounds truly awesome... I don't think a human songwriter could have done any better
This is happening to me too. I have a song I’m working on that abruptly changed how the chorus went, and also stopped using English and I suspect that a big thing is I’m zeroing heavily in on a part of the latent space that’s hugely fed on Scandinavian power metal. But still, I need consistency. I really suspect that lots of people who are already producers/musicians will be editing stuff heavily
Vocals, specially female, are again out of tune very often. almost every chirp now (web version). it happened in discord too as time pass by. tried few days ago, was about 50%, now it's maybe at 90%. very frustrating, no idea why it happens every time the longer the version is running
it's really bad. I hope someone checking this. like out of tune big time. listen to 2 short examples I got just now, it's just continue like this and almost every female vocal, at least on rock and metal
wasted money at the moment. 400 credits on out of tune stuff, all vocals, different styles, just broken vocals every time
I've just blew 100+ on clip after clip that aren't generating any vocals whatsoever... all instrumentals... its like whatever does the vocals just completely broke
finally got vocals after 10 attempts.. and they are horrible quality female singer
Check it out! I just published an article on Linked in about Suno.ai https://www.linkedin.com/pulse/power-storytelling-our-marketing-keith-hanshaw-wqpue%3FtrackingId=%252FfIzBgloQpqofG%252BJxTiqYw%253D%253D/?trackingId=%2FfIzBgloQpqofG%2BJxTiqYw%3D%3D
Any way to have it make music with the classic hawaii lap steel guitar or is that too niche for it to possibly know?
I got Hawaiian guitar yesterday, my prompt was something like novelty Hula rock song, steel guitar, and I put steel guitar in brackets in the lyrics too
Huh. Ill experiment with this. Thanks for the info!
Is there a list of GENRES that you're able to input that will work?
i.e. Rock & Roll, Alt Rock, R&B, Blues, etc
i see some people citing actual bands, musicians
does that work?
When I put actual names I get an error
yeah me too, but then i see others doing it so i don't get it
I think sometimes it doesn’t recognize it so it doesn’t error, but then it’s not doing anything so it’s sort of pointless
I think one of the devs was saying it’s still sort of a black box
wow, that's kind of lame
but thank you for the info
and to someone that commented a while back, i also sometimes get no one singing at all even though i input lyrics. i will also get errors where it'll skip over some of my lyrics randomly, or start the lyrics very late in the song, or even sometimes, start in the middle of the lyrics totally inexplicably., like in the midst of the paragraph.
Is there a way to force vocal change? Like:
[Verse]
(Female Vocals 1) Blahblahblahblah
Blahblahblahblah
[Chorus]
(Male Vocals 1) Blahblahblahblah
You can run it as a continue and put the vocalist in the style info, that plus putting [female vocalist] at the beginning of the verse mostly works
Go to the #🐣┃chirp-beta-1 channel to use the /chirp command, and check out the #🐣┃how-to-chirp channel for step-by-step instructions for making music using Chirp on Discord.
man i really hope that chirp is gonna be able to hold the original voice better. the amount of difference there is between the voice of it and in continuation #5 (for like a 3-4min song) is really high. there must be some marker that can be set in the future that makes that happen. also tags for [voice 1] or [voice 2] would be nice for duets etc.
Is there a way to actually blend the clip you continued from with the continued clip? Or do you just need to download them both, and use an external deal?
Also. Is there a way to get it to not do pure instrumentals? Every other generation, it just craps out, even though the others do follow the lyrics. Even if it's not perfect.
I don't know what's going on, but the generations I've been getting lately sing as out of tune as a crow.
Hey!
How to make a pause on my clip? I mean... I want to generate more melody between singing.
Any way to do that? How should I type to get such result?
/charp
Thanks
/charp
wow, thx
will [Percussion solo] work too?
My clips are metal/rock at most
okey 🙂
where dots should be text?
oke, will try
uhh
it generated additional random-unrecognized sing
when I added lone "percussion solo"
so gonna try with dots rn
the next one with dots, just duplicated clip before
oke, will try lone solo xD
I might have a problem, coz Im free-sub. SO its only 50 credits xD
okay, iv got something similiar to what I wanted to get, but it still starts with singing duplicate of previous clip
Im thinking about a fact, that Im making songs in my language, and adding an english phrases like: "instrumental solo"
wow, useful. thx
How to generate a full song? I mean anyone have the tutorial for that?
Click "heart" icon on prefered
Hello. I bought a membership with money. But I cannot use it. It keeps giving errors. Can you help me please?
