#📚┃suno-school
1 messages · Page 23 of 1
Sanbei I think you are confused about how this works, the noise is coming from the model itself. It's not something Suno is adding in the process of rendering the audio. If your generation contains noise the best you can do is run a new one. The one-shot capability is already mind-blowing, but just like image or text generation you're going to want to generate multiple runs and pick the best output. A trick you can use for suno if you have an almost perfect gen, but goes off in a direction you don't like is to continue from than gen some point before it goes crazy.
it seems like there is limits to the genre. my guess is that genre that is more copiryght free are possible, others not. what you can try is in lyrics add precisely what instruments you want to plat in brackets.
what i tried is to put in the beginning of the verse either [male singer] or [female singer] to free mor space for style of music. but it doesn't works sometimes.
Thank you for telling me that I'm confused, I'm just talking about snare which apparently have the reversed phase and which generate a latent noise, apparently by dint of annoying Suno, they still reacted (it seems), because when I generate a song with the same criteria, it seems that the noise in question has disappeared. For the rest, I'll let you philosophize quietly on the confusion of both Microsoft and AI in general.
and for quality the only genre I've seen is glitch.best for virtual samples and instruments. 80's style help to hide the quality of the voice. also the prompt in style of music : Eliminate muffled or unclear sounds : helps to limit the quality of a song so that it doesn't go lower on a random generation.
Hello. I have issue, I'm a pro subscriber I've spent all my credit, a few days ago I've received 500, I've spent that too, and after that nothing. 0
On the website, click on your subscription link, then on "Billing Details"
it should show your past invoices... it's possible maybe the last billing cycle didn't complete (but you got credits anyway)... just check the invoice to see if there is nothing 'incomplete' or 'pending' (if there is you should see a button to retry payment)
if the problem isn't that, there will also be a link to email
billing@Suno.AI
I'm having a helluva a time ending the song I am working on. I am providing (end) after the (instrumental) bit, but it just keeps going on and on and on. Any ideas?
send the song
And here's an example of a never ending ending. Would be great to just end it after the last 'Shit load of Blues' vocal! lol https://app.suno.ai/song/dc74ff11-699a-4dc5-9a94-b087ae98450b
[guitar solo] then [End]
Pretty much had thet. But the (end) was not capitalised (End). maybe that makes a diff?
But I'll try it. Thanks very much for your help!
I've found using [finale] works a lot of the time, not always though.
In one part (of a multipart song) the singer started using falsetto. And, now he's stuck there. He's even found a higher falsetto range. How do I get him to go back to his normal range?
is there an API for suno v3?
Not yet.. it's on the roadmap, and I have no idea when it will be available.
/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.
huh
thanks. And is there a paper or article about how suno works, what did they train the model with etc
No, it's not public
/bark
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.
is it already known that "solo vocalist" or even "soloist" seems to influence the occurrence of the overzealous chorus?
acoustic dynamtechnicaltimate midwest trap, numb post-male black 1920 anti-solo post-vocalist song. Listen and make your own with Suno.
this one is a pretty clean example
acoustic dynamtechnicaltimate delta blues, numb post-male black 1920 anti-solo post-vocalist song. Listen and make your own with Suno.
and a country song
they take audio pipe it into semantic representations (using hubert/wav2vec), an acoustic model (encodec or DAC by descript) and use a pretrained text encoder decoder probably T5. Train the T5 model's decoder to output semantic rep, then train a conformer (fast conformer by nvidia) on semantic in, acoustic features out, then decode the acoustic features
its from a paper called audiolm by google, likely v3 is using the conformer from a paper called soundstorm
or likely they use some sort of CLIP model trained on audio and text descriptions
To match the melody of verses 1 and 2, the same syllable as verse 1 was added after the combination, but verse 2 did not come out and the first verse was repeated with a prechorus and then ended. Do you happen to know how to pull out verse 2 with a similar scale progression for this part?
Don't suppose there's a way to tell suno to start a song with instrumental?
Reliably, I mean.
If you put [intro] in front of the lyrics at the beginning, it starts with an instrument with a small probability.
Gotcha, thanks 👍
you could switch instrumental on at first, then switch it off and continue from some point
hello, is this suno technical support? if not is there a way to contact them?
The team is notified. They are aware and looking into it.
Suno is down?
See above
I've played with Suno for more than 3 hours. I love it. Very nice work for all the team behind. 🥳 I think the servor is down, I can't use it anymore. Or maybe I used to much in the last few hours 😅 I will try again tomorrow
all my song magically disappeared, not sure what happened
IMCEAMAILTO-support+40suno+2Eai@namprd14.prod.outlook.com
Remote server returned '550 5.6.0 CAT.InvalidContent.Exception: DataSourceOperationException, proxyAddress: prefix not supported - MAILTO; cannot handle content of message with InternalId 20139101664855, InternetMessageId CH3PR14MB6251B91F474042A2623D68D9E3072@CH3PR14MB6251.namprd14.prod.outlook.com.'
yeah that support email doesnt seem to be working
Is there an option to add negative prompts?
and now my song are back haha oh well thanks everyone back to music crafting
Is your Suno web page down?
nah but all my song were giving me 404 for a bit
Hello, is there a possibility for negative prompts?
Looks like it's not implemented yet
Does clicking thumbs up and down affect yout next output?
I hope I can manage the history if it does
seems we cant manage history messages ,i cant find my last song yesterday
love your song "bring me mercy"
I mean the thumbs up& down history, IF it affects the output
Can suggest it in suggestions channel, it is a possible future feature
So the 2.0 model has definitely changed.
And in a good way. It's not me. I've been testing it on older lyrics, and getting results that never would have happened on the older renders.
- The voices sound less processed, more organic. Vocal fry has been appearing on my recording for the last three days in a new way. And it's fantastic. I like the processed sound of it.
- It's belting more effectively, and in what seems like a more controlled and even way than it used to
- The instrumentation sounds more like 3a, than the way v2 did before 3a.
- The verse calculator for v2 has also changed. It's a lot more flexible than it used to be, and more like the v1 model. Better than 3 for reads.
Here's an example. I'm still working on this one. But you can absolutely hear the difference.
Everything after 22 seconds is v2
https://app.suno.ai/song/70478673-7127-434f-b074-fed4314675d6
{G4}, quirky female, edgy pop-grunge, dobro, southern rock, surf-rock, jazzy, contemplative, eclectic song. Listen and make your own with Suno.
I haven't been noticing a difference, but I probably wouldn't have since when I use V2 these days it is in clips combined with other V3 sections so it is pretty far from how I used V2 back in the day. In your example as well, there are two big confounders. The first is that when V2 continues from a V3 audio clip, it's starting off from a base that is different than any clip you would likely generated with pure V2 from scratch. The second is that when we used V2 originally we didn't have access to the custom Continue From feature, and changing continue times has a huge effect on the audio you get. In particular when you do something like continue from 0:22 instead of the end of the clip, the style prompt is vastly stronger than it would have been when continuing from a typical end of clip originally. TLDR: 🤔 🤷♀️
Yes, the difference is really noticeable.
In my own testing it seems that there is a lot less processing/noise on the vocals.
Another aspect with V2 to me is that the melodies for verse and chorus are more melodic and with better hooks. Is that just my imagination?
Downside with V2 is that instrument sounds are nothing as good as V3.
I actually got Suno to do a duet, but my goodness the quality of these vocals are not so good, and country is usually one of the higher quality genres.
https://app.suno.ai/song/c7b1318b-f78c-450b-a88a-2ee303434ff8
And the AI really, really sucks at writing lyrics. But that was more to test prompting.
How can I generate a music without lyrics, and fade out in the end?
https://app.suno.ai/song/4a9c5133-e264-4afc-bb36-f14b7189eac5
some weird failed experiment
energetic anti-folk mathzarthoven, complex arrangement, vocalized chamber orchestra, jazz prodianist song. Listen and make your own with Suno.
I'm curious how you use v2, I haven't touched it in a while and was having a listen to my first album, and certain things in it remain pretty nice. I feel like I may be leaving some performance on the table by not using it. What is your method/do you have any examples you could share?
Most common use to break a pattern in V3. This could be feedback/distortion/noise, a choir, but also just to get the music itself to be go in a more different direction. Also I like to start V3 songs from a V2 clip, it gets things launched in a more diverse direction, without requiring prompting super hard
I think people should become accustomed to breaking their songs up a bit more ie use style prompt more and [] less imo if you want granular controls. In this instance, putting [end] as the whole style element at the point you want it to wind down is, in my experience, the most reliable way to get a consistent ending.
ah I see
so if you're getting a noise "crescendo" but its undesirable
V2 is also more reliable on some nuts and bolts stuff, like prompting "end" as a style, [end] in the lyrics, if you just really want a quick ending. That works better in V2 in my experience.
or can't escape chorus
you can try
[fade to black]
[end]
that worked when I used in v2
but that was ages ago
in dog years
if the dog is a machine learning model
OK, but fr though, I haven't tried it in v3, you may have success putting [fade to black] as the only style element in a continue
V2 is also more laid back. It's not constantly turning into an epic song or a radio hit, can stay chill longer. Though I haven't really done any only recent pure V2 songs, it's mixing with some V3 clips almost always. (maybe a few V2 only instrumentals)
for that kind of stuff I've found certain prompts that really put a "dampener" on v3's epic tendency
and keep it lower key
stuff like "numb, lofi, recording, 4-track, cassette, acoustic"
depending on scenario
would a metatag like [repeat once] work?
it's mostly bc modern music is so explosive
stylistically
whereas the 80s were kinda low key
V2 can actually give you bad singers and musicians, which maybe is not a good thing for most people, but I really like to have that in there. In V3, to get that, I find I either have to start from a spoken word and turn into singing, or use 'atonal' or prompts in that zone. Which are sort of not exactly what I want (a literal bad singer who is trying to hit the right note)
at least, that's my assumption
ooh, I like that
Probably not, but is there a more music specific term that means the same thing perhaps? I'm drawing a blank but feels like could be..
to be fair though, I am able to get musical innacuracy as a method, but i see your point
in v3 I mean
I guess loop?
post-anti-dynamintricate-bachzartoven mizrahi-grand-piano composer, ghostly post-creep cassette song. Listen and make your own with Suno.
still too talented tho
like, too lively
Loop is like 'boring music' from my memory, as a style. (Which can be so repetitive it does sort of loop by accident almost.)
ah
I didn't try [loop] in the lyrics though, could be different. It was as a style.
