#📚┃suno-school
1 messages · Page 56 of 1
Was it a DNS attack
Great!!!! 🙂
probably someone mad they cant create music as good as suno!
its back
Oh yes. it's up. I can listen to my weird EDM crap.
choral Celtic, Choral angelic prayer wall, song. Listen and make your own with Suno.
@wet minnow that was fast, its back up
same panic people had about ,, cable,, beta, 8 tracks, cassets, dvds,, blue ray... barf
I take great pride in this tune
Totally a kubernetes cluster thing huh. An update, change, etc. That's my world.
linux is the way
That's my life.
same
up again for me it seems
So suno people, if you are using RKE it'll be EOL in a year. Switch to RKE2
No
"Your having illusions"
"It was never back up 🥱"
k8s is a different world. How you do everything including DR and backups are different.
Is suno down for you guys too?
Sorry for the delay! You technically don't need to credit Suno if it's made on a paid account, though they mention that it's appreciated.
Hi tech chat 
thank you!
Oh crap -- was it actually down 👀
But… the real life dlc is bugged
I went outside the other day -- the graphics are RIDICULOUSLY good.
Too bad the other players aren’t so nice sometimes
"...did that woman just call me a 'griefer'?" - said by a rude person, who clearly was griefing
I gotta remember to give context when I quote imaginary strangers.
True
There just standing there like NPC's
I swear I saw one go flying
I think I fell asleep to this last night
Being hit by a car
From the technical side. You guys kick so much ass. I'm a play-just-to-relax sort of musician. I can upload the kids greatest hits and make new stuff from it on Suno. In their voice or style. OMG that's so freaking cool and fun, the kids love it. Just want to give a huge shout out to the people that make it possible. This is really fun!!
All those folks with the purple names -- if you have an iPhone, I'm sure they'd appreciate having a good rating on their app 😄 It's a small gesture, but it's an easy way to contribute to a metric that they're paying attention to and likely to make them happy in a small way ❤️
Aw man
So I called 911 before it blew up
Did you grab your pickaxe though?
Yep
If I want to help Suno and I'm super good at k8s and mass deployments of nodes, how do I do that? What is the best group to be in?
Don't want a job, have the best job in the world. Just want to help in my free time.
I don't think that's something the team are looking for
damn i got excited that was a v4 announcement 💀
Is there where I drop a technical issue regarding SUNO? I'm trying to access it on my phone on Google Chrome and it worked fine for the last few months even today until just before 11 am or so Pacific time I kept getting 400 bad request.
my brain overloaded i cant handle that much bandwith
Here is the right place.
But it was down today off and on.
I'll check later to confirm
Is there a way to filter your library by uploads? If not, that would be a helpful addition
I wish. or a search function to search hidden songs that are not public.
But if you remember the name of the uploads, you can search that way
You can search for your own songs in the library search bar, but if someone doesn't make their songs public, they don't want them to be searchable.
Unless I misunderstand?
oh my god I'm glad I checked the .wav version of a track I have for my album over the .mp3 version because the .wav version is all messed up I gotta redownload it or something lol
i created a small local application for myself to auto-queue songs so i can have constant generated music while i work. does this violate suno TOS? im using their song creation endpoints and my auth token for generating
okay this is really weird
the base song that suno generated is fine but if you try to extend it and add a ending that shortens the song to remove a long instrumental at the end it makes the song all autotuned and stuff and I dunno why
Using Suno in an API-like way like this does would be against ToS.
ah ok, i meant to say i was "thinking" about creating one. hopefully they add a similar feature in near future
I see it is "public" on your GitHub.. So I guess you are going to take that down now?
There's a known suggestion for this 🙂 https://discord.com/channels/1069381916492562582/1253766525458841610 -- if you want to throw an upvote onto the post linked here, it'll let the devs know that people are still interested in an API! 💪
i don't really want an api. just the ability to continuously create songs in the background
@radiant tangle thank you for removing 💗 🌼
I understand, but the way that would be made available would be through an API -- otherwise it's just the "playground-style" usages of the website and app at the moment.
that moment when you generate a new version of a song you made with Suno and it makes a better version XD
what i mean is natively supported in the Suno application, rather than having to create something yourself to support the capability
or when you upload a recording of you playing an instrument and it flexes on you
pffft
I've done that with a vocal recording of me humming and it just goes 'how about we make this 10 times more epic'
thanks Suno
#💡┃suggestions-web might be a place for cool ideas like that, because I'm not sure if I've heard that particular use-case suggested before -- I know the 10-concurrent generations is a limit, but it sounds like you're just wanting a single generation one after another. That's a cool idea, like your own ad hoc generated Suno radio
ah, right, exactly
Encouraging the release of the API would allow others to make a similar idea, not that you would need to make it yourself, but that's something that the tech-chat here would have a field day with, so we could make the cooler ideas the Suno-mods
I'm glad Suno decided to ignore my extension instructions and just do 'clicking' or silence for 5 seconds and 3 seconds respectfully, thank you for refunding my credits for that but... what the heck?
Out of curiosity, was it an extension from silence / the end of the track?
yeah but I thought the end of the track wasn't silent
okay I fixed it had to go back a fair bit
I have a few songs hidden from the public, and i would like a feature to beable to find them is all. ;p
That's what this is for! It finds all of your songs, except for those in the trash.
yes i can type in "Hidden" ?
This is for finding other people's songs.
and it will show all hidden songs?
No, you need to type any word, metatag, or style prompt -- but it needs to be the exact word
"song name, style, or lyrics"
By "hidden", do you mean songs that you haven't set as "public"?
It will search for those as well, but you are not able to "filter" your search results.
You do not need to know the name of it, but if you search for any word included in the lyrics or the styles, it will bring them up.
"rock" will bring up all rock songs, public or private
oh i know that... ;\
but still makes it difficult
need a function, "Show non public songs"
.<
are non-public songs visible inside a public playlist
if you dont have it set to public and i goto your profile, i will not be able to see it
afk
Example let's say I have 2500 songs and 200 - 500 are not public. In my own library I'd like to click "show non public songs" and see them all.
Of that makes more sense >.<
If**
I understand what you're suggesting -- filtering options for the library have been suggested many, many times in #💡┃suggestions-web and #🚧┃suggestions-layout ❤️
I'm just letting you know how to use the current search system we have
Oh ya, I know that, lol I've had to find lyrics cuz I've forgotten the names of songs
Glad we are on the same page! 😄
Hey if you've had to delete your album on distrokid before it releases and you've already selected one of the one time payment bonus things, when do they actually take the payment or will I need to pay twice for it since I've had to cancel the album because of one track lol
No idea. But ouch!
Its the loudness normalification one
O
Yeah I'm hoping the answer will be no I won't have to xd
And only pay it once becauzw aaaa
hmm, i opened a public playlist in an incognito tab and was able to view a non-public song i added to it
"public" vs "private" only means whether or not it's searchable on the social-media side of Suno -- the songs can still be seen in playlists, they can still be shared by link
That's the only difference
ah, thanks for clarification
Oh hey harber, can I dm you a song I quickly made that used one of my own original songs?
Always, dude! 💜
I'm watching BattleBots with @delicate idol at the moment, but I'll throw it on once the Robot Fighting Time ends
No worries, there’s no rush, just wanted to share something
We chatted about it in DM, but of course I need to let others know your song rocks 🙂
It’s been awhile did we add defining a key signature
nope
are your referring to like a note or metric or like tempo?
i use tonic / and octave to control my overall sound
tunning works also
i duuno about if any one have had sim
I can't recall seeing an instance of being able to control key with prompting -- but you can control a LOT with audio upload, @leaden lark -- do I have permission to share the song you shared with me?
Or if you want to share it, either or 🙂
She uploaded 16 seconds (presumably made in FL Studio?) and then used style prompts to turn the initial clip into the full-fledged song -- this way she was able to control both the motif/melody, the tempo, and the key -- a bit more effort required than what's possible through prompting alone (from what we've observed), but incredibly powerful.
idlike to see this
Sure
EDM, energetic, atmospheric, glitch song. Listen and make your own with Suno.
If you watch the video embedded in discord, you can see when the videos change, indicating a new extension
First 16 secs is my own audio
you can regen the video to omit your tags after you have cleened the lyrics i belive
But the song above is instrumental so I don’t really care if it’s seen
yuh thinking if you were to put on showcase . very nice
And yes I made this in FL Studio
It's a simple and elegant demonstration of audio upload -- thanks for letting me share 💜
No problem! It’s been sitting for too long so I just let Suno complete it for me
my GarageBand library staring at me like...
I have like over 30 flps just sitting there
LONG BET : Disney Pixar releases a movie in the next 5 years about unfinished samples that are waiting to be turned into full songs, only to turn into a crazy "Sausage Party"-esque movie about how "being picked" means you get turned into something unrecognizeable, and the moral of the story is learning to love yourself or something.
I’m at a standstill at work rn, I called maintenance to come fix an issue twice and still haven’t come and it’s holding back everything else
@charred merlin Yah I’ve got some good results with just my iPhone voice notes guitar and vocals and it does stay in Key so very legit approach
Is this the kind of situation where you get paid to do nothing, or is this a situation where you can't get paid until they do something?
First part, but someone’s fixing it now
Nice! Admittedly, you might need to edit the song externally in a DAW to remove the initial audio upload, but as for right now it's the only way to ensure some "guarantee" of granular control.
Drats -- of course I would never suggest sabotaging something so you can get paid to Suno for longer...
I only have 25 credits though
Though I shouldn’t sabotage the work, I’m a postal employee so….
If only we could figure out how to pan on Suno the music could be more dynamic and little less slot machine.
@untold briar and I tried some spatial audio experiments that I posted as a playlist over in the pool party > #1256301881966198805
It doesn't really seem like something we're able to do in Suno through prompting -- however, if you separate the stems using https://www.bandlab.com/splitter or something similar, you can pan the audio yourself in an external daw https://www.youtube.com/watch?v=2lyUGKj7J58
Whew works slowing down alittle now
I made a few tests today and all I got out of it was center channel madness.
Yeah, I haven't seen it being promptable, it seems more a roll-of-the-dice for when you've picked a genre that would have included spatial audio anyway -- my indie rock songs often have guitar playing in one channel for a section before following up in stereo.
Hi. I don't know which chat thread to post this in. Since it's something "technical", I decided to leave it here. I think this question has been asked before, and it's probably somewhere in the FAQ, but I couldn't find it. My question is simple: is it possible in the free subscription to limit to one generation per request (instead of two)?
Not even possible in the paid plan, every creation costs 10 credits and gives 2 variations (this likely is so they can act as "control samples" for one another, so you have less chance of getting a "weird" result) -- though there's a known-suggestion for letting users pick how many songs are generated. https://discord.com/channels/1069381916492562582/1254418510885687366 -- the post has currently expired so you can't add your upvote, but at least know that the devs are aware of the issue. 🙏
Thank you very much for the detailed answer
❤️ 🙇♂️
Not a problem at all 🙂 Just noticed that you're still a greenleaf! 😮 🫂
Welcome to the server, and feel free to swing by anytime & ask anything on your mind 💪
If you have questions, just ask us!
And should I also send questions about improving generation here?) However, I think they are painfully stupid x)
Frankly, the previous question was actually my 1st message here, although I joined the server quite a long time ago. I just muted it and put it in my far pocket. Now I'm crawling around here, looking at things to improve my generations x)
hi so some one sent me a list of meta tags they were just the tags i have added descriptions to them. i dont know if all work but iv seen most of them used in suno. have looksy
How is there not a "delete all songs" option on my account? I want a fresh start. But it only lets me select one generation at a time to delete. A process that could take hours.
Click on a song, then CTRL + A to select all songs on the page, and you can send an entire page to the trash at once.