What errors are you getting? There is a moderation system that makes an error, and there is sometimes an error where the songs are very short and it says "credit refunded"
Hi, I was reading the TOS. Doing chirps I did found an interesting groove, then I altered it and replayed it on a synth, can I publish the final result without giving rights to Suno-chirp? How should I behave when it comes to attribuition on the "collecting" paperwork?
Thanks in advance ^_^
with paid subscription your chirps get a commercial license and you get all rights to the song (can edit, publish derivatives and monetize)
With free subscription, Suno retains the rights (non-commercial use + attribution)
Hello, I wanted to know if there is a way to get more tokens in my free account
Thanks, u're right I did't detailed the questoion enought.
The point I'm confused about is this: with a payed subscritpion should I give atribuition in the P.R.O.s paperwork? If yes at wich terms?
no, with paid you get 'all rights' to the song. You do not need to attribute Suno.
ToS talks about this under (from memory) the section on paid subscriptions, not the section on license (paid alters the default license).
Thankz
Is there a way to isolate the instrumental from a clip with vocals? I've tried using AI splitters on the mp3, but it's far too destructive.
Did you try the beta of FL studio?
Its vocal splitter is often way more "decent"
than LalaAI
(still not perfect, for some results LaLaAi works better)
No, but I've tried SpectraLayers, and Ultimate Vocal Remover 5. Not sure it could possibly be better than them.
I have tried LalaAI, though.
SpectraLayers isn't top, lalaAI is often bad (strangelly it often isolates guitars well)... ...FL beta... ...is often good preserving vocals
TRUE
Tried Mouses.ai, too. It all results in the Acoustic guitar ending up muddy, and watery.
How do i know which part of my prompt is wrong
Nothing seems to work. I'll try FLStudio, but I'm not expecting good results.
Is the Beta available for the free, trial version?
idk
I think yes
*I did noticed that 21.2 is out (not more in beta...)
so technically... ...yes it should be available also as trial
(I'll stop because I'm going offTopic 😉 )
Here's the Suno.ai album I'm working on! https://youtube.com/playlist?list=PLFfJ7drQlvTkUZhPGXDlh3iC_FpNSJOId&si=xo4DCPJBIMO62K1D
Welcome! For help, check out Suno Help
Welcome! See Guides and Resources for help using Chirp and Bark. Email us at support@suno.ai if you have questions, comments or feedback.
Y’all need some more training data for female singers, for real. I keep getting off key warbling and hallucinations on prompts that work fine for male vocalists.
Anyone have a way to reuse a prompt without it bringing the previous lyrics along with it?
So any news about the out of tune vocals, this is getting worse and worse. almost impossible to get good vocal without "yodelling"
🤔 copy/paste...?
(it's what I do)
Well, I meant using the same voice/music, but using new lyrics. You can do it, but when you click the 'Whole Song' option, it imports the old lyrics and then your new ones.
Ah! Yes, this has gotten me in trouble once.
I had to edit it (which was hard because of how well it continued.)
Some of the better Chirp hallucinations in second half here. Specifically "Banger, across the sonic lights" a nice little variation on earlier lyrics.
Banger, in every beat, a revelation
(start hallucinations)
Banger, across a sea of ___s
Banger, across the sonic lights
Banger, across a sea ___
Banger, unless I kind of see
Banger, if I'm not __-ish in a low fee (low-fi?)
Banger, if I'm not ___ what is wrong with me
Banger, if I'm not ___-ish in a low fee (low-fi?)
Welcome! For help, check out Suno Help
Welcome! See Guides and Resources for help using Chirp and Bark. Email us at support@suno.ai if you have questions, comments or feedback.
ERRO MEU
why do i have 0 credits? I just found this app and logged in with google. Shouldnt there be 50?
I have 0 credits too!! (I've used mine)
Try logging out and logging in with Discord to see if it changes anything?
Cancion
Hello, I'm encountering a problem. I'm not sure if you are aware of this, but I have two clips that sound great individually. However, I'm unable to merge them into a 'whole song' without experiencing a glitch in the middle of the music. Each clip sounds great as a separate song, but together they're not seamless.
just in the minute "1:00".
Actually, the issue occurs at every 1:00 minute mark in the sequence. It's not a major concern because I can merge the audio, but I just wanted to report it.
maybe also post this in the bugs channel. That could be just a fluke with those particular clips but could also be a software bug in the merging. Especially if you can merge the clips manually and they actually sound mostly seamless, then that looks more like a bug.
does SunoAI have API access?