I might give it a shot
this would ne closer to the "bad" sound, but again, it grows too large
https://app.suno.ai/song/a159bee1-a1fa-4ecd-8c5a-529472f02840
experimental midwest trap slacker lofi cassette 4-track homemade amateur, post-creep, autism song. Listen and make your own with Suno.
I'll probably try out some v2 continuation on my next month
"field recording" specifically is sort of close https://app.suno.ai/song/f41a17f8-f1d8-42a7-a927-9db2802ef7ae/
field recording, field recording, indie acoustic, sound, field recording, b-side male, raw field tape, bootleg acoustic song. Listen and make your own with Suno.
yea, it gets closer to a real live one
I wonder if specific venues would get less produced music
A few very unproduced sounding V2 samples dug up browsing through some old clips
(I believe the lyrics to Baby Shark are public domain, it's actually an old song)
https://app.suno.ai/song/2dc5348d-6b3d-47e9-8556-f46aae993c63/
https://app.suno.ai/song/175db5b9-02ca-48f0-ae89-9ba6be9a8fc9/ - no sung words but it sounds like a child banging on an instrument
Yarborough-Zasso Yarborough-Zasso Yarborough-Zasso Yarborough-Zasso
song. Listen and make your own with Suno.
Apple-Oatmeal Apple-Oatmeal Apple-Oatmeal Apple-Oatmeal Apple-Oatmeal Apple-Oatmeal Apple-Oatmeal Apple-Oatmeal song. Listen and make your own with Suno.
haha, baby shark
Yea v2 can definitely produce some interesting melodies
I think partially it's the result of v3 being much more strongly adherent to style tags
whereas v2 just sometimes goes off making a beat out of random sounds
the only reason I have been hesitant to use v2 more is bc it seems to me it has slightly less complexity at time in terms of the melodies
v3 def is able to make more intricate
melodies, basically. Like, v2 really focuses on one element
while v3 can have a few things going on at once
but depending on the music, v2 can be desirable it seems
I was using a random name generator for V2 style prompts. In V3 random words (words with no Google search results, not just picking a random word) are more likely to have little to no effect, but in V2 they could sometimes be specific weird styles
lately I've been smashing words together
since presumably the way it reads the tokens
will still cause it to find the fragments
like dynamtechnicalt
as an adjective
bachzartoven
moved it closer to an "edge" between styles
and his sister gen:
https://app.suno.ai/song/f4f8db21-3f35-43e2-9b20-19138a9b54c8
OK, so that's clown music
https://app.suno.ai/song/d3d1a9d7-dff6-4f64-92d1-64836dcd344f
Might just be random but some nice unsettling vibes from "countree" no commas when she hits the third "Baby Shark" https://app.suno.ai/song/c9bf39d0-56f0-4e96-800b-0f3a9553d9b6/
countree countree countree countree countree song. Listen and make your own with Suno.
this make me sirv my countree
the random words are interesting
I will say yolkhead does push it for sure, but I don't think it has to do with my gens
I suspect it has to do with some acoustic tracks some Swedish person made under the moniker like 8 years ago
but yea, it's def less randomly influenced
mozarella beathoven would probably be good
often featured catchy melodies, simple arrangements, and a blend of traditional Indian instruments with Western influences.
Maybe using bagpipes is cheating but V3 was impressibly painful in parts here: https://app.suno.ai/song/a6c9843b-dbb2-42a7-8c40-8037d7c92f88/
A moon made of bagpipes [bagpipes] > > > [bagpipes] oh moon of bagpipes [b[bagpipes][moon] [bagpipes][moon][bagpipes][mo song. Listen and make your own with Suno.
Oh, I forgot to share this to technical
I know a lot of people here love metal, and I made this by mistake
anti-anti-{intricatewave creeptronic gothcore, deeply sad dance, extended cut, sound fxfxfx}.asmrASMR song. Listen and make your own with Suno.
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 ,
Today I have added lyrics and submitted to generate,but it's not generating and showing generating but more than half an hour,it's still not generating in Suno and also in my server too??
Rgds
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.
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.
What version of Suno Chirp is used on this discord?
v1
What are the main differences between v1 and v3?
The differences between Suno's models, V1, V2, and V3, primarily lie in their complexity, capabilities, and features for music generation:
- V1: This is the basic version, primarily used on Discord. It's suited for simpler song generation tasks and is the foundational model on which subsequent versions were built. It's more limited in features and the complexity of the music it can generate.
- V2: An advanced version compared to V1, V2 offers improved audio quality and more features. It can generate longer and more complex pieces of music. It's available on the web app and gives users more control over the song generation process with additional inputs and parameters.
- V3: The most advanced and latest version. It boasts the highest audio quality and the longest maximum clip length. V3 incorporates more nuanced AI learning, making it capable of producing even more complex and expressive music. It's better at handling detailed prompts and offers enhanced features compared to V1 and V2.
Any plans to push v3 onto Discord?
No. The team is thinking about it, but tbh I as volunteer think if it happens, it will take a while. The team is burried in work since Suno is getting extremely popular.
But I don't see the problem tbh. Why don't you want to use their WebApp? It's way easier that using it via Discord
Thank you for the details. I do use the WebApp and I am really enjoying it. I have a decently sized Discord server with a dedicated channel just for Suno Chirp. I was curious regarding the versions since users are starting to really enjoy it as well
Ah yes.. that makes sense 🌼
Can we keep the promo spam out of this channel 😛
q: How many seconds before the time you enter is used for the ontinuation? I'm often having trouble getting the continue just right and spend a bunch of credits adjusting the time start
Docs, Guides & Resources | Notion Excellent Resource! Many thanks!
I connected my discord and bought a subscription and for the past 12 hours only getting a spinning wheel on the create page and can’t produce anything. I’m extremely frustrated. Please help
so I've never used "neon" in promoting, but I have a theory now
around why people see it "improve" their music
neo-classical and the like
are weighted on that token
or rather neo being on there, contextually, is giving it a particular sound
similar to anti-
being strong
thank to anti-folk
we are all experiencing the server slow issues....
I'm sorry this is your experience right after paying, but it is not just you...
How to keep the same melody all along even when it's progressive ?
u think they have a better results
never used it
about to test some stuff
but I'm low credit
When is your reboot contract date ?
28th
damn
yall might like this. prompt was pretty easy once I committed to writing a blues song
trying to get close to skip James
blind lemon jefferson
lead belly
et
etc
delta blues, acoustic, male vocalist, introspective, lonely, melancholic, raw, minimalistic song. Listen and make your own with Suno.
final
Any news on the actualization Suno would give the community about HOW to use the app better? (HOW to prompt better, WORKFLOWS...) Thanks 😄
what's the metatag name to change the style ? ( switch ? )
Lots of info here:
https://www.suno.wiki/
Is this new? How old is it i mean haha
the wiki is constantly being updated.
ok, ok. It´s just that i heard from a person here close to developes that they in any moment will say something about this, maybe is that wiki i guess 🙂
which symbol changes the intonation?
If I understand you, adding a word in ALLCAPS will emphasize that word.
In Russian this messes up pronunciation, so I have seen Russian workaround using like French accents...
https://app.suno.ai/song/24ffed63-cafe-4fd5-b78a-52a51751fd02
Got Vocaloid instead 🤣
Japanese super eurobeat male singer intense punchy synthesizer melody song. Listen and make your own with Suno.
Info like that (adding a word in ALLCAPS will emphasize that word) is in the suno wiki displayed like that (instruction manual)?
Might make a full song out of this because it's very funny
i meant like certain letter in a word. with unique names. like u make cap letter with ^ symbol after it or smth like this
there are probably 1000s of individual tips that float around this Discord. Not all can be verified, and we help with what we can.
No one is withholding information from you, it is not a conspiracy.
Well, most of us is still trying to see what works and what doesn't works
Can you just type the example you are looking for? Instead of making me try to guess and give broadly generic advice?
Like many AI black box... We don't really know how much prompting and what prompt works
I dont think there is a conspiracy haha. I´m trying to put together a "instruction manual" to use the best it can be used so before wasting days of life im asking if there is a place where i can find info about this (if posible, from suno developers), thats all haha
The community uses the bot more than the Devs do.
The wiki is an attempt to compile info posted here from the thousands of users.
The Devs do not even know what all the bot can do. It is a LLM like chatgpt. It knows millions of words.
what's the metatag name to change the style ? ( switch ? )
idk for example the word "into" you can put pitch in "i" or "o". so to put pitch on "i" you do I^nto, for "o" its gonna be intO^.
i may be bad at explanation
someone was using [fast switch] the other day, but I don't know if it was working...
I think the 'real' switch is to Continue from Time, then continue from Time again at like 00:01...
the 2nd clip will have no memory of the clips that came before. (or rather will have a 00:01 sec memory)]
I don't know a way off the top of my head, however a work around would be using phonetics almost. So rather than writing "into", you could write "eentoo" on "n too" things like that
It will take a bit of experimentation really of breaking down words almost
I would use in-TO
or inTO
or inTOOOOO
tested some neo-
neo-trap jewish niggun, indie neo-klezmer-dance-punk mizrahi song. Listen and make your own with Suno.
to note, it did NOT want to let me make that
i was flagged several tines
times
likely because the word niggun (it's a yidish singing style)
sort of, I mean, when you say it it generally refers to like, a melodic progression, sometimes with words, sometimes without, and is generally sung by just regular people
so kind of like folk music
Because there seems to be a variation in what is created, I thought I would try and be clear about the song.
I created this prompt:
Title: "Spring's Gentle Call"
Genre: Country/Bluegrass
Mood: Uplifting, Hopeful
Instruments: Acoustic Guitar, Banjo, Fiddle, Harmonica
[Intro]
Description: A lively banjo intro, setting a cheerful tone that feels like a bright, sunny morning in spring.
[Verse 1]
Lyrics Snippet: "Winter's chill fades away, as the dawn ushers in a brighter day..."
