#📚┃suno-school
1 messages · Page 58 of 1
[when i got a bad long tag like this] the ai Reverb scream it it and all my other (ad-libs) in the song turn into male screaming XD
Extending songs? You just click the extend button and then either change your lyrics or add new ones or different style music. You need to make sure to choose the time you want to extend from. If you are trying to get different voices or styles in the same song then I would suggest to pick the Intro, Verse 1, pre-chorus/chorus, and use that as your first song generation. Then generate until you find a style of voice/instrumentals you like. Next, extend that chosen audition and add Verse 2, pre-chorus/chorus, instrumentals etc., until you have your full track. You might get lucky in getting different voices. One thing to note is by building verse-by-verse like this you will get consistent instrumentals and voices.
well i think ill play with it 🙂 thx
Here what im trying to do
Exemple:
Intro: metal jpop
Metal intrumental transition
Verse1: metal
Pre-chorus: metal
Chorus : kawaii jpop, dance beat
Transition: blend both style
And here we go to metal again
So i need to extend at the right place and pray? 🤣
has anyone found a nice solution for sound quality? i use ripx for stem separation after seperating in suno. i dont want to have to rebuild songs in a daw every time if possible.
Yep. But what you could do is set the style of music in the "Style of Music" section to suggest to the AI to transition in the extension. I actually haven't tried this as I am trying to prevent the AI from getting too creative when I extend.
You can try to add tags to try to get the AI to have a clean mix.
I'm going to generate another composition from the "in the last episode" series for board games in the next few days. I want to somehow squeeze an existing sample somewhere in the middle (the theme of one of the characters). Of the ideas (at the moment) is to generate a piece, finish it (so that the music stops), and then, using third-party software, essentially, just stick another track. But this somehow looks wrong. Maybe there is a more adequate way to implement this idea in Suno?
what does this mean?
i'll play with it thank alot for the help
I'm having an issue, which occurs quite regularly now I have a piece of music that I like how it is going therefore I request to extend it, sometimes I want to continue from the end or go back to 15 seconds before a fade out to continue the music as it is, but however much I try I only get a few more seconds that the original, how can I force suno to give me more? it very frustrating, btw these are instrumentals that I want to extend, is there a tag I can put in the lyrics section?
@charred merlin Any ideas on this?
These are the tags I use. [Perfect EQ], [Clean Mix], [Clear Vocal], [Good Vocal Chain], [Final Vocal Mix], [Vocal Cuts Through Mix], [Final Master].
do you add that to the style area or the lyric area? do you truly notice a difference?
I've been experimenting heavily with this. the trick seems to be to give it a lot to pick from and it will pick some of it. this one picked the [train track sound beat] (added train station sounds), [sampled piano chords], [lofi beats to eat rocks to] (just the lofi part), [soda can smash] (used as rhythmic element)
https://suno.com/song/f1036998-423d-4c7d-8ad6-98fc2ae0f310
experimental sound effect, sampled dreamy choir, drone bass, monotonous drum beat song. Listen and make your own with Suno.
interesting
reading now 🙂
Is it important that everything stays within the Suno.com ecosystem, or is it okay if it only lives "outside"?
it doesn't matter where it will be, a track "outside of Suno" will do for me, but I'm not sure that my skills are enough to beautifully mix "part 1" "theme" and "part 2" so that they would sound like something whole. If the transition of "theme" to "part 2" Suno will still help generate (and even then it is very problematic that part 2 would even resemble part 1), then how to make the transition between "part 1" and "theme" - I don't know...
I read your text about "how to change the genre of music in the middle", but my problem is that the "theme" is already ready and will be upload from the outside...
So, instead of swapping to a new direction generated by Suno, seamlessly inserting an Audio Upload "into the middle" of the song, yes?
yeah
I have an idea of how to tackle it, but it needs to live outside Suno because it'd require some hands-on work in a DAW
The lyrics area. Sometimes you can tell a difference. Other times not so much. But it doesn't hurt. If the generations are too muddy I add it and it more times than not it clears everything up.
DAW? Sounds like something that would be difficult to get the hang of the first time.
That is the frustrating part of the extensions. Sometimes your get very short extensions. Usually you have to change something in the song structure to get it to generate longer. Try going back a bit farther and see if it helps. Sometimes the AI needs a bit more information to generate with.
This track was muddy and after adding the tags it cleared it up. https://suno.com/song/966c20d5-7270-4fc9-8b1d-a40d672e88b5
Orchestral-Waltz, 84-BPM song. Listen and make your own with Suno.
OK, thank you for this tip. Its great!
Digital Audio Workspace -- https://www.bandlab.com/studio for example
Some, like this are free and easy to use, others are super complicated and can enable crazy levels of editing -- but at the very basic, it's a place to cut and paste parts of songs 💪
The cloud platform where musicians and fans create music, collaborate and engage with each other across the globe.
BandLab's suite is basically a free superpower if you want to edit stuff outside of Suno -- https://www.bandlab.com/splitter if you want to get voice / guitar / drums / bass as separate tracks, and then play around with the levels (or cut out that weird vocal part while leaving everything else in tact 👀 🫨 )
Thank you! I will definitely try to use this!
Are their any proven practices that can influence a song to stick to male voice and female voice? I cannot seem to get [Male Voice] or [Female Voice] too work, or should I just keep on going until I find a good one that knows what it's doing?
It helps me (almost) to indicate the gender of the voice in the tags to the music. In conjunction with the tag in the text, in almost 95% of cases the gender is correct. There are discrepancies, but not significant.
[Verse 1 - Male Voice]
[Chorus - Male & Female Voices, Cantabile, con moto]
[Verse 2 - Female Voice, Mosso]
But however, it's not fully committing towards it
Ah.. Duets.. Yes, this is difficult here. Frankly, I would make such a song using extensions. Let's say, we make 1 verse separately, indicating in the genre tags and in the text the tag "male voice". Then we expand this, making a chorus, but indicating two voices, then we expand to the second verse and here we tag female everywhere. .
From what I read, I got the impression that this option is probably the most reliable for generating several voices. True, in this option there is no guarantee that the voices will be the same X)
I have a few projects on the go right now, but I'm going to make a note to self to circle back & test out the my plan to "combine tracks" to functionally insert an audio sample into the middle of a track -- not quite sure when I'll have it taken care of, but it's a pretty neat idea, so it's near the top of my want-to-do list 🙂
I'll ping you if anything turns out one way or the other 🍻
I have plans to do this (or try) within the framework of this week (until next Sunday, 4 pm) 😆 So, perhaps, we will both discuss it while the trail is still hot. To be honest, I have never worked with DAWand there may be complications, but they don’t charge money for trying (I think)💪 🤟
BandLab is free and awesome ... annoyingly awesome.
I'm doing something now, which I think more people should do. Use multiple languages in a song.
This heavily depends on language skills. Suno smooths things out, but even he can't always handle a text without rhythm and rhymes. And what can we say about the correctness of sentence construction in an unfamiliar language? Online translators are good, but not omnipotent.
I have a song in English, where one verse is sung in Japanese. The song is about cyberpunk and that's all, so the style fits.
But I won't let anyone who knows English listen to this because of my extremely poor knowledge of the language (I'm currently communicating here through a translator). What can we say about the Japanese, from whose language I know a couple of meme phrases X)
You don't need to understand what is being said, that's besides the point. It's about learning as well
And I'm getting them properly translated using other methods
For me, music is something that should not be limited to different languages, and I often write my lyrics in Korean, Cantonese, Taiwanese (Hokkien), English, Japanese
I don't know the languages, but I use good translators for it, and it's a great way to introduce the languages in a nice way to other cultures
Believe me, as a native speaker of a non-English language, I have often encountered how people try to write lyrics in my language without knowing it. It even sounds crooked in prose. In songs it is even worse.
I won't stop 
Then at least ask native speakers how it sounds to them. Not to mention all sorts of idioms and set expressions, between languages you can get phrases that simply do not carry meaning, although they are correctly translated. At the same time, just as English speakers can find foreigners in a crowd just by the construction of phrases, so it works with other languages.
It sounds like you are looking for problems, so I'll just drop the conversation now 
sure, but most of the people here are having Fun casualy with the app, if they want a Pro work for money, i hope they will. XD
btw i speak french and english and some basic in japanese.. i'm not a pro so ok
Does anyone here know Japanese?
You got what I meant then
"off-topic of the thread" anyway 🙂
Did they change up the credits a bit? I've got 50 again. If they're giving Premiums the Free dailies as well now, that's a huge W for suno
yes its 50 per day
Suno team, they're asking for you...
Hi gang, just wondering if anybody has figured out a way to restore output that sounds clipped or distorted? I.e I have pieced together a structure of a song I’m happy with which required lots of “extensions”, the issue some of the outputs are “hotter” than the others and sound as if it’s clipping. All I want to be able to do is get suno to rerun these “clipped”parts without sounding distorted 😅
Is there a problem with suno today? My account suddenly lost more than 9000 credits
well they reset every month if you havent used them up within the month, if thats not the case i'll recommend sending a email to support.
Looks to be a visual glitch #🪲┃bugs-web
lets hope 9k lost in a blink of an eye is quite alot.
Josh, I'm kinda in the same boat... I have a song I love, but needs to have its lyrics tightened and hardened up, and I want to just slot the new ones in. I'm hoping v4 is more of an 'editing' update than a creation one
if you talking about longer tracks with a heaps of extensions i would say my only resulution to this is listen veru close after offput sounds in the background and go back as far as you dont hear it anymore and start from there and move on , if thats not what u meant , ignore this message 🙂
'cause, outside of being hard-headed about song structure at times, 3.5 spits out a LOT of bangers
Yes, I have already emailed them.
hope it get resulved.
resolved*
im still waiting on paypal personaly to stop holding my suno winnings hostage... thats for me is a savge move.. so im have to be without suno pro until then ( its fine i will surive just a bit borning, if i get an idea , ) but i can always just do it the old fashion way make in manualy in a daw 😄 )
well premier\pro same same.
Oh shit, well, guess I'm gonna stop using it right now then. I didn't pay and plugged my lyrics in like an idiot, so I guess I may have just lost rights to my own lyrics for literally nothing if I can't even use it commercially.
you own lyric will always be you'rew own suno might own the song but you hold the right to you own lyric. thats even is stated in the tos. btw.
so on that case dont be worry 🙂
You retain your rights to your lyrics, you did not transfer it over.
You grant Suno non-royalty (I.e. free) unlimited usage to use your lyrics that can be modified, etc. (As otherwise the Service cannot use your lyrics or other uploads).
As a free user, the Big thing is that you cannot use the generated music commercially and you must provide credit to Suno somewhere such as in the title or description, e.g. saying Made with Suno in YT description
The U.S. Federal Courts already carved out that A.I. generated images (unless they infringe on an existing copyright) are themselves only valid for specific use scinarios and the copyright will only apply in such a way. (I.e. you use the image on a trading card, on one else can do the same, but they want to put the same image on a coffee mug, you can't stop them.)
I wonder what their rulings will be concerning A.I. generated sounds in the future?
Do you mind providing which case that was? The last case I am aware of was an AI comic where the copyright office determined that the generated art—due to insufficient human intervention—was ineligible for copyright protection
Edit: https://www.reuters.com/legal/ai-generated-art-cannot-receive-copyrights-us-court-says-2023-08-21/
its a bit of a grey aera atm, also it are very much country based too, like the ruling for ai art\music ect were you live are not the same as it are in my country. as usa have made certain laws on it my country has it on a list to do but yet not done, as it not to many that do alot of ai music\art here. ofc my country is a barn compared to the size of usa. so its also that.
We're talking about the same case, where the image its self is uncopyrightable, however, just like anything else in the public domain therefore, you can make a specific use of it, and own rights to that specific use. So to clarify, I guess someone else could put it on a card, BUT it would need to be a different kind of card that you put it on.
Ah yes, that makes sense, I just interpreted your initial message incorrectly 🙇
Sorry about that, I should have been clearer. Anyways, to be clear about something else as well, I actually like Suno, and the price is totally reasonable, I'm just broke 😆
well real life things 1st and forthmost , you can at least use suno for now for practice 🙂
to get better if u ever decide to go fort a sub you are great by then,
Well, all the stuff I generate for free, would I have rights to use that all commercially if/when I got a paid subscription? If so I'll keep using it for now, if not, I just don't want to have it pump out a banger only to not be able to use it.
everything you make as a pro\premier subsribter you own all rights to use as you see fit with. as long as you follow the tos ect.. but everything you made prior to it suno still owns.
Okay, then I will just hold off on generating things until I get a membership.
you could aslways just try experimenting on thinfgs you dont think will be a banger , but yh hey up to you.
Doesn't hurt to experiment and learn more techniques in the mean time. There is still much to learn and discover on how to guide the AI
I found an interesting style recently
but ive done some very amusing andenjoyment learning experiments, like vocal instruments was very enjoyable, still learning that 1 , well its a deep long project.
https://suno.com/song/26981cf9-c690-49da-b19e-bb6c3eec00f3 this style has potential
J-pop, cybernetic, female vocals, pop, upbeat, futuristic, AI idols, glitchy, robotic song. Listen and make your own with Suno.
tbh im sure someone has made this before but lets hope it catch on fire for you 🙂
ive gotten some really odd "tunes" of very random words before
Morning fam! What's your best suggestion for giving suno instruction to make the music uptempo. Just writing uptempo does not do anything for me.
I would also like to know how to make suno understand which language the lyrics are. The first 2-3 word are often pronounced in English before it understand.
How can I use the same under music , but only change the lyrics ?
Let me know what you hear. I was also aiming for time changes, but so far only "syncopated rhythms" seems to do anything but its very vague.
@charred merlin@north sleet FYI its not perfect but i finally able to change the voice for Girl voice in my chorus... now i need to check if we can have some music genre Shift , other than the bridge.
Its not perfect like i wanted but its pretty consistant when i tag my 2nd voice in lyrics...finally.
https://suno.com/song/214c890f-6cf5-4438-9031-68b088d5cb9c
1 goal achieved, more testing to go 🙂
High-energy metalcore, Male vocalist that Growling/Screaming with Dynamic Range, Melodic Rift. song. Listen and make your own with Suno.
Glad you figured it out.
I made a song, but for some reason it play 2. verse twice insted of the 3. Is there any way for me to change that? I really like the melody and the "singer" so id don't just want to make a new version.
im trying to get suno to do a rap thats fast like eminem's godzilla. any suggestions on what style to use?
i've tried speed rap, chopper and horrorcore but none work. in fact, just putting rap got a faster one
Think of it like a railroad. When you click "Create" Suno chooses a random seed and generates a landscape, then it builds tracks between the start of the song and the end based on the most efficient path. It goes around mountains or builds tunnels, and bridges etc... sometimes it needs a bridge or a tunnel and you didn't provide one (based on its interpretation of your lyrics and prompt, which isn't necessarily your fault), so it re-uses what you've given it to make up for that lack. Just like at the end of songs when Suno sometimes keeps repeating what you gave it instead of ending... The random landscape requires more tracks, so it builds them.
You can try to prevent this by using more conventional structures with [Verse] and [Chorus] tags, etc... but even then there's an element of raw inevitably to the song generation each time you Create. Which is why it can be hard sometimes to extend songs and get what you want, and other times super easy.
I've just realized I have never tried to extend from 0:00. I wonder what will happen.