Does anybody have any tips how to avoid in rap to have the chorus sound the same as the verse? I mean, I would love to have chorus being sung or a different background etc
Some tips for song structure prompts here (not specifically rap)
https://www.notion.so/Metatags-22a69f63b1814aa9afbec9d21f7b55c4
so bark is trained on music. but is it trained sepret? like the music only and the vocals only?
Would love to have different song style for chorus
Gotta be careful with those vocals. I threw in an Ohhhhhh-- once and Chirp decided it liked it so much it became part of the song itself from that point on.
How do you download the mp3 file of a song? I did it the other day but when I try today, I'm only getting the mp4 file. Is there something I'm overlooking now?
Are you using the Discord or the webapp?
@neat nexus Discord
That explains it. The Discord version is still using the earlier version of Chirp and is due for an upgrade soon but that is not yet available. In the older system the MP4 is correct that is what you're expected to get. If you go to this url: https://app.suno.ai/me do you see your clip or song there? That should give you an option to download the mp3 there.
@neat nexus Oh, that must be where I downloaded the mp3 from, but I did it in parts and then put them together in Audacity. How come the whole mp4-3 is not there, but just the sections of the completed song?
Did you stitch the song together like instructed in #⏩┃continuation ?
@neat nexus yes I did. In Discord I have the full mp4 video there. All sections in one.
OK, wow.
Let me test this quickly.
(Trying to have it stitch something I did a long time ago so I can see if it shows up. May need to aactually create some parts to see if it's not working.)
restitch didn't show up at all. Now making something in Discord brand new to test...
@neat nexus Thanks
I had to wait a few seconds but the stitched together file did show up and I was able to get an mp3 file from it. (Don't hate the music it was entirely randomly generated.)
Disregard.
I'm sorry I didn't realize that was not the stitched together file.
Good catch! This is definitely a bug of sorts.
I just posted about it in the #🪲┃bugs-web section.
@neat nexus So, I'm guessing you didn't see your stitched render there in the online app? ....Great! Thanks for posting it in the bug section. Thanks for your help!
Yeah I thought I did then realized that it was just the other option for the part. I listened to the part and it was the same deal. I went to https://app.suno.ai/create to see if the part could be combined there but no joy. The option doesn't even show up.
I can tell you from personal experience that there are apps that can allow you to convert the video to MP3. I use one on Linux called Sound Converter, but I don't know what would work best on your operating system.
@neat nexus Thanks for the suggestion to convert it. I'll look into it, an app to use. Or I may try using Audacity again. Thanks for your time today!
Not a problem, I like to help. I just wish that I could have done more.
The good news is that the upgrade for the Discord version of Chirp is coming very soon now. We just have to wait a bit longer for it to be done.
@neat nexus Well, you did steer me in the right direction. And gave me insights to it all. Thanks again, and enjoy the day!
Welcome! For help, check out Suno Help
Welcome! See Guides and Resources for help using Chirp and Bark. Email us at support@suno.ai if you have questions, comments or feedback.
Hi, I would like to ask if you have a publicly available list indicating what styles, instruments, and emotions can be input in the "style of music"?
chirp/ football is a beast x4 rythmic beat
Have a look through this community list
https://www.notion.so/Style-Genre-List-f7d1bb28193c4eac910f03db3f602508
Quick question in regards to subscriptions. So the free version provides 50 credits/day, and the next plan up provides 2500 credits/month, which equates to around only 90 credits/day.
Is the monthly an amount of credits provided monthly in addition to the free 50 daily? Or is it like... one or the other?
So, my first attempt at a language other than English (Thai) gave me a really good song but why doesn't the mp4 have text available in all languages?
Welcome! For help, check out Suno Help
Welcome! See Guides and Resources for help using Chirp and Bark. Email us at support@suno.ai if you have questions, comments or feedback.
||היי||
Hi everyone. Is it possible to invite chirp bot to my server?
still no 100% strategy to add intro and outro without hallucinations?
there is no 100% strategy. The bot is designed to be random.
here's some tips I use when dealing with hallucinations:
https://www.notion.so/Voice-Hallucinations-4a2e52000c474888b38132f1b70b1a25
yep, i already tried everything from this articles
Hi, thank you for the information you provided last time.
I spent some time researching.
However, I would like to ask about the command list for singers. Currently, I only use the terms "male singer" and "female singer."