Still painful to go through a ton of pages, but WAY easier than song-by-song ❤️
We're really glad to have you here 🙂
Even the people who found Suno on that meteorite that crash to Earth are still learning all of the cool things it can do (and sometime surprise-drop new features 👀), so none of us are experts, and we're all learning as we go -- especially as each new model basically acts like it's own new "thing" (#📚┃suno-school message)
I'm going to give a very biased suggestion to stop by the pool party and check out some of the experiments that myself and other Sunovators have contributed to -- not everyone learns the same way, but I really wish I could have looked over ideas like these, tested in isolation, so I could better understand how some of the metatags / style prompts I was using interacted. > #1256301881966198805
[pace]: The speed at which someone speaks or reads.
And how, for example, to use this? It looks as if it is necessary to somehow designate the value of speed. Both "slowing down" and "acceleration" come right away. And there is also the degree of these effects...
It allows me to select the entire page of songs. But how do I delete after I have them selected? If I click the trash bin icon, it just takes me to the trash folder. If I click "delete" nothing happens. If I click on 3 dots, it unselects the entire page.
I'll record a quick video 🙏 Are you on mobile or desktop?
...heck, I should post these in a workflow guide or something.
Desktop. I'm using Chrome. 🙂
You're a saint.
I'll definitely check out this thread. It looks very interesting. I'm not sure how well these or those solutions will work for me though - I generate songs in languages other than English, which I feel could introduce some distortions into the generation.
like i said i dont know if all them work. try your self i was give a list of tags all i did was add description
suno cant change pace or tempo on demand
unless you upload your own beat and mix
Just making it pretty -- I'll ping you when it's posted 💪
Amazing! Take your time. haha
Obviously we speak of desktop only: can hold CNTL and click on each as well -- in one page. Then you have go to the next page and repeat. Faster? Hold SHIFT, click first, scroll down and click the last.
Until this point, I solved this problem by crudely stitching together pieces I liked from different generations (through extensions, of course, otherwise the voice changes catastrophically)
After the second extension, do you "get whole song", or extend from the extension?
Stitching? Offsite you mean? If they have stitching on site I wanna know about it.
does suno keep a model of all your work in it head, can you reset this model if true
i dunno know how to explain it. i know with conversations in a gpt i can help it learn my structures and styles i have imputed in to it
I usually generate the first generation until I find the right vibe. Then I fix minor flaws using extend. Having received good pieces, I stitch them together using the most stupid and linear software I found for free (I think it's audiocity)
"stitching" I do outside the site. But from pieces received by the button "extend"
Right. Audacity.
I would say , no. It's process is to learn immediately from everyone's input and it's own outcome to improve and output what is "likeable" but I don't think you can steer it into a style as of yet. We can get songs that are close and often by arranging the length of lines and style of writing which "forces" suno to relate music to fit that style and , ta!da!, you get back something that sounds like you intended. But that's about it I think.
makes more sense when its put that way
Sometimes it turns out to be so bad at controlling the prompts that at the end of my generations (after a short pause) I received a piece of generation of a completely different song (rhythm, style, lyrics)
When I put in my own lyrics, provided it's not close to a standard, 4 line, 8 syllable per line, stanza, (or rap) it spits out (60 % of the time) a song that is 80% or more close to how I originally produced it. Create more gens and I finally get one that's like my original.
So ive been taking down my originals in soundcloud and replacing them. Lol.
Not on demand, per se. But ive gotten new tempos or , say, it switches from pop, to swing, by clicking edit lyrics, REMOVING all lyrics and style, click extend, put in "extended" lyrics -- or just click instrumental for a long interlude our outo-- put in your time code , click create. Of course you should copy lyrics before deleting so you can paste back in.
No, Suno doesn't learn from you -- the only way you can "teach it" is by giving thumbs up and thumbs down, or ...> Report > Bug > ... -- this doesn't give immediate results, but helps train the next models like v4.
Ahhh, I see, I have to hold Ctrl down while I click on the 3 dots to delete. That works! thanks a million! Appreciated 🙂
Until the "2nd verse coda of death" is fixed, that's what I'm doing now to keep it from repeating things, but the side effect is that I can add new prompts and it gives a different, if only subtle, flavor to he extention.
Hrm, I don't seem to need to on my browser, but the important part is that you're now able to organize your library up to a potential 12x faster 💪
Cool. Glad it works for you.
will reporting bugs give you credits back?
No, you only get credits back for generations under 8 seconds -- the benefit is that you help try to resolve issues in the next model so it helps everyone.
i mean some time the song is fo bar and i report it
"volunteer service".
ahh make sense
What do you call a bug. Hallucinations , laughter, giggles, spoken words, babble-scat and mumble-rap are not considered a bug.
It is like voting, "if you don' vote, you can't complain" -- if you don't like the audio quality, that is the most effective way to "help Suno learn" what you don't like.
i mean its literally swearing at me, and i didnt put in there
I agree that it might not be a "bug" per se, but I also had the same questions when it came to reporting -- especially when I was trying to balance 2500 credits each month. Shades of grey for different degrees of technical savvy, perhaps 🙏 💙
My most frequent problems (apart from Suno sometimes incorrectly putting stress on words in my language (this is easily fixed by setting the stress symbol correctly in the text)) are:
- The intro often blends in with the music or even sounds like part of the music. I can count on the fingers of one hand all the generations when the intro sounded like an intro.
A recent example: the intro "ladies and gentlemen, I present to you our new guest" generated in my language sounded like someone was mocking an electric guitar, and not like words. - Completely randomly, in the middle of the chorus, Suno can start singing the 3rd or 4th line of the previous verse. This is wildly upsetting at times, but, to be fair, quite often after this there is a repetition of the chorus (not included in the text), where everything is fine.
I haven't gotten swearing yet. Not even "darn" or shucks, but it does seem to babble the word really close to "funk" alot.
I'll have to send you something, then 🤣
My guilty admission is that I'll never report those hallucinations as inappropriate, because they make me laugh harder than ANYTHING.
Never.
lol
I'm sure it exists. I must have click the NSFW pre-set in my profile. Lol.
I've got a 4 minute mumble-rap song that suno did that I listen to constantly and am learning the mumble words. But just using fuggin' at the appropriate times. I'm tempted to try to phonetically write down the illegible "words".
Sorry I’m just seeing this, but feel free to ask anything!
I had the perfect song and the the AI dropped the F bomb in the last 20 seconds. I tried to reroll an ending, but Suno kept giving me 2 minute endings. I gave up, flagged it, and moved on.
Oh my god, my wife @delicate idol and I did that, there's a mumble rap that yells a couple legible swear words out of nowhere, then goes back into rapping -- we both wrote down what we thought we heard, then "rapped" eachothers lyrics to eachother, and I think we both almost peed ourselves laughing
I have apparently a hmmmm... Japanese babble song with a Hindi-babble ( apparently a featured guest artist) interlude. Coolest thing. It's hard to tell. But since I use bollywood and jpop in my prompts, that's the way I'm leaning as a guess.
My most recent song wouldn’t end so i tried using V2 and it surprisingly worked to end the song
Well heck. I legitimately never think to use v2.
I tried using V3 but kept giving me 2 min extensions so I tried v2 since I was low on credits and it worked
My personal suggestion is to start your songs as instrumentals, and then add the lyrics as an extension afterward -- this way the musical intro helps guide the style of the vocals, and can help getting less "samey" voices. Also, this way you can choose EXACTLY when the lyrics start by when you choose to place your first extension.
I’ll have to try that.
Suno is trained, as I would understand to emulate what we humans are listening to as popular songs. If swear words are prominent in post-2000 songs, then it will be digitized and spit back out. Obviously though, I think there are some parameters in not even having babble and hallucinations say anything racist, for instance. A good thing. Like, we don't need to yell fire in a theater just cause we can.
Though maybe I’ll try to start a song in 3.5, extend using v3 and try to end with 3.5 and see if that works
- Are you trying to create the entire song at once -- all verses, choruses, etc all as a one-shot generation?
The more you ask Suno to "juggle" at one time, the more likely it is to mix up what goes where, or start to hallucinate. If you find that it can't handle a certain amount of words /tokens, give it less to handle at once, or go back before the song starts to mess up and place another extension point to continue from & overwrite the mistakes.
There's a thought. Sounds like an interesting test.
I'm afraid my knowledge of English + the power of Google Translate is not enough to explain the problem of introductions properly. I could give links to specific generations for understanding, but I'll get to my personal PC in 8-9 hours =-=
I’ve tried to replicate a song where I got a duet for a pre chorus (only for pre chorus) so far no luck
I should do more of that. I'm waiting for front-end trimming. No that sounds wrong. Ummm.... trimming the beginning of songs. Once that's in place, one can just extend from the end and trim off the original 3 minutes.
Interesting solution. I'll definitely try it. But an interesting point: when I generate a really LOT of text - such hallucinations do not happen (a lot is when it does not fit into 4 minutes even at the speed of rappers)
And when the text is short - this is a more frequent occurrence. For me it is usually 3-4 verses of 4 lines + x2 pre-chorus and chorus also 3-4 lines.
I'm heading to bed before too long, but I'm happy to answer questions in DM if you'd like ❤️ Feel free to message any time, and I will reply when I am able 💪
If you give Suno something to do in the Intro it will then pronounce the first verse without extending the intro into gibberish.
~~https://www.manscaped.com/~~ Sorry, I misunderstood -- yeah, I hope to see something like that in-browser as well ❤️ It would be nice to edit songs without having to take them out of the Suno-ecosystem.
https://suno.com/song/0ab36074-83de-4de0-8bb9-71bcdb78df7e so im try to replicate what I got around 1:18 with the duet, any tags to possible try to coax more of it?
[[Fusion]], alternative, rock, pop, fast paced, hard rock, [female vocals], futuristic song. Listen and make your own with Suno.
Right. I think it's got some leaning towards fitting a song somewhere between 3 and 4 minutes as most songs on radio are presently defined as songs by that time. I'm just spit-balling. But yea, then with few lyrics, it tried to keep going and fill in the gaps until somewhere after 3 minutes.
to me it literally sounds like ```[female voice] [intro]
Ladies and gentlemen!
I present to you our new guests!
[verse 1]
lyrics```
upd: with the difference that the text is in my language
This is interesting 😄 It is remarkable how much even the choice of lyrics by themselves can influence the song (https://suno.com/song/65a2e520-e54a-4def-be67-d483a08a49eb -- no style input, only the lyrics "I'm so angry right now" makes a powerful rock song that sounds "angry") -- we are all learning, and what works well for one person might not work well for another.
I believe we are all building out own "English-to-Suno" translation dictionaries in our heads, and the translations are a little different for everyone depending on what they're trying to "say to Suno".
A song. Listen and make your own with Suno.
I chased harmony last month and could never figure out how to get it to trigger.
Hold on. I'll type something. BRB.
@viral moss described Suno as "a vibe interpreter", and I like that description. 🙂
I'll try something.
I got extremely lucky for even just that pre chorus
Have you tried the [Spoken] tag?
I found out about its existence about an hour ago, in this chat, so no. I'll try it tonight.
Is there a tag for generating music on a single instrument? In terms of, say, only violin music. Without drums and other stuff. I tried several times, but I couldn't find anything worthwhile.
https://discord.com/channels/1069381916492562582/1253341197917884530 This is sort of a tricksy-topic.
The post above is a known-suggestion that you can upvote to let the devs know another person is shaking their first to the sky and demanding "negative prompts" -- because Suno doesn't understand negatives, when you say no drums, it hears drum. drum me. drum me now.
Man, negative prompts, in painting, that would improve my stuff so much
Hmm... Well. Let's hope and wait... In my language there is a beautiful sad song about a musician with a violin, but unfortunately there is no version of this song with the violin in the background. Only a piano. It's somehow strange for my taste x)
https://suno.com/song/c04e54d2-93e4-45a1-a535-82441e3baf71. so here is suno creating it's own lyrics to get a Duet. Others try the same thing with less luck as I understand. But just to show it works on the first try sometimes.
https://suno.wiki has a pdf that has hundreds of tags. There is this other wiki https://sunoaiwiki.com that has tips on tags.