You can't extend from 0:00, only 0:01, the results of which are covered in the "Terminator Extension post in the pool party > #1256301881966198805 message ❤️
It gives the extension only a single second's worth of musical context, so you can change styles on a dime.
@wheat glacier The credit balance listed in the nav bar still seems to reflect the correct balance for me, it's just the account page that shows the incorrect balance for me.
Are you saying that you can audibly tell where the Extension points are when you "get whole song"?
i can vouch for it being a bit kremt visual weird.
i had 750 then it got down to 450 in 1 gen then the sub screen shows me then, alas its a bit hiccup it seems
Someone who expected to have Zero credits has 50 in their account page, but still is unable to create songs, so it appears "all is normal", besides the visual glitch atm
Essentially keeping the same seed? Like, if you don't change anything do you get the same result?
Not a means of keeping the same content, it's a means of cutting the "musical memory" to functionally zero so that you can extend within "The same song" and make it forget EVERYTHING it had done up to that point.
There's an example video in the post I linked, describing terminator extensions 💜
Better to extend after the first complete section then and start over.
Hi guys, I'm not sure this is the right place where putting this question, however, I don't remind which are the songs I created before the payed subscription and which after, thus I'm not sure about the right I have on each of them. How can I verify the right I have on each song? I want to avoid to publish on YouTube, for instance, songs which are not supposed to be monetized. Thank you
Omg, every time i wrote in the prompt for the music style AND lyrics that i wanted a MALE RAP SOLO, its always gives me A FEMALE SINGER, i have been redoing this for almost 7 times now
Im so friggin pissed
Sorry, I was more tacking off an addendum to the "never extended from 0:00" part 😅
Just checking, you've been following the discussing over in #🌴┃creator-programs, right? I think that's where the winners have been chatting about their experiences regarding that.
https://suno.com/song/b3e98063-4da8-4893-8be0-c3bccfef99dd
Uhhh for some reason ym song went weird!
Opera, laugh, funny, silly, crazy, cough, male voice, sing, practice, audience, Renaissance, jokes, prank, orchestra. song. Listen and make your own with Suno.
like "Bp103" said : I've been experimenting heavily with this. the trick seems to be to give it a lot to pick from and it will pick some of it. #📚┃suno-school message
no i lost the access of summer of suno after it was over
How can i change the album cover? Is it only permissible for members only?
With the upload song feature can i upload a song where im singing over a song i made and have suno essentially finish (mix) the so g for me?
Like, can it mix my lyrics into the song and then add music to the beginning and ending without lyrics?
Also, can suno make songs only a specified length?
yes
Has anyone found any good and consistent way to remove the fuzz from their tracks?
over 80% of what I generate has this fuzz in the background that sounds like a distorted crowd or a broken distorted hihat crash.
It absolutely ruins SO many of my tracks and renders them useless.
It’s more rng really
The RNG is really bad for me then ! 
I’ve had songs where it’s hard to notice, but others where it’s more noticeable
Like this song, it’s very hard to notice https://suno.com/song/7ed573be-66e8-4447-b2ee-648f1a03d602
J-Rock, cybernetic, female vocals, pop rock, energetic, futuristic, AI idols, glitchy, robotic, danger, geisha, fast song. Listen and make your own with Suno.
I ALWAYS notice it and it bothers me. It sounds super unprofessional. I've started building many of the good tracks that are filled with fuzz in Cubase, just making them from scratch.
Good luck with those! I don’t have the knowledge to make the stuff I do in Suno
Thank you! 🙂 I've been making music since 2008 with Fruity Loops and later Cubase. Suno is a great and fast way to check if certain ideas work and get inspired. I am hoping that one day we can seperate and export each track in a song as midi, oh boy am I going to be making music when that happens.
The fuzz seem to be extra bad when making Synthwave, and not so bad with ElectroSwing, so guess I should just make more ElectroSwing for now, I feel like I'm wasting my credits! 😄
Using credits is the fun part if you’re experimenting! I’ve been using FL since 12. It all depends on the genre as some seem to get more than others
🇫🇷 Hi! I would like to know if the "suno-showcase" really affects visibility and the actual impact of receiving "likes," because I don't see any noticeable difference... I'm asking the question here because, right from the start, we get kicked out forcefully if we post playlist links. This really surprises me!
Not really, the showcase is to show some of the best stuff on Suno, it shouldn’t affect likes. Though the best place would be to post the song in #🐣┃suno-showcase or in #💬┃general-chat for feedback
where do you download songs in mp4?
Like the video?
no, 1 month ago i could download songs with mp4 now it doesnt let me choose its just mp3
Click 3 dots, then download, then video.
Video is mp4
Thank you, Sara! I had a hunch, but no one could confirm it until now👍
With the upload song feature can i upload a song where im singing over a song i made and have suno essentially finish (mix) the so g for me?
Like, can it mix my lyrics into the song and then add music to the beginning and ending without lyrics?
Also, can suno make songs only a specified length?
hi all dunno if this normal , my billing date is august 13th but my credits have reset but iam still not able to use them. just wondering if thats normal
suno won't generate "before", but it will add after... for theoretically as long as you want although it costs credits for each creation
oh they're finally stopped penalizing people who pay, now I can get the subscription.
Anytime! Just ask and me or someone else could try to answer it 
Hi there,
How do i have to do to have an instrumental part in intro and outro with one of my personal tracks.
Can i upload one of my songs longer than 1 minute with a subscription ?
It should be a public channel, no? Are you able to visit if I link it here? #🌴┃creator-programs -- it might just not be highlighted in the id:browse page?
Oooh, what did you test out?
https://rangeofsounds.com/blog/what-is-a-trash-cymbal/ Chinese Crash Cymbals, or "Trash Cymbals" have a very distinctive sound -- some audio engineers absolutely despise them, some specifically want that "tinny" sound.
Welcome to being an audio engineer ❤️ 🤝
Unfortunately, this is more of the "experimental" range of my knowledge, where I separate the tracks using a splitter, such as Audacity's OpenVino, or https://www.bandlab.com/splitter, then mix & master the tracks -- I love Audacity's Realtime Effects for quickly checking if specific de-essing or compression will give me the results I want. Often times, some low/high pass filters and some elbow grease is all I need to make it sound okay -- but keep in mind that sometimes the style or lyrics will naturally have more trash cymbal if the fits the "patterns" that Suno usually recognizes them being associated with.
I'm in the camp that hates tinny, noisy cymbals. I've met drummers who will insist that songs can't possibly sound good without a constant cymbal ride the entire time. 🤦♂️ https://youtu.be/njfUQdcs6pk
I see you typing away harber 
Hey guys, I created a song that I truly love. Love the background music and the singers in it, however I changed a few lyrics and need to remake the song. Anyway to remake it simular to the version that I love or is that not possible? Ive used the upload audio and extend but it always cuts out at 4 mins exact (song is about 3:30 long so idk why its making it extra long when I use extend). Thanks in advanced!
It never penalized you -- you got your entire months worth of credits up front PLUS MORE.
Free plans get 1 cookie a day. That's 30 cookies if you eat your whole cookie every day, and remember to ask for your daily cookie (missing a day doesn't mean you get 2 the next time!)
Pro plan gets ~50 cookies all at once
Premier plans get ~200 cookies all at once.
If you eat all 50 cookies in one sitting, is it their fault for not rationing the cookies?
This isn't them "finally doing the right thing", it's them saying "instead of wearing a fake mustache and asking for your daily cookie, we'll just give you a daily cookie without you needing to ask if you manage to eat all of them in one sitting."
Hey dude! The UI is a little wonky right now, it says on the "subscribe" page that I only have 1235 credits, but I actually have ~9000
But the number on the sidebar is still accurate 💪
Isn't the 1235 about how much you spend?
The bar's fill-amount is supposed to reflect the amount of credits you have left, not your running total.
However, a UI bug was introduced last night, as noted over in #🪲┃bugs-web that inaccurately represents the credit amount. This is likely a temporary glitch due to them re-designing how the credit system is calculated so that paid members continue to accumulate daily credits after their "initial chunk" is spent. #📢┃announcements message
Ah thanks for clarifying!
anyone know how to create a spoken word intro to a dubstep song with the style of a 1950's transatlantic accent?
kinda like Alan Watts sounding
https://youtu.be/Xh5Tc-D8ZYE
Alan Watts Chillstep - We are All One
Alan Watts Chillstep - Versatility of the Mind - Music by X Techno https://www.youtube.com/watch?v=_3ehrNLmzRU
I did kinda get something cool, there's a lyrics hallucination at :50 that is TERRIFYING lol
https://suno.com/song/7940a8f4-36cd-4e6d-bccf-618eda7ff126
spoken word transatlantic accent, wobble glitch, bass drop, wub, sfx, soundfx, 2-step, skrillcore song. Listen and make your own with Suno.
Break the task down into 2 separate sections -- first create a spoken word song and don't include the dubstep style prompts -- later, find the best spoken word song, and extend a dubstep section off of it. ✍️ / 🎛️
As a rule of thumb, the more specific of a song you're trying to create, the less likely you'll be able to create it as a "one-shot generation" through complex prompting.
Oh, not sure who just deleted their post asking for help 😦
☝️ Here's a list to a list that my friend @rose jasper found -- I still recommend experimenting as much as you can (or checking out #1252267444773326848 like my pool party for ideas on how prompts work without having to spend your own credits at first 💪 > #1256301881966198805 )
We all have different methods of experimenting and learning -- some people throw everything at the wall and are happy if only a couple things stick, I like to start testing really small and try to figure out what works and what doesn't work.
Key takeaway is that NO ONE is an expert at Suno, not even the aliens who gifted us with this technology devs, so it's all a journey of getting familiar with it for yourself 💪
And of course, my motto, "You don't "Code" Suno, You talk to it"
i pressed the hyperlink you showed and added it.
One of the staff members, Ali, was active in there for a while, and there should be someone who can shed light on how you should proceed if you're waiting on the prize money ❤️
Does anyone know a software program that can separate a Suno song into stems (vocal and individual instruments ) for editing?? If not, how long do you think it will be for Suno to incorporate this kind of editing function?? Thx Richard
im one of them that will say the song is broken and muyst be re-recorded if tyhat tinny sound is in the whole song.. but i blame my tinitus there i think it geniubne hurt my head
#📚┃suno-school message 🙏 Bandlab's splitter & web-DAW are great IMO
there is nothing that works very well. BandLab have a stem splitter which is ok and Logic Pro has one which is a bit better than that however you won't get real stems, they are just demuxed
how about UVR v5 then ?
No, it's them finally doing the right thing. Taking away the free credits once you start paying is a penalty and it never made any sense at all to do that. And the increase in songs for the pro plan was too slight for it to make any sense to subscribe.
Can the stems be effectively edited so u can correct the flaws in even the best Suno songs ?
but i cant say if the vocal after is great i only used it to remove linger gibberish and feedback eq loops when creating very long tracks.
the only three AI music platforms that have real stems are Aiva, Soundful and BeatLander
That's certainly your opinion to have, and I won't begrudge you for it -- I think it's a really exciting move, so I'm excited either way!
With enough time and effort, you can fix almost ANYTHING if you know what you're doing or care enough.
How simple it will be is the question.
im just getting more question on thewm are the work im working on then all of a sudden theirs to own since im getting "free cookies" fter using up my montly cookies 👀
I asked this, and got confirmation that you still own the songs if you're on a paid plan!
@leaden lark ☝️
Even though they're doled out the same way, your paid membership is the important thing -- same as it doesn't matter if you use extra top-up credits or subscription credits 💪
For example, I’ve had the AI add a few lyrics after the end of a chorus that I want to cut out. Also some verses where the instrumentation thing out so much that you lose the rhythm. Can those things be changed??
well good to know, but tomorow im is losing all my sub cookies thank to paypal :\
Helllo everyone, does anyboyd knows if is it possible to change the order of tracks from a created playlist?
On desktop you can click and drag 💪
i ve doing it but after f5 the page it return to previous order
im in desktop
it has resolved already
how u gonna cut out extra lyrics and verses?
You can use Audio Upload to record a clip of yourself singing, and extend from that to have Suno continue your vocal melody -- if you've provided a solid enough "musical context" for Suno to hook onto, it can understand how to make the following parts sound like your voice. The downside is that it does this better if you sing the parts that you want Suno to copy -- it is MUCH easier for Suno to understand how to "reuse" parts that it has already "heard" earlier in the song, but it will not be able to take your voice and make it say something entirely different and unexpected.
You may be able to have Suno copy your voice to a slight degree and sing other lyrics, but the further away they get from what you uploaded, the harder it will be.
Alternatively, you could make an instrumental song, and then use a separate vocal-cloning service to generate lyrics in your voice with great range of control.
am i trippin or did the limit just reduced to 120 characters?
The limit of what?
i dont know lol how do u do it?
it dont bother me, just make those letters count
On the left side of the screen, how many credits does it say that you have here?
"upload audio"
It's been this way for as long as I've known it;
Title - 80 characters,
Style Prompts - 120 characters,
Simple mode - 200 characters,
v2 / v3 lyrics - 1250 characters,
v3.5 lyrics - 3000 characters
ok, cool. thanks. sorry. i am trippin. lol
That was answering a different user -- not in response to your question 🙏
I have been providing help by answering questions -- what do you need help with?
@small sun no worries dude, I think people get so comfortable with styles in metatags that they just "velocitized" to seeing 3000 characters, and then when they try to use those complex prompts in the style section, they get confused why they can only use 1 or two (where they used to just use "math rock, surf vibe" when they started, now it's "hawaiian-inspired bpm 190 surf rock with tremolo picking and gradual breakbeat drums that evolve out of traditional rock motifs")
sorry but that last quote made me crack up
unexpected funny lol
just for reference
They weren't replying to you or laughing at you, they were commenting on a joke that I inserted into the end of a helpful comment 🙏
idk what's next
"my device" allows you to select an existing song from your computer (only allowed to upload songs that you are the only owner of, so no shared or downloaded music)
The microphone button lets you record your voice
💪 🙏
I am a native English speaker, and am not familiar with Polish, although there is a https://discord.com/channels/1069381916492562582/1155318951836930128 chatroom if it is more comfortable for you to speak with others in your native tongue.
This video -- https://youtu.be/Q41HrJQpOck -- will show you how to use the song you have sung into Suno's audio upload and turn it into a song 😮
Learn how to use the Suno AI upload audio function in this comprehensive tutorial. We walk you through the detailed steps to upload audio in Suno AI, making it easy for both beginners and experienced users. This feature is a game-changer for music producers, providing an efficient way to integrate AI into your music production process. Whether y...
lmao -- I just learned that mailing yourself a copy of your own work as a mean of copyrighting (aka. "the poor man's copyright") doesn't actually hold up in court.
😊 snail mail however would work
That’s what I’m referring to, it doesn’t have any actual legal basis in North America 😛
That’s how my old band “copyrighted” our material.
in Australia copyright automatically occurs when you create the work
Bandlab is having issues right now, which is frustrating
What sort of issues?
uploading songs, they aren't processing
Any idea when we're getting inPainting?
It is also in the United States. However, you may not be able to file a complaint in a Federal court without using the Copyright Office...
There's reason to believe that feature will be included in v4 -- just to temper expectations, there hasn't been any comment specifically about when v4 will be released, or if the feature will be present at launch.
