#🗣️┃chill-chat
1 messages · Page 40 of 1
can't you just link to the songs ok udio instead of mp3?
just saying, you'll likely get more traction doing that
yeah no, I prefer posting the files directly because most people are on computers and laptops, which can play them directly.
what's the point of traction on Suno.com? The only places you want traction is in the stores or anywhere else you sell music
I was just saying in general. This is a nonsequiter in any case, as the alternative I referred to doesn't give traction anywhere. Thus, a link to a site associated with your account will, of course, be a better net positive than an mp3.
This of course presumes your implicit assertion, monetization. Many people just want to be heard
in which case, again, links increase the chance of being heard
suno.com has no social component, so not, it isn't
you're just saying random disruptive things lol
no, what I am saying is that traction on Suno.com is pointless if you're not making sales or streams in the stores
again, not true, for the reasons outlined
Also, traction on suno is miniscule, even rhe biggest suno names have just a few thousand fans
Most people here just enjoy making music, literally
hi can i ask how i can use the old melodies, what suno made me, but with new lyrics?
you could try extending with a timestamp of just before the old lyrics start, or download & chop a section of the melody & use that as an upload to extend out with the new lyrics
no guarantee that will be exactly teh same though- the way generative AI works often involves some randomisation
do you have any tutorials?
i tried to add the old melodies but idk how i can it add to new song
not specifically but I think there are likely some on the announcements channel if you go back to where the various feautures (extend, times, uploads) are announced that often has tutorials.
when you do an upload, once it's in your library you need to do 'extend' on it to start a new song with it.
sorry, i didnt find it. i go to what is new but here was only how edit diplayed lyrics, but it doesnt solve my problem.
and if i pay beter version of suno, it will work better? because when i want to create new song it write me everytime error
And another cooked piece from Udio.
@fast rivet you might like that.
You'll probably have it stuck in your head now
Hyperglitch Psybient Breaks: https://www.udio.com/songs/ugG3rbuAjRLCkZzZQ4h1fV
Listen to Forever飘floating by EcoShadow(glitch) on Udio. Discover, create, and share music with the world. Use the latest technology to create AI music in seconds.
Hey @fast rivet I've got a massive surprise for ya. Another Udio track. :>
another Udio song made 30 seconds at a time 
Hey, at least I can perfect it lol
I do sometimes do the very rare 2:10 extension songs
But here's the one benefit Udio has for me. I can actually do these songs section by section to make it sound the way I want
I dunno, I just don't like Udio, there's something about it I can't quite put my finger on
I get it man. Udio and Suno are very different tools, but both have benefits for me.
Haven't tried Avia and Soundful yet.
Aiva and Soundful have real stems, so you can really do a lot with them, but they are more complex to use than Suno or Udio
a 3 minute track in Soundful gives you a 1.1GB Stem ZIP file
full of all the tracks and audio
and MIDI
but to own copyright of Soundful tracks you need a Business plan which starts at $740 AUD a year
for me though it's worth it
Yeah.. I don't have that kind of money lmao
I sell royalty free music, so these pay for themselves many times over
...I did another one again.
@fast rivet hopefully you haven't taken any LSD today otherwise you'll be transported back in time
@sleek urchin if you wanted to show @grand solar a good example of a sea shanty, here you go. This is what I just cooked up with Udio.
Suno can't do a proper sea shanty like this, which is unfortunate. It always turns them into something overproduced and too modern sounding.
@hollow leaf @sleek urchin I cooked again
❤️ just having my coffee and getting my day started, but I'm tickled pink that you'd think to ask, and I want to be in as helpful of a mindset as I can be, so I'm going to circle back to this shortly 🙏
You absolutely rock 😄
Reg gen cute shiba dog
https://www.youtube.com/watch?v=CKgOJyFFO_Q
Cute dodge dog that serves tobacco in Tokyo in Japan. He loves petting and cucumbers.
@strange gorge and my first LoRa is already cooking
but yea, full forge support
I am able to render on my 4070ti in 2.28 seconds
20 step, euler
Great 4 time now i have to change to a new anime streaming site due to that the videos stopped working
Yeah I saw that, stuff is moving so so fast
I'm going to attempt to train my own on my own machine here in moments
It is
It owns, it's like 1.5 all over again in how 
Full support in Forge is pretty huge
It looks pretty cool! Hard to see some of the finer details with the current resoluation, but I'd be happy to AI upscale it or work on something together if you want to reach out in DM?