I want more variety, for example, voices like Japanese high school girls or deep, magnetic male singers.
but still can't get short musical intro before first verse 🙂
best male voice I get is with prompt
sultry male vocal
(I think 'sultry' is the key word, it gives a mostly consistent voice)
someone described their vocal as 'squeaky cheerleader' in the style and it was amazing, but I never got the same result...
Someone also got a child voice, but I never did....
I think the song-style description has a big influence on what the voice will do, as the Suno 'singers' seem to follow an underlying voice that comes out of the AI music-gen (and can be heard during hallucinations, sometimes sounds like backing vocals or fx on the vocals, like the Suno singer is singing over it) ...
I think the 'trick' would be to get that underlying voice to do more, maybe with some in-line prompts.
[emotional] works...
I haven't experimented as much with voices, most of my stuff in instrumental, so I haven't found that much vocabulary...
OK, I'm trying it out. Thank you very much for providing guidance. 
Copyright only covers your lyrics
Suno uses Licensing, like the rest of the music industry.
and software industry.
they voice?
Chirp does not use celebrity voices. the voices are generated.
Ok, read the ToS...., but it will give you a headache.... 🤕
your lyrics are your copyright the moment you wrote them, however submitting them to Suno you grant them the right to use those lyrics.
under a FREE subscription, Suno retains the rights to the song. There is nothing preventing you using your lyrics again elsewhere, but you granted them the right to use your lyrics in the process of making the clip, hosting it, and making it available to download (yadda yadda, forever).... Under FREE sub, Suno owns the song with your lyrics in it.
Paid sub, Suno changes that license and gives you "all rights" to the song, and a commercial license to monetize and make derivatives.
(you still granted them the right to use your lyrics to make, and host the song)
they know what your account was at the time the clip was made. That license stays with the clip. If you cancel or switch to free tier, those songs still have their commercial license.
I have not heard of Suno actively searching music for matches in their database... but I guess in theory that would be a way to prove it...
errr, not sure about that... they would have original audio files I guess.
Yeah 😅 it wouldn't make sense to go after individuals with a youtube account just to be vindictive... Lawyers cost money.
But once you start doing 'business' with music industry (and corporate advertising) proof of license becomes a real thing...
Like imagine you sell your song for a tv commercial then Suno says they own the license and the company owes money... The company would cancel the ad, and blacklist you for wasting their time....
Hi, I have another question.
My friend wants me to create background music (BGM) for him, but I'm not sure how to adjust it without lyrics.
I know that I can use commands like [Introduction][Verse][Chorus][Outro] for lyrics.
However, if I want the BGM to have instructions like [Introduction][Body][Climax][Transition][Conclusion], how should I input these commands?
these all sound like good ideas... I haven't tried those words specifically so thanks for suggesting them!
I have a page on Instrumentals on the wiki, but it sounds like you are on the right track.
https://www.notion.so/How-do-I-make-an-Instrumental-352abcb0bf6a4f77b8fcf76a768f18af
😋 when I have done corporate jobs they were not demanding I show certificates...
But when you listen to some of the songs, they are easily good enough for corporate presentations... As artists we forget that real money is B2B and corporate licenses and marketing....
We are still at the point where an 'artist' is going to get pushback for using AI, but if an ad agency or in-house marketing dept could just dial up 50 songs rather than paying for a content library, they will.
Hi mom 👋
Sounds exactly like Britney's real voice. what's the issue?
sounds funny, but not a pop song lol. and the second clip, idk what is ist at all =))
🤷♀️ 😂
band names and celebrity names don't work in Chirp.
It does not clone songs. you need to use genre and description like "female singer" etc
pop song genre doesn't work too?
have seen "eminem" prompt and it was
I'm guessing all the other info you gave it confused the issue
try keeping it simple, at least at first
previous result of almost same prompt were pop songs tho
ok then use one of those?
It is AI. It does random stuff sometimes.
just wanted to report an issue to dev team, actually
not that i upset too much about result 🙂
Ahh, there is a https://discord.com/channels/1069381916492562582/1082775596767117312 channel
(but I think some random weirdness is probably a feature for many people)
oh, ok 🙂
and the work around is to describe like 'female singer' etc
so bark is trained on music. but is it trained sepret? like the music only and the vocals only?
Bark, or Chirp?
(I know Bark has some music, but idk the details on how it was trained)
chirp uses bark.
bark is an ai model that can do tts and singing, suno made bark
yes.