Description: Acoustic guitar joins, painting a picture of the world waking up from winter's sleep.
[Chorus]
Lyrics Snippet: "Here comes spring, hear the call, in the whisper of the leaves, and the rain's gentle fall..."
Description: Full instrumentation with a joyful melody that captures the essence of spring's renewal.
[Verse 2]
Lyrics Snippet: "Fields of green, skies so clear, the songs of returning birds, are music to my ears..."
Description: Maintain the hopeful tone, focusing on the signs of spring that bring joy and renewal.
[Bridge]
Description: A fiddle solo that evokes the feeling of a gentle breeze and the freedom that comes with longer days.
[Outro]
Description: Slowly fade out the instrumentation, leaving the banjo to end the song on a note of anticipation for spring's adventures.
Additional Notes: This song is a celebration of spring, emphasising renewal, growth, and the beauty of the natural world as it awakens. The lyrics should be reflective yet filled with anticipation for the new beginnings that spring offers.
However, I found out I could not input such a significant prompt. It seems the less you can add, the less of what you are looking for will come out, so you use a lot of credit.
I also found out that you can not modify the song created. It seems you have to go for a new creation every time.
So, can someone tell me how, based on the prompt above, I create a prompt to produce what I'm looking for, please?
Is there a way to make it sing a song in the genre of vocaloid? Because I tried it, and it didn't come out with the right genre. If it cannot use that genre yet, then how can I make an electronic song/electro-swing?
Any and all help is appreciated! 🙂
Can you drop the wiki?
search this discord for 'vocaloid' many examples posted each day.
My BIGGEST problem with Suno it´s ben at the moment to continue a song i have almost always the same problem. I select a time to continue, then for example i put a [guitar solo] and then i repeat the [chorus] with the same lyrics created before. And suno just jumps to the chorus and then halucinates and mixes the guitar solo or doesn´t do it AT ALL. Any solution to this? Anybody has the same problem?
how is this so much better than any music AI generator I've seen from big companies? It makes music that is actually good
i'm blown away
too much people on Suno now ^^
You got to let the previous section or go straight to the word. It work 100% like this. what i meant is, if the sentence is under [verse] section, lets all the section. If it doesnt works, it gonna work a the last word u put
Another problem, i´m trying to continue a song BUT changing 1 WORD of the chorus. No matter what i do Suno keeps repeating the first chorus ignoring my lyrics changes 😦
SO i gotta stick things when a complete metatag like a [verse] ends completely to add anything rigth? pretty sure i tried it but...
it's maybe skipping because the same word is where it jumps
for avoiding that, you need to let all the section yeah
exemple if its skipping at that sentence: " Spanish word marijuana " from : " [verse]
the birds are flying high like mary jeanne,
Spanish word marijuana
But Juan doesn’t like Esmeralda,
Like the bossue of Notre Dame"
and jump to an othere section, you got to let : " [verse]
the birds are flying high like mary jeanne,
Spanish word marijuana
But Juan doesn’t like Esmeralda,
Like the bossue of Notre Dame" or continue from :"high like mary jeanne ".
It depends on the case
How do you do chirp? Gonna see if u can try to recreate a song my friend made. She had a generate lyric from some words and I want to see if I can redo it
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.
did u succeed ?
Does anyone know if its possible to continue or remix a song and be able to keep the same style/sound of the original and charge a few words around?
The Suno AI website isn't working. It was yesterday just fine.
restart your modem will fix it
Alright.
Still having problem. Might have to check back later and see if it's working.
Brutal
yeah.. or wait.. His isp just didn't update some SSL creds.. rebooting helps
He could try ALT+F4..
read above
ok ty
Question, how should I use the command [Intro - Spoken], and how can I tell it what to say and when to stop?
is SUNO still down ? my library is completely empty, which kinda scares me...
Try this:
[Male/Female Spoken Intro]
Type your spoken words out here
Thanks!, Ill try it when suno is back
@red mauve You're not the only one.
Check this link for more information: #💬┃general-chat message
???
They are moving databases... check my message above you
👍
Rip
Do we know if they have given us a time line for how long
A few hours.. Please check this message: #💬┃general-chat message
Could be up to 3h or longer
or shorter.. but it will take a while
Depending on how long it takes them to fix the Variable and databases
But you NVR know
is the ai broken?
its down?
Yes
ok
Read his for more info: #💬┃general-chat message
they probably got a bigger computer
From coding experience It might take them awhile but not long bc one a database can the variables that store certain things like tokens and text
And even Coins
Such as what's referred here
Points
How do you make song
what happened tho?
Type in a few words
The AI will do it for you
Make song about happy dog
isnt it down?
thats sound really cool to see
How
Yes accept use it in #🐣┃chirp-beta-1
And run /chirp
A prompt will pop up
And type in a few words
It won't be any use at moment
Though
So I wait a good bit
Until there system is back up
When
Fr
Possibly 3h or longer and could be shorter
Oh ok
Are you always here?
Yeah
Nice
bet they're staring at some big migration SQL query wondering how long it's going to take and knowing if they cancel it now it'll take twice as long to roll back 😄
i wish i had those skills honestly Ai make being creative so much more fun if you don't know that stuff
More likely yes
it would be interesting to know what infrastructure they run on, have they ever said anywhere?
Yeah it takes time but it's cool to learn
I took a year to get certificates
On JavaScript , HTML, C++, C#, Python, Sunflower
Etc 🥰
I wouldnt know that part
But I say it can be very minimum
To the creativity
Bc there's not too many out there
Like for a real musician it can be a real source of inspiration and a base for much more sophisticated songs
yes it can be and it can help them look at song form different precatives it can also help inspire writers or painters in a more imaginative way to create
/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.
Hmm during moments like these, having an auto backup to a cloud like G Drive or Dropbox would be nice to have. Or maybe an option to mass download/backup to a hard drive
Is it just me, or does the AI seem to integrate some form of Power Metal or Symphonic Metal into a song that's any other sort of metal often? Especially for the vocals? It kinda ruins otherwise-good songs I generate when I'm doing something like metalcore or industrial metal.
That kinda happen to me when I tried using the Rap Rock genre lol
But it became a banger so I didn't complain lmao
I suspect this model may be trained off of power and symphonic metal more than any other genre of metal, lol
Hopefully that isn't an issue in V4
is /chirp working for you guys?
I'm not sure what it is. but I don't like it. Messes with vocals, adds harmony that I don't want and gives a double-track effect that I don't like. I would like clean, unprocessed vocals with all genres.
Very much agree with this statement. It really does screw with the vocals and the flow.
also /info, doesn't seem to work
so nothing is working at the moment?
I keep getting "DNS points to prohibited IP"
can you guys extend on music?
im trying to extend, but the "continue" button cant be pressed
test
Check if your lyrics/style prompts are completely empty. The button is not clickable until you type at least something (even just a space)
Are you able to download the songs u make?
thx
extending songs cuts it off before it finishes?
Newbie question, lenient help greatly appreciated:
I like to extend a song, continuing in the same music style, rhythm, melody, singer's voice. For the life of me, I just cannot achieve to have the extension sound like part 1, instead it sounds like a complete new piece. Thx for hints in advance.
Try leaving the style of music space blank when continuing the song (when its blank, it uses the style it detects from part 1), and make sure it's continuing at a part that's instrumental, or sounds like it's gonna transition to a new verse.
Also if you're gonna add new lyrics, don't include the ones in the first part, otherwise it'll just repeat the song again. Only include the new ones, and make sure they're structured correctly.
That's all I can offer.
All chirps down right now?
The website should be working (new url https://suno.com/ but old url should forward) but Discord bot is still offline
Suno
Is there a way to use a specific voice from a generated song to create another new song?
Are there ways to improve the quality of songs in post-production?
/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.
just me screwing around
neo-baroque jewish niggun, indietronica neo-klezmer-mizrahi song. Listen and make your own with Suno.
I'm going to subscribe! Why? I think I've figured out how Suno works! It works best to think of the song you want to create as a musical journey or story. Once I changed my outlook, the songs worked so much better!
Here is an example:
In Simple Mode, that is, with Custom Mode turned off, I added the following prompt:
A serene acoustic melody that echoes the tranquillity of a quiet lakeside at dawn, with soft, reflective lyrics about finding peace.
This prompt suggests a calm, peaceful theme, guiding the AI to produce a song that captures the serenity of a lakeside scene.
If you visualize the essence of the song, you wish to create and consider the emotions, narratives, or atmospheres you want to evoke. Simple Mode thrives on the power of your imagination, translating descriptive language into music.
So pleased it made moGronik's day! 😏
You're just jealous lol
All my songs are cut at 1:58 min. even when the text is already said it repeats the chorus and then cuts off cause 2 min are reached i guess
I paid for this... Any help here?
Hello! Does anyone know if there is a prompt to get the same (instrumental) theme already exposed earlier in the track?
Good question and not sure off the top of head. Also interested in the answer if somebody knows the method. First guess maybe just tag both with similar structure like [hook] or [hook: blah] and reuse that tag exactly.
I'd like to have the same instrumental theme that I already had previously in the song, the problem is that I only used [instrumental] and if I reuse it I get a new instrumental phrase and not the same theme. I also tried with [resume the same theme from minute 01:02] but it does not work
Same, I didn't pay though lol. Another issue is "extend" doesn't work right and the lyrics get cut off half way through a word instead of cleanly finishing a verse.
E.g. I had a lyric which went "Their reach knows no bounds, their story yet unfold."
It cut off "unfold" to "unfo"
So, my next lyrics put into the extension started with "ld" and then the rest of the song. Firstly it didn't finish the previous verse and just cut off "unfo" and went to a drum fill, and the tune of the song wasn't even the same as the original song in the extension completely destroying it and making it sound terrible
When extending, you can put the whole current verse in the extending lyrics, it correctly continues from the right position. If it doesn't work, try continuing a second or two before the cut-off.
What I have problems with is choice of instruments. Is there a way to discourage the generator from putting a piano in EVERY single of my slower songs?
I was exploring the BETA options. It sounds like some amazing things are happening. I found the sitar and drum. Are these available now in Suna V3?