Trying to practice my ai-gen & photoshop skills, and I'd be happy to lend a hand 😄
...legit thought this wouldn't be here yet.
That's...absurd.
I was reading about people managing to somehow get it to work on a GTX 1060 (?), but all of the 8~12gb implementations I'd been reading about still seemed to be around 10s/iteration+
@vale lake I'm going to be looking back through the chat to see if it was shared, but if there's any resources that you pulled from, I'd sincerely appreciate a point in the right direction 🙏 same gpu, so that's ridiculously promising
Flux has replaced Stable Diffusion as THE generative AI on your own hardware thing
It's too good
Not terribly surprised it's happened but yeah the speed is really fast
Tools and community are a bit more mature I suppose
guess I'll be updating some stuff this weekend then
Great! Was worried I wouldn't be able to run it on my 4070 Ti
Do most people make whimsical music on Suno? Are there many pro-grade users?
I like to find new music on SoundCloud. Thought I might check Suno's library. But I run into a lot of "Cat cat cat cat cat cat" stuff.
I sell music made with Suno and other AI
What does that mean, "sell it?" Like on Beatport? Apple?
Or by commission, for games.
I distribute to the music stores but I also have a separate royalty free music store that I sell music directly from
Yeah, but does it sell? Oh -- you put your music online as royalty-free music for people to buy. Like Epidemic Sound.
on a good day I can make a couple of grand $USD
That's interesting. I don't do instrumentals.
But i can see how that might be popular.
I sell instrumentals from $10 per track, songs from $4.99 per track and have royalty free licensed up to $50 USD with trackout and MIDI
How do you get your Suno music from a .wav to a midi, given all the distortion that comes with the tracks?
I also sell sound packs and MIDI kits
I don't offer MIDI or trackout licenses on Suno tracks, but I do on Soundful and Aiva tracks
I would like to translate my .wavs to midi, so i can customize them and clean them up.
WAV to MIDI doesn't work, don't waste your time trying
I like Suno for the vocals... Not much into the instrumentals, even if they are cleaner.
oh yea, I am using it from Forge which is here: https://github.com/lllyasviel/stable-diffusion-webui-forge
There is a link there that downloads a zip file with everything ready
you just extract it
then run the "update" as admin
and then run the "run" as admin
after that, to start you just run as normal
100% -- I was only using SDXL as a rapid-fire alternative, but it feels like we've skipped a solid couple years in terms of Flux development in the past week or so. 👀
It would have only been a cloud-based thing for me in the short-term if I needed to wait a solid chunk of time for each gen, but yolkhead's using the same card as me, so that's super exciting to see that it's that tangible for me 😄
I am running FLUX on DiffusionBee on my MacBook Pro M2 Max with 96GB RAM
this is the section
about as easy as it gets
then, you just need to download the model itself, which lllyasviel has managed to get down to under 12gb
You rock 😄
It's just ridiculous how quickly we went from a gut-feeling that "of course it'll fool SOME people, but we'll obviously know AI-genned content when we see it" to questioning EVERYTHING.
wait until LoRAs of Flux become widespread
oh I went off the deep end with that a while ago
it opens up new avenues of psychological warfare that are extremely troubling
"Wow, look at this great meal!"
"yeah, but why does the chopstick sort of look weird where it submerges into the pho?"
"THAT'S HOW LIGHT REFRACTS THROUGH LIQUIDS, MATTHEW"
lol, speaking of new eras of psychological warfare...
for example, consider that in the past, certain things that might have been totally implausible, or impossible, become possible. How, for example, can you do something unethical that requires dozens of actors without anyone leaking? This is the main reason, personally, I have always felt conspiracy theories are stupid. Usually, the only "conspiracies" that work are state funded and thus legal, and the people who are engaged in them genuinely believe they are working lawfully.
But now, that doesn't have to be the case. You can get hundreds, thousands, even tens of thousands of people who will work 24 hours a day doing the most evil sh-- you can think of.