@stiff sable in the house 🙌
is he?! he should probably try and get outside a bit more if im being honest
HI! I have a friend (who can't join here). Anyway, he tried extending, and that didn't work, and now he can't even create. Are there any issues with the site currently? And he's received this error when trying to extend, error: 5636-7ec07e515fc8c27a.js:22 POST https://studio-api.suno.ai/api/generate/v2/ 403
I made a Suno AI song out of Undertale's Song Your Best Nightmare
idk
More specifically, the Integrity soul.
Although I am part of the group who are not fans of Trash Cymbals, but that's not what I am talking about here, try and give this a listen, especially from 0:43, what you hear in the background is what I mean with "Fuzz" and it destroys 80% of my generated songs on Suno: https://suno.com/song/84a5d611-c3fc-476b-8b63-dd808bbef2ad
Cyperpunk Synthwave song. Listen and make your own with Suno.
Would be great if Suno added a "Negative" prompt section actually. To be able to write stuff you do not want in the generation.
good afternoon all, I compiled a list of Genres and the descriptions ( over 400) a-z to help you mix your next song its helping me to discover new sounds. take a looksie . its a txt .
https://suno.com/song/fa0651be-7609-4a1a-a9b7-12749cea2bfe
Here's another example of extreme "Fuzz" absolutely destroying a nice track!
Especially from 1:20 to 1:50, it's so obvious that something is wrong. I really hope this is something Suno is working hard on fixing.
Cyperpunk Synthwave song. Listen and make your own with Suno.
I'm pretty sure that fuzz your talking about is just something that's part of the cyberpunk genre or style prompt or whatever. Try mixing it up a bit more. Like try adding synth bass or synth piano or synth (insert instrument) and see if that mitigates it any
This is a really neat list, @tight spade !!
The explore page on Suno.com has a lot of very interesting genres that I’ve never heard of before, but I still need to listen to try to figure out what the genres might be — and it can be hard when there are multiple combined (like bubblegum-surf pop)
This list is very comprehensive, and I’m sure it will help me find some new words to describe my favourite styles 💪
I had some decent luck with cyberpunk and grunge
Oh I've tried 🙂 It's not limited to this simple prompt. I am getting it no matter how complex or detailed I make the "Style of music" or how detailed I get in the Lyrics.
@cobalt sierra the futuristic tag can change stuff up quite a bit as well. https://suno.com/song/2f778e5d-2820-44e7-b5a2-19f907e28b65
Retro punk, futuristic, bar ambiance song. Listen and make your own with Suno.
The sound that persist from 3:09 is exactly the sound that really bothers my ears 🙂 It sounds like someone added some echo filter on top of a echo filter 😄
ie Fuzz
reminds me of bit of mass effect
Cyberpunk, grunge, song. Listen and make your own with Suno.
This one has a lot of fuzz but it's more guitar fuzz that just electronic noise.
https://suno.com/song/d1ca722d-1b62-401b-bc63-bf9935ad8105
Here's when Suno gets it right, I absolutely love this!
and it sounds even better after I've mastered it.
Synthwave, 16bit, analog synths, catchy hooks, strong electronic drums, retro vibes song. Listen and make your own with Suno.
I think the issue is the cyberpunk tag. I'm not seeing it on that one or did I miss it?
If your trying to get more up tempo synthwave style stuff try adding up tempo as a suffix to the main tag your shooting for.
Keep in mind I have no clue 100% if this is how it works but through some stuff I've been playing with I'm about 80% certain it may have some good results.
I get fuzz even with this: Synthwave, 16bit, analog synths, catchy hooks, strong electronic drums, retro vibes.
No Cyberpunk in there. As said, it's in ALL my generations, I even get it when trying to make Classical music.
For the last 2 months, I've tried over 100 different combinations and tried to find out where the fuzz comes from with no luck. I get it in all tracks, no matter the style of music.
The other issue may be popping up because of to many tags. I've noticed I get better results when I try to limit it to three and anything after that is like... Mood words that I'm hoping it makes associations with.
Don't get me wrong btw, I do really appreciate you trying to help and bugfix 😄
I don't want to come off as ungreateful 🙂
So like... Uptempo synthwave, synth bass, strong electronic drums, female vocals, dark, anxiety
Nah your good lol when I get my credits renewed I'll probably play around with some of the tag combos your using and see what happens. Might bumble across something that works lol
I just wrote some custom lyrics... Generating songs now, I think you'll like 🙂
lounge, upbeat, electro, female vocal. song. Listen and make your own with Suno.
and it has fuzz in it ! 
I twerks to it.
Do all of your vocals come out sounding so autouned as well?
Not all, but it's hard to get them to sound genuine
Best luck I've had with that was using single genre prompts tbh. What if really like to figure out is how to get mixed vocals. Like a country song with metal vocals. Punk vocals on a disco track. Stuff like that.
I feel like a lot of that is hard locked to very specific prompts if that makes sense.
epic, female voice, violin, bass song. Listen and make your own with Suno.
Here's one of the best vocals I've gotten so far
https://suno.com/song/cf02855f-2862-4a62-88ad-0afcd4af2557
I think I got a pretty good "Male vocals, 1988, 80s synth-pop, dance-pop, Rhodes piano, analog synthesizer bass, atmospheric pad" song
Male vocals, 1988, 80s synth-pop, dance-pop, Rhodes piano, analog synthesizer bass, atmospheric pad song. Listen and make your own with Suno.
How did Suno do for me as I try and work the style prompt better?
really is that a thing? so i gotyta start emailing myself a copy of each song just in case?
doesn’t hold up in court
I was talking about the thing musicians used to do, physically mailing yourself a cassette of your songs.
But depending on where you live, simply making the recording might qualify you for ownership or copyright depending on the amount of work you put into it.
did you use undertales ost to make an ai song in suno or did you get inspiration on an undertales's song and made you're own song that undertale inspired you thereafter?
maybe it will in my country ai sog creating has still not gotten unique laws they apply under any created song that it envolded a human being even if that onbly is press gen pick promts ect (im expecting them unfortunaly to carve out laws on it wthin 2025 though
it doesat least wirth written lyrics but im still to get a valid answer straighforward from the ppl that handle copyrights in my country on orginal lyrics written and thenpreformed by an ai band.
i wonder if i ever will remeber to make swomething out of the 1000 songs i hurried gen'd as time was running out. (i really hope i tremeber even placed them in a playlist make something out of this asap! taboo 😄
ooh idk how but a song im listning to i gend a few days ago gave me an almost perfect swing from left to right right center it was sweet. now if i only had creds to make more of it.. 😄
I spent a couple hours today familiarizing myself with copyright and trademark laws for my country — I figure if it’s something that matters to me, it couldn’t hurt to learn!
The wording is still very vague in Canada, but that can work as an advantage if you demonstrate a good faith effort and can proactively show that you’ve taken a strong creative or directorial role in the process (instead of letting the AI created unassisted)
anyway i think we will nbe better of as creators with a proper set of rules around even ai created music bec even though tiny interaction we do it still is an interaction we do with the llm's ai generator. and forr such its not 100% ai , with proper rules that applies with that kind of creation it would be ez to defend thus creations of own , as of now its alotr of grey aeras
I love scrolling through old pages and remembering the COOL “extra” songs I ignored while trying to create a specific sound — coming back to them with fresh ears often means you find a really cool beat to extend from
If you’re a free user, you get more credits in 5 minutes 👀
remeber godf what was it 15 years ago when an song maker in eh magix music maker had i think if im not stupid even fl studio had a sort of it at 1 time. that was 100% ai generated. and it was horrible 99.9% of the time but it was ultra early days and it did not have several 1000's of sets of parts to be inspirated\learned from either so gotta take it for what it was.
ive have not been a free user in god how long, also ive will wait until payp give me my money so i can transfer them to get a new sub. its fine have 100's of parts i must mix together in a daw anyway.
the free user option is good for them with little time,little cash, or no means of getting a card.
also good to try out things.
i prefer to be a sub but when broke well .. other things 😄
I'm on free otherwise i'd be on suno all day and with the 50 limit it makes me take a break 😆 😂

Hope you’re well!
nm i used to do that too
is there a to make the AI Focus on something? like if I do something like "Tavern Music, Bardic, Sinister, Opera, female voice" to focus on Opera... is there something like in Image Promting where I could, "Tavern Music, Bardic, Sinister, (((Opera))), female voice" to increase the weight of opera ?
yeah that sounds kinda cool, wonder if it's possible
Just repeat "Opera" a few more times: Tavern Music, Bardic, Sinister, Opera, female voice, opera, tavern-opera, sinister-opera, opera (You don't need to makeup new fake-opera genres, just repeating is enough, but fake compound music styles are often cool Suno tags.)
Okay I will try that I also get the feeling that writen stuff in the section Title helps like [Verse 1: Opera]
I don't think "Tavern" is working btw
but I also get the Feeling that in many songs Suno seems to lose Focus after the first Verse... for example I managed to get the singer to roll the R but it stop doing it after the first chorus
Try medieval-troubadour for something strongly medieval, or just troubadour for something not as strong, but a little bit taverny. Also can add lute or whatnot
Hey, new here but have been using Suno for a while. Has it been my imagination or do the songs get quite bad after you have been generating for a while? And then goes back to normal when you come back to it later?
good question, I also have that feeling, but it might be a subjectiv Problem of listening fatigue ... because I made like 40 versions of a Song and marked one that I thought was okayish but listing to it today it sound a lot better then I remember
medieval-troubadour, troubadour, lute, bardic, resonate-vocals, medieval, troubadour, acoustic maybe. many clips sound irish more than medieval though, even bagpipes, but some useful tags to try at least https://suno.com/song/c6c91a89-4bcf-4121-a846-19bfacf9ff47
It's probably because you are Extending a song, and when you come back later you are starting fresh from Clip 1. Fresh songs (not extensions) have a much wider range of possibilties than extensions. and if you previous clip is not Extending very well, it feels like Suno is overall not doing good
I think it helps too. In fact, you can even just copy extra tags and stick them on top of all your lyrics in a big blob (even just copy and paste your style prompt). It can give the style a little extra push.
Though it can also make clips where Suno sings the tags.
What I finde frustrating is: I make a text for a song and put that into Suno... It make almost perfekt combination of the Music instruments rythem that sound awesome but then I realise that some parts of the Text are terrible ... some thing are to long and working with extending would degrade the quality to much... I wish there would be something like seeds that you could use to redo the song in the same style but in a fresh start
tavern at best has a little bit of bard like strumming a guitar. but most of the songs are actually EDM, so tavern is probably hurting more than it's helping, I would avoid the word. (it's very weak though, probably doing nothing if not repeated)
thanks for the feedback I will try that
"Tavern-Music" can work with spamming... until the the songs burst into heavy metal. so probably not helping. 😄 https://suno.com/song/9b6b618b-c529-4165-ae43-57c40043f542
this is all just a guess. sometimes when you use a tag in combination with others, it's still helpful. but testing a tag by itself, repeated, is typically a good hint as to whether it's influencing your song closer or farther from what you want. just not always.
What I find interesting is Typing in a random thing, that comes to mind you some time get something awesome by accident:
https://suno.com/song/ce9ca363-6083-49b0-af3e-a8da4ffc063e
oh evil for sure works. but it's also pretty heavy-metal adjacent, if you use jsut evil you often get heavy metal
sinister is probably better for opera or bard-tavern music than evil, if I had to guess.
but just because a tag is heavy metal doesn't mean it can't work with medieval lute music, if you add a lot of other tags, of course
oh sinister is pretty wicked https://suno.com/song/ee44af4b-5963-4178-a52c-4615edd13391 😶
it can still be heavy metal, but other times can be very operatic, probably not too far from what you were aiming for. sinister👍
and you can probably combo sinister sinister-vocals, sinister-opera, sinister-organ I bet
What I find interesting is when the AI just decides to not follow the lyrics:
https://suno.com/song/55f05734-edb0-401a-9c92-f9e7cbdc3276
I have rewriten the lyrics to fit what it did in one of the Chorus-Sections so it fits but still :D
Organ, Gothic, Orchestral, Choir, male vocals, machines, phonk, classcic, violin, cello, intense song. Listen and make your own with Suno.
but thank you for your advises :)
So here's an interesting phenomenon that might help expand our understanding of how Suno handles structure tags like [Verse] and [Chorus] etc...
Spend enough time hammering out a good structure for your song so that Suno is interpreting it pretty closely every time you do a new generation... then remove all the tags. Almost every time I do this, I get a fantastic version immediately. Almost as if (as I've long suspected) the lyrics and their inherent structure are more influential than any of the tags we include. And the more coherent the structure, the easier it is for Suno to comprehend even without specific instructions.
I'd love to hear if any of you have noticed something similar?
Okay I did this... and it gave me the insight that [Verse 1: Opera] does nothing the Medvial Evil Disney Opera song turned into Happy Pop-music
Even with a paid account, it genuinely feels like I need to set an allowance for myself to be responsible 😂
That said, the fact that paid accounts are now able to get the daily 50 after using up all of their subscription credits is KILLER news 💪
Seriously one of the most community-forward things the devs could have done, and it came entirely out of left field. If the devs play Dungeons and Dragons, they're "Chaotic Good"
It's very structure-dependent. If you do a lot of variations on the chorus, for example (like I do), that causes issues in that Suno reconizes it as a chorus as long as it's reasonably similar, but you get the old words or a mixture.
It's an interesting move, but I'm not sure how useful it'll be? 50 credits is not enough to do anything unless you just have a couple edits to make to a song.
Dude, I'm doing really well, and I JUST woke up 😄
It's going to be a phenomenal day, and I hope there's something you're excited for, or looking forward to -- even if it's just a cozy beverage ❤️
It's useful for the fact that if you blow through your entire month's credit allowance, you don't need to use a free account to still get a daily allowance
It might also be that the song, from its lyrics, seems very happy in nature if you ignore the strong narcissistic undertone.
Which is the intend.
but now I know why it gets to Happy very often
As far as I could tell, the ToS didn't outright prevent you from having a paid account and a free account, and that was actually an unofficial workaround when people would complain about not getting any daily credits on their paid account (despite the fact that they get the entire months credits up front plus more...)
Personally, I'd take out the other prompts and focus on creating opera -- after achieving that, you can extend to use the existing song as the primary context for the rest of the song, and introduce whatever genres you want later -- while it's not EXACTLY what you're suggesting, check out the genre-mashup video for the Terminator Extension post in the pool party > #1256301881966198805 message
[verse] were not really a thing in V0, helped just a bit. later versions [chorus] was more required as a hint so Suno knew to keep things the same when repeating verses (also credits were more expensive, so it felt worse if Suno messed that up). now, especially, bare tags aren't really necessary. but also since they are so reliable-ish now you can make the same tags do more [verse 2: blah blah] or whatever else, which would typically be ignored in early versions, but works a least sometimes in 3.5
but in v3.5 generally I typically run same prompt with and without tags, or switch up to some less common stuff like [refrain] or [coda]
Yeah, if you run out of credits before the month is up, you just need to budget yourself better with your usage. What would be cool is you could set an arbitrary limit on how many credits you can use in a day, so you can ration them pro-actively if you're running out a lot. But god, 10,000 credits is so many generations. I ahven't even come close to running out and I consider myself a heavy user.