Lol how did I get a song in a diff language
are there any ways to mark or force timing? theres times where the chorus bleeds into the verse or lines get smashed together that I want to avoid
if you have some lyrics that break the pattern (not verse, not chorus) label them with [Pre-chorus] or [Bridge]
Chirp will handle them better, give it something original and not a weird extra-long 5th line of a verse...
is there a way to have a longer verse? like for example this style
[Verse]
A
B
A
B
C
D
C
D
do i have to label it twice?
when i try this format it combines AB and CD
in this one it combines the 2 lines
where are you now and hiding away somewhere should be separate phrases
you can see the original lines i capitalized
when you type, use RETURN to make a new line
don't use LINEBREAK (like discord)
?
Chirp seems to respect RETURN, but ignore linebreak
do i have to type RETURN after my line
yes, that is what I am suggesting
ill try that when i have daily credit
also why is the web version better than the discord version
it was generating 1:20 of the song discord one only 0:40
wait, not the word"return" just use the RETURN/ENTER key when you type
Discord is still v1, half as long...
oh i see
they upgrading it this week
i did heres what i typed
[Intro]
(Ooh oh-oh-oh oh)
(Ooh oh-oh-oh oh) Yeah![Verse]
Where are you now?
Hiding away somewhere
Well I just found out
You weren't very fairDesperate intentions went in my directions
Hoping that I would take the bait
But in my decision despite the derision
I figured that you weren't worth the wait[Chorus]
Left you for a reason
I made no damn mistakes
What you did was treason
Shit you can't make me takeYou acted like a demon
Made me stay in my place
But now it's hunting season
And it's time for my big break
maybe it was just long for the 1 clip so she sang it all run-together?
(a problem with the shorter running time)
maybe i should try marking [pre-chorus]
would it behave better if i used a continue and added the chorus and second verse
yes... but it might fit all in the 1:20 v2 on website...
ok ill try that version when i can ty
v2 on web is much better, longer and stereo
does the bark one have longer output on website too?
yea. it is very good. but how do i “correct” weird pitch tuning?
🤔 i don't think they put Bark on the new website...
maybe I didn't see it
that's a 'feature'
the end of the verse has strange singing every time i generate it
yes, for that use [Pre-chorus]
it did pretty good with 3 line group structure tho
I think it sounds good...
does the music generator genre input accept things like time signature and key?
people are trying it...
they also putting it as a prompt in the lyrics, sometimes a key next to each line...
I think if you search 'chords' or like a specific chord 'Am7'... you will find some discussions....
(i don't know how reliable it is)
interesting. but if i put it in […] its always a guide and never voiced right?
right... supposed to be...
sometimes chirp will sing it, but not usually
so i can try putting [C] [Am]
cool 
or, maybe you find a more reliable way, but that's how I saw them do it
would be cool to see stuff like midi guiding like synthv ai too
https://youtu.be/Yrs4rF_V92s?t=40s
i use synthv a lot so i would be used to it more
Listen to "Not Yet,” a demo song for our new singing voice database, Synthesizer V AI Eri.
Eri is a Japanese female voice database endowed with a unique versatility. Her rap expressions range from a breathy and velvety decadence most distinct in the lower tones, and a figure of mature and commanding power under the "Aggressive" vocal mode. Besi...
yea im really interested in that also
would be cool if i could run it on a cloud server and have it make music endlessly like an AI DJ
i could leave it running and have it stream songs by just typing in a playlist prompt and it can make songs that last forever or multiple songs for a playlist
and use gpt4 to make lyrics
sorry, i didnt finish my message bc i was leaving somewere. so i think i found a way to make audio sound hyper realistic. since it trained on music (with vocials in it) it might sound lower quality and unrealistic voicials, but someone tested this and it worked, maybe train the model more on vocials with no music, we tested this and it worked and it was crazy realistic. maybe try that
how do i get chirp to do stuff like scream
i tried using [Scream] ...
but it wont listen
not sfx
screamo
like in this song
it didnt work in the first generation
the main thing is i wanna control where the scream is
like i want them to say "im dead IINSIIIIIIIDE"
not "dead inside"
Is there a way to recreate the exact same song with different lyrics?
The* "Continue" *Option continues previous one.
I want exact music that I liked from one of the 2 generated clips. "Reuse this Prompt" doesn't give exact music.
I can not find. what main commands does he understand Suno?
you can try continuing but download only that portion
The problem is it's the one clip from the first 2 clips.
The intro music is very good But SUNO AI has generated some extra words/lines that were not supposed to be there.
So I wanted an option to recreate the exact clip but with different lyrics (or no lyrics at all)
I continued that clip and downloaded whole song then manually deleted the unnecessary lyrics.