Guys can someone help me i try do create a dark country intro (allready got that one i think) and want to switch the genre now to a dark djent/ modern Deathcore vibe... there are few generation that fits to go into metal but not this kind of genre... maybe i need another prompt or so? Could a prompt expertise guy help me?
is there any way to have the extension set to a particular time limit? For example if I just want it to end the song without being cut off?
try "[end song]" or "[outro]"
Thanks!
Would this work in Complex Mode, Custom Mode on, under the Style of Music: Uplifting pop with orchestral elements, featuring sitar for a unique blend.
No sitar in created song.
I'm not sure where is the best place to ask questions... Can anyone suggest a place?
I want to know if Suno can work with BPM, Key, scales such as major, minor, Pentatonic, Microtonal, raga, etc. Can I use other aspects in the song creation prompts, such as melody, harmony, rhythm, timbre, etc.?
BPM and Key, no. (Maybe they could work, but I don't know a reliable prompt for them). Major/Minor probably not but some users report mild success. The other terms you mention are the sorts of words that are more likely to work though I haven't tested every one in your list, but that category of music terms I mean. In some cases you may have to repeat a word because the effect works, but is just very weak, something like "country, atonal, atonal, atonal-country"
Does Suno support prompts like [Chorus], [Intro], [Outro] etc.?
Yes, those specific prompts (song structure tags, also [verse]) are recommended as defauluts actually, they tend to help the AI. Sometimes [verse 1] [verse 2] can be good too. You can do other things like [guitar solo] but that is much more hit or miss (but fun to try)
Noice!
Many thanks for taking the time to reply! Discord is a strange place!
I suppose as Suno develops more will be added to influence the outcome. I was thinking that as it uses a LLM, it may already have such elements built it, but it comes down to what Large Language Model is used.
Thanks again! Have a good day!
As an LLM, you can try almost any thing like [rising aggressive strings] to type a random example. I have seen people pull some wild things off. But that level of tight control is hit or miss. Also can do things [chorus: angry rap] and it usually at least has an effect on the music, even if it's actually following the instructions
I've just been reading some of your posts on X. Very good! The video ChatGPT apologies was magic! I think you may be able to help me with my fund-raising challenge, which I'm getting ready for starting in August. I'll leave you peace, time to have a break. All the best, and I hope to chat again soon! 👍
Haha, if you check my songs I post in Discord they may not be the best examples to learn from. I love to use prompts which make very strange things, or confuse the AI, have it hallucinate, etc. Do as a say, not as I do, lol. My songs are filled with [absurd bracket things].
does anyone know how to play a section of the music with just 1 word?
cause I want a chorus to play but I don't know how to have it do it without going over the limit
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.
also where does it work in discord
I'm not sure what you mean by 1 word. It's only 1 word but it's also too long and over the limit?
It seems to be something wrong with the continue from track. It only shows expand now. And one of the tracks I had to day was a regular pop song, but the sound was quite distorted when I downloaded it. Is there any way around that? A bit sad when I've been playing around with words and stuff and it sounds too bad. I mean just the sound quality. Most of the time it's pretty good, but not distorted.
still chirp is not working from morning?its not creating songs
chirp beta is not working from morning??
still chirp is not working from morning?its not creating songs
whether chirp beta is still down??can anyone confirm it
Yes, I confirm. Doesn't work for me either.
ok ,any information on when beta will up
Unfortunately no.
Is it mp3 only for the songs or can the they be dowloaded as WAVs etc? Thx
Your bot doesn’t work
like have a designated chorus then do something like (chorus) to have it play that (that doesn't work but example)
Are there keywords to force pauses? like comma or a dot? i saw that () forces background voices. is there a list to make the text as clear as possbile for the AI
found out that [Half rest] works each generation (4/4) before a verse but for some weird reason [Whole Rest] worked only on 1/4
Also most of the Songs end abrupt. Can't really use any of them. Thinking about a refund tbh
Hoping someone can help me. After many many many tries my song came out perfectly. Except. I need 5 more seconds at the end for it to fade out. It cuts off RIGHT as it's about to fade out. What are my options? I don't want to extend it another 30 or so seconds.. just 5-10 to fade away. Edit (rare "I figured it out" follow up): I "extended" the song from the end and only put "[Instrumental break]
[Fade out]" in the lyrics. Worked great.
I'm running into something similar.
I can post a link to the song, if that helps
can take the lyrics, make a custom song, and maybe add [Fade] in there?
Good idea, but I don't want a new sound. I love this version and I just need it to end properly.
Oh. Yeah, would probably make a new song out of it-with a new sound but the same lyrics
Figured it out. Edited my question above.
I've been trying to extend a song that i made on the website, but it's not working, it just generates a whole new song. I guess i must be doing something wrong.
Make sure you are clearing the lyrics out of the box on the left when you click "extend". Otherwise it will use the lyrics in the box when extending the song. You have to think of it like "what comes after X time" you are extending. For my example, I kept the "extend from" time to be the very end of the song and only added some meta tags for [outro] and [end]. This worked for me.
However, on another song, the song cut off mid final chorus. I extended the song using the two metatags above. It extended the chorus and kept the song going a bit more. It ended about halfway through the "extended" clip, so I'll have to go in and make a small fade out on the "whole song" final version.
I hope that helps.
Thanks, i will come back to you, you're a lifesaver 😄
Why when i use /chirp does it say this interaction failed?
But the continuation sounds very different from the original. So, am i correct to assume that it won't ever continue with the exact same style?
It should sound the same. Make sure you have the same style settings. Can you post a screenshot of what you see when you click extend?
I think i figured it out, because i had an instrumental part at the end of part 1, it continued by sampling that, i got a much better result now
Awesome! Nice work!
🦧
Is suno broken today? I tried it didn't work. Is it me?
Appears to be down atm
works for me...
For Suno. When extending a song and go into the second chorus will it use the same chorus (lightly adapted) again from chorus 1? Or does it create a new one?
I used [Intro] > [verse] > [Chorus] > [breakdown] and would like to continue with > [Chorus]
Suno doesn't work for me a whole day? What happened?
I've had no problems. It's been especially good for the past couple of hours.
they changed DNS with the update yesterday.. and some ISP haven't activley pushed the new "route".. In my case a VPN helpd, but restarting router helps too sometimes.
That's strange. VPN doesn't helped at all. I've got one and the same error "Interaction error".
Will try it tomorrow again.
Thank you!
MP3 only
Yeah, lyric generation is no longer working on my end.
you can't reuse prompt to continue the song atm
how would you focus on singer 1 being male voice and singer 2 being female voice for generation?
to generate 2 distinct voices for duet
There is a message over in #💬┃general-chat for one approach:
#💬┃general-chat message
I asked for a song in a 12-bar blues format and it didn't work.
why can it let me type in a prompt?
The music has been having a lot of skips in them for the past hour or so
wont work
every single time i type /chirp and press enter i get this!
this was in a DM with Chirp too
yep
Thanks for the reply. I don't think Suno can work with these parameters at this moment. I'm sure it will be here soon. Have a good day!
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.
Good Morning all from sunny England! ☀️
Well, I've been checking out the BETA options on Suno.ai. It looks as if Suno is continuing and developing its excellent song creations based on the common elements of music: melody, harmony, rhythm, and structure, along with common themes. But it looks as if they are looking to add the many amazing music genres, styles, themes, and instruments from around the world.
At the moment, the songs from Suno often have a similar sound, structure, and output.
However, if Suno can, and I'm sure they will add World Wide Themes to play with, the possibilities will be so, so different!
I can't wait!
Have a great day creating!
Can I please ask you a question, before I start wasting credits.
Can I use classical terms eg, Largo, Moderato or Andante etc, inclosed in brackets?
Many thanks
will there be a way to add more lyrics in the first attempt without having to extend? I find the extend function really clunky and doesn't do what I'm asking it to do
You can try. I'm not 100% sure that they make any difference. I have tried in Style, without brackets. Capriccio, I think makes a difference and also Cassation seems to work. As with anything in Suno, experiment, experiment, experiment!
why doesn't the bot work?
same. Did you figure it out?
I want to repeat the chorus, but i want to repeat it "stronger" with more energy and a bit louder music, which meta tags should i consider?
i am making instrumental music, is there anyway to extend the music to make it not repeating the same melody over and over again? How to improve the quality of the extended music?
- Use unique and specific style prompts, full of many diverse words, instead of something short and simple.
- You can guide the instrumental in the lyrics field somewhat with things like [guitar solo] [glissando trombone] [ominous, dark] but the effect is not stricly followed (even though the words usually do have an influence)
Someone got an idea how to get only one guy singing ( i don't want echoes of other people )
Hello all! I have a question, I suspect answer is No but just in case:
Is it possible to "fix" lyrics in a song but keep the music?
"male solo" / "female solo" seems to work for me most of time
Can male and female choruses be formed?
how can i direct message suno?
Direct contact with the Suno team is via mail support@suno.ai
No matter what it always shows interacton failed
hey its my first time using suno, can you force the ai to stay closer to your lyrics?
For me its just luck lol sometimes the song follow and sometimes not
Interaction Failed
aight thanks haha. can i somehow changer the ai to only produce one song?
it says everytime interaction failed by me when i try it, can anyone help?
me too...interaction failed...😒
Hello hello all, just a question regarding subscriptions for Suno, is it not possible to use PayPal at all for it?
I am getting this error too. Did you end up solving the issue?
Guys how do I tell it to make an instrumental break/bridge in-between lines/sections of a song?
Hey abobeek, Were you able to get chirp working?
no
Thanks, thought it's up running. First time using this. Thought I am doing something wrong. Sorry not much help on your ask
Consult about a problem of song editing.
I composed a piece of lyrics in suno custom mode. The structure is:
1- Verse
2- Chorus
5- Bridge
6- Chorus
Then I got a good result.
However, the structure was trimmed when I wrote the first time as a trial. The whole structure in my mind should be:
1- Verse
2- Chorus
3- Verse (same melody with different words)
4- Chorus
5- Bridge
6- Chorus
Then I use "continue" at 1:09, where point 2 ends, and rewrite the whole lyrics.