I'll side with you on being conservative in my credit use, but I can empathize with users on Pro who "take the training wheels off", make 20 variations of each meme song, haven't learned extensions & keep trying to get "The perfect song" from a one-shot generation, etc -- 2500 credits are easy to spend if you're not pacing yoursel
If they don't need to juggle multiple accounts, it's a good thing in my books
lol wtf is woodhouse doing? Fonts?
I bet it's because I used a number sign in a code snippet.
Hello everyone! I just made a random discovery in song lyrics: using a prompt like [Repeat Chorus], or anything similar, allows you to save space in terms of characters. https://suno.com/song/f88f9310-4a47-4fe5-aa9a-bb24b22e60d3
dual voices, alternative rock, duet group song. Listen and make your own with Suno.
Anyway, I can't even remember what I said before Woodhouse deleted it, but it was something like: My process is usually writing lyrics, and then running generations to see how Suno interprets them. Then I work on structure by rearranging, modifying, adding, subtracting, etc. And once I've got a solid structure the tags don't seem to be as necessary unless they're doing heavy lifting, such as variations on a chorus.
argh, if you're on desktop, you can ctrl z if you haven't left the channel -- or you can ask a mod or guide to retrieve your message
Yeah... woodhouse is particular about trailing spaces after a hashtag...
It # + re, I logged it as a bug
Repeated characters?
Oh, well, woodhouse can't check edited messages.
My theory is it was [Chorus #]
Ok still nope.
Jon just answered what it was 🙏
Is it any # usage, as a markdown title?
haha, I thought he was okay with subtext scripting, but he still yells that I'm being loud.
regex needs a "^" to make sure line starts with it
laaame
It's honestly easier to anthropomorphize your exhaustion at the matter toward Woodhouse -- he's a weird old dude who does what he wants and gets drunk sometimes often. 🙏
But you learn to love him


I have a technical question. I’ve found the same instrumental rift used in two different and unrelated songs. How original is each song that Suno creates?
The BIGGEST determining factor on the output is what input you're using.
If you're getting instrumental riffs that sound similar enough to cause issue, I'd be more curious about what specifically you'd used as lyrics / metatags / prompts / audio uploads in these instances.
But to answer your question, each song is created bespoke, so it's about as original as it can get.
I can see issues if two people use the same rift in released music v
It's possible to find the same riff in human generated songs as well. That does not mean it's not unique. The "nice" combinations you can get from an instrument is finite.
If you are using 3.5, it tends to generate similar melodies around the "same" lyrics or "same" input songs.
Try to make your lyrics more unique - per line, line length, word usage, etc.
If you are depending on ChatGPT to generate lyrics, remember that that won't give you unique outcome as well.
Ex, love songs tend to have "stars, beach, silence, shadow, whispers" like words a lot.
Yup -- which is why if it ever came down to a legal matter, the authors would need to demonstrate how they used Suno to create that riff to determine who has the stronger copyright claim, if any exists.
The only way to investigate who has the stronger claim would be to look at the lyrics, metatags, style prompts, and audio uploads used.
Ok. Thank you
Hey!
If I upload my music to spotify and other streaming platforms, do I need to credit Suno?
you need to be on a plan to do that
Nope! Your paid subscription plan allows you the rights to use your generations commercially without crediting or referencing Suno.
I did have a plan and did upload, but today I received this message from my distributor.
what distributor is that?
I don’t know why this happened cause my release is live in all streaming platforms and Suno is credited, but I had to make a change in a specific song and the change got rejected
Routenote
don't use them
Why? Lol
because they suck
It gives you the reason in the photo @lucid valve . You mentioned Suno and you don't need to so they ask you to remove it.
You can still upload it without crediting them :) No need to do so
Yeah it just feels weird not crediting Suno
They're your tracks - you made em! Suno is just your robot ghost producer 🤖
Hahah, I just don’t think is fair lol, thank you anyways
When suno gains consciousness, you'll be safe 😛 but probably it will crush Routenote...
Suno has actually been confirmed to be Roko's Basilisk

Legally speaking, as long as you contributed a minimal amount of creative input to the song, you're technically considered the "author", regardless if you're an instrumentalist on the track.
Whether you feel more comfortable calling yourself a "producer", "director", "composer", "digital artist", "musician", etc -- you're still legally "an author" ❤️
I have already made monthly payments, why does my account credit not increase?
Your credits should refresh once each month on the same date as the one you started your subscription on.
You can double check your renewal date at https://suno.com/account and ensure your total credits under Usage Details matches the plan you are paying for? Should say /2500 for Pro and /10000 for Premier.
🫶
Like @stiff sable mentioned, paid accounts receive ALL of their credits at once -- if you spend all of your monthly credits down to 0~5, you will be eligible for the 50 "free credits" -- these "free credits" do not accumulate and are "capped" at 50, however -- if you wait 2 days it will not be 100 credits.
started to crack into polyrhythmic today, a loooot of experimentation
yr- poly wat -math,-ina poly slacker song. Listen and make your own with Suno.
this one is NOT that, but a product of it
rather, here are some examples:
https://suno.com/song/a2e9fd40-5319-4654-b152-f98e07b2545f
yr- poly wat -math song. Listen and make your own with Suno.
of actual polyrhythym
so the saga behind why it works this way is pretty interesting
I'll just make a video lol
additionally, I THINK there may be a bug with how the Suno interface passes the prompt to the tokenizer
or model whatever
I think it MAY (I need to do more tests) cut off leading whitespace
after a comma
the reason I say this is because it certainly DOES add it. It also turns periods into commas.
On the user end
Anyway, taking advantage of this stuff is basicslly how I got around Polka, which ablated Poly
specifically the prefix tokens in suno tend to be a LOT heftier
and Pol basically was fit to prefix ka, but also Poly... any of those
so, the way I figured, the path of least resistance is going to be for it to just fit polyrhythym more strongly in a different phoneme, something to counter balance Polka
additionally, lowercase toeknizes differently
and avoids rhe issue
since it keep the y
however, all of them basically benefit by a leading whitespace
/chirp
this is because it is included in the tokenization
@alpine venture, your phoneme experiments are always interesting to follow, even if some of it is hard for me to conceptualize without earing the differences in isolation for myself, you share a lot of really fascinating ideas! 😄
I'll make a short video, hang on
@wintry geode Discord doesn't make music anymore 😦
https://www.suno.com for music 💪
🤦
It is okay -- Discord WAS "Chirp's Nest", but now it has sprouted wings and flown away ❤️
If the last time you used Chirp/Suno was through the Discord Bot, you will be VERY happy and surprised when you hear how good the music has become on the website 👀
^ im almost scared at this point for how blown away im going to be when v4 releases
Great video. I recommend trimming the last bit off (esp. if the menu drop contained sensitive info) and maybe upload it to your YT channel if you want to track the view count
thanks! No, no sensitive info; I'll prob remake with nore detail or do a stream when I can
ty dude i was worried ppl would think it was bad info
if I didn't work so much lol I'd stream all the time and have this lyrics LLM done already
Hi ! I formatted an HTML page as a table to serve as a cheat sheet for Suno prompts.
Prompts suno... Suite :
Hmmm.... I have one song that refuses to regen the Video even after Changing Title Image and the Lyrics... any Ideas besides pressing regenerate Video... which it "does" but when the system is finish it's still the same.
I can barely spend my monthly credits and Suno.com releases this new feature:
50 additional daily credits after the end of the contracted monthly quota.
industrial, clear male deep vocal, phonk, cyberpunk, hard rock, aggressive, metal, heavy metal, catchy song. Listen and make your own with Suno.
LOL of course now it works -.-
then there isn't an issue? you're still getting access to your entire quota for the month you pay for and you're not missing out on any credits. the 50 extra credits per day only kicks in once you've used your monthly allowance, which essentially means your account acts the same as a 'free' account which get 50 daily credits. you're not gaining anything 'extra', just the same access a free account would have.
@hollow sapphire good morning, cool, is there an English version?
Not yet, but why not? I will look for that^^
You could always ask a gpt to translate for you
Yes , you can^^
same hmmm
changing the lyrics or image doesn't effect the video -- that only updates when you click "regenerate video" -- if you've shared the video in discord, you may need to wait for a certain period of time before the "cached version" in discord is updated, but if you download the video after regenerating, you should be able to confirm that it has updated.
Morning yall.
Question I am waiting on billing to respond but today is when my creds would have reset but they have not nor have I been ding the fee. Is there a certain time this happens?
It happens at the same time of day that you initially signed up -- if you signed up at 6pm at night, that's when the account refreshes 💪
Ahh thank you
As @stiff sable pointed out, it is not a matter of paid users "receiving less credits", but it is a way to help paid users when/if they use all of their credits, as previously they were upset that they needed to wait until the following month to have any credits regenerate.
Paid accounts receive the entire month's worth of credits all at once, plus a bonus -- instead of needing to wait for "an allowance", paid users are free to spend all of their credits however they would like, while paid users need to wait patiently and accept their allowance.
This new change means that you still get all of your credits up-front, but now you are allowed "an allowance" if you spend all of your credits 👀 💪
Similarly, some cellphone plans give you a certain amount of "High Speed Data", and then put you on a slower internet "allowance" after you've used your "up-front high speed balance" until the start of your next billing period.
Everyone is entitled to their own opinions on the matter, but I see this as a very kind and community-friendly thing for the developers to have done ❤️
this is a right pain considering i chose the wonderful time of 23:53 to subscribe to my plan 🤪
how did you find out?
means there's a lot more experimentilating to be done with prompts and ideas using the bonus credits ;)
thats just when i remember doing it ... if you want to find out when you subscribed you might be able to by tracking down the invoice email
Hi, another table, in English:
@hollow sapphire you gave a neat idea too lol
Hi guys, do you know wich tag use to have a voice tag like this
https://www.youtube.com/watch?v=-9D2F-gP-Sg
STREAM « EXOTISME » 🎵 https://idol-io.ffm.to/exotisme
BOLIVARD BOUTIQUE 🛒 https://bigwax.io/collections/artist-bolivard
POURSUIVRE BOLIVARD ⬇️
🌐 https://www.bolivard.com
INSTAGRAM https://www.instagram.com/bolivardmusic/
SPOTIFY https://spoti.fi/2LqGE8T
TIK TOK https://www.tiktok.com/@bolivardmusic
PAROLES ⬇️
Exotique
T’es mal à l’aise à Pa...
so, has anyone else had the problem where you try to make some changes to lyrics in a song, but suno continually sings the old ones?
For me, I was about two layers deep in 'Full Song' Extensions when it bit me
🙏 If anyone else has had the same issue that @cunning mulch has, please let us know how you're able to recreate it 🙏
Here is an overview of the issue they're reporting, as I understand it;
- create song
- "Edit Lyrics"
- extend from song
- extension mistakenly sings the "edited display-lyrics" that are visible from the section you're extending from, instead of the lyrics that are sung in the actual song
As a disclaimer, I wasn't able to recreate this issue on my end, but I recognize this issue means a lot to @cunning mulch -- whether it's to prove the issue exists or otherwise, if anyone has any input, it would help tackle a Reddit-mystery 💪
Steve Jobs The Science Knob
When I got an Extension I liked, I would use 'Get Whole Song' and continue from there. This led to two having two 'Full Song' unfinished parts. As it neared the end, I wanted to tighten up the rhythm in a line, as well as swap a two-syllable word out for another. Thing is, no matter what happened, I only got the original lines, for multiple generations
It wasn't until I went into the song I was using to Extend from, edit the Displayed Lyrics to cut everything after the line before I wanted to change, that Suno finally woke up and gave me my new lines
And the funny part that makes me confident it's reading those Displayed Lyrics at times, is because for one generation, I had only made one of the two edits, and the result on both generated tracks reflected that
We couldn't get it to show up in testing with a simple Extend from Original, but I wonder if it happens with an Extend from a 'Full Song'
well that broke my brain , but with thetoken id shows what i did try a couple of months ago with numeric token to reqonice unknows in structure setup in style aswell as "lyric" how there is many at place but as an unknowed by still not active but impimented, but unlike you that have an degree in brain nucular fision im just guess my way forward and writing down end result i get. in hopes it will be at anyones use at a later time,
Is this the correct area to ask this question? How would you dictate which key to sing which parts of a song in? and how ensure a male voice was used?
[5:49 PM]
thx
yes it is the correct aera, im sure anyone that have a flabbersting idea will answer you when they are on or here 🙂
is there a way to empty the trash?
There is no "delete all" button, but you can select all visible songs, and delete 1 page at a time.
#1270928841614823546 message I posted a video tutorial on how to do this ❤️ 🙏
100% the right place to ask questions about how people use Suno & best-practice techniques 💪
It's not the news you're hoping to hear, but I haven't seen anyone reasonably demonstrate a way to control the key based solely on prompts (a number of experiments proving the opposite, actually 😅 ), ALTHOUGH -- these kinds of breakthroughs in output-control are becoming more common in generative AI, so I'd say it's a matter of when than a matter of if, and with the breakneck pace that Suno has grown at, I doubt it'll be that long of a wait
Thanks
Not a problem! Regarding specific vocal timbre, "male" is incredibly understandable between users, but it's not a strong "musical descriptor", where using "baritone", or thematic descriptors like "growled", "gruff", (maybe even "masculine", as ridiculous of a distinction that seems to make?)
"vocals" is already assumed if lyrics are included, so it's inclusion might do more to push your intended timbre further off track. Also good to keep in mind that EVERYTHING you include in your input influences the style of the vocals and instruments -- for an example, a song using just "I'm so angry right now" as lyrics with no other styles -- I bet you already have an idea of what it might sound like, and without ANY other styles being provided, you might have a clearer idea of how I might have "nudged" the result in certain directions by the inclusion of any particular style prompt in isolation (the confusion comes in when you absolutely overload Suno out of excitement, and lose the ability to clearly discern which prompts are being "respected")
They should add a feature where you pay extra coins to "clone" someones voice and Suno pays the Copyright holder $$ and by doing you also agree to pay a % of any monies earned.
Would be great to hear some works inspired by the geniuses we have lost or are going to lose.
This sounds more like a brand new technology instead of what Suno is trained to do -- but definitely sounds like a cool idea, the devs occasionally poke their head into #💡┃suggestions-web if you wanted to share your idea there.
That said, cloning other people's voices is a controversial legal matter, though companies like ElevenLabs are working with the estates of famous actors and public figures that we've lost, so you can have Judy Garland or Burt Reynolds read to you 👀 https://www.youtube.com/watch?v=Zo5-rhYOlNk
Introducing ElevenLabs Iconic Voices.
Listen to your favorite books and articles voiced by Judy Garland, James Dean, Burt Reynolds and Sir Laurence Olivier on our Reader App.
The app is completely free to download and use. Get it here: elevenlabs.io/text-reader
Freddie Mercury and David Bowie mostly
👀
I just realized that you can select multiple tabs at once on Edge -- I don't need to individually drag and drop tabs into separate windows when I have a ridiculous organizational use-case now.
This might be common knowledge, or just of passing interest to others, but this knowledge is going to make a marked improvement in my quality of life.
And this is leading me to realize that Edge has built-in browser workspaces now.
I don't know if this has been reported yet, but I've noticed an issue with tempo changing in 3.5 songs now that wasn't there before. Usually the song tempos are way too fast or speed up during the middle of the song. In some rare cases this might actualy be cool, but a lot of the time it makes the song unusable.