But it would be great it was possible to redo just the lyrics part while keeping the generated music (instrumentals).
Hi, I just got a $30 subscription to Suno about a week ago. I'm having trouble finding my earlier generations. I believe the website used to save them, now they're all dissappearing within a day and they're also not located anywhere on discord. Is there another place that these exist or is this a bug?
at the MY CLIPS link, there are pagination links down at the bottom.
(I think this is the only page that shows that)
I see it thank you
when Discord gets upgraded to v2 (maybe this week?) the website is also getting a small update to help with navigation –– I haven't seen the web update so not sure what it will include... I think the idea is we can generate on both Discord and Web and share the clips
(a system of tagging and searching, albums, and deleting clips is also coming, but not sure how soon)
Just a UX suggestion for the Page numbers, maybe it could say 1, 2. Only the 1 is highlighted and 2 is grayed out. When you click to the next page the highligted letter changes.
Or alternatively, highlight the arrows on the right and grey out the ones on the left to indicate that there's more to the page. Might save some time on a few inquiries, if possible.
But also I could just be blind and everyone else might notice it lol
It's definitely an evolving UI... 😓
And sometimes hard to know if the ETA is in 'Dev time' or 'ready for the public' time, lol
Understandable, it can be a lot to take on. I think what you've got so far is pretty great. No complaints from me!
Hello, when I generate a new AI song, can I isolate the instrumental and take the vocals only, then remix it? Can I have permission to do that?
remix doesn't change the license.
Free subscription is non-commercial+attribution. Fine to post on socials with credit to Suno.
Paid subscription give commercial license and all rights.
I use UltimateVocalRemover to demux because it's free....
Oh thank you, I am a small artist who makes few musics. I thought of making a song with lyrics and vocals made by AI but I also thought that I should have permission. Thanks.
I pay one month and the song is mine BUT what if i stop paying the monthly fee? Suddleny the song is no more mine?
the license 'sticks' with the song...
if you later cancel or switch to free, your songs with a commercial license keep their commercial license forever.
how do i make it so that there are no lyrics at the beginning of the song?
it doesn't work with [intro].
interesting, so even if I pay just two months and then nothing more the commercial licence sticks with me perpetually?
What if I pay the licence after I generate a song? I explain myself: today I generate a song that i like and tomorrow I want to pay the monthly fee. What happen with the song I generate the day before?
tru with [instrumental] I am experimenting with it right now
[instrumental] at the beginning of the song doesn't change anything for me
Many more predictions and I will have to charge money for communing with the future.
yes, forever is forever.
No, the songs you create under Free Subscription are owned by Suno and can't be monetized
(but you'd need to have a hit before Suno hires lawyers to come after you, lol.)
(but they can also just cancel your subscription for violating ToS)
thank you
Thinking of upgrading but i am a bit confused from this message. I want to go from the 10 dollar plan to the 30 dollar plan. if i choose the first one, will it extend the current one with 1 month or will i have to hastily waste all of them within 13 days i have left before it runs out?
(Preferably I hope for an official response so that I can 100% be sure of the answer)
I have another question maybe a rethorical one:
what if I am inspired by a part or a general atmosphere of a generated file and I elaborate it in a more personal way?
I mean this actually can happen even with "normal" songs, right?
yes. the rights include derivative works, remixes, edits, live performance
for paid tier, the ToS says 'all rights'...
you can do these with free tier too (but you still can't monetize and have to credit Suno when you post on socials).
obviously inside the limits of plagiarism I guess you can do everything
Suno is coming out with a friendlier explanation for all this, lol
I think it's coming with new docs and guides....
(idk when... sooooooon™)
I created this song using Suno AI.
https://www.suno.ai/
(Verse 1)
In the neon city lights, we find our way,
Chasing dreams, we'll make them stay,
With every step, we'll dance through the night,
Underneath the stars, our future's so bright.
(Pre-Chorus)
We'll paint the town in colors so bold,
Stories to be told, our hearts of gold,
Together we'...
I created this song using Suno AI.
https://www.suno.ai/
(Verse 1)
In the quiet of the night, under a starry sky,
I'm searching for the answers, don't know where or why.
The moonlight paints a picture, of dreams and endless hope,
I'll sing my song of life, through every twist and slope.
(Chorus)
Oh, this journey we're on, it's a melody unknown,...
My Lucky Song will be released on Spotify on January 7th, 2024.