Suno AI Wiki (Suno Wiki) provides beginner tutorials and tool resources, gathering advanced usage tips and prompt examples.
The workaround is if you can think of any musical phrases, thematic descriptors, etc, that would suggest specific musical styles -- for instance, you can't directly prompt "3/4 time", but if you prompt "waltz" correctly, you might get it. You likely can't say "vocals only", but "a cappella" works really well (check out the a cappella playlist in the pool party #1256301881966198805 )
Some of the style prompting that @naive quail has used makes me wonder if "violin-audition" might work, especially if it weren't overpowered by the other choices of style prompts (some are definitely more strongly weighted than others)
Interesting, was there any specific tags or styles used?
a million words of gratitude!
Look at the lyrics. Yes. [Verse:male]...... [chorus:both] , etc.
Nice. Will have to try that one out too.
Alright, I’ve tried stuff like that before and had no success
Yup. Your mileage may vary, even if it works for me the first try from suno , it might not work the next time.
Might also vary based on the genre you’re trying to do
I originally instructed: "alternating male and female voices about a boy and girl chatting over dinner". So the word "alternating" might have been a force majeur.
I didn't type any genre. It chose pop.
So if your doing your own lyrics. Throw in " alternating male and female vocals" in the prompt. Who knows.
Instead of "duet" or just "male vocals, female vocals"
Then metagtag in the lyrics as discussed
literally yesterday I generated a variant with a duet for two voices (male/female) In the verse and choruses they sang 2 lines in turns.
I generated an intro of 2 lines, then the text began. Before the intro there was "male voice", and before the text - "female voice"
The unintentional duet really surprised me x)
If anyone wants to try this style prompt out, it gives an incredibly fast and energetic beat https://suno.com/song/1e095941-cd2f-4b80-abc1-b70aff459aa9
電脳和風EDM: 高速和太鼓+篠笛シンセ+三味線ベース. 琴グリッチ効果, 尺八ドロップ. 歌舞伎スクリーム+演歌ラップ. ネオン浮世絵サウンド, 侍テクノ, 花魁トランス. Dubstep, Trap song. Listen and make your own with Suno.
It’s insane
@north sleet -- I don't think we've been formally e-ntroduced? I know my memory is absolutely terrible at the moment, but I can't recall having bumped into you in chat, though I've mostly kept my head stuck into tech-chat.
What sort of music do you like to make?
Yes. Definitely got a fast one on my use.
This is how I was able to find headphones at work X)
on the first link (before 00:23) - an example of a good intro
the second link (before 00:11) is one of the best examples of how it was generated for me all day yesterday. At least I understand the words here x)
https://suno.com/song/5b5d3076-f8a3-4d0f-9a2c-fa9da80634df
https://suno.com/song/883ba4e4-ec02-4a8e-b691-b6fc91d6c052
rock, electro guitar, metal, hard rock song. Listen and make your own with Suno.
rock, electro guitar, metal, hard rock, female voice, cyberpunk, sadness song. Listen and make your own with Suno.
Welcome. I
I think we have crossed paths in #💬┃general-chat
I make music for every type of style. 😉 though I gravitate towards electronic styles even though I have done a couple of waltzes and other genres.
I try to make instrumentals but then I end up writing lyrics. Which means I have to build verse by verse to get the target BPM or Suno tries to cram all the lyrics in the 4 minutes by singing too fast.
Profile: https://suno.com/@caxi
Hello Suno team want to adress a situation with the video generated with lyrics, it takes forever to generate and when complete the lyrics do not follow 100% the music and vocals sometimes stay still without change then it corrects and jump to current time, also want better outcome just the lyrics without the rest of the txt in the song like brifge, chorus verse etc
as it turns out, it's useful to browse thematic discord ^^
I browsed a couple of topics and talked a little. I corrected the tags and after 6 generations I got an (almost) perfect version ^^
Nice! Finishing it up now?
Personally, I don't like that in one single place of the chorus Suno distorts the stress (although in all previous versions there were no problems with this place), but it's not such a significant problem. "for myself" - sodyet anyway. And publishing it is useless anyway: the audience will always be limited to my "non-English" language, even if many people speak it x)
UPD: P.S.: and besides, it's not really an original text. More like a cover-interpretation-translation x)
Hi I am new to Suno, and I was wondering if there is anyway we could remove Suno mark on video downloads and whether i could save the voice that I just got! I love the voice
If anyone know...
?
If you are still interested in the final version, I have uploaded it to the channel #🐣┃suno-showcase
upd: #🐣┃suno-showcase message
you can't remove the suno watermark on the videos and you can download the vocal stems and use that as the audio base to get something similar
Hi, I just wonder that could we change another voice of “singer” ? Cuz I feel like they just have some voice for specific styles of music and it make boring evens how cool the music u create 🫤
Suno!!! Noooo!
"voice detection" helps me. something like [husky female voice]. Voice +/- matches this definition, but is noticeably different in each generation. Although I can't guarantee that this will help you...
It seems getting stems don't work properly. It doesn't seperate distincly the vocal and instruments. May be the combining algorithm must be defined at the beginning
isnt there a community summer of suno or something like that in the discord? cant find it
do you mean #🌴┃creator-programs?
Thank you! I will give that a go! 🙂
Since your creating it in udio, I think you would be fine, but not 100%
I'm just being cautious tbh
As long as you have the rights from Udio, it should be fine. But I haven't specifically looked at Udio ToS
Hey does anybody know when the credits reset for pro plan? is it a fixed date or tied to your supscription date?
I supscribed at the 8th of july, does that mean my credit resets tomorrow, or should be resetted already?
Yeah, thanks, do u have experience with change the vocalist?
Is "clear-voice" a good prompt for getting a... erm... clear voice, or is there a better/recommended prompt for this?
I had the same problem, so i started to just use "crisp" and "clear"
removing the voice part
like for example "synth-driven experimental, male-singer, clear, crisp"
hope it helps
Legend, thanks for the advice!
Anytime!
@charred merlin I attempted to recreate your bpm test with the addition of rythm tags. I do not completely agree that Suno does not respond to these tags, but I do concur with you that we cannot control it by hard steering. Maybe the random generator fools us at times, but I believe like just in image generation, it is all about the weight of the prompt. Words occuring with equal frequency, words that are associated with certain properties are weighted heavier and some associations are yet to be discovered, like that someone found with vowels and consonants affecting the style of music. My experience is that bpm and rythm does very little in few worded prompts. However the more detailed you are, the higher the chance of the BPM and rythm tags to hit their marks. Adding generated lyrics made it more precise with the rest of the tags, I believe. This goes against the prompt weight theory however.
I first tried just "country, rythm, tempo" as instrumental. Very little effect when changing. When prompting "metal, rythm, tempo" the changes were more noticeable. When adding a bracket to the tempo, I believe it weighted heavier as I got a song that was not metal at all, but had the tempo.
Perhaps I will continue the trials the next time I have credits over before renewal.
https://suno.com/playlist/180082d0-421f-4ffe-a484-54cfeb56bbdf
You need to use email to reach Suno billing billing@suno.com
This is phenomenal work!
Admittedly the BPM playlist took a bit longer to curate because I wanted to test the BPM of each song, then added the results to each title, so people could look at the playlist and take away the gist of whether or not the experiments worked, without requiring them to listen to each track critically.
I’d love to feature this playlist in the pool party if you’re okay with me sharing it there?
Absolutely, let me know how. I am very new to Discord.
No effort required on your part, my friend!
If I were to message you with a list of suspected BPMs for each of the tracks, would it be too much of a pain to ask you to edit the names of the songs in the playlist to reflect their resulting BPM?
(I realize how nitpicking this is, I’m just ecstatic to take up to this kind of community science, and I’m excited to be a part of it somehow 🙂 )
Sure, I can give it a try.
You rock, man 🙂
I’ll DM you when I get up and about 💪
@charred merlin One thing that has started to edge my brain is how the training of the AI works in real time. I do some pretty subsubgenre stuff where I am very exact with my lyric themes and like the songs that hit what I want. Some tags and formats are as such very specific to my music. It hit me when doing this trial playlist that some of the songs actually featured segments that are very unique to my subsubgenre. I am wondering if my serious attempts at making songs has actually trained the AI to respond with those traits in tags that are not meant for those properties.
As example, one of the tempo/rythm combinations is one I heavily use on an album that features metal with fiddle interludes. To my surprise one of the test songs had exactly such a fiddle interlude without any tags related to it.
Not me
But same problem
Sounds like you need to contact your card issuer/bank?
If you're attempting to make a purchase, and they're blocking the payment without your request, you're a customer of theirs and need to resolve the matter with them.
Also as it mentions in that thread, please make sure that you're attempting to make purchases specifically for suno.com and not one of the scam sites pretending to be Suno.
Again, email billing@suno.com for an answer without hearsay, because as the product grows to millions of users -- statistically there's bound to be at least one person experiencing ANY issue or random quirk that you're experiencing at the same time, and it doesn't necessarily indicate a larger trend, or suggest that the established advice should be discarded.
Quick Q. When writing prompt, using a word such as “Cannons” I keep getting the error “user name detected” will there be a patch for such issues? I must admit it makes writing these things a bit tricky when other people use common words for their user names
@true warren Re: your suggestion for more space in the style field, I initially felt that same way, but no longer do. We have all of the Lyrics field to further refine the style(s) by using tags and descriptors within the tags. We can even take a song in a different direction just by using tags. Suno's more likely to disregard some of the tags if you ask for too radical of a change in the style, but sometimes... https://suno.com/song/674e2ba9-f2a4-4dc9-b3ec-71bcd75742e8
Folk, Blues, Flamenco Guitar song. Listen and make your own with Suno.
Can you please send a screenshot of that error message? I would laugh so hard if the devs were cooking up ways to kneecap the "diss track" nonsense.
Gotchu fam. It’s not just this word specifically. It’s many others in every language
I do find it ironic that one of the front songs is “Suno Please Fix”
Oooh, I'm familiar with Artist name, I thought they added a new one for User Name to get me to stop making disstracks about @naive quail
NEVAR!!!!
It’s a bit annoying.
Automatically halting song production because some chap decided to use a decent word as an artist name
Any chance "cannon" works? Have you tried experimenting with different ways of describing the sound you're looking for? "Artillery-fire"?
Yes but it doesn’t flow the same way.
You're putting "cannons" in the style prompts of custom mode, right?
Since "cannons" isn't a musically descriptive term, and would act more like a thematic decorator, can you explain the breaking change that occurs? I would imagine it not fitting if you were trying to say "cannons" in your lyrics, but I'm not sure I'm grasping what's happening here.
hi Is there a possibility to keep the background music same while revising the words. When I try to revise the word after it is done, I keep getting a different background music even though I upload it back from. my computer as instrumental only
I too would like to keep the same beat whilst changing words. Sometimes you just get “The one” version that hits just right but some of the wording just doesn’t fit
Hi, I've a question, not sure it's still possible up to day, but when you upload a song containing a "verse" and a "chorus" melodic section I want to use the both as ref for the IA, can I specifically tell to the ia something like [verse from the first melody section (with timing)] and [chorus from the second melody section (with timing)] ?
Or, upload an extract for the chorus, another one for the verse, and merge the both in one extend (shorter, multi-uploads/sources extension we redirect for the ia in each section, would be great). I can make it separately, but in that way, I lose the link/context/continuity in between, it would be better the AI makes it itself directly in one whole song.