However, phooey, the whole song turns out messy! The order of the lines of the lyrics was sung preposterously. And the panel of lyrics displays such structure:
1- Verse
2- Chorus
5- Bridge
6- Chorus
1- Verse
2- Chorus
3- Verse (same melody with different words)
4- Chorus
5- Bridge
6- Chorus
which means it just merely adds them up! But what I need is to insert a part in the middle, not to continue by the end.
So what should I do to acquire the result that I suppose?
Btw, when I did it yesterday, "remix" was missing after the update. So did the "remix" serve my purpose?
Please do not double post your questions in more channels, as stated in our rules here: #📜┃rules message
To edit the song structure in Suno, consider these steps:
- Start Fresh: If a song’s structure isn’t what you intended, it can be easier to delete it and start anew, inputting the complete structure from the beginning.
- Use Precise Timestamps: When continuing a song, be very precise with the timestamps to ensure correct placement of new parts.
- Be Aware of 'Continue From' Settings: This feature might retain settings from previous clips, affecting the new generation. Adjust accordingly if you notice repeated or short sections.
- Try Different Inputs: Experiment with different lyrics or styles, as varied inputs can influence the length and coherence of the generated sections.
- Consult Community and Support: If challenges persist, the Suno community and support team can offer additional tips and solutions, as they regularly deal with similar issues.
And "remix" is now renamed to "reuse prompt"
How do you make a song start with an instrumental intro? and how do you tell Suno to insert an electric guitar solo in the middle?
As for starting with an instrumental, I typically make the instrumental first then "expand" into the first verse from there.
it stands to reason that you should be able to "expand" from a verse into a solo as well, then back into versing afterwards.
If I'm planning to extend a song, should I put the full lyrics from the very beginning or should I add the rest of the words later when I extend it?
Sorry for reasking, this message just went up too fast in general chat, this one is slower.
Best to only do 1 verse at a time
Thanks
woah thank you, yeah it actually even says on the website, but I think it said like 8 lines?
for the intro yeah. I've found that 1 verse of 4 comma separated lines works best, especially if you plan background comments "()" at all. Technically that could also be construed as 8 lines.
do 1 verse and 1 chorus. i tend to do chorus for every part in order for it not to forget the chorus. make it diffiqult as sometimes i get a good generation but it is too long for the generation to handle which is sad
definitely an option. I very rarely see it complete both in one publish attempt. YMMV, of course!
i recently got a very interesting result out of trying to do rock and roll
listening to genres can be quite illuminating
Firstly start from where who want the new lyrics. Secondly I’ve found I often get more consistent with earlier section type results if I do it section by section. Even if you leave a chorus lyric in and it’s a second chorus, sometimes it affects the new verse that you’re starting with.
Mhm ok 1 verse of comma separated lines is sort of what I defaulted into but good to know that’s actually a thing.
Any concrete knowledge or even ideas about how to guarantee suno to use a specific instrument? 3 separate prompts for a specific, common instrument in instrumental and normal mode and it can just never use the instrument.
what if i write /chirp on che chirp-beta-1 channel and nothing happens?
i am so new to this - how can I lower the music and raise up the words.
"midwesinger/songwriter"
as the start of a prompt
produces high clarity on a set of prompts I'm checking
could help with others
I found that doing extension with 2 verses or 1 verse and a chorus work very well, the extension is often around 40 to 70s
Doesn't it make a veryyyy long intro ?
not if you choose where to start with the time markers. I usually dont do more than 30s
as butter 🙂
good idea, including for solos in middle of songs i guess
Do they expect to extend a little the size of the prompts we can put as styles ? to be able to add more precision on voices etc ?
mid-song tends to be a bit hit or miss as the AI seems to have enough free will on deciding if it cares about your request. 😛
maybe better asked here....
If I write my own lyrics and I want the voice in one part changed from male to female, how can I do that?
Sometimes extend do not conclude the song. Never tried to add [Outro] or something, will try it next time
I tried an extend and changed the style to Male tenor voice for example and it worked (80%, the girl game back to alternate with the man) , it was satisfying
i use Outro, Chorus, Pre-Chorus, Bridge and Verse tags on everything I do.
By tags you mean [Chorus] for example I guess
aye
I would like to find a way to put tags that says who is singing to generate dynamic verses with alternating voices
i've found that you can tag [Solo], [Happy], [High Energy] and other sorts pretty reliably. But then again, the AI decides if it wants to listen or not
and how did you do that?
I tried [female] but that didnt work
short question... is it me... or does the quality improve from a suno song if you wait a bit longer to hear the result? 🧐
Did first the girl with alto voice and extended with male etc . Example: It's in French but you will see https://suno.com/song/9bae2dbd-c3e8-4b36-a5db-42535566bbde
Suave blues alto (Girl), acoustic instrumentation, chill rythm, passion rising song. Listen and make your own with Suno.
No tags, use the style to extend with a different voice
i put also in tag but not really what changed the deal
say what? but it changes back and forth... did you extend it everytime it changes?
Not sure I did duet , not sure I kept them but it worked
I will probably try it out tonight.
I've found that over the span of a day the AI's voice seems to grow weaker with every generation. I generally get better results the next day if/when that starts happening.
still getting this error when I type anything to the bot. this was after typing /help
Anyone found a way to force a change of chords (or chord progression) in instrumental electronic tracks?
Apparently they're mono "chord"... bassline is always one tone
No. Chords aren't recognized. I've never had any success with chords.
do you mean in insturmental, or with lyrics?
Both
Mhhh
Normaly if you put the same chorus lyrics in it it will recognize it today i found out the solution :3
But only for vocals... not instruments
A chord progression or a change of key?
Chord progression
Can you show me how you do that?
mhhh
oh yeah
here:
Modern Nu Metal, Hip Hop, rap, heavy, bouncy, angry, scratches, scratching, sprechgesang, screams, [women solo] song. Listen and make your own with Suno.
After the first generation when you need another chorus in your second generation then it will work as intended
for example in this song the chorus will be generate two times so it will remember even better the chorus
I don't see where you put in a chord progression
chorus should sound very similiar right?
And that it does! 😄
even with uplifting the last chorus use a whole crows like it should because community theme and so on
Ah, I always get my choruses to be the same but I don't put in a chord progression
ccould you explain to me what you mean with chord progression? Maybe i misunderstand the term right now .___.
Yes, I think we're not talking the same thing! A chord progression is, for example c major, f major, d minor g major
ohhhh! Now i get it... no i saw in the technical document and here a word from author that this isn't possible right now
You mean "harmonielehre" (in german :P)
Now i understand
Yes, same as BPM or setting the key, not possible, very hit and miss
do you count syllables? i try to get the ame amount in every line because otherwise the vocals becomes offbeat or too slow or is streching some weird stuff
depends on the genre... when you use rap and hiphop it will work very well
[break]
for the vocals
Even here the same example my song https://suno.com/song/bb5f6a07-a665-4090-becd-e78955652b55/ xD
Modern Nu Metal, Hip Hop, rap, heavy, bouncy, angry, scratches, scratching, sprechgesang, screams, [women solo] song. Listen and make your own with Suno.
look at the lyrics [break] and then "text"
ja das geht besser
deutlich
aber hinter jeder Zeile Break schreiben ist auch dumm 
Aber es wird funktionieren! 
https://suno.com/song/45d39f10-cb34-4e45-acf4-d777c893f312
Hier bei dem Song ist alles bis auf den Chorus soweit gut...
Aber da verhaspelt er sich in der letzten Zeile...
oh... nun das wird warscheinlich wegen dem chorus nicht funktionieren, versuch mal 8 Zeilen anstatt 4
Aber der rest sind doch auch 4 zeilen
Jaund? 😄
Dann dichte ich nochmal 4 dazu
😄
ja ich hab doch kein plan, mensch 😄
Jetzt ignoriert er die Breaks 
Egal, ich mach Morgen weiter
Is anyone know how to do instrumental version of the song with vocals based on the generated song? I tried to extend the song with [Refrain] and [Reprise] and even tried "instrumental version" in Music Style. It failed.
/chrip
Is anyone else experiencing distorting and reduced quality overall? Is the company aware of it? I feel like I am wasting my credits when things are so muddled right now.
hey when type "/chirp" its saying the bot is failling any ideas?
Sorry, the Discord bot is offline. The focus has been on the website which is also a newer version.
Yeah, all good. It’s been a while since I came back here. The website is way better. I wish I came back sooner.
@river lodge is this the official wiki?
is there by chance a list of metatags i can get a link to or directed to?
nvm, the wiki has it there. should have looked before i asked :3
Chirp is down
It's unofficial but written by many people who are Discord mods or guides, which are kind of official-ish in a way, though just volunteers. But it's not a guide written by Suno employees.
how can i create only one song from the lyrics instead of two?
Is there a way of choosing if the singer will be a man or woman? I didn't find it.
Chirp not working
If you're using Custom lyrics, I've seen someone say that applying a masculine/feminine name to a verse or chorus (eg. [Chorus] [Paul] or [Chorus] [Paula]) tends to work.
If you're just using Simple, then add "male singer" or "female singer" to it
The Discord chirp bot is offline, but Discord was Version 1, and website suno.com is Version 3 and it is working.
Thkss!!!!!
why suno music clip is only for 2 min ?
Many thanks for the reply! I am experimenting, but there must be some consistency. There is no guidance on how or what to add to produce what you want.
I was just listening to Whiste Along, number two in the Trending Charts, and I noticed that you can Reuse Prompt. So, I gave it a try. It took me directly to the Complex Mode (Custom On), where it added the details automatically and clicked Create.
Amazingly, it seemed to use all the lyrics, and it sounded pretty damb good.
In the Lyrics section, there is a specific tag, [whistle solo], and in the Style of Music section, the prompt was for a unique blend of whistle music, beatbox, Melodic rap, and nu metal. I'm assuming this was the original prompt that it was trying to recreate.
However, there was no whistling! Does anyone have any idea?
there is no telling how many tries the original song took to get the effects it has.
Quick heads up for experienced Suno prompters. A lot of 'technical' prompting talk ends up taking place in #🗣️┃chill-chat , surprisingly, spinning out from other off-topic discussions. So that channel is worth a skim or search as well.
hrm.. anyone else getting: This interaction failed repeatedly when trying to use /chirp?