Pro rata. Eg, I had used 2300 of my basic subscription, upgraded to the 10,000 and balance adjusted to 7,700. Note: why would people pay $30 for 4,000 top up when you get almost double by changing tier?!
I created this song using Suno AI.
https://www.suno.ai/
(Chorus)
Breaking the chains, setting each other free,
A symphony of heartache, lost in the sea.
Finding ourselves, beneath the starlit skies,
But in the echoes of silence, a new dawn will rise.
Goodbye's the hardest word, but it's time to say,
Farewell to us, in this bittersweet ballet.
B...
Is anyone else having struggles using the “continue this clip” feature?
I've been using it no problem, but it's been a few hours since using it
I don’t blame you it’s not the easiest
Go and listen to my 3 songs guys
I assumed it’ll continue to add more to the original song. Am I wrong? It looks like I have to add lyrics and a song description, but that doesn’t seem right. My first time so sorry if this is old news.
I tried searching the community some as well as Google, but couldn’t find anything. My own testing didn’t help either because the tests failed and credits were refunded consequently.
Yes it'll continue from that part where the original clip ended.
So if I want the song to continue, I just hit “continue from this clip” and it’ll generate more lyrics and music without further input from me?
I just tried and I got a 1:00 not much and a 0:05 generation. 😂
when you click on "continue this clip" then you can keep the "Style of Music" section as it is or empty.... but you should put something ( e.g new verse etc in lyrics section)
it's random . if you are not using your ownb lyrics , then generate new lyrics in the Lyrics section of new clip...
if you want your own lyrics use tags like [Verse] [Chorus] [Bridge] etc
and [Outro]
Okay gotcha; thank you so much! The one I’ve been trying to extend is so catchy it’s been stuck in my head all day! 😂
I have got one too .. where it generated a very catchy intro with flute and the overall music is really good. but made a mess of lyrics so m waiting for update where we can re-create same music but change the lyrics ...
yeah it sounds good.... try using the continue option a few times ..... you ill get a very good continuation this way......
You think future chirp updates will play nicely with our existing stuff? Like, if they improve the audio quality or add in some "in-painting" feature, can we then retro-fit our past creations? If yes, then I'll keep using it and stash away demos for the future. Otherwise I might hold off until more features come in.
Just curious.
is there a way to use the continue from option in chrip but on custom tracks like self production stuff
Nope
Hello. Is it possible to keep the same voice in all my songs?
Short answer: Nope
Long answer: You can let a song end and continue from there to start a new one so you can keep the voice, but It's not practical and the quality degrades over time, so yes but nope
I'm running into an issue on the web app when I'm trying to continue a song and the output always has Credits Refunded. Am I doing something wrong or is this a bug?
listen to the end of the clip you're continuing from.
It probably has a second or 2 of silence at the tail of the clip because the song 'ended'
the Continue clips are continuing the silence, then cutting off because there is no music.
no workaround for this yet, but some new features are coming that will allow us to continue from a time before the problem happens.
Alright that make sense. Thank you for the heads up. I did notice that they were silent toward the end but right up until the last second there was an instrument so I thought maybe it would keep going.
this is one of the continued clips.....
Can someone give me a breakdown on how to make quality songs with good continuity with chirp? What should I be putting in the genre / other boxes
If I wanted to start with an intro, would I put that in the genre or lyrics generation?
Here’s an example (please keep in mind this isn’t the only way to accomplish this) One thing to take into consideration is that bracketed information in the “Lyrics” section will be used as instruction, not as sung lyrics.
Hope this help 🙂
This is cool. Do you have a guide or do you know of one that goes into detail how this technique and the others work?
And what's possible
Is there any way of creating a few words in a song, like in dance or EDM you won't necessarily get a whole song, but just a few vocals.
Use the Create(Custom) link on the website.
it uses 2 prompts, one for style/genre the other for lyrics.
repeat short phrases,
and repeat vocalization words like
aaah aah AAH ah-haa HEY aah ehh, AH-haa
vary them and add some commas so the lines look a little different, then repeat those so you have a block of them like a verse
Okay thanks, I'll give it a try
Can you create a song in custom without lyrics? Like a theme song of sorts? Just with some choir at some parts and mostly prompting the instrumental of it?
Absolutely! Look here for tips: https://suno-ai-wiki.notion.site/How-do-I-make-an-Instrumental-352abcb0bf6a4f77b8fcf76a768f18af
Do annotations like that actually work? I feel like it just ignores any instructions that are even remotely detailed (whether it's inlined in the lyrics or in the Music Style input). In my case it even sung the "[no drums, soft piano]" part out loud lol.