The problem is today, when you upload a song with a verse and a chorus (or two distinct melodic sections), the AI uses most of the time only one melodic signature (and it makes it very nicely most of the time), more rarely the both (it happened one time for me only). And the reason why I want this feature is because this is my own piano composition, so I try Suno to use my melodies for chorus and verse. I'm pretty sure we're close to this feature in the future with this awesome new feature, it's like a game changer for me. It's the only real limit I find now in Suno to make everything with our own melodies.
The only way to get a whole song is to extend it first? a 3-minute song, which is great, no need to extend it, to get it in full, do I need to extend it?
outstanding work btw.
@foggy wing just used you tags and incorporated them with mine. interesting effects.
inpainting is one of the features that seems to be getting teased a bit for v4 so fingers crossed we might get that ability soon 😁
I understand, thanks bro
The closest thing to that is what you suggested, editing the audio yourself to use as an Upload. If the last part of the Upload is unchanged, the song should still be continous and you could merge the audio back into one manually. I've suggested a feature like this, doesn't have a specific name, sometimes I called it 'virtual song history'.
has anyone had success in subtly controlling the song tempo by mimicing the word/line/meter often associated with songs of like type? for example patterning /formatting lyric structure as metal but the prompt is ballad or lullaby?
I've been trying a couple experiments (as I was diving into music theory and how some patterns like that can drive Suno to make music that follows patterns that way) when I have spare credits. Sometimes Suno just broke and gave me flat instrumental with no lyric, but i did get some screamy slow paced lullabies, and some lullaby metallic... LOL. i wish had way more credits to experiment more. As well as more musical theory to build from.
Thank you. Yes, with some editings and tweakings, but I guess we'll arrive to this feature since its still in beta. I give just an example of command here. I'm testing it for experimenting with the IA hehe, but i guess it doesn't really know what is part 1 (the audio upload) and timing today.
If it's just 1 part, that is the whole song -- the "full song" stamp doesn't mean that it's special, it just means that it's the result of multiple parts stitched together so people don't try to pass it off as a one-shot generation.
In a way, NOT having the "full song" stamp is a different mark of pride, because it means that you were able to get that song without having to put "extra touchup work" into it. 💪
would be nice in version 4 to have a version number next to gens with th esame title instead of typeing one every time
im not sure if this puts me in the minority, and obviously everyone uses suno in different ways, but unless im in a really experimental mood, i usually try and get the full song in the single generation 😅 i wanna just put my descriptors in, lyrics and see what it cooks up
sry too budge in but i tend to use the shot gun method when im experimenting. see what sticks what dosnt . lol
for sure , thats what i love about it tbh! everyone's creating their own music in their own ways, they have methods and ideas that work for them and its all a big experiment ✨
True! I found nice ways for my chill atmospheric pop
100% 😄
I don't mean to say that there's anything wrong with one-shot generations, it's entirely different styles of cooking. Some of my favourite generations have been one-shots, but the ones that I put more time into definitely resonate with me stronger or for longer. 💙
I believe ChatGPT calls it "Virtual Memory" or something similar
This is a one-shot gen that I was just testing how Suno interprets ascii https://suno.com/song/855bed2c-ab78-4bda-92fb-9c62261720c4
[Fusion], Dubstep, Fast, Energetic, Intense, Grunge, Dark song. Listen and make your own with Suno.
Still have no idea how the ai interprets it
For my part, PayPal is holding my balance for identity verification! 20 days of waiting, they say! That's all ^^ 😎 So, i'am a thief!? Behaps!
In regards to paypal, do you mean a Summer of Suno payout?
Yeah yeah that's it! But since I haven't linked my credit card with PayPal, they're not happy, happy, so we'll wait!
https://suno.com/song/197cffc1-da6e-4fb6-afe7-362b62f0c84b I still have no idea what it was trying to sing here.
stage-performance, myspace, tennis, live-recording, sound-fx, sound-effects, oral-noises, improvised song. Listen and make your own with Suno.
I suppose that makes sense -- without being able to verify who Suno is sending the money to, PayPal needs to make sure that everything is "above board"
He doesn't sing, he farts the pig!!!😂
Yes, i suppose too, but i don't care about this!
Hello, what's the best process to have a foreign version of a song I created ?
My ears at the beginning lmao
Can you detail the process ?
sure, can you give me sample of what you trying to covert
use you favorite chatgpt, or copilot say convert the lyrics to this language wile keeping my stucture, : copy paste lyrics here
sure, can you give me sample of what you trying to covert or use you favorite chatgpt, or copilot say convert the lyrics to this language wile keeping my stucture, : copy paste lyrics here
mind my spelling
and here it is if you say to keep note structure
What about the song ?
what are you trying to acomplish ? can you be more clear . thanks
I want song A in language B.
sure then copy you song into a chat gpt and say what language to convert it while keeping structure then copy past back into suno
what language do you speak other then english?
Right so you joined yesterday.
i rejoined the server yesterday yes, but iv been around the block a lot longer
@charred merlin hey with your spotify artist pfp how did you end up doing that?
distrokid , iv done the same
I know that but what part of distrokid lets you do that
I can't find it lol
or is it post album release etc
when my spotify artist profile goes live with them I intend to put a professional portrait picture of me as that profile's picture, but I wasn't 100% sure if I was able to do that and if I was how I'd do that on Distrokid's end
I'm a first time artist lol
np
For me, I use this as my online alias since I don’t want my real life appearance known to keep my privacy
Idk if wesogo is still here, but......
One extra request from chat gpt is, if needed, instruct it to keep the intent and meaning but translate and write lyrics with no more that 11 syllables, preferably 8.
Per line
ahh intresting
When I translate from English to Japanese, at least, that's needed, else it's like a 15 second gatling gun of Japanese words just to say "I need you".
It will re-interpret and write the lines meaning the same, but less "words".
French, Korean, whatever the language, the same can be done.
If translate from Language to English, as your native language, then you have the option to see what AI can do or just rewrite yourself to better fit the meter.
So.... of WeSoGo ever comes back that is...... otherwise, fugetaboutit. Lol
*if
Upload audio IS dead ?
What error message are you receiving?
Thank you. We all want to issue AI commands, but it rather listens to our feeeeelings. Anyhow, still fun (and sometimes not) to figure out what makes it tick. Let me know what you find out by using the tags.
If I add notes, for example "
[Chorus]
[slow piano] "
do I still add style of music or no?
You don't need to, but it's where you can choose the overall stylistic theme, or you can let Suno decide for you based on the lyrics
Yes, but it is still a part of the random generation matrix. Some generations will have it, some will not. It heavily depends on the amount of syllables in your lyrics. I have found some success by adding "[interlude] (slow piano expressing longing)" or just [piano interlude]. The latter will usually have greater success rate, but lower ratio of getting the feeling you want.
Thanks peeps <3
https://suno.com/song/65a2e520-e54a-4def-be67-d483a08a49eb here's a song that only uses "I'm so angry right now" for lyrics, with no style prompts.
A song. Listen and make your own with Suno.
I strongly recommend you guys to write your lyrics in haiku format. Suno loves it.
damn and it slaps hard
Ill try with and without will see the outcome
😦 wtf, why did it start saying that just for my account
GROS this shit slapssss, best decision to not include style of music
am I alowed to share it here? or do I need to put it in #🐣┃suno-showcase and then copy the hyperlink to here?
Is there a way to attach Extended tracks together, or is it assumed the user will do this on some other software?
Yes, ellipsis and then option "get the whole song"
This will stitch all the parts together into one track
ok any one know about free ai voice cloner and dubbing?
my anniversary is this month and wife and i have been talking again for a bit i wrote a song i want to dub it with my voice. (i cant carry a tune ) lol
anyone know a prompt or genre that will make similar music as this ? mainly the drumline style https://www.youtube.com/watch?v=WHeOTKfbf2o
BUY BEAT HERE💰(Instant Delivery): https://bsta.rs/a5d981b74
BUY BEAT HERE💰(Instant Delivery): https://bsta.rs/a5d981b74
BUY 2 BEATS GET 1 FREE!!! (Add 3 beats into your cart to activate)
Use Code "SOBOL" for 10% off
|-------------|
THIS BEAT IS FREE FOR NON PROFIT ONLY
Any use of this beat REQUIRES CREDIT (prod. by Sobol)
CONTACT: s...
im wondering about if u can use the same voice ai mad but just fix the text more is it possible?
Is there a way to regenerate a song that I want to maintain only the chorus part? I really liked the chorus of a song that I generated but I would like to remake the intro and verse part, is there a way to "save" the chorus?
You can extend something as a means of "adding" after, but so far adding before is not yet there. But if you extend from 00:01 or as far in as possible, it has kept the chorus at times.
Extending after with new lines, then download and just trim off the beginning and save is also possible.
Remake the into too? Is it your own audio?
*intro
I mean, I would like to change the intro and verse part of this song, but without changing the chorus part
how do I force the AI to use my dang lyrics?
its hallucinating lyrics
its completely off script
so found a realy cool i mean really cool free ai song and voice dubber that runs off your cpu or gpu from mit
no where a good as suno , but the voice over dubbing is free
Thanks for the help! You made this soulful blasphemous bluesy thing possible and I LOVE IT
Southern Gospel, mysterious blues, majestic song. Listen and make your own with Suno.
I see that someone answered your question about the profile pictures, but feel free to shoot any questions about the DistroKid process ❤️
I'm testing out releasing some tracks through LANDR as well to familiarize myself with their processes as well 💪
Apologies for the delay -- literal catnap 😻
oh huh
neat lol
I'm still gonna have to figure out the profile picture thing though but that's later on when the stuff gets put onto spotify lol
By "distrokid process", I also meant the fiddly stuff getting familiarized with the "Apple Music for Artists", "Spotify for Artists", "YT Studio" side of things 🙏
I just wonder that anyone tried to use kits ai for changing the vocalist or clone the voices for the song yet?
👀 ...it's not fair just how many cool services are popping up these days.
Admittedly I'm only familiar with ElevenLabs just from their reputation as an industry-leader, but do you have any experience with kits.ai? I've been away for a bit today so I'm just "starting the rounds", but I'm going to look into this shortly -- just some spitball questions to help give ideas if you have anything to share ;
- is there a free plan?
- if yes, what limitations?
- how have you used it so far?
- any before/after examples?
Actually, I just tried yesterday and the free one limited like they said:
“Kits AI offers a free plan that allows users to explore its features for music creation, such as AI voice models, vocal blending, and mastering tools. However, the free plan has limitations, including access to only one vocal slot and no ability to download audio files.”
So I tried basic premium for clone voice and blending to the song, but still feel some complicated with me first time, that’s why I asking if anyone tried to clone their voice for the song yet, will try again later.
I also thinking about elevenlab, but it’s just clone and not blending for the song right?
Admittedly there's no "clean" way to handle this request at the moment.
If you have a strong enough musical context to "Extend" from, v3.5 really excels at continuing song structure -- if you have enough music before the lyrics start, v3.5 can "understand" how the rest of the music should likely sound, and while it's going to be hard to guarantee a vocal melody that matches the other language's melody, this is most reliable way to handle the issue using just Suno-generated content, without considering the control that using Audio Upload can give you 💪
@mystic flume, I'm just realizing that we chatted about your attempts at making English / Japanese versions of Celtic Moonshine's anime-style album, but I don't think we actually got into the nitty-gritty of how you handled it, or what sort of challenges you experienced?
When uploading songs to publishers, do you use your legal name? You can still use the artist name separately, but I think that's just a matter of making sure that no one can steal your music.
I did before, but after those couple songs, I used my alias name
I've found that if you ask to change the language, suno will also pull in the cultural traits of that languages singing styles. So take your English song, add prompt to sing in Japanese and you get more lilting falsetto voicing throughout, of course in Japanese. Or if preferred, singing English with a Japanese accent and with the accompanying cultural accoutrements mentioned.