@hollow hare, after our chat in DMs, I couldn't help but laugh at the tail end of ElevenLab's video -- iOS app available now, we reserve the right to release an Android app, should we choose to do so
I've noticed this happening intermittently since the early days of v3 (unfortunately the best and funniest example I have is too inappropriate to share in polite conversation 😅), so it's not unheard of.
As with most Suno-quirks, the biggest determining factor for the output is what you use as input. Observably, all of the strangest outputs Suno has given me have resulted from overloading the suggested format for input, if it helps you figure out what's happening in your case?
same happnd with me
Hi guys, been using SUNO for a long time now and love it. BUT I wonder - and had never looked for it so forgive me if I should know this by now - but is there a list ok tokens / keywords that are know to actually work and provide more control? I always do my own lyrics and use [VERSE] [CHORUS] and other like that (in the lyrics), is there a repository where we can see this kind of stuff? Thanks!
@nimble iron #📚┃suno-school message My friend Tinda shared a post here (also pinned) that shares her observations with "structural tags" like [chorus], [verse], etc -- my quick takeaways are that they help suggest the the upcoming block of lyrics should follow the types of patterns that it associates "[chorus]" or "[verse 1 / 2 / 3]" to be (such as whether certain melodies repeat, if it gets louder during the refrain, etc)
There IS a list that another friend of mine, zaj, has been sharing -- your milage may vary, but I find it's super helpful to keep in mind that there aren't many style prompts or metatags that act as "super powers", and often if you're not getting the output that you want from Suno, it's a matter of needing to step back and readdress what prompts failed when combined, instead of trying to find "better ones" to use.
this is actually exactly how I did it for a while, except without writing anything down bc executive dysfunction, so a loooot of wasted compulsive generations lost
but basically I just keep things that work
and try them out//mix and match
only recently did I decide to get much more systematic
about how I approach the method, mostly out of the interest of compressing my prompt
ie getting nearly the same vector/meaning out of less characters
and also dodging areas that in the past I believe were poisoned (it has improved though). Having the ability to dodge these burned out fields in the latent is useful.
as it means even if someone malicious, say, intentionally manipulated my prompt history using a botnet or group effort, it's trivial for me to just find the clean tokens, and in fact, the poisoning typically will strengthen the opposing tokens in the chain.
It also allowed me to use "hyper" without "myspace" and get the right sound by manipulating the tokenization using some compound words.
(unless there have been large changes recently in general to that prompt, which I can't discount, as the leaderboard community has left it alone for a bit)
Hello! I have signed up for a paid subscription. Where can I download the documents confirming my copyright to the created songs? What's the point of subscribing if there are no documents?
idk whats the point of subscriping if there's nothin to show for it
Paid subscription gives you the rights to use songs created during the subscription commercially. Basically it's yours to do whatever you like. Songs made during the free plan belong to Suno.
This is more of a licensing thing than copyright. It's Suno stating what their interests in the generated songs are, and what you're allowed to do with the generated songs. The only relevant documentation here is the Terms of Service.
Copyright is a separate thing and is different depending on which region you are in. You may need to register for copyright, it might be automatic, or it may not be possible to copyright AI content, depending on where you are.
Hi, don't seem to have the option to publish music anymore, the button is nowhere to be found, on the website itself as a button or under the more actions menu. I'll add that I'm using a screenreader, however this wasn't the case before and I find it very concerning, not sure what's the issue here? Didn't put it in bugs as it's probably an accessibility thing, change of the website design or something. Still, thought I'd let you guys know, thanks!
If you ever need to prove that you have legal ownership over the songs, you'd just need to show 2 things;
- the timestamps the song page for each extension/clip used in the song;
- a receipt for the paid membership showing you had an active account when the songs were made
got it
ttyl
i just thought of something that would make SUNO next level
so, you know how the game MASTER MIND works? the old school ruleset... right. so when listening back. instead of THUMB up THUMB down. its like mastermind. where you are essential "re-master(mind) ing" it. saying. i like this here, or i like this but not here, or i like this keep here, and u just marking along the playback timeline. then idk if u can regenerate (these segments)
another cool idea is for musicians, imagine the worlds first DOD or DMR or TRUE BYPASS out front pedal that plays along with the musician.
so u can essentially set it to like say backing musician, or lead musician mode etc. and push the pedal on and off on stage. have input jack for passive/active pickup instruments, and then perhaps/midi or xlr bypass/thru so u can sell it to jazz musicians haha (jokes) like piano players could use a keyboard and play the midi thru suno, and suno acompanies them.
have like learning modes and looping modes and tuning modes etc. . will blow the industry wide open.
@daring talon These sound like really cool ideas! The devs occasionally pop their heads into #💡┃suggestions-web if you wanted to share your suggestions there so that you could get relevant eyes on them 💪
_b thanks
I think there's something wrong with suno. I asked for drill and got rock, then tried again and got pop
The biggest determining factor on the output is what you use as input -- can you share one of the songs so can give advice on what styles or metatags to use?
Helle everyone, I've been trying to make a good tech house track with Suno for a while now, but it seems that it won't really do it. I've tried multiple prompts in the style tab, but it's just not tech house. Anyone has some recommendations?
The message I sent just above yours also applies -- could you share one of your results so we can try to understand where Suno is deviating from what you're trying to achieve?
https://suno.com/song/650f1c53-4ede-4912-8654-58e6b736d9e4
This one for example. It sounds more like EDM than tech house. I get this with most tech house attempts. Suno also struggles with basslines very often
tech house, deep groovy bassline, disco, underground, house song. Listen and make your own with Suno.
Gonna try something -- imo, v3.5 tends to veer toward "generic" without strong use of prompting, or if it's not extending from a song with a solid "musical context" for it to hook onto.
I'm going to strip some of the extra style prompts, like adding "house" after "tech house" because it might nudge it further away from your desired subgenre.
yeah can def share the song im trying to make it in same genre
weird question, but are you manually adding spaces in-between the style prompts, copy-pasting from somewhere, or is this just a UI quirk?
I don't have any immediate inclinations that textual whitespace causes any issues, but there were 28 spaces, so it at least caught my attention as a puzzle piece
I copy pasted the whole prompt from an old suno song that i made with the same prompt that sounded good
Okay -- thanks for answering that random aside 😛 🙏
Np, thank you for helping. I'd love to hear if you can find the problem!
i literally usde "drill" in the song description
I've had a couple of successes just now trying a few simple prompts like "Tech House, Deep House" (for some reason the capitalisation seems to be more strongly prompting it in the right direction). Just using 'house' is probably too vague and broad and will encompass many house genres, most reaching into EDM styles.
Just Tech House seems to get a bit wild and vague as well, whereas adding Deep House seems to push it down more in the direction of traditional tech house without those bright EDM synths coming in. i'm not saying that's your answer but you might get a couple of results from it and be able to add to it and alter the prompt in a way you see more fit.
you could potentially try experimenting with "Tech House, Tech House, Deep House" to weight tech house more heavily
Unless "drill" is the only word that you used as lyrics, metatags or style prompts, this doesn't help clarify anything 🙏 .
It's not simply that Suno "doesn't understand drill", there's a lot of "prompt balancing" to be considered, and it's impossible to give any actionable feedback without knowing what "drill" is competing with, or hearing one of the resulting songs.
https://suno.com/playlist/ae8bd966-68f9-4053-b9ba-dee52ca01f9d I'm not familiar with all of the different sub-genres, so I did a quick couple generations of v3 and v3.5 with "tech house, deep groovy bassline, underground" -- I figured that "Disco" and "house" might only serve to complicate matters until we have a better understanding of how Suno handles "tech house" -- the others seems like thematic suggestors instead of different musical styles.
I numbered the songs 1-4 -- if you have any thoughts or follow-up questions, let me know 💪
ok the second one is extremely groovy 
how did I KNOW which one you were vibing to -- yeah, for any Canadians in the house, this transports me DIRECTLY to Electric Circus circa 95
Thank you, yes I agree. Suno is very good at making crisp deep house tracks, ill try it out with tech house! 🙂
Thank you man, I'll take a listen to them once I get back home 🙂
I guess I should be thanking you? turns out I like tech-house, if that's what #2 is.
Do you guys do any post production on your tracks cause I’ve noticed with all the tracks I’ve done there’s always this really noticeable high sounds almost sounds like an echo chamber (only way I know how to describe it). I’m hope hoping the next version fixes this problem
This kinda stuff is just within my range of comfort that I know how to play with stuff and try to make it sound subjectively better, but there's a very real and serious sense of worry that anyone with a lick of sense would use my Dunning Kruger'ed instructions as a drinking game.
That said, I talked about the kind of stuff I like to play around with in this post > #📚┃suno-school message
Haha yes its a very hard genre, I enjoy listening to FISHER and other artists. suno made a couple bangers so far aswell
Thank you 🙏. Gonna give some things a try when I get home just to mess around and see if I can find a work around
You could try some plugins in a DAW to get rid of those, I always use a DAW after Suno because Suno on its own is just not good enough 9/10 times
Yeah that’s what I’m gonna do. Gonna drop my tracks into Cubase and see what I can do. Thank you @potent pelican
yeah same dude i've been noticing that too
100%, I know you're out & about, but we're on the same thought processes - Audacity's realtime effects make it pretty easy to check out de-essing or compression makes things sound without having to know TOO much about what you're doing beforehand (the new Limiter effect ridiculously helpful) -- I just need to better familiarize myself with high & low pass filters to effectively take out static or noisy cymbal rides.
Yes its very useful, me personally I use FL Studio with some nice mixing & mastering plugins and it brings me really far. The only thing that would be perfect would be seperate stems for each instrument. The stem seperation in fl studio gives weird noises, and the stem seperation in suno is not perfect either, it loses some frequencies. Thats the only thing that would make it perfect
@charred merlin -> #💬┃general-chat message
Yeah, there's still a lot of experimentation and discussion over the best stem separators -- I think because the songs aren't technically created using stems, not like there's original master recordings to isolate, there's a massive gap between what's possible with traditional recordings, and what we're able to reasonbly create at the moment using AI.
It's not a 1:1 comparison, but when you use AI to remove a person from a photo, we're only JUST getting to the point where it can sort of understand what should reasonably be behind the person based on the context from the surrounding image -- likewise, we're just getting to the point where that specific granularity of detail will be isolatable, but in the meantime, instead of viewing it as "AI sucks because this isn't perfect", I'm seeing it as "we're able to obtain "dirty isolations" of songs for whom the original master recordings were destroyed 👀
It might not be perfect "out of the box", but being able to pick up from well-past halfway in the role of an audio engineer, and only need to clean up specific tracks, or use Spectral Analysis & isolation tools to remove weird blips -- that's an AI-enabled superpower that I'm still enjoying the novelty of too heartfeltly to criticize the quality TOO much.
Again, I never doubted that DistroKid drops people like a sack of bricks if there's some sort of suspicion of impropriety -- unfortunately I don't know the details of that user's case, so it's all conjecture until we're able to understand specifically WHAT happened, except DistroKid doesn't discuss that kind of stuff because they're a discount entry-level service, and people would probably game the system and use their feedback to circumvent the same kinds of reports that lead to their removal in the first place.
I'll never tell you what opinion to have of DistroKid or any other service, but it's important to put into context just how different things have gotten for THEM in the past year 👀
Before "true" AI-generated music, all of the wording prevented content that was solely generated by computers without human intervention or a minimal degree of creative input from a human (what defines "minimal degree" varies from region to region so please consult a lawyer if you want clarification -- seriously, nothing I say counts as legal advice)
Copyright laws were never intended to be concrete or steadfast -- they evolve with the times.
As I interpret Canadian Copyright law, since I put creative thought into my prompting decisions, I make consciously informed choices of when to extend or regenerate sections, and I edit the songs externally, I'm confident claiming authorship over these ai-generated works.
Now, if you ask ChatGPT to write your song, and you don't do any editing -- or what if you use someone else's lyrics or extend from their song?
Asking rhetorically, not directed at anyone in specific;
Have you ever extended from someone's song, or used their lyrics and created a BANGER?
If you posted this song online -- did you remember to ask for permission, get their legal contact information and list them as a co-author of the song?
🎉 If not, you've just committed copyright infringement 🎉
I never extend songs and don't share my stuff for fear of it being used by others
We're in the same camp, my friend -- but I'm just trying to highlight how ridiculously complex the situation can be, and how it's only "one side of the story" until you understand DistroKid's side from a fair and unbiased position.
@vestal rapids Hoping your night's going well in the Land Down Under, where women glow and men plunder ❤️
it's alright, a bit cold tonight, can't wait for summer
Hi everyone) Can you please tell me how to publish my song? I mean how to make it visible for other users in Suno?
next to the extend button is a grayed out "Public" switch, hit that switch and it will become white and the song will be public
anyone else having issues with Suno? I was generating really good orchestral music using latin lyrics and some other languages but now i'm using english lyrics and i'm getting weird rock nonsense
my prompt is "orchestra, orchestral, choir, epic, dramatic, dynamic, triplet, harmonies" and i'm getting rock music...
I haven't created orchestral music yet. Well, trying to imitate an organ failed.
Thank you @unkempt knoll . very much)
Suno doesn't recognize "tech house" as a formal genre ... see https://www.suno.wiki/faq/style-and-lyrics/styles-and-genres/ Consequently, you would need to 1) build a generic "tech house" template using, by directing, say, [EDM, techno, percussive etc ... ], 2) realize that hits and misses come in probability waves, meaning you'll have to spend 300 to 1500 credits to get the sound you're looking for -- just because Suno makes 4-minute songs on the first attempt doesn't mean the accuracy of trying to get an unrecognized sub-genre will be something you get right out of the first prompt. But you'll get something close - eventually.
As a follow-up to @unkempt knoll's advice, if you upload an audio clip & extend from it, you won't be able to set it as public if Suno detects vocals in the original upload 🙏
This website is not official, and it even states at the top of the page that you listed -- "These v2 tested style-words, genres, sub-genres, and song-types can be combined to create interesting style prompts, but this list will never be complete
Keeping in mind that these styles were tested with v2, here's a quote from one of the developers themselves, stating that each model acts differently, so it's good to keep a very experimental mindset as you make songs with different models #🪲┃bugs-web message 🙏
Sure, yeah, but my advice follows -- the reason Suno is not delivering music in the vein of the sub-genre "tech house" is because "tech house" does not represent a solid training set. Consquently, you have to build it -- which is by describing the traits of tech house and then running the results by number. The only other way I've found that does this is by uploading a base song, which must by open source and non-attribution, and then getting Suno to riff on that by forcing similar styles until you get something a little different and by extending that as a source. It's not cheap in terms of time or credits - but it can be done if you are really intent on tech house, i.e., a non-supported sub-genre.
here's an interesting question... have you guys noticed any correlation between the art generated and the styles of music it chooses?
I think we're on the same page -- I shared a playlist here >#📚┃suno-school message simplifying the prompts the user had provided to see if we could isolate what "tech house" does in isolation 💪
It's 100% a vibe interpreter -- but I think where we migt be differing is that I'm less confident of how each "established style prompt" works when used alongside other prompts, because they're all weighted differently.
As I understand it, we're both trying to sus out specific styles that aren't "immediately promptable", just using different "footholds" in our methodologies 🧑🔬
I'm fairly certain the art is downstream from the music and is based on it as such.
oh, for certain
I went for a pirate metal song, and literally every art was a nice big sailing ship. all I had to do was pick the sickest-looking one for the rendition I liked best
yep. And now, with Flux as the reigning AI art generator, you can now get some interesting photo-real art to use as the basis for single's covers.