Everytime I type /chrip I get “this interaction failed”
anyone else have htis issue?
ah nvm got it to work
We’re all learning together on this. What I’ve shared is an example of my most recent test. There are a lot of other options to learn though. Check out #1151542418991239168 and https://www.notion.so/Metatags-22a69f63b1814aa9afbec9d21f7b55c4. Also scrolling this channel and #💬┃general-chat you can learn a lot.
To a degree but very much in the testing phase of all of this. We all are. Feel free to share what you learn as well 🙂
Glad to hear! If you have any questions along the way, feel free to ask 🙂
Yeah, this would be amazing. I wonder if it’s possible to give music + text input and not just text to change for example entire melody to violin.
Hi, I'm Time4Coasters and I am a coaster critic! Please subscribe (= https://rctgo.com/downloads/members/32385.
I have Instagram too https://instagram.com/time4coasters
Probably better off in #🐣┃suno-showcase but no problems for now.
Or did I misunderstand and there was something specific about the songs there you wanted to discuss?
What could be reasons why Chirp refuses to generate a song?
The person inside the Chip machine got tired of writing songs
No 😭
It gave up singing on my side too.. 🤣 Maybe will be back after a coffee break...
Oh dear
when it comes to splicing, is there a way to reuse a previous recording within a continuation? Not only would it be nice to save a credit, but might like the chorus to sound the same
/chirp
Go to the #🐣┃chirp-beta-1 channel to use the /chirp command, and check out the #🐣┃how-to-chirp channel for step-by-step instructions for making music using Chirp on Discord.
Suno came up with this beauty. Too bad it got cut off right in the middle of the vocals.
Well this did not cooperate at all. I at least got a decent instrumental out of it I think.
Have saved the first clip in my 'Future fixes' directory against the hope that some day with being able to adjust the continuation on the clip I might be able to take it forward.
I think the way the first clip ended basically screwed me on vocals in continuations.
The few that I did get were horrible.
Loved those horns though..
Like the vibe on this. What was the style prompt?
Style: traditional, dream pop, downtempo, chillout, catchy hook, male low pitched vocal
Dang. That's a lot.
lulz.
Considering it completely ignored attempts at vocals might be able to skip that part.
😶 me, wondering if 'out' is a generic suffix, like 'core' and 'wave'....
(i think i've been at this too long. everything starts to feel like
A HA I AM UNLOCKING LAAAANGUAGE! )
Given that an earlier attempt used slowcore instead of chillout....
Dammit...you're going to make me try that one aren't you? The curiosity is eating me up....
what is this...? Grooveout?
That one was: traditional, cookout, grooveout, catchy hook
These are: cookout, grooveout, catchy hook
These are pretty excellent
I know, right?
😄
It's like we discovered a whole genre.
Credit goes to wetcircuit of course.
GOOD WORK TEAM!
I've been shoving random-ass words into the style of music box for a bit, who knows what stuff can be summoned
grooveout seems like a hit
I'm running one now without the catchy hook to see if that's the lynchpin...I suspect if it is then the next one will be flat. We'll have to see.
This is just: cookout, grooveout
It may have lost a little something but it's still noticably itself...
This is grooveout.
it's like funk-lite...
diet funk
so pretty!
grooveout dietfunk
You guys ever have a song that does not want to end?
This is the cookout style.
That first one is a bumper on Nickelodeon circa 1994
Like RIGHT before it switched to Nick at Nite
So how crazy do you think I am? Yeah....
cookout, kebab, bratwurst, hamburgers
I'm hungry now dammit...
🤣
LISTEN TO IT!!!
A MAZE ING
🤣 i can't.... what is it singing...? omg ow, my side, ow
It's definitely not Irving Berlin, that's for sure!
Summer before you ??????
it's so wholesome. and sincere
🍹
I am officially just doing word salad now.
😎
Up next (for better or worse) is: grooveout, bonghit, tokeout, catchy hook, mellow daddio
No, I don't know what the hell that last thing is...but holee sheet!
MINE!
That first one 🤌
Yeah!
That one I'm going to do something with even if it's just an instrumental.
Instrumentals rule though also I bet you can throw words in it later anyway
Tipsy Grooveout
(is the name I register under in Beverly Hills)
I added "German new Wave" along with "grooveout, bonghit, tokeout, catchy hook, mellow daddio"
Think your word salad has a real effect
OK, here's a question though--have either of you been able to get vocals with it?
🤔
I hadn't tried but lets see