Dangit, this reply made me smile. It got me right in my feeeeeelings. 💪
Well the artist might be Brad Paisley, but the writer is Bradley Douglas Paisley. Use your legal name.
Though, all of Elton John's songs say the writers Bernie Taupin and Elton John ( not Reginald Dwight). But to be safe. Use your legal name.
I think the worry there is that someone else could establish a digital presence with the same name and attempt to claim ownership of it, but it's harder for someone to change their name legally to match yours and try to claim that they're the same {$hoshino.firstName} {$hoshino.lastName} as the one who directed those compositions.
I also do have the songs in yt content id, etc
Just listened to a song written by " David Foster and The Tubes" I guess even a band can be listed as the writer. Share the credit, I suppose.
I didn't know that, but I feel like Elton John did that as a power move -- "Come on, try me. Someone else claim to be THE Elton John. Dethrone me."
Now if I can get suno to change keys and modulate several times in the song like The Tubes, Squeeze or a number of Broadway tunes, then we'd have something.
See how I brought it back 'round to stay on subject?
And that’s how I will know if someone’s uploading my songs without permission, and I can see how many claims there are
check this out https://www.weights.gg/
Create with our AI tools for free. Generate AI voice covers, text-to-speech, and more. Join our community of creators sharing RVC and AI voice models.
we should get refunds for bugs
heavy Metal, Anthemic, Dark, horror, Electric Guitar, 80s, meatloaf song. Listen and make your own with Suno.
'spreading fear "eternlesslsemight"' is not acceptable
Uh is there any advanced prompting stuff that can be used in suno?
https://suno.com/song/9dd0e79d-d6cd-4749-90e8-472eacc3f381
I don't know how to fix this lyrical issue
heavy Metal, Anthemic, Dark, horror, Electric Guitar, 80s, meatloaf song. Listen and make your own with Suno.
Repeating Verse 2, Pre-chorus, and Chorus is a bit annoying. The song runs out because of the repeat.
Try rearranging your structure to avoid that, based on what result you're getting.
My railroad analogy: #📚┃suno-school message
Can you tell us about it, or elaborate on why you’re suggesting it?
Before I click on an unfamiliar link, I like knowing what I’m getting myself into 🙂
i was looking for a way to clone my voice and well replay does that but weights.gg is collection of community voices that you can ues with replay to dub your music or basically any song
you also have a default avatar, which is super sketch
like your existance isn't even worth a pfp
it cost me 1.50 $ to clone my voice so i can gen on my pc instead of useing web sites and paying an arm anad a leg to too do it
if the site is decent, you could in theory use stems to change the voice
weights.gg is a open ai model that allows the community to create pretty much anything. If you search HuggingFace you can see different models.
well i wrote a song for my wife and i wanted it to be my voice but budget is reall friggin tight till monday
thank you
I was answering @charred merlin concern of what weights.gg is.
yes i know
i was appreciating it because it was the answe she was loooking for
shhesh
?
hard bunch here today lol
😅 I didn't mean to imply you didn't have your bases covered 💜
default avatars are typically burner accounts, its just general internet safety when someone shares a url
lol what ? are you smoking
im a community manager for 12+ years, this is general consenus of fact
im also aussie, it takes more than that to offend me, all g
whats a community manager?
I run and manage Google Firebase on discord
oh well this is suno.. usally if i want people to take advantage i wouldnt use my suno account name
a name is a name, you could be a compromised account, hijacked account, or simply a burner with a similar users name.
these all happen
It’s fine! 
eeek easy there
Sorry for the delay! If you're sharing a song for the purpose of soliciting technical feedback, or ideas on how to approach a problem, I'd say it's 100% within context to share -- if it's something that you're simply proud of and want to share, #🐣┃suno-showcase or #💬┃general-chat if probably a better fit just to keep the channel "on-topic".
That being said, I'm not "the host", just an active user sharing what ideas make sense to me 🙏
Stratuss, I'd appreciate it if you could keep conversation polite -- they're voicing reasonable concerns.
sry i will i do aplogize
if I delete a track after reporting it as a bug, does that ruin the bug report?
All's good, my dude 💙 it's a party -- there's bound to be people bumping into each other and misunderstandings since everyone is partying for a different reason -- some are here to dance
, some are here to meet people 🤝 , some just noticed there was a party going on 👀 -- I'm happy that you're hanging out with us here ❤️
Not that we've been informed, but that's a really good question that I'm going to urge one of the mods or guides to follow up on? It might be a bit until we hear back, but that's something I'd be curious about as well... @unique stone, are you able to fly it up the flagpole at some point? 🙏
adobe audition is there a free brother other then audacity
if you can tolerate stuttering and lag, NCH Software is capable, I rather prefer it over audacity. its just clunky to use
Record and edit music, vocals, sound and other audio with WavePad Audio Editor.
I recall using it back in the day, it was perfect. but yeah, I can't say its as stable now, but then again, I also get other types of lag in audacity
maybe my intel cpu is just dead
i ll check thanks
In free accounts, most likely, except they're not audible to humans -- just to computers, like the copyright detection systems used by streaming services.
In order to legally claim any sort of ownership over a piece of content owned by someone else, you need to provide a "transformative act", and can't simply share their content as your own with minimal changes made. That's plagiarism.
If the only change you're making is removing a digital watermark that makes zero audible difference on the resulting track, that's not going to hold up well in court. If you're making that change specifcally for the purpose of circumventing digital rights management, you're basically paying to pay for the plaintiff's lawyer.
I know it seems hyperbolic to talk about court stuff, but it's RIDICULOUSLY hard for us to actually comprehend how fast all of this is moving -- sharing stuff often leads to some form of agreement that you made, where you swore to have complete ownership of the content and agreed to be held liable if you falsified information. We're having fun in the eff-around stage, but it's a genuine disservice to "future-you" to pretend that the "find-out" stage isn't waiting around the corner with an excited grin.
man suno has opened a can o worms to music and apps and ai, and and and , chedder cheese
my god, im sick of fighting the lyrics
You got it 💪
Pro & Paid accounts are 100% yours -- I believe I've heard something about a generalized watermark to indicate whether it was made with AI, kinda similar to how pictures have metadata, but that's way outside of my knowledge range.
Reviving souls where shadows are!
becomes
Reviving souls where endless might!
Yeah, you can upload stuff that you own, like if it's from a pro or paid account -- if someone uploads a song from a free account, it's only a matter of time before it's detected as "stolen" and the account gets bricked.
By transformative, I mean how people can remix classic songs and get them on the radio, or how "react videos" technically aren't simply just showing the videos, but they're "providing transformative commentary"
air quotes
It's so rewarding going back to the songs you used to work on, and improve them ❤️
night all
Yo @quick agate, I just noticed that you're a greenleaf -- welcome to the server! 
While sometimes Suno will hallucinate lyrics just because it wants to keep us on our toes, I find that it's more prone to these hallucinations when given a large amount of "tokens" to handle at once, such as lots of style prompts, lyrics, [metatags]. etc.
While it's one thing to simply suggest placing an extension point before the song hallucinates lyrics and continuing from there, I might suggest trying to only "feed" Suno half as much lyrical context, and then extend the best result with the second half of the lyrics afterward?
i've been here for like a month
on second month of premium
as for the lyrics, im not sure what you mean
im keeping it super simple
but its using older lyrics from previous extentions, etc
seems more of a bug than a feature
I didn't mean you were "new" per se, but users have a "green leaf" beside their name until they've made a certain number of posts -- I think you're still ~<50
Also, I didn't think we'd spoken before, so I was introducing myself. 👋
its even using lyrics from rejected extentions, seems like a memory leak or something
hi 👋
Can you share a song for an example?
By "super simple", do you mean not many words, or just "plain lyrics"?
Is this your only account?
just plain lyrics, im not some pro audio-phil, frankly I don't care for the minute details
here is one example
https://platform.openai.com/tokenizer This is what I mean by "tokens" -- it's not a matter of how complex it is, but how much you're handing to Suno at once.
After a certain amount of "tokens", it can start to hallucinate, just the same as any AI, like chatbots
The closer you are to filling out the character limit, the more likely it will be to hallucinate.
not jibberish, I understand some concepts with ML and AI
but I did some lyric tests playing with "dark apocalypse" for example, and I went back to my original, ignoring the extension
but it then used dark apoc in the new versions
from the total song?
im only looking at 380 words to finish the song
Okay, that doesn't seem like a crazy amount of lyrics, but how frequently is it mispronouncing the lyrics?
99% of the time as of today
I found one version where it stayed on track and I went from there
I'm making 4 copies, just to double check which lyrics are being mistaken almost every time?
Reviving souls where shadows are!
becomes
Reviving souls where endless might!
that was another, yeah. its been a struggle as its mixing some chorus together
legit surprised that "meatloaf" is allowed as a style prompt, btw 👀
ikr?
it used to also block queen but its allowed? not sure
im looking for that grand power rock/metal style
but im to dumb to know what else to put
It's not that they're "allowed" now it's that Suno is trying to be less restrictive so you don't end up with a block on common words...
Better to describe the specific thematic elements you associate with that music -- doesn't need to be "musically accurate", it can even be random "vibe words"
im too 'artistic' for that, lmao
No one wants to write a song about the Queen of Sheba only to find out that the word Queen is blocked.
#1256301881966198805 message either way, using artists names doesn't really seem to work anyway. 🤷♂️
im coming at this as a consumer of music, not a creator. i don't have the mental capacity for music
Yeah, you don't need to know specific "musically-accurate terminology" -- as @viral moss points out, Suno is a vibe-interpreter.
The challenge isn't finding the right artist's name to simply "ape their style", you need to think about ANY way to describe the music, even if it doesn't seem to apply musically.
so the issue with suno is if you have a long lyrics and you want to cut and extend from the middle, any previous lyrics are also preserved in the memory heap.
so I can't later replace a typo or a word without it using the old one over the new
Like Lagwagon is "skateboard punk" in my mind, and that will likely get me a closer result than typing a specific genre most times I've tried.
heavy Metal, Anthemic, Dark, horror, Electric Guitar, 80s, meatloaf song. Listen and make your own with Suno.
check this one out, I realized I had a double lyric 'grows' and 'grow'
I fixed that with flow and grow
but it refuses to use the new lyrics
[shout - Heavy Metal, Electric Guitar]
With every breath, his power flows,
With purpose and courage, the risen grow.
In his hold, we belong,
In this anthem, forever strong!
[pause]```
I also changed necromancer to his
this is all part of the outro, it never showed in the previous clips
so its 100% a lyric mem dumb issue
maybe the lyrics input aren't cut off when extending?
https://suno.com/song/3fe6a22e-b103-427b-9d93-daec366ac8c4
https://suno.com/song/36ea3a5f-5db9-4aa3-bbdf-2b537a12d750
https://suno.com/song/1c81312c-35f7-4664-8d69-3214665ad413
https://suno.com/song/9dfa8118-d092-4e00-a8e3-79b05de27a58
So, it got that line right 4 times out of 4. My 100% success rate doesn't seem to match your 1% success rate -- either one of us is misunderstanding something drastically, or further research needs to be taken into the matter.
are these extended or just fresh?
Fresh
then there wont be a lyrical issue
its sampling lyrics from the previous song, even if they were cut before the extention
🙏 Please provide the song I should be extending from, which timestamp, and any additional details I’d need to further test this. 🧑🔬 🔬
and i've found it use lyrics from another extention, when its not related at all
this is the current one I have:
https://suno.com/song/df8d5717-d090-496d-8504-93a8b26195c8
extend at 00:52 or 00:53
heavy Metal, Anthemic, Dark, horror, Electric Guitar, 80s, meatloaf song. Listen and make your own with Suno.
I actually lost the other segment but its certainly a problem i've encountered
found it
heavy Metal, Anthemic, Dark, horror, Electric Guitar, 80s, meatloaf song. Listen and make your own with Suno.