The Discord bot Chirp is down, but the website version is working. And probably the website is preferred because Discord is version 1.0, and website is version 3.0.
Oh ok
When will /Chirpbe up and running once again?
So whoever it was just regenerated, regenerated and regenerated!
Are there any specific settings/inputs that give a known outcome?
The AI has strong random properties. It's designed to never give the same thing twice.
With specific lyrics and prompts it will give a similar song, but most subscribers would not want everyone else making an exact duplicate. We would not be sharing prompts if that was the case.
Thanks for that. I can see where you are coming from. I suppose this is what gives you the buzzzzzz when using Suno! The expectation of what will be.
Voice
Impossible at the moment
I think the biggest challenge is the voice that are becoming metalic or choir effect
Because it turns a good music into something sometimes very difficult to hear compare to the first 2min
have to agree. Almost every song I do is plagued by that 'effect'
I wondering what is the problem, because if it's beautiful, why is it losing his brightness by time passing.
is it when it try to change the flow ?
doing add libs ?
dunno
I find it interesting when the AI seems to be losing its breath. The singing gets all faint and the sounds like a whisper.
what kind of music do you generate ?
is there no way to at least guide the random? Like if i want the song to be high tempo and open with guitar harmony, i cant really instruct it to do this. I can maybe put [Guitar Solo] before the first verse and then set a genre before wasting 200 credits until i finally get what i was looking for
Prompt is the only input. That is where you should focus.
Yes you can guide it with prompts and metatags, and especially your lyrics...
https://www.suno.wiki/faq/metatags/
When are you going to FIX THE /chirp command?
The WebApp will be taking over.
I quote: We are going to archive these channels at 12pm ET on Monday (4/14)
Nice, thanks!
See my post above
Hi all ! 😄
Someone have a good lyrics generator for song ?
i try some of Gemini, GPT, jasper or jarvis,
but I find that the sentences that come out to me are too abstract, too childish
if someone have a good BOT i take ! 😄🌺
a human. (not joking)
yeah ahah i know XD I've been writing about my life and my feelings for over 5 years now... and I confirm your answer x)
Except that my depressive texts are censored most of the time ^^
Technical question - internally, suno's AI generates spectrograms which are then converted to audio using the Griffin-Lim algorithm, right?
I was thinking that could explain the bizarre swooshing artifacts that occur when the ai tries to make noisy audio (where most of the frequency bands are high amplitude, such as with snare build-ups, heavy reverb and sweep pads)
when i click on " get the whole song " nothing happen
those pesky artifacts I cant even
I'm looking to try and replicate the low to high pitch the voice uses in this Audio clip; but I have no idea what the voice technique is called
Does anyone know?
anyone know how I can add sounds in a song? like a bark, lightning/thunder, bouncing ball, squeeking shoes, etc?
it's bad at anything that's not music. you can try sound effect but never tried it myself
how?
maybe add bark in style of music or sounde effect : bark or try in lyrics [sound effect : bark]
Hi /chirp is not working, I have a subscription and added the bot to my server. Can anyone help me?
Manually tuned Voice.
Awesome, thank you
Ah, so it's not something natural, you gotta voice tune it to do that that?
Can I specify a band or song, so I can get a similar tempo or voice?
[Allegretto] and the likes in the beginning of lyrics. and for the voice you have to be precise and concise in style of music. (like reverb/delay in voice,etc....)
Any suggestions to get clear vocals, every time I generate a song I get reverb or it sounds like the mike was on the other side of the room when the lyrics were recorded?
suno v4
I can send you a link to the song I made and another of the one I want it to sound like.
How do we finish a song?
the best solution and the least worst option is to add in style of music something like 90's voice or choose the date.
ok thanks
generations are throttled to just subscribers right now...
lofi beats to study to, but without lofi prompting
"anti lullaby" prompt giving me success lately, but doesn't mix with many other styles well yet
/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.
Wassup, guys. Sorry if it was asked earlier, can't find. Is it possible to reuse generate music style on website ?
post-anti-instrumentalized anti-hyper-gothllo trapelloized jewish nigun40404040404040, accidental song. Listen and make your own with Suno.
well that's enough messing around lok
lol
hey bro ! I listened to some of your sounds, I find them great!
Where do you find these prompt ideas? they are lunar 😍
loooots of experimentation
oh okay x)
in general, if you notice a prefix in any genre tag on music you like
mess with it and try to get a feel for how it impacts the soind
when I generate, I start with an idea of what I want, I draft a probable win, and then adjust
I spend the most time on first gen
like, the start point
same for me but i never have your prompting skill x)
Is there any reliable way to make a call and response style song? As in, have a male singer ask a question and have a female singer answer?
I can't seem to download my songs as a video or audio file to my phone gallery from the website. Please advise.
I am trying to make a part of a song with a female, happy, high pitched voice
but whatever I do its always a male voice...
is there anything i can make to force it to be female?
i cant creat song .. iam new heree
i managed to get it to female by writing like 10x female in the music style but now I put the original music style but now it wont turn back to the original singer or even a male voice
In custom mode, when I have generated a song before doing the next generation, can I edit the lyrics while maintaining the melody intact, or trim some rear part of the lyrics for amending it in another generation?
you can change the Extend from time, with that you basically cut off the song from the time you specify and continue there
you cannot edit the lyrics while maintaining the melody unfortunately....
Only way I found is to cut off at 0:01 to keep the melody with different lyrics... not the same but closer
I just tried your method by cutting off the song from the time I specified and adding nothing. Then generate and get the whole song. It turned out that the whole song displayed the remaining lyrics duplicating themselves for the second time! But I only need one time to amend new lyrics later!
Could ask the dev to train the IA, on Haitian Music pleaseeee 😢
Definitely! 😎
How do I end an instrumental song? For songs with lyrics, I just need to insert [End] into the lyrics, but I don't know how to end an instrumental song.
Howdy folks; is the /chirp functionality down now that the site is live instead? Seems like "/chirp" in the chirp-beta rooms error out.
My goodness I just remembered why Woodhouse was such a familiar name - what a great show
2nding this one
Ok so I'm not sure if this is legit or not but I got it to work @bronze birch
Before they updated the UI, I noticed that setting it to instrumental just puts [Instrumental] in the lyrics. So I kept instrumental off, but put
[Instrumental]
[End]
And it worked
But the quality seemed lower
might just be bad luck
Dunno if this is where to ask but how do we adjust what is on the profile? it's seemingly random and not the songs I'd like it to showcase
Is there a list of prompts somewhere I can see?
Is there a limit of the length of lyrics in custom mode? How many words?
There is but if you max it out it'll never get there
b/c time limit
best to keep it to 8-12 lines
I mean I amended part 3 but couldn't input more words any longer.
like couldn't type or stopped talking
when you continue back up the time on the continue it depends on final beat in the song when you continue. That's if it's not talking not if you can't type
And could I use a metatag to simply repeat the same chorus instead of retyping it?
i retype and put [Chorus} above it
here's one of my songs for example how i structure https://suno.com/song/392747bb-b9c8-4802-8e6e-95011decbc0f
Indy pop rock punk female singer song. Listen and make your own with Suno.
But that will exceed the word limit.
it won't read the [Verse] or [Chorus] lines
I never hit word limit I do 8-12 lines per generation not counting those tags
I directly copied and pasted your lyrics and it did exceed the lyrics box, failing to display all. So how could you make it?
thats many generations to make it all
this is a single generation
[Verse]
Sittin' in the darkness, starin' into her eyes
Mischief in her heart, a devil in disguise (disguise)
Playing with her toys, plotting all her schemes
A tiny ball of fury, causing chaos in my dreams
[Chorus]
My evil kitten, oh, what have you done? (what have you done?)
You're cute and cuddly, but your mind is on the run
You'll scratch and bite, leave me black and blue
My adorable nightmare, I'm a pawn in your debut
[Verse]
Sittin' in the darkness, starin' into her eyes
Innocent eyes, a little disguise
But deep inside, there's mischief and lies (lies)
Paws so soft, but claws are sharp
You purr and play, but you're breakin' my heart
then i extend to continue
Do you mean if the input exceeds, then I should generate more parts with newer lyrics?
yeah just do like 2-3 verses then continue on next generation
By the way, if I got a excellent melody but the lyrics have typos or I need to insert a part before the bridge, how can I make it?
I did. But when I got the whole song, the lyrics was wrong since it rewrote the whole lyrics from part 2 below part 1 instead of inserting it in the middle.
may of got confused sometimes that happens
Can I edit the lyrics to trim the redundant parts without changing the music at last?
should be able to if you set your start point for extend generations
you just lose other parts after so have to recreate up to that.
How to set start point in the lyrics box?
like this
where it says extend from
you put only new lyrics you want
old lyrics already in the start of song
I meant, for example, I composed a piece of lyrics in suno custom mode. The structure is:
1- Verse
2- Chorus
5- Bridge
6- Chorus
Then I added 3 and 4 in the middle to expect:
1- Verse
2- Chorus
3- Verse (same melody with different words)
4- Chorus
5- Bridge
6- Chorus
Then I used "continue" at 1:09, where point 2 ends, and rewrote the whole lyrics.
However, the lyric panel turned out messy, displaying such a structure:
1- Verse
2- Chorus
5- Bridge
6- Chorus
1- Verse
2- Chorus
3- Verse (same melody with different words)
4- Chorus
5- Bridge
6- Chorus
if you pay for GPT i made a song crafter to help https://chat.openai.com/g/g-HK6Jg1Eo7-song-crafter
I have it set up for suno song creation
Does it have a relation with chatGPT? I just need the lyrics displayed correctly in suno.
yeah this is a chatGPT model I made to structure songs for suno I use as start point and then build off them. If you have the lyrics already you just plug em in suno
I have a chatGPT acc but it asked me to upgrade my plan.
Simply, can I write the code or metatag in the lyric box to solve this problem?
yeah just write it yourself
But I don't know the internal mechanism of suno. Thus I don't know how to write the code in the lyrics box to make it insert lyrics correctly. Do you know how?
just [Verse] or [Chorus] for starters it's a trial and error to find what works best
but those always work
[Verse] for the first part of your song that isn't repeated and [Chorus] for the parts that get repeated throughout the song. If you use [Start] then [Intro] below it, it will majority of the time give you a nice opening instrumental part to start your song.