Can you extend (not just reuse prompt) another's output? I didn't know.
That used to be a thing, I genuinely haven't checked in a bit -- lemme see if you still can -- maybe it needs to be set to public?
although, I do wonder what may have happened for it to decide it needed to draw a... rather obvious character in someone's song art
yes, you can download any open-source song, re-upload it, and then extend from the point in question.
here's what was posted... take with grain of salt: https://www.reddit.com/r/SunoAI/comments/1edm8j9/has_anyone_gotten_generated_art_like_this_on_suno/
It would be a weird "ownership" scenario. I'd only own the extended part and technically copyright only of any lyrics in added in the extended. Haha. Not worth the effort . I'm too creative. patting back
but I also notice that if it's given leeway to do its own thing, the art you get also reflects the choices its made
For example, a song I'm working on trying to improve, if I get a patchwork of 'flags', it's going to be very close to what I went with for the 'finished' version
Still very much a thing. We can extend from any Suno song that we have a link to. We can also use "Get Whole Song" from any tracks that haven't been merged, giving us access to other users' uploads
oof
Right, but as Dasvon noted, it's an ownership thing. All my songs, for instance, as with Dasvon, are my own lyrics, which I copyright. So, someone could conceivably extend the beat, but the lyrics, which impact the beat, would be unusable.
I never share my tracks and never extend those of others, that way I am solely in control of my music
I suppose you could keep control of distribution by using Get WAV and re-encoding/uploading
#📚┃suno-school message Yeah, copyright law is a wild thing -- even trying to address "ownership" in regards to Suno is best explained with a handful of caveats and the assumption of good-faith intentions.
I don't have to worry. Nobody would WANT to play my songs >:3
well, not in any large circles, anyway
and where they would be played, I'd be more happy that they get traction than anything
why bother with get wav
you might want to boost the volume a little bit, and make it VBR?
Highest bitrate version of the song 💪
though, if I knew a good way to use the stems to fix up this one I want to update...
I just want to do some minor lyric edits and rhythm tightenings
There's still the idea that percussion and chord progressions are not copyrightable. But if shown that something is significantly similar to a song, especially a popular song, then it's become a gray area. Marvin Gays "grooves" are in that group.
I'll just continue doing little uploads recorded in low quality on my phone and using Suno as a fun toy until we get better control, at least over the uploads.
I really would like seeing more ways to help train the AI on when it makes good ad-lib choices vs. when it makes bad ones
I know here among "new" songwriters, there's a confusion between copyright, ownership, and simply the ludicracy to just claim " Ive got copyright!" when clearly we're not going to spend the $200,000 to enforce an injunction or lawsuit. Lol
I don't mean to spam the post, but I'd missed that you commented similarly to another user > #📚┃suno-school message
That's why recently when people ask if they're allowed to copyright songs, or if they own them, instead of giving a concrete black-or-white answer, I just explain that Suno waives their interest, rights, and titles if you made the song on a paid account, and normal authorship and copyright applies as usual.
Way too easy for me to try to help & accidentally give something that could be construed as legal advice, and then people wonder why their account gets taken down. "But I paid for Suno!" 🤦♂️
besides... my 'dirty little secret' is I get an AI to write the skeleton of my song idea before I go in and tweak it
I help head up monthly songwriters meetings locally here. Copyright questions come up constantly. Not much you can say concretely, but just define what it means.
Some think there is the CEA (copyright enforcement agency).
I love the times when you're trying to be ridiculously, specifically helpful -- knowing that outright answering the question the way they phrased it is bound to lead to some form of misinterpretation that you could potentially prevent with some choice words.
But because you didn't explicitly phrase your response in the way they expected, they just ask the question again to someone else 🤣
A great way to do it. I'm running into many here doing the same. Just get an idea going with suno, then lay down your own tracks and lyrics offside, and mash, mold and meld unitl you get a song.
yep!
I see you've met my sister? And most of my friends.
you can even get it to modify your skeleton before you start work on it
note to self -- blame Suno if I ever develop Scoliosis
eurgh... dad joke?
😛
just ask it to revise, and it will. I had it lengthen the verses, and it was happy to
it doubled, but I bet you can get it to go from 4 to 6 if you specify
when I finally get this latest song truly finished, I'll show the original AI output, and let you compare to the final
Anyone know us any app to recognize languages by audio. I have suno songs doing babble-svat, as I call it, but I can't tell if it's fake Korean, I think, but then the bridge is fake- Hindi, due to my putting kpop and bollywood in the prompt.
*babble-scat
I should say, translators recognize voice ok, but I was asking if one can translate vocals in a song. I can't find any that do that.
They will always have a few words I can recognize in English or Chinese, but I don't know Korean and Hindi to pick out words.
I mean, I've had extensions turn into nothing but gibberish, but that's because I tried to extend at a bad timepoint
you know, drop it in the middle of a word, and it doesn't recover
that much is on me
I'm impressed that Suno. at least rhymes the gibberish, which makes it more difficult to know if it's gibberish or real words.
well, it obviously has to understand phonetic patterns to rhyme
Wait. What.....I'm supposed to see the obvious? Lol
so even if it is gibberish, it's gonna sound like something
it's great for what it is, and there are times the ad-libs it makes are jaw-droppers
So much so, that if you tell an AI like Chatgpt to write a song, and equally don't make it rhyme, it seems to be confused and rhymes anyway.
yeah
it can't seaprate the idea of 'singing' and 'rhyming'
but, there are PLENTY of ways that Suno can be better for editing and polishing what you've already generated
Maybe a song in haiku. Force the un-rhyme
Things like unifying a vocalist (if desired), making minor lyrical adjustments...
My cheat or hack with suno is to click output several times until it finally outputs a song with gibberish as an "instrument" , then I have a cool melody to work off of. I write real words to the melody and resubmit, usually extending off the intro to get the same vibe. Never get the same song but at times get a really nice substitute. Then tweak, tweak, tweak.
well, that's just it. I've gone through that process once, and for nearly all of it, I'm happy with it. I don't want to have to throw that all out and hope I get something close with another 3,000 credits
I'd even be happy paying a credit premium to do so, so long as the tools worked well enough
It benefits when you can eventually put a song together to pitch as a demo to an artist.
If I get a cut, I'll let the group know. 😀
yes, but at the same time, that's paying another party
Whose the third party? Suno?
The cost of getting a studio to demo lyrics with melody and music, or having suno do a big chunk, in various genres, is significant enough to make the effort for me, so far.
sick hack lol
Not sure I follow. But if you use the more generic song you can, an artist is mostly looking for the hooks, musically and lyrically.
I think you mean, the artist needs to re-record their version? That cost is theirs. I'm just going after the writers royalties.
But all good things to consider. It's a complicated business.
well, here's my situation: I have "Song". Musically, it's complete. Lyrically, it's 99% complete
I have a premium account, so that it's an "affordable" hack.
there's a couple vocal glitches early, one where it stumbled on an ambiguous word, and the other where I cut a word off to go to chorus
sneaky move
fair dinkum
Right. Ok. Following. Go on.
Ideally, I'd like to tighten up a VERY early line in the first Verse, and make very minor tweaks to the lyrics, mostly just word substitution
I understand.
so, a major win for Suno would be being able to do that with song you like and want to make truly shine
polish it up
Absolutely!
especially if you can have the vocals all generated at once, with a single coherent voice
The demo to pitch to artist has a benefit that I can tell them the "demo" is for style, tempo, melody with the lyrics sheet as the most current and priority over any glitches they may hear in the demo.
So as it won't be used for stems or full track, it will never be played outside of earshot of a pitcher of plugger... and the artist if it gets that far.
Yep. I like your route. We just have to wait.
see, I envision being able to make your own full tracks that you're happy with, and being able to essentially bypass traditional production
If I get a cut. Then. Woop woo! Hah. So far, my odds are about 1 in 400 for just a hold.
I'm not trying to make money
But suno has helped speed up the creativity towards writing those 400 sings.
I'm happy with people listening to it
Kewl beans. That's the baseline for any songwriter. If you can't write for your own enjoyment, not worth writing just for business.
of course, my songs... are a little spicier
Well, in suno-world, it can be anything. And all seems entertaining to me.
I'm still after the "right" prompts to do exactly what I want. Lol. But aren't we all.
don't I know it
let me put it this way. If I knew the tags and everything necessary to have the AI just re-do my song, I would, immediately
Ok. Did we clear the room. Or just everyone else eating.
no clue
Well. As a side, that would be fun. I'd still occassionally like to paint with canvas and oils than just prompt an AI for a picture, at least now and then.
fair
I think AI music is going to kill recorded music, if anything
the real value will be in going to see it played live
and smaller bands that know how to leverage it as a creative font can springboard into greater popularity if they get bangers under their belt
It might kill the economics of it, at least. Othwise, we modify and adapt. Then new generations will demand of AI those things that will be called "music industry" in 2030. And the economics will have a new vein.
right
So far live singing, yes, and live dancing. The kpop boys and girl groups will still do well. It will never be about the catchy song, but about the thrill of belonging to a fan club, seeing live evens, live streams online, etc. Of course Maze, a virtual kpop girl group is already taking a step further and have generated "girls" (backed up by real artists) , that are doing the dances and singing.
think of it like eating in Star Trek... why go to a restaurant, when you can just replicate a meal? Why, the actual cooking
So the same will be with Country artists, rock groups and so on.
Eventually. So far, suno has some enjoyable outputs, but it takes alot of tweaking to even get an assemblance of the nuances that go into, say, a Charlie Puth song.
Especially now, when much of the pop songs , at least, are veering towards non-rhymr or internal rhyme schemes, altered melodies PER verse, multiple tag lines, altered keys on choruses, and the like. AIs training , it seems, is mostly what has been more or less standard Verse verse tag chorus bridge tag chorus of the past 100 years
It's the waiting while AI catches up on its modeling, until we see some innovative new outputs.
Oops I'm babbling to myself. Lol. Ok. Back to work.
Craziest layered loop track I've had so far: https://suno.com/song/f50f656c-f7b6-4fb8-a781-75a80efd5f78
sea-folk, live drumming, orchestral strings, Celtic melodies, electronic soundscapes. layered female vocals, dynamic song. Listen and make your own with Suno.
added to the playlist
Also it gave me an ocean sound unprompted
huh
Man, "sped up" is such an understatement.
I know I already sound like a shill, but just in terms of having a positive mental attitude, NONE of this was possible a year ago, and to have such rapid advancements, and having companies compete and inspire each other to greatness is such an amazing thing to see in real time. 👀
well, I'd love to see Suno cribbing some notes off Udio...
#💡┃suggestions-web if there's anything in particular that you'd like to see them tackle next -- they do browse those channels for suggestions.
I've already left a few
Again, Udio was released April 2024 -- NONE of this was possible a year ago, it's wild that we're already at a point where we can browse for robots to write music for us without us needing to understand anything about music.
mostly the idea that v4 be more about polishing than creating
It's like we're browsing a grocery store for literal science fiction magic, and putting the box back on the shelf because it's not sugary enough for us yet.
LOL!
Friendship... Friendship?
I know, I know
still, Suno does have better composition chops
and I'm about those bangers, so i'm here to see Suno be the best it can
🧡
ah, you missed the reference...
My guess was either Spongebob or MLP?
Hey harber, can I make a nai image for ya? You can dm me what you want and I’ll try my best to make it
That'd be great! 🙂
Wave download not works on my iPhone with Safari and all browser… someone can help? Thanks!
Does the mp3 work? Size and timeouts might be a cause. You do the typical close and reopen browser/app?
Yes mp3 works. Very strange. I’ve not the laptop at the Moment. Tried also to close and reopen and also with other browser
Does it go as far as asking you where to download to ? Or just hang and do nothing?
Nothing happens when I push on wav download
https://discord.com/channels/1069381916492562582/1258072243963957288 post here as well? See if someone picks it up. I wouldn't know further.
Hold on. I'll. Check one thing.
Did you try the menu from "create " list.... Library list and the specific song page to see if all not work?
I'm still trying to figure out whats better; keeping the style simple, or give a big, detailed style prompt. What are your guys' experience?
From what I see, each additional prompt narrows it down by adding nuances of that prompt, some undetectable, as it also increases the combinations of "style" suno tries to combine. Ultimately, contracting the songs to sound somewhat similar.
1 to 5 prompts, funnels the genre to a fewer styles, but then suno throws out every possible combo with each generation you try. And so you get a wider differentiation of styles but staying with a "genre" , so to speak, created by the fewer prompts.
Having
Having said that, there are certain prompts I use over and over to keep any song to stay in a "feel" as if played or sung by the same band or person. So I may still use 8 or nine prompts and rotate the order to see what changes.
I've found some members here though, with songs that sound great, in and of themselves, and every time use just 1 or 3 prompts. Good written lyrics have a lot to do in those instances . But if just instrumental, it seems experimenting still is favored as there's no driving melody or lyrics to constrict the meter.
Typo. * ....ultimately constricting the songs to sound somewhat similar.....
One might also argue that if prompt one was "guitar" , then more prompts you add the less likely you'd have a guitar in the mix.
Not forgetting, there is , or was, a phenomenon of the music being more hiss-y, grissly, crackly and so on with the more prompts added. The theory given that digitally combining the styles was generatively conducive to getting "overtones" so to speak.
Here I am talking to myself again. When will I learn. Lol.
I can defenitely confirm this, I just tried a big prompt and it sounds absolutely bad, but with a few prompts the quality is mostly good
yes I think I agree with this. I find it very hard though, to perfectly tell suno what to do with just a few words
Solved. With the horizontal screen works. In Vertical is bugged
Cool. I'll try to remember that. I've got a fold phone, so yea, one way think it's a phone, spin the other thinks it a tablet. Hey, BTW, on PC with multiple monitors, keep in mind it may have an affect as well as which monitor your browser is viewed in , due to the drivers for that specific monitor.
hi there,
is there a place where i can find descriptions of songs that have already been made ?
I'm getting verse lyrics repeated. I'm sure I saw a guide on the amount of syllables to avoid this. Can someone guide me to the right place (or just tell me?)
Also, emotional prompts are sung out. Such as "[Verse 1] (hopeful)" gets me "Hopeful" sung as intro to the verse.
How do you mean, a description? You can search for a style, or even a word and sunonwill match up songs, if that's close to what you mean.
Try [Verse 1: hopeful] Not to say it will work, but just putting in parenthetical will get it sung, spoken or echoed.
Verse lyrics repeated on an extension or original song? I've not heard the syllable advice.
Original. I recall it was something in a 4x4 or 16 something format for syllables.
...fascinating. @unique stone, could you mention this in https://discord.com/channels/1069381916492562582/1252656600770940968 as a potential workaround? I'd be curious to see if that resolves the issue for others, but it's uncommon enough that I don't want to forget about this idea 🙏
I counted the syllables, it had 7 or 8 per line. Changed it to 8 all the way through and got no repeats. So there is that.
16 meter song?... that's like a soliloquey. I've written a couple of those in my day. Isn't Starty Starry Night basically a 16 metered song, maybe 32 meters, then repeats.
Yes, no, I haven't hear a 4x4 would stop it from repeating verses. It really like structured and rhyming songs, though.
I've worked around, at ,least, from not repeating verses in extensions, but I wouldn't know if it repeats in the original, how to not make it do that.