I wasn't happy with the outro part so I wanted to cut and extend from it
so extending before it, it wants to preseve the 'grows' part
Okay, extending from 0:52 -- specifically what should I enter as lyrics?
[chorus - Strong Male]
Across the lands, his legions roam,
In moonlit night, their power grown.
Spreading darkness, wide and far,
Reviving souls where shadows are!
[Pre-Chorus - Inspirational]
With ancient words and power vast,
He raises armies from the past.
In shadows deep, his strength we see,
Spreading fear eternally!
the other segment is what I finalized on, once it played right
Man -- v3.5 can really get stuck into a groove sometimes, eh?
Sometimes it decides that a song should go on for another 5 minutes despite it being presented with the perfect chance to end a song
it really does, its not bad at making music
Other times you've given it everything that it needs, and it decides that it only wants to spit out a single minute of greatness then call it an evening.
Whoa -- I just noticed that the track you linked is "part 4" -- do you extend from "get whole song", or just extend from extensions?
That makes a HUGE difference 👀
Like one of my ascii songs that 3.5 decided will only be 1:47
lmao, it might have just gotten tired
https://suno.com/song/855bed2c-ab78-4bda-92fb-9c62261720c4 this one, I might try to extend once I get more credits
[Fusion], Dubstep, Fast, Energetic, Intense, Grunge, Dark song. Listen and make your own with Suno.
Assuming I can figure out where to extend from
I don't use get whole song because it kinda sucks, but its something I was about to test if it would bake the lyrics or not. the problem is it keeps previous lyrics that were cutoff with extentions
Can I dm a song? It’s one of the songs I remade recently
no change, its starting off with the pre-chorus instead of my outro
heavy Metal, Anthemic, Dark, horror, Electric Guitar, 80s, meatloaf song. Listen and make your own with Suno.
lyrics set to:
[Outro - Heavy Metal, Electric Guitar]
With every breath, his power flows,
With purpose and courage, the risen grow.
In the necromancer’s hold, we belong,
In this anthem, forever strong!
Always. 🤘
I got 2/2 results with correct lyrics -- how many separate generations have you made where they're sung the wrong lyrics?
about 15-20
burnt 800 creds
20 would only be 200 credits?
not sure but the monthly rolled over and im down to 1300 left from 2500~
so maybe more, lol
got some grief from suno
I had one extention, where it just repeated 'no more, no more'
was weird
Usually I'd suggest ... > Report > Bugs > Doesn't Follow Lyrics, but if it's happening this frequently, and it's not reproducible on other people's end, it might be best to take all of the bad generations, compile them into one list, and send an email to support@suno.com, asking for guidance on why it's failing at that particular spot so frequently, or so early into the song when you're not providing "a large token input".
i've reported a few
its why I suggested confirmed bugs, should get refunds or extra credit. been at this for 4 hours. lmao
should have just learned vocaloid
Since you've already used 4 separate parts to create 52 seconds of material;
1- How long are the generations usually?
2- When in the songs do they usually start "hallucinating"?
most of them die after 1 minute
they hallucinate within the first chorus/verse typically, or after the first minute
It's not a bug unless it's reproducible -- the first step is being able to reproduce it on your end, but it's not reproducible on my end at this point.
sometimes they use the wrong verse, but it sounds ok, so I roll with it
To be clear, I'm not throwing my hands up and saying "there's no issue buddy", because clearly there's something funky going on, right? but in order to figure it out, we need to "get a foothold" on something that we understand
I wrote to you on this issue ❤️ 😺
could be fixed with a infill/re-lyric feature honestly
Basically an inpainting feature
ok, its not using the correct verse here but WOW, that vocal range
https://suno.com/song/99cc09d2-2324-493c-b3a8-54d6dca625fd
heavy Metal, Anthemic, Dark, horror, Electric Guitar, 80s, meatloaf song. Listen and make your own with Suno.
❤️ Yo! Sorry I've been all scrambled today, haven't had a chance to comb through DMs, but I'm gonna put yours at the top of the list 🙂 Be with you in just a few
NP! ❤️ I guess you're in high demand and just decided to raise the flag where you're most likely to see it 🙂
Can I dm a song of mine that has a crazy range as well?
you're in high demand
😅 still incredibly low demand, just got distracted helping @delicate idol plan her raise-request tomorrow 💪 YOU GOT THIS, BEE
why dm? sharing is caring.
Rock, female vocals, uplifting, ballad, slow, fusion song. Listen and make your own with Suno.
Like an icy-phoenix, this song returns from the trash can thanks to @hollow hare 💙
I can only hope to get a voice like this again for a later song
interesting song, I was referring to a bug with my lyrics where it lost all style and went 150% free range
is this a working tag?
[Suno.ai: dolby atmos]
I think it helps a bit
the impossibility of having something like a "voice seed" is just killing me ТТ_ТТ I have a need to make a series of songs about one long story. In theory, it would be good if one voice sang... At the same time, I have two songs that are perfect for the same character (from Nastokli) but there are two different voices ТТ_ТТ
sob-sob
other music gens in the past have had them, its certainly possible but would take an overhaul to preload that feature
played with one about a year ago that allowed you to upload a custom voice and it would auto tune their vocals to match
since we have stems, you could in theory do similar
theoretically, the "black box" of our AI has a principle of choosing a voice for one generation or another. That is, somewhere inside it there is this very "seed". Perhaps in some implicit form...
https://discord.com/channels/1069381916492562582/1252789683176345641 If you give the post in this link an upvote, it'll show the devs that this feature is still highly requested! 🙏
is it normal that the ai read what was between [ ] ? kinda funny was wondering was it was saying at the begining of the song and realise it was just reading the thing I wrote between these [ ]
I also caught such a mistake. I put "speed up" in the hope of getting a sharp increase in the tempo of the composition. Imagine my surprise when in the middle of the song Suno took it and read it X) In the end, I left this mistake and took it as the final version, but it was very unexpected x)
I mean personally it didnt ruin the song, cuz it was at the begining but still made giggle when I understood what it said
drop me the song if you dont mind I'm curious
It happens more frequently if you include more than 2-3 words as metatags -- something like [backflipping drums] is probably more likely to have a direct effect than [drums start off quietly and gradually crescendo]
That being said, Suno gets drunk sometimes and will say things because it doesn't know any better -- @naive quail often uses the style prompt "enraged goats on welfare", and whenever Suno decides to randomly sing that in the lyrics, it never gets old
that's hilarious
at the end of one of my song the ai just became sentient and started having a convo randomly I was very confused
F 3.5 and its love to sing!
working on the outro, and it decides to add 'more' to the end and extend it
someones got some sass
😛 I've seen that a couple of times -- admittedly I'll never report them because it's still the absolute funniest thing in the world to me
so much so, that I would pay for a "glitch mode" where all of the horrible traits that have been trained out could be re-enabled for kicks & giggles
that would be funny as hell
"What's that? You're tired of songs just ending the way they're supposed to? Turn on "dramatic ending mode", and the song will "re-collect itself" to really end the song in style. And then continue to do so until "dramatic ending mode" is concluded. Leave it enabled during a 20 minute generation to annoy your friends and family bad enough they'll consider legal action against you."
you crazy man
big ayoye I love it
its jammy tight
said drums gives you a CLEAN ass saxophone riff
I literally could not have phrased it better myself.
OMG with a piano tooo
It's the tightest of jammies.
welcome to the life of a french quebecois in the infantry ahahhah, write funny in english
Like, at least 2-sizes-too-small.
🤦♂️ I didn't even see the Fleur-de-Lis! ❤️
my pirate career is starting ahah
I'm Moncton-based at the moment 😄
Dont worryyy! had the communism symbole but had to remove it they told me I could get removed from the force if I had anything subjective
🇨🇦⚓
There's another user who lives in the North end of the city, lots of users in the area from what I'm starting to learn! -- I stayed with a friend in a small incorporated-town near the citadel (?) when my wife and I moved out East
do you like it?
iz it more frenchy north or south of the city?
since NB is really mixed ahha h
Actually, it might not have been the citadel in QC, I think it might have been somewhere in montreal old city, because he lived on-base -- hell, he probably fixes your arms now that I think about it😂
this is a chain of generations of 3 pieces. I had the final audio file somewhere, if you are interested.
In my case, the additional surprise was that the song itself is not in English and then suddenly, completely unexpectedly - a phrase in English x)
if it easy to find shoot me it !
Maybe a homie of mine cuz havent deployed yet sadly, but soon they tell me in about 6 months
Total overshare below;
I absolutely love Halifax, and I would have spent the rest of my life in that city, except the fleabags (who bought my building on the cheap after the owners sold due to covid) made renoviction attempt after renoviction attempt until they literally broke us mentally & we agreed to take a payment and vacate.
We moved to Moncton because my wife had family out here, but going from the most pedestrian-friendly, culturally-forward city in Canada to Riverview NB was a bit of social-whiplash 🤣
That said, we're happy enough here, and despite being snuggled right into Dieppe, we haven't struggled too much with our 20-year rusty French 😅
I find the ai's voice sounds better in another language than english ahha, Like I put russian lyrics and it sounds better than english haha
kingston in ontario is hella pedestrian friendly, like your literally first always its fun
eh if you need a refresher in french I'm here !
NFLD, PEI & Halifax are definitely different experiences altogether -- instead of well-reasoned road systems with any form of forethought, they just paved over the original horsepaths, so you have 7 way intersections, and somehow everyone who's lived there longer than a panic attack or two just understands intuitively who goes when -- and the pedestrians are just given right-of-way, because people have just acquiesced to getting to their destination late no matter what they try.
I decided that for your understanding it would be useful to get a translation of the text. I did not try to put it into rhythm. Here is more a translation of the meaning (and that is through Google)
legit described what other nations think canadians are ahha
I absolutely adore Kingston -- I only spent a short time there visiting friends who attended Queens and really enjoyed my time exploring campus -- and that brew pub near the library(? god this is so long ago) was amazing, dragons ale left an impact (though I swear to god, if anyone folds up a leather jacket and slams it behind me again, I'm going to have a vietnam-style flashback. wtf is with the engineers there? seriously? are they a cult?)
UK mens eh ahha
the instruments are weird as hell, but yooo this slapssss hard!
also me has 1/2 of my songs with russian lyrics but eh ahahh
were the zzt zzt afterpost?
music tags is "techno, glitch, glitch, female voice" and [dubstep][glitch] after almost every part tag (verse/chorus)
ohhh its the glitch that's making that sound, damn that's special !
^^ 90% of my songs are based on or around events/characters from role-playing board games. This one in particular is about a girl who can interact with PCs and networks like wizards do with magic and ether. It seemed to me that glitches were absolutely necessary here X)
what's the board game? seems like a cool idea
Board game "Shadowrun". This is one of the character classes - "technomancer". In the game's lore they are called "children of resonance" x) "Resonance" here is a term at the level of "ether" among magicians
thanks taking a look right now
seems interesting and has multiple version of the game, looks interesting ! what did you name your character?
since the lore of this character says that she was once found by mercenaries during one of their missions and later raised, I decided that she shouldn't have documents. In the game world, the document is called SIN.
Character's name and surname: Jane Doe
Street name: UnSin
Was chosen as a clumsy form of "Sinless" and consonant with "ensign" despite the clumsy pronunciation.
jane doe haha I like that, also Sinless is just pure
I wonder if the idea of "sequential generational play" has been brought up here before? Like, I open "part 5", press a button, and it starts playing from part 1, moves to part 2, and then on through the expansion timecodes.
found that editing lyrics from full songs can help reduce the ghosted lyrics when extending
"Get Whole Song" on any of the parts ❤️
Hmm... nothing happens to me after that. literally...
A "full song" will appear in your library, you might need to refresh the page to see it.