I mean, how to insert a part before the bridge in the second time generation?
Ah gotcha. Yeah I've had trouble getting it to work for me on that part. Seems like the couple times it worked for my husband was a fluke. Now it won't do it at all.
it's all luck at that point
LOTS of generations
there is no catch all make instrumental
So do you or anyone know how to solve it?
solve what it's how the ai works. you can manually edit the music that's only real solution
there isn't a add instrument here repeating chorus there. it's ai. you generate a lot till you find something useable for what you want
even repeating chorus's aren't perfect in it
Tip: If you want an instrumental intro in your songs, put "Instrumental Intro" in the style of music. For a lot of generations this seems to work best.
the beat alters sometimes or reads it different
he wants instrumental in middle
Add (Instrumental) anywhere in the lyrics
once you have a chorus you like does anyone know how to prompt it to do a double chorus?
But I expect I could type a code in the lyrics box like:
[Insert begin]
lyrics
[Insert end]
to solve it.
gotta pay suno the big bucks
yeah i do lol
I never use all my tokens
wish there was a 5k plan 10k I never use up
😮
i'll take your points
i have to pace myself otherwise i blow through 2-3k credits a day
You gotta experiment with Suno and burn hundreds of credits just for a song or two
it's blocked at my work so only get to mess with it for like 2hrs a day
though more like 500-1000 for a song usually for me
Go the Pro subscription and top up when low. Top up credits are flexible
Premier is a much better deal than topping up on pro
Yeah, if you can use up 10,000 credits a month
I never use the 10k but it was cheaper than topping up for me
since there's no middle plan
it makes sense at 6.5k+
If one can afford 3 overpriced Starbucks coffee, they can afford Suno Premium
yeah plus I like them the devs are pretty nice even though they never fixed my discord to web issue I had lol
Suno is a small team and they are just trying to scale. GPUs are expensive
yeah. still best music ai out there. the new udio has better vocals but the beats are kinda crap
The key to get a clean generation is not to press the play button right after Suno generates it. Give it 3 minutes or more to generate.
I expect I could type a code in the lyrics box to solve the lyric insertion problem. Does anyone know how?
Nobody. You gotta burn credits and experiment yourself
depends on genre and how you write lyrics so it's unique to you
yeah even so it's not as clean vocals.
though suno is by far still my fav
I just need to amend:
1- Verse
2- Chorus
5- Bridge
6- Chorus
to be:
1- Verse
2- Chorus
3- Verse (same melody with different words)
4- Chorus
5- Bridge
6- Chorus
But it turned out:
1- Verse
2- Chorus
5- Bridge
6- Chorus
1- Verse
2- Chorus
3- Verse (same melody with different words)
4- Chorus
5- Bridge
6- Chorus
So how to fix it?
are you auto generating?
Add a refrain and hook
just write your own vocals or copy paste into a notepad and write how yo uwant
That looks like a ChatGPT format for lyrics
they give that many?
I've written my own GPT for suno song writing so haven't tried generic chatGPT for so long
No, I drafted the lyrics.
Or you can reach out to your Weibo friends
then why can't you just organize them how you want?
not sure what you mean how it turned out if you're typing em
like how did you go from 6 verses to 10 if you are sectioning them off yourself
I'm confused
I drafted lyrics:
First version:
1- Verse
2- Chorus
5- Bridge
6- Chorus
Part 2 (continue from 1:09), and the lyrics display:
1- Verse
2- Chorus
3- Verse (same melody with different words)
4- Chorus
5- Bridge
6- Chorus
But when I clicked "get the whole song" it turned out:
1- Verse
2- Chorus
5- Bridge
6- Chorus
1- Verse
2- Chorus
3- Verse (same melody with different words)
4- Chorus
5- Bridge
6- Chorus
While I expect the whole song to be:
1- Verse
2- Chorus
3- Verse (same melody with different words)
4- Chorus
5- Bridge
6- Chorus
So how to fix it?
why would it of cut the song unless you change the time to start from last creation.
i don't know what 1:09 is on your continue is that after it said everything from first verson
if you start before it says a verse as long as not right before it should skip all after and start new if too close it will still say it
private chat
just a question, does the audio download allow anything more than 192 kbps or is that hard set?
May I directly chat with the tech staff who is responsible for the lyric panel? What is his/her username?
I believe its pretty much hard set at 192kbps unless it can be specified in the style of music you are generating. I honestly couldn't tell the difference when putting the kbps in the style area so I can't say yeah or nah on that if it really works.
Nwnw thanks for the info, I'm just assuming those hardcore music ppl will hate 192 kbps since they can hear the quality drop lmao
No problem. I like hardcore music myself and haven't heard any quality drops. lol Maybe I've just gotten lucky in my generations.
video file editor or a music producing software cut it and loop it
Fast-paced, Rebellious, Operatic Style, Chinese Folk, Somber, Reflective, Heroic, Lute, Erhu, Zither, Dizi, Cymbals. song. Listen and make your own with Suno.
Which dataset was used to train Chirp?
I’ve been creating more songs in my Learning Suno Journey.
Although, it creates some fantastic pieces, there are lots of frustrating enteritis within.
I’ve been working with a simple song. The first problem: There was no pause before it started playing the song. I therefore added [Intro], but it didn’t work.
Second: It doesn’t seem that the tags, e.g. [Verse] [Outro] etc, work all the time, as you’d expect. I would have thought that during the training, these would have been set and confidently usable.
Next, I did find leaving off the full stops and commas at the end of my lyrical lines strange. This seems counterintuitive, but this is only a tiny issue.
Next, As with the [ ] inconsistency, there is a more significant problem with the ( ) options.
I’ve tried many options to find the ones that work and those that don’t. Here are some examples:
(Rhythmic Pause)
(Increase in Tempo)
(Bright, Upbeat)
(Build-up, Crescendo)
(Slow Fade)
(Stress on “future")
There are a few more strange, unfamiliar ones here, but this is something I often note when writing a song. Is there a list of consistent Tags, both for [ ] and ( )?
Well, that’s my input for the day; now I’m off to play…. Have a Great Weekend!
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 have found a most interesting bug...
not sure if it is common... or.. how even it could have occurred.. but, my video has subtitles on subtitle track one , id'd as Corsican, but is properly in ukrainian, as the music is.
and it is displaying on the video when i enable it, and it is NOT my music, but some babble about leprachauns and wishing badluck to someone name Sam Greenfield
neo-bachzarthoven indie post-anti-instrumental, anti-folk song. Listen and make your own with Suno.
How do i make the insturments last longer before the lyrics start?
I would make an instrumental, then extend from the point where you want vocals to begin maybe
oh thats possible?
amazing thank you so much
is there a way i can make it longer than 3 minutes?
its hard to blend the extended parts
I mean, just download them and chunk them into audacity, that's what I do. usually.
seperate it into parts
for example, above actually began with this:
https://suno.com/song/f4ae7b60-cfe2-42c7-82f4-d6e65c015b4b
neo-bachzarthoven indie post-anti-instrumental, anti-folk song. Listen and make your own with Suno.
coool
then I did a continuation off it with this at like 33 seconds I believe
i see
midwesinger/songwrier indie lofi-whispered vocals, anti-folk, intimate close voice, whispered in ears, sad rich living song. Listen and make your own with Suno.
@untold briar any ideas ? that is just weird, and odd language, since it is actually ukrainian. i'm extracting the srt and going to see what it says in full in a moment.
I don't have any guess. If you find anything useful, can also post the information the bugs channel?
i tried to extend
and i added the lyrics
and generated it
but the lyrics are not added just instruments
what's the style tag?
additionally, create a space
before lyrics begin
like, at the top of the page
ok
yeah, I was asking in here really cos, I do not know why it would even have subtitles added, when it has the lyrics placed in the video itself, seems odd to even have that enbaled, furthermore, it is NOTHING at all from this song or its lyrics, in any relation
remove any reference in the style tag to heavily instrumental genres
if you continue to have trouble
wondered if there was some backend use for subtitling as a stream embed that was used for some reason
i removed it
and I'll see what I can do
Blend of Ambient and Cinematic music, incorporating Folk elements to connect with the natural, raw elements of the game song. Listen and make your own with Suno.
here is what i tried to extend it
😄
part 2
that's just one gen, just do more until you get what you want
with that
ya ya
For prompts, best I can recommend is to look through popular songs
and see how they prompts
I do some unconventional stuff
btw, I like this one
the ambience is nice
yeah i tried to generate like 5-10
in this case btw, my thought process was that this genre doesn't have many vocals, so it might be struggling to understand where to put them in
but adding operatic and orchestral
kind of help it orient
and understand how it should go about it
Chirp bot on discord is going to retire.
oh
Please use the webapp
k
@untold briar I figured it out, so. this is a vlc issue it seems. Apparently if the filename is in ukrainian, it causes some problem, and loads "*" as the srt file, I did indeed have an SRT file, i just didn't realize it, from when I had downloaded captions for an english movie, in ukraininan ... a long time ago.
So this error occurs with ukrainian character filename, followed by . before .mp4 so I changed the name of this song to Медик..mp4 and Час би посміхатися..mp4 was named with the ..mp4 because the title had a period in it on download it named it as such. This error is in fileparsing for name matching, since it is 0 width, it matches the available character based language (ukrainian in these track titles titles) srt file, probably a misvariabled title name, that ..mp4 breaks, while the language charset is not Медик.<>.mp4 but Медик>..mp4 while the srt title search is terminated by the previous by the . in both directions, so the srt title search finds '' which i guess is defaulting to a greedy search. probably some regex goofery with 0-or more * instead of + so it is likely matching any title that contains .srt no matter the title, and is likely due to something with the ukrainain language breaking that character termination... not really sure, but if I add the same ..mp4 to english title mp4, it does not happen. or if I change the ukrainian titled ..mp4 to 1..mp4 for example, it stops doing this.
And for some strange reason, the sub stream is listed as corsican lol.
anyhow, okay. so I figured that weirdness out, now to go make a bug report with vlc
neo-bachzarthoven indie post-anti-instrumental, anti-folk song. Listen and make your own with Suno.
out of credit, but I have something fun to start for the next one
here's the song that lead to this weirdness, since the title had a . in it. lol
Does anyone know how to use metatag [Continue] to fix my problem?