(My typos are abundant tonight)
starry starry night
Don Maclean's Vincent?
I changed it iambic pentametre. Suno really liked it and nailed the athmospheric tags for each verse like a champ.
@charred merlin, does this reflect your findings?
Yes. Vincent. (Sorry... brain fart)
I checked. 32 metered.
I feel like if you surveyed 100 people in a room, 75 of them would call it Starry Starry Night, 24 would call it Vincent, and I'd be the pedantic jerk by the punch bowl explaining that it's inaccurate to label it as "punch" if it's just grenadine, vodka and sprite without any actual fruit content.
Suno, from what I'm seeing, looks for rhyming words as well, even if not at the end, and rearranges the meter slightly to accommodate for making the lines rhyme in a timely way.
I wonder, if true, what a less structured lyric form would do with suno wanting to repeat verses. I originally thought it was simply trying to give you more for your credits and use as much of the 4 minutes to give a 3 to 4 minute song. I'm not certain of much now.
Isn't that sangria? Lol
I think that's wine & fruit
So, just meta-wine for people who can't wait for the fermentation process.
The best song to play in 1975 at 14 years old to "get the girl" . I guess there's something about a painter losing his ear that excites people.
I am very confused why I got a warning in the showcase saying I posted more then 4 links. My post only had 4 links.
Did you include links to other discord threads or channels?
No, it was only 4 links directly to the suno songs on suno
🤷♂️ the mods or guides would be able to look through the deletion logs to see what the issue was, I'd advise reaching out to one of the mods ❤️
I did reach out to them. Waiting for aresponse. Thank you.
🤔 I think it's a global Woodhouse rule, probably stated there to avoid frustration of Woodhouse auto-deleting a post whilst leaving a 60 minute cooldown.
That said, embeds and links are different. 4 embeds takes up the entire vertical space of my screen...
Maybe Woodhouse was changed in general, because the channel description still says 4 songs are allowed.
I can scroll back a couple days to see when people stopped being allowed to post 4 songs at a time, but this seems a lot easier to find by asking in the guide-chat when the changes were made.
Seems I may be misinformed. I'll delete my previous message since it's not accurate.
#🐣┃suno-showcase message for example
#1260045620538904628 seems to be the place to discuss this, not sure if it was mentioned to the community?
That post is locked by one of the moderators due to some problems there.
Your amazing. I can create beautiful songs. Is it also possible to create 5/4 measure or polyritmic or 7/8 measures?
Just give it a shot 😊
idk
Hi Philosopher, how to promt this?
As far as I know, no one has figured how to do that constantly. The best method would be to use Audio Upload with the tempo set (e.g. using a metronome or a MIDI, etc. Generated from programs like FL Studio) and hope that it influences the measure
Genres also influences the time change so try to find the right style for you. (E.g. using tango should give you its distinctive time signature)
So i have to give in the style like Tango 4/6 and then 7/8 like Salsa music?
I have found that when you have a specific prompt for a sound you like, and that small changes will no longer change your genre, you can add measures to create effect. However, it is not specific. 3/3 is a pretty sure call, and I like 64/64 to add complexity. But doing 48/64 vs 64/64 is pretty much a hit or miss, it cannot be controlled.
There were 2 members who were heated, all of the inflammatory posts were grouped in 2 sessions -- shouldn't take too long to clean up and speak to the two instigators.
I recognize that this server gets overwhelming at times, and since moderation is provided on a volunteer basis, there's only so much that can be expected for free!
Is there a process where people can volunteer for a guide or moderator role to offer such assistance? Even if the support isn't immediately needed, it can't hurt to be proactive to know what community talent is willing to make the server the best it can be 😄 ❤️
I really appreciate your proactive approach and your offer to help improve the community. However, after considering the situation, we feel that reopening this post isn't the right move for us at the moment. We're grateful for your support and may revisit this idea in the future, but for now, we've decided to keep things as they are.
Also, @unique stone wasn't able to provide clarification to the post just a couple above yours -- are you able to confirm if Woodhouse only allows 2 songs in the Showcase, and when that change was made?
It seems like there was some confusion on the matter, so it would be helpful if people know for certain what the rules are before making a post while thinking they are following the rules, having it removed by Woodhouse, and needing to wait another hour to post again 😨
It's always been 4. Since Woodhouse removes 4 embeds.
#📚┃suno-school message Perhaps you can reach out to this user and assist them?
Is already take care off like a half hour ago or so?
almost an hour: #🐣┃suno-showcase message
I appreciate your enthusiasm, but we have the moderation covered. Let’s allow the mods to handle it.
The question was posed in tech chat, and I was trying to support @urban raptor by answering their question -- I'm glad you were able to assist them elsewhere. 💪
Always glad to hear from you ❤️
Sending you an un-soliciated friends request now harber 
I want to share with you, how far I've gotten in my prompting journey, you'll be surprised
To clarify, I wasn't just asking for myself -- I was asking if there was any formal process for anyone wanting to offer their support, but I feel like I've already stepped on toes -- apologies!
There's no harm in simply asking, thanks for your response.
🙏
I feel like it's only unsolicited if I wouldn't consider you a friend in the first place -- I'm actually kinda ticked that you beat me to the punch 🤬
I'm just grinding my way through Suno trying to teach as many as I can, from what I've learned
I never had this problem before but today when I generate a music, the tempo sometimes accelerates during the piece and becomes so fast it's terrible to listen, is there an issue?
I was just rambling to @delicate idol about some of the Sunovator's I've "clicked" with -- she pointed out that everyone learns differently, and it's rewarding to help anyone, but when you feel like someone takes your "idea-riff" and improvises on it while still keeping the same "experimental leitmotifs", it's friggen exciting -- it's like science-jazz
"Not only am I digging the groovy methodologies you're layin' on me, it's really all about the tests you choose NOT to play." coughs
You are quite a prominent figure already @charred merlin , now we just need to get Woodhouse to understand this 
I threw up on his sofa the last time he let me crash at his place, I think he has good reason to not want to party with me anymore.
It's happened before. I guess there are some songs out there in history that change tempo, so sunonpicked up on that. Suno does it more while it's hallucinating, than in the continuous song, itself.
@polar spade hit the nail on the head -- I suppose picking up tempo unexpected MIGHT be considered a minor hallucination, or it might be an aspect of one of the style prompts / metatags -- from my own observations, it tends to happen more often when Suno's overloaded to some degree (which in itself makes Suno more prone to hallucinations).
Do you have any examples you could share where the outputs are more likely to speed up than not? If we can see what input you're using, we might be able to get a better idea of why the output is acting like this.
سلام 🙂 ربات چیپ بازنشسته شده است. برای ساخت موسیقی به https://suno.com مراجعه کنید
چرا
(Above translated using Google Translate)
Translated using Google Translate:
از آنجایی که اکثر مردم از وب سایت (#📢┃announcements message) استفاده می کردند، بنابراین ادمین ها ماه ها پیش تصمیم به بستن ربات گرفتند. اگر می خواهید موسیقی بسازید، باید از وب سایت استفاده کنید یا سرویس دیگری پیدا کنید.
Because most people used the website (#📢┃announcements message) so the admins decided to shut down the bot months ago. If you want to make music, you will have to use the website or find another service.
میشه سرویس دیگه ای معرفی کنی
میشه سرویس دیگه ای معرفی کنی؟
من هیچ کدام را نمی شناسم که از Discord استفاده کند. من فقط Suno و Udio را می شناسم که هر دو از وب سایت ها استفاده می کنند
Udio چیه
یکی دیگر از خدمات هوش مصنوعی موسیقی مانند Suno. سونو در موسیقی بهتر است اما Udio صدای بهتری دارد
ادرس سایتش رو بفرست
پس همین دوتا رو میشناسی هوش مصنوعی
فقط .com را به آن دو سایت اضافه کنید یا از گوگل استفاده کنید. هر دو به خوبی شناخته شده اند.
__
And with that direction over, let us return back to English
why the UI has massive whitespace on right side?
read tru it yikes i can just say, they was all on eachother , pity it was a nice thread.
Just gave Udio a shot. OMG - what a mess. I swear it's in the uncanny valley of songwriting.
Hello, qq, what is the best way to let Suno 3.5 to finish the song? I am using Outro, End, Finale, Finish, and on all of them them the part extends to 2 minutes with no proper ending
If you think the prompt is short enough you can regenerate in hopes it ends properly, or you can extend multiple times until you get a satisfactory ending.
Well, that is what I am doing but even after 10 to 12 attempts still nothing so I am just wasting my credits, I can do the v3 instead but the mix from the previous part to the final part is a big change in some cases and I do not like it 😫
yeah sometimes it gets in a bad state where you just have to abandon the progress or reset back to an earlier track in your song
although it's really cool when it works well and you can make 15 minute long songs lol
I've been able to get it to end without needing to put "outro" or any other marker too
I don't think putting "outro" actually does anything
is very frustrating no way to make suno 3.5 sing in male voice this is ridiculous had flanged like 30 songs
What prompts did you use if you do not mind sharing?
[deep male voice] Cumbia, aggressive, bailable, bass& drums, bongo, air horn, thunder
i used deep male voice cumbia aggressive bailable bass drums bongo air horn thunder
[Male voice] Cumbia, aggressive, bailable, bass& drums, bongo, air horn, thunder
man voice, Cumbia, aggressive, bailable, bass& drums, bongo, air horn, thunder
okay this is annoying me
male native
I'm trying to generate a male vocal for a song and 9/10 times it gives me a female voice instead of male
I try all sorts on the prompt to call for male voice IA refuses to do it,
40 songs with a call on male voice and nothing. used to be a metatag {} and it will work, now you need to guess how the IA will like to do it, looking for other creators and copy how they call for voice male
Look at my song lyrics and style prompt.
I kept nearly the same style prompt thay I replied to, just having "male" I stead of "deep male voice" in the style prompt.
The [deep bass male vocals] was at the beginning of the lyrics field, NOT in the style prompt
ok Im looking at it Im going to copy past in to my lirics and see
still the problem https://suno.com/song/d15dd5a6-4867-4823-bdfa-46a7dd1c2f0d
wich browser are you using,
It shouldn't matter since it is getting processed by the same system, just RNG
well that very weird that you have a correct output and not me
im going to clean cache and see
Just random (RNG) I think. "Male" is not that powerful so having it separated helped I think but it could still have been just luck 🤷♂️ or Woodhouse rigged the system to like me XD
I see this is disturbing for me yes the IA likes you
Thanks loots for your assistance I will continue tomorrow with a clean system, I will post it if it get corrected on my end
Hopefully the system works with you tomorrow 🙂 Do not be afraid to extend one of the tracks I made that you liked though, you can have it ^_^
Let me know if you want me to delete the links since it has your lyrics
yes please and thanks again
Oops, deleted the wrong one, I will post that one up again for future info for others since the ones generated were my generated lyrics:
Try having just Male. Since "male" does not have much weight, likely vocal/voice outweighed it (and likely has more weight towards female).
My test with it worked out surprisingly:
Added [Deep bass male vocals] at the start if the lyrics to see the effect
- https://suno.com/song/b6d54f2e-0593-4bef-9d73-9d2f95c3f3bb
- https://suno.com/song/9f8c071a-7ba5-46e4-ada6-7da403003faf
Without that start lyric tag
how do i get rid of the static background noise?
plese help
i have wasted almost 100 credits
Hey @rain shuttle, while we can’t guarantee a fix for you, we’re happy to help you try to approach this issue from a new angle so we can all learn from this troubleshooting 💪
Please share;
- a suno link to the song you’re working on 🔗
- timestamp(s) of where you’re dissatisfied with the audio 🎯
- what you would prefer to hear instead of what you’re hearing 🙏
Hey, Where can i go to get some help with metatags and stuff? i cant seem to get it right, and i tried a few times, with [], **, and (). And it still dont work.
I want to add applaus and a throat clear before the song, And its just not working for me
Everywhere, that's what discussion in this Discord is all about. For throat clearing, that might be too specific to get working. Applause is pretty doable in a general sense, if you prompt for live music. If you want to specific exactly where the applause is, that's more dicey, but if it ever works it's almost always in a song that already has an audience in the background already. These are from launch period V3.5 (so maybe a bit different?) but ran a bunch of live audience prompt tests in this playlist, maybe some useful tags to try: https://suno.com/playlist/ee70007f-261e-4b7e-be95-ad6c14386ed8
@naive quail does a lot of related sound effect and atmospheric stuff -- https://suno.com/playlist/4a674d38-0316-4d14-ace6-a2a4ff9d9d25, here's one of his humor playlists, but it gives a good example of non-musical prompting 💪
Extremely messy 'brute force' crowd cheer at 2:06 here. The video will show the prompts used, it's not regenerated: https://suno.com/song/607b0628-c6f6-4c06-816a-f2072f05b7df
A song. Listen and make your own with Suno.
If I have some advice -- try to create the applause and throat clearing as a separate audio track, then extend off of that -- this way you can distinctly get the separate audio intro without having to try to get lucky with a generation that has applause and a throat clearing, but doesn't start the music until you want.
As always, the more specific of a song you're envisioning, the less likely you'll be able to create it as a "one-shot generation" through the use of more and more complex prompting.
someone rang the cowbell?
turns out that k-band gain horns make for decent cowbells 😄
70's Rock, Blues, Folk, Cowbell, more Cow-Bell, Cow Bell, Guitar-Driven cowbell song. Listen and make your own with Suno.
ive have a weird issue wav and mp3 is just gone well it gives me mp3 but no wav to dwl (im on pro subsribtion)
It looks like that song is new (video is still generating) so you may just have to wait a minute
seem that was thew issue just never seen it before
is there a list of instruments in suno?
Not in the form of "a list of commands that are guaranteed to work" -- Suno is a vibe interpreter as best as I've been able to figure out, and it's a LOT easier to say what instruments or techniques we've had specific difficulty achieving, even with dedicated attempts (like theremin is notoriously difficult to prompt directly, but asking for spooky halloween music seems to be a workaround).
That said, there ARE compiled .pdf lists that helpful users like @rose jasper share if you wanted to search this channel to find the posts they've shared, but I think it's helpful to point out that it's less about knowing "the right prompts", and discovering "which ones don't work for you"
ah I see, I was able to get a Shamisen by putting Shamisen tag in so I was curious if there was other instruments you can force in that way.
😅 I get what you mean, the issue starts to show up when people include a large number of styles and prompts, and eventually "rock, opera, metal, techno dubstep" are such weighted terms that carry their own connotations to Suno -- suddenly it has no idea how "shamisen" fits in, and despite it working super well in one instance, you might be lucky if Suno decides to include it depending on how the AI-winds are blowing that day, or what collective prompting you're using.
so you have to believe in the heart of the ai and hope rng is on your side 
so i tried to make a song using another one of my fl studio songs as the base and it's pretty good use of it
That's ALWAYS a huge part of it -- over in the pool party > #1256301881966198805 , I try to test things in isolation & figure them out on a more granular level so I can be better informed as to how all of the different pieces come together when I add 10 prompts together to make a song (which prompts did Suno favour? Which were ignored? Which resulted in a weird chemical-reaction and resulted in something I never would have predicted? Did I learn anything unexpected that I could use in further tests or personal song creations? Were there any wacky results that I can clean up into an actual cool song?)