It appears where a newly created song would be -- it costs 0 credits
oh... now I have 6 "full" versions =-= Somehow I can't enter all of this =-=
When you Extend, you only use the single previous track as "musical context", so Suno is only able to see that far back in order to understand what the rest of the song should sound like.
If you "get whole song" to extend from, you're letting Suno see the ENTIRE song as context so it can continue from it easier and keep everything sounding the same.
If you extend from an extension , it's like photocopying a photocopy
That's okay, "get whole song" is free -- just delete 5 of them
https://discord.com/channels/1069381916492562582/1270928841614823546 I just shared a video tutorial last night on how to delete multiple songs at once ❤️ 🙏
You should notice a difference in results when extending from "get whole song" and extending from a short extension.
If you're looking to explore helpful ways to make the most of Extensions, an intermediate-level read is my guide on "Terminator Extensions" -- a method to end a song whenever you want, or make a song swap to a new genre in as short as a single second. #1256301881966198805 message
any suggentions ai generated music videos?
There hasn't been such a need yet, but I'll definitely check it out
Not necessarily for everyone, but I try to collect helpful tips in that thread if you're ever looking for new techniques to grow into, or want to figure out how a specific prompt works in isolation 💪
Oh... I got very disappointed in Noisee.ai at some point. I was inspired by the review on YouTube, but in reality I couldn't get anything relatively intelligible. And this is not even looking back at the fact that by the time I found this AI - some of the functions from free became paid (like generation by image)
Now I don't even know if it's worth sticking my nose into it at this stage. I'm not sure that it's possible to adequately recreate the idea from my head in a reasonable number of iterations... Maybe I should wait a while and let it grow =-=
I haven't played around with different video models much, but I didn't mind the results that I've seen others use resulting from noisee, but I realize that's pretty subjectve,
the results that I have encountered in one way or another are of a very abstract nature. As I wrote above, most of my songs revolve around very specific characters. So, in Shum I tried to make a video sequence "like a cyberpunk rock concert" and described the character's appearance in great detail. I can even share all three results. Everything is pretty close there, but there is no smell of a single character, for example.
what is the audio though? if it's from a licensed source it could be flagging it
thats not what a flagging error looks like
is it original tho
Si
have you tried refreshing, or a seperate copy of the file?
could be corrupt
the audio is fine, i used other ones too and they didnt work
wifi or data?
is it compressed?
i uploaded it from capcut
do you have any kind of browser adblock or anything
yea try it it could be that or its capcut
but i listen to the file and it sounds fine so i dont think its corrupt
try to upload something from somewhere else like a voice recording or anything else
yep on google im getting the same error
you can listen to it but it doesnt mean its compatible with the uploading. if its capcut maybe thats the issue
try any other audio file for example record yourself tapping on a desk or anything at all
the record audio thing is working
teamwork
use bandlab or other workstation
ive used capcut a million times before tho
Reaper is also "free" for 60 days
capcut has had a million outage issues for me nefore
hm idk i used a mp3 converter
maybe it is capcut idk
Is there anything you can put in that prevents the song from just ending abruptly? Few tracks will be great but it just cuts off without a proper end.
well thats not "technical" related try in general but yes
I was worried i had the wrong channel earlier, thank you for letting me know
I would think tags are under the technical category considering workflows is part of this category too.
How bad is it if you often get flagged for explicity? I write deathmetal music and i often get flagged... worried here 😄
and its also very difficult to understand where the line's are drawn... some things work some dont.
if the audio format is in aac, it has errors, try converting it to mp3 or wav. That worked for me
"Hello, I have a paid account. I can't log in; when I enter my phone number, it says, 'This phone number is taken. Please try another.
I have solved my problem.
I'm trying to download a video of one of my vocal tracks for the first time using Regenerate Video, custom art, and edited lyrics. However, the downloaded video shows the wrong art and the original lyrics without taking any of my edits into account. What am I doing wrong?
Did you regen the video ? It can take time to update after you do that
It seems I got some timeout messages in the background that disappeared before I noticed them. After about a dozen retries, it finally worked
Why this issue?
Am I right in thinking that the uploadable audio is essentially used as an intro to the song? Or is there another way to use the uploadable audio other than an extension?
1- 100%
2- Not at the moment.
@empty saddle pefectly suited for tech-chat 🙏 🙂 Asking for advice on best-practices for using Suno is what we're all about here 😄
Gen-chat is great for "taking the temperature" on a matter, but since tech-chat moves a bit slower, it's a great place to dive a bit deeper into specific questions like this 🙂
But! You can still use it, to get the same desired result throughout the song, by using proper prompting, let's say you are using a singing part at the start of the song, this would now mean that you have "the seed" (Whilst this not possible within Suno, it's the closest we can get to maintain an old style).
I actually did this just last night with one of my older songs, and the result I achieved was actually much more pleasant for the ears, as I now know more about proper prompting.
Old version:
https://suno.com/song/0b6021d2-534e-475e-945c-7b3293c1a97e
New version:
https://suno.com/song/a62a8224-9014-432a-bb25-e18c9d3cacc5
I used the Extend here, and implemented what I recently wrote about, and you can hear it "kind of" keeps the same style
japanese pop rock song. Listen and make your own with Suno.
japanese pop rock song. Listen and make your own with Suno.
Can you elaborate, or share an example song so we can see how the songs are ending suddenly?
v3.5 tends to crave music coherance, so sometimes it can "get stuck" on a particular sound, but whatever your issues are, I am confident we can work them out as a community 💪
this would now mean that you have "the seed" (Whilst this not possible within Suno, it's the closest we can get
@rose jasper, I haven't even been out of bed for 30 minutes -- give someone else a chance to earn "most awesome person" before taking home first place ribbon so quick out of the gate 😂 ❤️

if you have a short bit of silence at the end of your uploaded audio (eg your audio therefore does not 'loop' perfectly on tempo), then a lot of the time, your generations will use your audio throughout the song when it deems necessary and use it as 'inspiration' (eg i uploaded a flute recording before; it used it not at the beginning but in the chorus and added extra layers to my melody) but you currently cant specify where or how it gets used.
While I'm sure there's a more accurate descriptor, I've been calling it "a seed" similarly, like a "seed crystal".
Thinking about it this way sheds different light onSuno's insistence that you don't use clips that you're not the sole owner of -- if multiple people use the same "seed" and prompt for similar songs, it's very likely that the songs will be much closer than other generations, furthering the chances of any complications.
All in all, a great pointer to see someone adding 😄
That was my suspicion as well, but I just tested capcut and didn't have the same issues reported. 🤔
If you're still having an issue, I'd advice contacting support@suno.com, as they're the only "official support" presence, but if you'd like I'd be happy to try to upload the clip on my end and see if I get the same errors? If it uploads fine, I can send the suno link your way 🙂
😅 In my earlier days I used to make "inappropriately funny" songs, and frequently ran into this issue. They haven't explicitly called me out for any of them or kicked me out of the alpha group, so unless the content you're trying to make is particularly egregious or troublesome, I'd imagine it's just "the guard-dog barking" and there's nothing larger to be concerned about 💪
Though I'mma sprinkle some "contact email support for clarification if you still have concerns and want to be proactive on the matter" on that, to taste.
ew, I just realized the salt touches his arm.
Why are you even wearing nitrile gloves if you're just gonna glom all over the salt like that, bae?
How long was your uploaded track and what timecode did you expand from?
Gives that extra "salt" 
So, what am I doing wrong that would cause suno to absolutely insist that it repeat lines 7 and 8 of Verse 1 before even attempting to move on to the chorus? I've thrown over 3,000 credits at this song so far, and I'm just about fed up
my audio was about 46 seconds, extended right from the very very end of the audio
3000 credits means you've clicked "generate" 300 times without getting the results you're hoping for 👀
I'm glad that you decided to reach out for clarification or assistance -- typically such difficulties aren't common, so I'd be curious to look into the matter to see if there's any misunderstandings we can help with 💪 ❤️
Can you share one of the results that outputs your lyrics improperly?
Heck, I can even show the one I extended from, that also decided it just had to repeat them again
now, I did change the Displayed Lyrics, because I've found that can, and very much will affect things if you decide to tweak lyrics later on in an Extension
(it will happily ignore any kind of changes you try to make in the Lyrics Input, in favor of what it finds in Displayed Lyrics on an Extension, if it finds anything beyond the time you've extended from. If you try to add a word, it won't change until you go back to the track you're extending from, and edit those Displayed Lyrics, to either cut out what's not used from the clip, or to change them to your intended new lyrics)
Would it be alright to PM the link?
Absolutely!
I mean, I wouldn't have any issue if it had run out of supplied lyrics and decided to start repeating lines again (that would be my fault), or if it did so in a manner that was actually appealing, such as much later in the song, but to have it insist on doing so only a quarter of the way in, on the first verse, to the extent that trying to use Extend to break it of that trend required essentially yanking the leash so hard it 'yelped' (and caused bad vocals at the 'seam')... that just got infuriating
I mean, sure, I did try something a little non-standard, but it shouldn't fumble that hard on 8-line verses
Also, is there any relation between the artwork generated, and the style of music?
How do I indicate to Suno to NOT create any backup vocals? I Just want a single clear vocal. Thank you!
Do you have any examples to show?
Can you add the ability to upload files to the general-suno channel?
Backup vocals aren't something that Suno adds without reason -- it's likely determined by the lyrics / metatags / style prompts; like zaj mentioned, it'll be easier to figure out what's going on if you can share an example of this happening?
This channel is for talking about how to use Suno, you want to make a post in #1252257127326548028 asking to make embeds in general-suno
Also, they took embeds out specifically to stop people from interrupting conversation with irrelevant content like what you just shared.
#🗣️┃chill-chat is a great place to be silly, though 🙏 ❤️
That and all kinds of other security reasons 
Thank you.
I'm just a user like you, my dude -- thank your friendly neighbourhood moderator next time you see them around ❤️ 🙏 💪
ChatGPT 4.0 is such a good method to improve already existing prompts, honestly
I was actually thinking about your meme #😜┃suno-memes message and can’t help but wonder why there isn’t a 5000 plan. 10k is too much for me, and 2k5 isn’t enough
There's extra credits, or you can subscribe to a second paid account if you want an easy way to have a $20/5000 credit plan?
gooooood morning suno-tonians !! my cursor disapears in the lyric fields and boxes and where where i have to select text. happening across chrome, firefox and edge.
only happens at suno, any one eles haveing issue
Have you tried the #🪲┃bugs-web channel?
yes in there now ,
any one know a way to stop suno from adlibbing ?
i just want the lyrics to be spoken, not some guy going aahhh uhhh yuhh hmmm ad the star of every song lol
Do you have any examples?
yup one sec
I second zaj's request for examples -- this isn't really a wide-spread issue with Suno, so it's likely determined by your lyrics/[metatags]/styles
The prompt is more of a suggestion for the AI than a complete blueprint. If you are using spoken word that might be the AI trying to add the intricates of speech into the mix. I would try using tags at the start to try to clear the mix. The following is what I use. [Perfect EQ], [Clean Mix], [Clear Vocal], [Good Vocal Chain], [Final Vocal Mix], [Vocal Cuts Through Mix], [Final Master]
Yeah, it could be so many things that gets put into play here.
I remember there had been differences in opinion regarding the use of ChatGPT writing songs from scratch, including huge blocks of text with very complex metatags (5+ words each), so I wouldn't be surprised if overloading metatags at the beginning of a song would cause a degree of confusion for Suno's ChatGPT interface to see a massive block of words, and then try to figure out that it's not supposed to sing the first half of what's been provided because it's intended to "instruct the song"
im usually trying to make it add more adlibs haha i need some vocal ear candy in my song
my lyrics are just parentheses all over the place
Can you show the entire block of lyrics before this is sung?
Yes please, and the Style of Music
Oh, you made it sound like explicit "aahhh uhhh yuhh hmmm"s, but this is just a vocal melody?