Plus, can I find a tech staff from Suno on this server?
is there a way i can insert normal text? like text to speech not singing it
i could help
I appreciate it, but no worries, I'll save it for later
Hey there. Is there a way to get the lyrics shown under the song correct? Mine are rather messed up in order and even showing text that is not sung. Example here: https://suno.com/song/b489e395-4e86-4c07-87c1-207194c811cb
A sad but also frustrated song with a deep male voice, quite piano, guitar and some drums. song. Listen and make your own with Suno.
Take this example:
[Verse 1]
Down this long forgotten road where echoes fade to gray
The signs they blur and bend like memories that stray
Each mile marker a story lost or found
In the rearview, shadows of my life spin round
Often the singing seems to blur from one line to another and doesn't sound as I wish it would. Is there any way that I can change, add or edit the text to influence how the words are sung?
I want to change parts of a song I made on Suno, but keep how it sounds at the end. Is there anyway to splice clips into the middle, or will it only allow you to extend forward?
longer text blocks like that tend to become 'wall of lyrics' without a lot of variation.
I try to break up big text blocks with punctuation and line breaks, and try to make sure there are internal rhymes.
Unfortunately AI lyrics tend to cheat their rhymes by only rhyming 'throw away' idioms and prepositional phrases... makes the lyrics kinda weak because the emphasis is on unimportant parts of the sentences....
only forward...
v4 may have the ability to 'in-paint' which sounds like the feature you are talking about.
@broken lotus Oh awesome, I hope it comes soon.
Many thanks! I guessed it might be you to reply!
'Wall of lyrics' is a good description.
I read somewhere that commas and full stops at the end of each line don't help, so I left them out, although this does feel unintuitive. So, does punctuation work within the text?
I'll try your suggestions out, so thanks again for you time!
Are you staff from Suno?
I...
try...
to break up the lines....
like this, sometimes...
I am a subscriber, and a mod for their discord
(not a Suno employee)
powerful drum and bass song. Listen and make your own with Suno.
Do you know any tech staff of Suno from here? Username?
yes. but they prefer not to be tagged.
Can I answer any questions for you?
This one. I elaborated before. About lyrics trimming and insertion.
have never heard [continue] do anything.
here are some tips how to use [metatags]
https://www.suno.wiki/faq/metatags/
Anyway, in v3 custom mode, can I do insertion or merely lyrics trimming?
for example?
what is this chirp about ?
So has anyone else noticed the crazy good adlibbing that's been happing on the v2 model this week?
I would post an example but the site's having issues
Thank you man
Interesting... Been trying all sorts of things. Anything in ( ) it sings. Tried [ ], some work, and others don't. I used the same text every time and the same style to try and keep some consistency.
It's interesting how Suno musically interprets the lyrics, almost regardless of anything else you add.
I don't think, at this moment in time, there is much you can add, alter or edit to make a difference in how the song is sung.
I think it would be good for us and Suno if they came up with their own unique tags to impact the phrases sung: Slow, loud, quiet, fast... I'm sure you get the picture. Make them unique to Suno and keep ahead of the game.
I'll try your idea... and see what happens 🤞
Many thanks and enjoy the rest of your weekend!
Do you know of anywhere else that is a good place to talk, learn, and share ideas about Suno?
I don't like Discord. I find it strange and disjointed.
But I think that is more due to the fact that I'm most likely one of the oldest ones here! I remember black-and-white TV and real shops on the high street. 😏
I don't sadly..., and I hear you about how chaotic it can be here (and other social media too)
I miss a nice forum I can navigate...
Found this map of music genres: https://regorxxx.github.io/Music-Graph/Draw Graph.html
Might be helpful in prompting
can some tell my why in chanel chirp-beta the command /chrip doesnt work for me?
it says application did not respond
See this announcement: #📢┃announcements message
suno, next version should be in stereo like udio!... i tought suno was great but trying udio... i realize the sound quality with suno is still very bad.. eventho its really food in composing but you cannot use it as professionnal material in a movie or things like that, with udio, the sound quality th e soundscaping is great youc an tell that each sounds have been panned, equilized and the final master sound great no mather the type of vocals or sounds you have.... there is no electronic feedback that keep destroying the song. for many style of music i had greatest result musically with suno but quality wise and final product udio is superior, also with suno i often need to create 12-15 render so around 30 clips to find something that sound good, also i did not tought i would prefer the 30 seconds clip from udio, but finally i realized that often with suno, there is 30-40 seconds of a part i like and the 20-30s econds after i dont like.. so i have to regenerate multiple times... the first 2 minutes also is too long i feel, should limits suno at 1 minute per part, again often you get the first verse very well but you dont like the chorus.. so you loos the first part you liked.
i think both tools will become the best but now suno know where to upgrade
i also like the fact that udio give the opportunity to select and create a part that goes before
example you really want an instrumental intro, but with suno even when you write intro.. often there is none, so by adding the opportunity to had an instrumental part before its great
Has anyone successfully trimmed both the melody an lyrics after generation in custom mode?
Suno's been in stereo since last November, v2.
@broken lotus yes you get a stereo output, but its the same on left and right... you dont hear distinctive soundscaping
what udio do is a proper mastering, sometimes a sound will be heared in the headphone in the upper left or things like that this is very important to avoid phase cancellation
thats why when you have an epic chorus with a lot of sounds, the vocals are always muffed or the vocals become very disturbing to listen to because its breaks you ears... when you ave proper sound scaping and a real stereo output you can avoid those things
Does anyone know if there is a working meta tag that will "turn off" doubling, tripling, quadrupling, kazillining, vocals? It's ruining the vocals and pushing me to another AI music app faster than a cheetah chasing a gazelle. About half of the songs I generate are unusable for me since I'm not interested in hearing songs sung by a choir in unison...
try simplifying the style prompt.
try NOT mentioning the singer or vocals.
Let me try not mentioning the singer. I usually don't, but sometimes I mention the vocalist's gender to make sure I get that as intended. Be back in a minute or so...
Nope, that didn't work. I got the Boston Pops choir singing a nu metal song.
Can't confirm, but it looks like adding the words "no doubling" in the style prompt is working. Doesn't seem like it should, but we'll see. Might just be a coincidence.
My latest style prompt rabbit hole is
hyper binaural insert normal stuff here singer-songwriter ultra asmr
all without comma
where singer-songwriter is optional I got a metal track whith whispery vocals out of that.. and other interesting stuff. thing is the tags have a lot of stuff they can bring in.. ultra asmr almost exclusively nets female vocals hyper binaural can give you natural ambience, birds, really loud noise or stereo effects (no true binaural stuff yet).
I am not done testing yet.. but i already wasted 1k credits because it's interesting how it behaves with genres and instruments... there are probably ways to achieve the same results without this cursed style prompt but I guess it's just really interesting for me 😅
And the style prompt looks funny
suno can output generations with different left and right channel. But yes udios seems wider and it's more deliberate probably difference in training data.
Very much a difference in training data. Their Eurobeat skillset is heavily lacking vs Suno but Suno sounds more cloudy
\ chirp
chirp is retired . Please use the website suno.com instead https://suno.com/
how can i listen to the complete song when i extended one?
on the last clip (with the ending) click on the [...] button and choose Get Whole Song (free, no tokins)
\ chirp
chirp is retired . Please use the website suno.com instead https://suno.com/
are extensions all the time only 1 minute?
yes max 1 minute
do u have an idea why my extension are midlyrics cut at 22 sec?
Greetings, is there was way to trim the ending of a full song once stitched together?
download it and crop it in audacity , use fade to silence over about a 2-4 second period up to where you want to trim it works perfectly
thank you for that suggesting... was waiting lol
Does anyone know?
Do you guys know how to get a good beat switch in a song? When I try it keeps the same beat.
Can you give an exxample of beat switch?
does your music your lyrics or instruments?
@grave pasture
For example going from orchestral music to rap music in terms of the instruments
There isn't a meta tag for that in my search but you can create a transition by describing the shift in mood, tempo, instrumentation, or other relevant factors within your prompt.
I get a lot of useful tid bits from chatting back and forth with gpt.
Hence how I found out about meta tags. Oh, another useful bit of information, use hash tags.
Switching that dramatically is generally hard to do, there some tags that may make it more likely, and it's more likely after the song has a quiet moment where it sounds like it might change. But the easiest way is to use a Extend From time that is very close to the start of a clip, with a new music style. I have some longer Discord messages explain in detail, I think I used the term "bridge clips" but just a popping in quick for now.
So something like Ex. 90's boy band ballad #intro
then transition into 90's rock n roll #middleeight
then conlude guitar solo #outro
This is a bit rough but you have to do it in pieces.
Then stitch together.
I see your point but don't make it work to hard. Try and have it be as clear and concise within the prompt and promise it will work out for you.
Ok thank you very much guys. I will try these the next time I use Suno. 🙂
I'm attempting now seems fun. Try to keep the Genres close enough where the instruments don't clash to much. Like Blues and Jazz.
Will do
good luck to you
Does anyone know how to successfully trim both the melody and lyrics after generation in custom mode?
Post production @pure bay
What?
Someone recommended audacity I just started using audition.
I've tried everything that I could think of, but found nothing that works.
There isn't a trim feature within Suno, atleast not yet.
When creating the 2nd song you can " trim " the previous song by changing the time but that's it.
Audacity and Adobe audition are 3rd party tools that allow you to really fine tune or " Master " to track. I'm not sure how this would work when using tools like fruity loops etc for instrumentals.
I hope this stirs you in the right direction.
But I need to also trim the lyrics in the lyric panel of Suno webpage, while it still displays the trimmed part.
as in your running out of characters or you can't see everything you wrote?
What do you mean?
I'm trying to follow you but I'm not sure what you mean.
Original:
A
B
I trimmed at the time when A ends, to expect:
A
But in the whole song that I got, the lyric panel still displays:
A
B
Now you understand?
I just get "Interaction failed"
Does anyone know why songs dont generate? I wrote something but when i went to click create its dark and wont let me
Did you run out of credits?
maybe, where do i find those?