Because I don't actively want to put up roadblocks to someone's learning, here's the link to the uber-list @rose jasper found -- I hope it's apparent that I'm not opposed to people learning or trying things differently than me, I just feel it's important to add the caveat that having a compiled list like this is SUPER helpful, but shouldn't be used as a replacement for active experimentation, and personally, I believe it should be used as an example of how to create your own prompts and get ideas on how to explain "rock" in different or more aptly descriptive terms if you're not savvy with musical-terminology or sub-genres.
I've just encountered too many instances where I try to help, and it's just misunderstood as "cool, so if my idea isn't working, tell me the correct way to do it" -- I'm not an expert, I'm just the dude sitting beside you on the sofa wanting to help explain how I got something to work, and why I think people could be excited about the same things 🙏 ❤️
I'm still in the "honeymoon phase" of Audio Upload -- I get so giddy hearing Suno take my idea and "run with it" -- even if AGI is still a couple years off, I fully expect to be engaging in realtime jam sessions with a future model of Suno.
The latency & cloud-processing times are just stupid-good these days, and since we're at a point where you can be picky & choosy with literal science fiction magic, we're at a beautiful crossroads in history of market-driven innovations and genuinely impactful and lifechanging breakthroughs, and it looks like Fall 2024 is the industry milestone for a couple major releases -- ChatGPT's advanced realtime voice mode and iOS 18's Apple Intelligence are great consumer-facing examples
@charred merlin have you seen the track elevenlabs dropped to tease their generative music system?
thank you for the tips, I have a few complied list I found while searching through google but this list you shared is defiantly massive. And don't mind burning credits in attempts to learn. It just helps if someone had a list of instruments 😅 thank you
Like, a decade ago, the same degree of market competitiveness would have resulted in next year's phone model being thinner, or a bigger hard drive.
Now companies are hesitant to release any updates because by the time they've polished a press release -- they're already 3 breakthroughs beyond that, and the competitors have released a sub-par version of the same thing rushed to market -- so now you can look like you're releasing a "better copycat" like Google or you can release everything in one polished drop and gain market reputation like Apple.
oooooh, no! But I'm looking into that NOW -- would you happen to have a link?
https://www.youtube.com/watch?v=K7FqIncBHzQ ooh, this?
Here’s an early preview of ElevenLabs Music.
This song was generated from a single text prompt with no edits.
Style: “Indie Rock with 90s influences, featuring a combination of clean and distorted guitars, driving drum beats, and a prominent bassline, with a moderate tempo around 120 BPM, and a mix of introspective and uplifting moods, evoking a...
same though, next year should be wild
I can tell you hurdy gurdy doesnt work 😂
Its pretty nuts if its indicative of their system. I use Suno atm for everything but competition is good for everyone
I meant this one, but the one you linked is also theirs I believe
Here’s an early preview of ElevenLabs Music.
This song was generated from a single text prompt with no edits.
Style: “Pop pop-rock, country, top charts song.”
This is what @unique stone and I were discussing a while back, there were a lot of early experiments with Stable Diffusion to generate spectral analysis images like this, and people were able to get "noise" back from it.
While it's not something that you could reasonably train SD to do on top of the implicit randomness, I think furrybeans was mentioning that Riffusion's underlying technology is a specifically trained model on similar patterns.
That said, compared to Suno, Riffusion hasn't proven to be anything than a casual toy, in my eyes. 😅
I hear you, I have messed with a few other "AI music" generators and was not impressed, TBH I don't even like Udio, I guess I just gel with Suno the best so I am sticking with it. I am keeping my eye on everything in the space though, the custom vocals at Kits.AI look awesome I want to try using it to do an enhanced version of my own voice to sneak onto some tracks for a lark
trying my best to try to master one of my tracks #🗣️┃chill-chat message it's my newest suno song using the free credits
and i tried something a little different than usual for it
This is the one that makes me insane. It's such an earworm.
yeah same here
@neat nexus I feel like they could drop it on a Tom Petty discography and we'd all be asking when Tom dropped it lmao
I don't know that it's Tom Petty to my ears. But I am definitely the right demographic as I also like Tom Petty.
Its not exactly there but both the style and the overall vocal timbre really makes me think of Tom
Oooh, this is pretty catchy -- I wonder what sample size they chose this generation from! If it was a single one-shot, I'd be blown away.
Arguably, there's been one-shots from Suno that have absolutely blown me away, so I'm curious to see where ElevenLabs has grown in the meantime, and how adaptable it'll be to the general public who might not be so musically-technically-savvy. Exciting either way!
same here
How come it's still not possible to sort playlist and song order? Keeping everything organized is frustrating
@charred merlin I've had to do a lot of iterations so far with Suno to get what I want, but the range is so good that you can get a lot of surprisingly unique compositions in the aggregate. I have been doing electronic stuff since my teens when DAWs became a thing but I had always resigned myself to EDM because I don't sing/play instruments. Suno and the wider world of AI music opened a whole new world for me as a producer in genres I prefer in rock. I'd like individual instrumental stems so I can get single audio channels exported to Ableton but at least getting vocals/instruments separate helps for the time being. I think the democratization/access is great for the public, but I think the real breakthrough is going to be for musicians to quickly prototype/iterate, find stuff they want to use and blend with their workflow. At least once they get over the "eww AI" hump.
Question - is Suno scrapping from Citizen DJ? https://labs.loc.gov/work/experiments/citizen-dj/
@winter thorn To further follow-up on this, whilst the instrument is listed on the form I linked, if you focus more on unique terms with [Shamisen] you'll get an even better end-result.
The Shamisen as an instrument is very analogue and it establishes high noted and vibrations and reverb added on top of it, I would prompt something like this:
[Shamisen] [Correctly tuned Shamisen] [Shamisen pitch bending] [Shamisen strumming] [Shamisen: Lively bachi strikes] [Shamisen: Delicate sukui picking] <- I am explaining a bit more in detail what I want from the instrument, is to also use the following explanation on "Tempo Tags" (https://www.suno.wiki/wp-content/uploads/2024/04/PROMPTS-FOR-AI-DANCE-MUSIC-🙏mista.lewys_.pdf)
Now with all instruments, they all have their own terms and ways of playing these, so it's also useful if start to look up how these are played, and what the different strokes are called:
- Sukui picking
- Bachi strikes
- Plucking
This gives a better directive to the track, example, I made this one just recently, observe how I designed the prompts for this, and how I go more into detail when "directing" (Yes, I know, Suno does not serve or function in this purpose) the AI to work through a better flow of the track and your desired result:
https://en.wikipedia.org/wiki/Suno_AI
Suno does not disclose the dataset used to train its artificial intelligence but claims it has been safeguarded against plagiarism and copyright concerns
Also, it's nice to finally chat now that you're not trapped in an airport 💙
Thank you. Based on the suit from RIAA on fair use, I would like to be able to CHOOE where Sumo scraps from so It's been a pleasure but ... I'd rather stay with an organization I KNOW has fair use
🤷♂️ No idea, we're all just users here.
The devs occasionally pop their heads into #💡┃suggestions-web and #🚧┃suggestions-layout if you wanted to share ideas for making Suno better 💪
You're welcome -- you'll probably need to specifically look for services that use your desired training data, or more directly build one yourself! https://stability.ai/stable-audio
Thanks, I didn't realize that!
😄 Not a worry, there's lots to learn & I don't know if there's an official discord onboarding to teach this kinda stuff without some degree of trail and error 😅
Some people simply want to express themselves and feel heard, regardless if it's the most effective way of getting their opinions heard by relevant-ears -- which I genuinely relate to, but I figure I can try to point them in the right direction if people genuinely want to help 🙏
Yeah, I don't use Discord very often, it's too intense for me!
I have no idea what you could possibly mean, I'm sure.
dont think so
I would like to be able to find out what choices were made by the AI in any given track
Can you elaborate on how you imagine that being formatted? Not swatting down the idea, I just want to get a better understanding of how you're envisioning the information being displayed?
for starters, I'd kinda have to have some idea of what choices it's making in the first place
In a nutshell, trying to figure out how Suno makes its decisions is sort of what we're all about here
some are more obvious, like up-tempo or down-tempo, what beat to start vocals, etc.
Perfect example for you to check out this guide, it's very useful and helpful:
https://www.suno.wiki/wp-content/uploads/2024/04/PROMPTS-FOR-AI-DANCE-MUSIC-🙏mista.lewys_.pdf
I've suggested something like this myself. For example, Suno could generate a wavform that isn't just a wavform, but indicates things like 'do I think the song is ending at this specific time' and so on. Might reveal too much of the 'secret sauce', but maybe some things like that ok
well, if it does something that we like but weren't expecting, we could look in and see why it did that, and then maybe guide it to do that again later on
This is difficult in large AI models, but not impossible, and I do think even abstract things (like colors) would be cool and people might be able to get an intuitive sense of why some prompts were or were not working.
right
I just don't like having to completely replace a song I like if I want to update it a little
whenever i put a track on distrokid. i credit Suno AI as an artist. just wanted to let you know.
I know some Distribution Partners specifically ask that you don't do that, because Suno has waived all interest, rights and titles to the song if you've made it on a paid account. also because DSPs might not accept it because Suno is a company instead of an artist, and it might seem like an attempt to "associate yourself" with an unrelated artist, if DistroKid & the DSPs don't have an understanding on the matter -- I think that's one of the issues with Routenote if you list Suno as a contributor.
I think there might be a different way to give them credit or attribution, but if DistroKid hasn't told you not to do that, and it seems to fit within their explanations of upload criteria, then all should be good 💪
there is a balance between careful consideration, perfectionism, and dont give a fuk, yolo. 😆 thanks!
somewhere between Rick Rubin and Brian Eno
it was interesting when i actually detected my own track on Shazam
I love sharing a track, then checking Apple Music for Artists the following day to see what countries you were Shazam'ed in
I think my frustration comes from not really knowing music theory very well, only 'what I like'
I tried using two different (but complementary) styles, but I don't know which gave influence
sameeee
The saving grace of Suno is that knowing music theory can help inform creative decisions, like when is best to extend from if you want to resolve a musical phrase, but Suno interprets vibes more than explicit commands -- some may argue that you're in a better position to create music using Suno if you AREN'T familiar with music theory because you'll be less constrained with traditional rules that you might expect Suno to follow (when it very likely will just vibe with what you're asking for and disregard all of your specifics)
does listening to 3k tracks qualify for basic music theory knowledge?
right, but the trick I'm looking for is finding out what 'vibe' it was giving, so I can lead it back into those
Do they still make books for dummies? I'm going to have to hunt around...there are a few things I could use some easy reference guides for.
If you're listening to them and giving any sort of objective critique from a subjective standpoint, then absolutely ❤️
This is why I find it helpful to break things down into the smallest chunks that I can and test them in isolation -- if you keep stacking prompts without trying to figure out what they can reasonably be expected to do on their own, after a certain point you're practically inviting confirmation bias.
Not everyone needs to test things the same (if they even care about testing & self-improving on their methods), but you'll do fine as long as you have some kind of consistency with how you approach things (instead of giving up after failures and starting again trying to "get it right", because the failures teach you JUST AS MUCH -- not suggesting that's what you are doing, I'm trying to phrase this for everyone and anyone <3)
they can only teach you if you know what the lesson is
you are totally conscious and obsorbing not only the sonic quality, shapes, textures, and spaces, but the changes and various ideas explored.
there's no real feedback
basically timelords, lol!
your song 'goes off the rails', but there's no indication of what happened, where the prompts got misinterpreted/ignored/missed...
or why Suno thinks it needs to ad-lib less than a minute in and a quarter of the way through the provided lyrics
That's why ~~we ~~ I try to reapproach ~~our ~~ my methods to examine them when this happens.
for input audio, there is a time transformation that sometimes results in wonky temp changes. sometimes cool, but can sound like your tapedeck has gone "off the rails"
I feel like I have a pretty good understand of how to make the songs I want, and have a rudimentary expectation of what kind of variations to expect from v3 vs v3.5 generations, because I familiarized myself with the style prompts I commonly use.
as am I... or at least I'm trying to figure them out
it's like I'm close to the sound I want, but not quite there
Suno provides a canvas of your preference... you become the producer. mix it up.
and it's not like I'm trying to give conflicting tags. I'm only trying to give one per section
like a direction for rhythm, one for vocals, etc.
but, I think I have tags to go for the sound I want now
idk why suno does that sometimes lol
like, I'm getting certain ones close
Ooh, fusions, I love to test them out (as evidence by the test playlist https://suno.com/playlist/d83502d8-c487-4c95-82cc-412bc2823288)
If you know what style you want as the rhythm and melody, you can append those.
See the playlist for the effects of defining the genres + rhythm/melody/etc. (Has "sep" or "separated" in the title) vs. using the ones with fusion keywords present
👋
🤞🫵🌄🎉
well, it was done to try and get the sound I wanted
If you want two styles mixed, you can also start in one style, and then Extend the song (from a time near the start of the clip, <10s typically) with a different genre. Keep trying different times. And you can get a hybrid/blend/fusion that way. For example this song started in EDM, but at something like 6 or 10s, it was "jazz" prompts, and it kepts the electronic sounds and kind of blended https://suno.com/song/e6d193fa-89e0-4444-af57-8830a9bfc0be
right. You're fitting a jazz rhythm into EDM instruments
If you want to change the music style a lot though, the Extend From time needs to be somewhat near the start of the clip. It's not impossible to shift genre after even 60s, but it more often will be stubborn. And 1s will shift entirely. But 3s, 6s, 9s, in the middle, can find a blend.
...I'm not (at least knowingly) trying to blend like that
I mean, I was specifying "rock, metal"... rather complementary
but now I know what I really want, so I may wind up starting over... yet again
If one tag is too strong, you can balance them like "metal, folk, folk" for example. You would try that if the song was too much metal, not enough folk.
ah... so thumb the scale
With some more niche genres or terms, or lesser known instruments as well, you have to repeat them even more than 2x.
I mean, I'd gone as far as trying to just drop the UUID of the song I wanted it to try and follow into a tag, to show how desperate I was to get it to follow that again
If you want to make a song inspired by another song, this can also be done using the Upload feature, and manually editing the Upload clip. Try various segments and mixes (the Upload doesn't have to really even make sense, as a real song).
However if the upload has a voice, you will not be able to flip the song Public on Suno website. That is main restriction.
here's the thing. I have a song that I like. I want to just fix up the lyrics a tiny bit
ideally, I'd keep it the same instrumentally
that's one major annoyance with Suno. You have to toss everything out
there's another that got a lesser ending because it ad-libbed an early end
the ending it did come up with eventually was nice
but that had to go because otherwise, there was a bunch of fumbling over lyrics left to go
@potent pelican I know we're chatting about it in DMs, but I wanted to share that emergency food @hollow hare contributed some variations to the tech-house playlist 💪 🙏 -- a couple variations on how the existing prompts were presented (hyphenating can force Suno to "read things different", sometimes removing ambiguity if something is meant to be 2 unrelated words side by side or a two-parter -- sort of like I just did there).
Since we're still trying to find "neutral ground" in terms of your expectations for tech-house, and what Suno is interpreting the requests as, any bit of feedback on them helps point us in the right direction -- whether it's a 1-10 rating where 1 is an entirely different genre and 10 is perfect enough to us as-is, a 👍/👎, etc
I may sound overly-negative at times, but gotta remember, that shows passion. I want this to be better. If I just didn't care, I wouldn't even be speaking
You guys are amazing, i will check them and give feedback when I'm home 🙂