It's very easy to do this though
Perhaps you could make the introduction instrumental only, then create an extension after a certain time so you can take granular control of how and when the lyrics or vocalizations are introduced. 💪
With those prompts, you are reaching deeper into what those "tags" should act and where they are normally used. Let's say [Soaring] this can be interpreted like a vocal performance for example.
I think "soaring" is suggested on the wiki or a pinned post, regardless if it works or not.
[Isolated Instrumental Buildup] <- This could work for yourself
[Occasional Vocal Breaks] <- This as well is a very powerful one to use
I remember someone asked how to make the a song "go higher" for the last chorus, and someone suggested very confidently to use "soaring", and told me that "key change" wouldn't work.
I was confused by that, and only understood where they got the reference from later.
I think the wiki suggested it as something to try, and they interpreted it as "code"
dont mind the [notes: mostly for me to keep track of things ang changes suno just does what it knows
For making the voice go higher, I would recommend you check out the manual here:
https://www.suno.wiki/wp-content/uploads/2024/04/PROMPTS-FOR-AI-DANCE-MUSIC-🙏mista.lewys_.pdf
Tempo Tags:
I have the problem where Suno vocalizes the tags.
I appreciate it -- I was more sharing the anecdote than requesting assistance though 🙂 🙏 ❤️
I used to have this problem before, but it all of a sudden stopped doing it for me. Try to give it more renditions, and it will in theory "learn" and "catch up"
I know, hahaha, I should've responded to loanlytiger instead 
when that happens for me i ake sure every thing is [ ] off and if it still does it it a sign that your spacing is off
Can you share some examples of tags that are being vocalized? While in theory this can happen with any metatags or style prompts, it tends to happen more if Suno is being overloaded or if the metatags have more than 2-3 words in them.
Try adding only Instrumental prompts, before the vocal part starts
smart tahnk you ill try that
I'd have gone one step further and not use lyric prompts at all, and just use the instrumental toggle, that way you're not trying to further exert control over timing and structure through the use of prompting, but making certain features "available" only when needed, creating an extension point to choose when the lyrics start
I did it for this one, I basically focused on how the guitar should sound like, and I had very few vocals on it, it was more of a guidance on how I want the guitar to sound like:
https://suno.com/song/c7f9d879-f0ba-4f15-aa31-5762a6e5db14
dark, metal, power metal, Arpeggio metal, racing music, electric guitars, grungy, angry, intense, angelic song. Listen and make your own with Suno.
Let me see if I can find one that does it a lot. It might be lost in the library.
True, only add the lyrics afterwards, and just do "Extend" from the part you're happy with, correct?
searching for any word from the lyrics in the library-search bar can help narrow down, but sometimes it's just a fool's errand trying to track down specific songs 😛
Don't be afraid to use the prompt literally to describe what you want the guitars to sound like. "Dynamic Riffs" "Palm muted arpeggios" etc
Everyone has their own methods for making songs, but this way I figure you're better able to ensure that the vocals don't start RIGHT at the beginning of the song like a lot of one-shot generations do.
Plus, if you use an instrumental-only intro, it gives extra context to v3.5 for deciding what vocals would fit more appropriately.
Yeap! I do the same with old lyrical pieces as well, so that I can maintain the same style
https://suno.com/song/9bd42b25-86c9-427d-9138-4a54a3531dd5 I got this sound from putting in the prompt "Glitchy guitar riffs" and "minor scale" - Headphone user warning if you have your discord volume on max lol
[Djent], [American Industrial Metal], 194 bpm, dark, heavy, glitchy riffs, cybernetic song. Listen and make your own with Suno.
Super helpful for guitarists and existing musicians to be expressive and have fun with it, but also just as important to note that you don't NEED to understand musical terms -- "chugging quiet guitar" will probably work just as well as "palm muted guitar"
Super!
Or even something goofy like "[guitar :jugga-da-jugga-jugga da-jugga da-jugga-jugga]" to if you have no idea what "djent" is 😛
okay, gotta try that
you won't, you're scared
interesting iv never had such problem for a song befor this , spent a few credits in frustration . lol
When you include certain tags, they tend to be used by other users primarily for X Y Z, so they are variable and ever changing.
okay, i guess that gif was too spicy 😛
Yeah sometimes suno plays nice, but 90% of the time it doesn't, and you gotta drill into it what you want
But the best part is to isolate for the current version
Also if you don't want to muck around, put [Instrumental Intro] into your lyrics, then do just your first verse, bridge, chorus etc.
Combined with your prompts you should get a lot of songs a lot closer to what you're hearing in your head
And hope to god that it doesn't repeate verse 1 twelve times in a row 
lol
Hahaha nah, if you extend after the first chorus it'll remember everything from part one and you can just continue on.
You're a monster
you should see me on job site lol, i tend too get spicy,
Okay, so it didn't give me djent as expected, but it gave this when I took the brackets off the tags, and I mean, I don't hate it. The "breakdown" at 0:50 is just cute, in context.
https://suno.com/song/c5047dcc-0dea-430c-af13-64af2a5da746
Dang it Woodhouse
This is honestly the best
hahahahahha Give it an indian style, perfect
Ok hear me out dude, pull the MP3 and reupload it with [Djent], [Industrial Metal], [Djent-step] as your prompts lol
lmao
Woodhouse is eating well in tech-chat this morning.
Why is this so catchy tho, what the hell
My ADHD mind now wants to make this into a song, damn you all ❤️
2 months from now, that djent-youtuber guy is going to make another video and he's going to have no idea why Hindu techno is trending on the Djent charts
lmfao
lol
Then just a whole video on basically making Bloodywood
this may sound weird, but how would I determine certain data for a song I got that I really liked, to try and re-capture it with tightened lyrics?
it did take a few Extensions to get though, so I'd need to probably prod it in a couple different directions
If I’m understanding right, you want a similar instrumental and vocal style using improved lyrics?
and preferably same meter
yes. Thing is, I... didn't exactly give it too much direction for specific style
so, I'd want to try and determine things like BPM, Time Signature, certain key notes, etc.
Best bet right now would be to download the song and use the audio upload feature, there isn’t anything to “reuse” a singer or instrumental yet
doesn't that just make it use the uploaded audio as-is?
I need to make some changes early on
as early as Line 2 of Verse 1
It uses the audio as a base, it doesn’t give the exact song back
I see
and if it's based on a suno generation, would that kinda load in the same 'seeds' as it originally got?
Hopefully
But it’s not guaranteed to get the same exact “seed”
But it could get something similar
im not clued in to who this is but i need to be 👀
see, as long as the verse and chorus structures stay the same, I'm okay
https://suno.com/song/e00f2a20-3c45-44a6-9be1-c06275361c0f this is an example of where the beginning 16 secs is my own audio and then Suno runs with it
EDM, energetic, atmospheric, glitch song. Listen and make your own with Suno.
okay, so... let's say I give it a minute with the verse and chorus in it. Will it pick up on those if I have it start early in the verse section?
so, 60 seconds of audio, but Extending from 12 seconds in
or will it just discard everything in the original audio past that timestamp?
Extending the song from 12 secs, or from the audio base?
If you start there from the audio base, it should hopefully give the ai a hint of how to do the verses and chorus
but will it USE that whole audio base in the song, warts and all?
let's say there's a vocal glitch in there
There could be a chance the new gens could get something similar
Though sometimes the bugs actually make the songs more interesting
the glitch in question is a mixup in reading "our": it came out as 'are-err'
I was going to just lean into the two-syllable one and have it read "hour"
I have noticed some words you need to misspell to get Suno to read right
it's ambiguous ones that can be a different number of syllables
"our", for example, can be closer to 'are', or 'hour'
For one of my songs, the ai couldn’t say kitsune, so I had to spell kitsuneh for it to say it right
yeah... don't tell it to un-wind, either. Gotta be un-wined
Any other questions?
well, I've never tried using uploading audio, so I'm not sure how that actually works
can't do anything for a couple weeks anyway, since I'm out of credits
Is there a way to upload existent instrumentals and add vocals over it
You use the "Upload Audio" feature and extend from any point.
hah it really do be like this to be fair fam
my brother recorded the shred ting not me lol
=LINKS=
https://www.patreon.com/kmac2021/
https://twitter.com/KmacTweety21
https://www.facebook.com/Kmac2021/
https://soundcloud.com/misterboy/
https://soundcloud.com/colmaband
https://thenightyouleft.bandcamp.com
=LINKS=
@cerulean kestrel presuming this might also get a chuckle from you, given the guitar-stuff
What does it mean when there's a red dot next to the picture for music you've just generated in the Create window?
If you generate Stems, that could be the reason
Stems? I'm sorry, I'm not familiar with the term, nor what that red dot is supposed to actually mean.
These are songs that you haven't explicitly clicked to play yet ❤️
OH! Is that a new thing? I haven't seen it before... or I somehow haven't noticed it. 😅
Yup, new feature!
Ah, ok. I was concerned because I just tried generating something and a message popped up about it being flagged for moderation before it was auto-deleted. Which is weird because I'm not using any new terms I haven't used before. At least, like, in terms of general language or what could possibly be offensive or something I think.
We're not able to "in-paint" over a song to add a specific sound while keeping everything as it was, but if you find a point to Extend from that still has a good chunk of musical context at the beginning , Suno is pretty good at continuing from musical context so that any new lyrics might follow the same melody and pattern.
It's an AI detection system afaik, so there can be false positives (where nothing was wrong but it was flagged) or false negatives (where something clearly was wrong but was allowed).
Perhaps it's just a matter of how you're describing your lyrics, or perhaps the arrangement of the lyrics puts specific ideas as "the focus" and might be more prone to flagging than if it were less prominently noticed -- like if a song starts off with a swear word, opposed to it being an accentuating word used in appropriate context.
If you're on a paid account, you can test out v1 of the "get stems" feature to separate the vocals from a song, and have a track that's ONLY the vocals, and a track that's ONLY the instrumentals, like a karaoke version of your favourite Suno songs 😄 🎤
I see, I didn't play my stems... and they all have the red dot
https://suno.com/song/24d6c086-a972-442a-a47d-053e7ffb46e4
longest intro before lyrics begin I think I've ever gotten
b-slowcore -ina -umble ot, otot -b
mys-shitty-fucking-slslslslslslslslsl revival-revival
#MySpace #myspace >... song. Listen and make your own with Suno.
it's a full 2 and a half minutes
OH! I see! That's honestly kinda neat, thanks!
before it just cuts
Yup -- the red dot will disappear when you click on the tracks. If the track rolls over to the next song and auto-plays, it doesn't remove the red dot. This is so you can be more aware of what songs you haven't specificially thought to listen to.
Huh, maybe. I don't use profanity or vulgar terms, so perhaps a false positive?
lmao, I always love trying to imagine Suno waking up from a nap, checking the "inbox" stack of requests on its desk, and seeing a form stamped "Yolkhead" as it just starts to enter a static of quantum-quandary.
2m17s of intensifying musical animosity, followed by 5 seconds of hard silence.
That sounds like the Myspace I remember
oh my god. i forgot about myspace. 
What do I do if all my credits disappeared? I had over 2000 and now it is reset at 50!
hit subcribe button then back to create
see if they reappear
2 questions;
1- Are you sure that you're still signed into the same account?
2- Do you still have an active subscription with valid payment details?
huh. doing an experiment where im using the prompt to basically ask suno how it feels about making all these random songs, using any genre. so far, all female vocals and all electronic pop. i know it's not like trying to talk to a regular chat ai, but i find it interesting 
....god i need a hobby
It's entirely interesting, imo -- AI is still such an amazing breakthrough in-and-of itself that ANYTHING it does is still mindblowing to me. I'm not wildly familiar with how the ChatGPT-side of it works, I wonder if you could just have conversations with the "generate lyrics" box 😛