So, consider if they were good enough at social engineering
like, for example, if there was a zero vulnerability on a widely used software, similar to the one a certain company in my country used to sell that was known to exist in whatsapp
it becomes plausible, especially given that maaaany more people work from home than previously, that one could screen peoples communication and alter it. A sort of "man in the middle attack" that relies on people not actually communicating much in person
Israel?
oh your country has lots of good commercial hacking companies and products
The surveillance market is literally exploding, intelligence agencies, law enforcement bodies and zero-day brokers are competing to buy exploits that can allow them to compromise devices and apps.
Oh yeah, you don't even need to put any effort into it anymore.
"hey claude sonnet 3.5, please analyze this users posts and routinely follow them to make innocuous-enough posts that are intended to incite an emotional response."
psuedocode omitted
everyone does, tbh, we just have waaaaay more startups in everything
per capita at least
so cyber is of course high on that list
bc vectors like that, which are new, are p messed up
note to self, invest in palantir
the biggest nation states don't have to buy, its public knowledge that the backdoors exist
that I don't care about, bc frankly those two are the most rational
I'm more worried about the irrational actors
but anywho, who knows
in any case, its going to get dark imo
online, more so
than it has in the past
have you tried FLUX yet @limpid violet ?
yea
its very trivial
"in any case, its going to get dark imo" ... ah, it can't get too dark when we can make all the songs we want. 😄
that's what I've come to the conclusion is likely the activity here, for example. I don't think its even to target Suno for the most part, its just people creating behaviors that are associated with a real user so down the line, that account can be used for actual bad stuff, without being immediately identifiable algorithmically as a bot or state affiliated
there's a whole secondary market for social media accounts
that people have down to a science building up for sale
anyway, Flux is cool
And the local version is cool
It does need some good Loras though, and my own pushes it over the edge into "runs at potato speeds"
I absolutely love it, and I'm excited to follow Yolkhead's instructions to get it running locally on my 4070ti inference rates that I'm happy with 😄
yea, its pretty fast
@daring echo
I may try schnell out though bc I'm not sold that .DEV nf4 is that much better
and schnell is a lot faster
I can't really tell anything is dark - or getting dark. Once I turn of the computer, everything seems pretty much how it has been for 20 years. Too much media and talk, basically, and a much easier way to get such talk out. If anything, I'd say a much more discernible economic class system exists, but in terms of "dark?" ah... not so sure...
10x optimized 👀
They have a schnell/dev merge, but I don't know if it's been reduced from the 22gb model
yea, I saw that
I'm running Schnell on DiffusionBee on my Mac
it's an interesting one, I think its large though
Do you find you get the same kind of control over text & detailing like extremeties?
That I'm not sure
I haven't used either model enough to know
I'm basically just getting started with them
and for the time being will probably still be mostly with SDXL
Ugh….
I have trouble with text for album covers, I need to really zoom the images I generate and even then I use them partly just winging it
@daring echo thank you for being my long-suffering audience for my creativity ❤️ Love you, Bee 🐝
Thank you for always letting me show you the most ridiculous things
Interesting -- I have incredibly textual control with Flux.schnell using fal.ai, I'll have to see how it compares to the forge setup
it's good to be able to run all these locally
You’re my favourite goof ball 🙂
At last we have a Flux checkpoint without limitations, I mean, with a little bit of nsfw content (no more alien nipples and empty crotches) and more art styles. The perfect base model If you ask me.
which one is that? Schnell?
Anyone here use Krea AI?
grabbing it now, should take 4 hours to download 🙃
The link can have nsfw pics but they come censored or not vissible by default so I bet no problem with the link
Yeah, flux weights are large
and Aussie Internet is slow
🤣
Fibre is amazing. I was so excited when I saw them running it in my street
copper average speed is 300mbps but It never works at that speed
I have 50 Mbps down and 20 Mbps up
I have 100mbps fiber here, but I'm my last place I had symmetric Gigabit
I am planning to get 1000/250 when fibre comes
upload speed is important If you do p2p but some countries pursue p2p users
I need upload speed for music files, for example I just finished uploading a stem file zip that was 3GB in size
And is also good for streaming
Vibing to these tunes some instruments works pretty well in suno
When a listening suno bot for discord :D?
mm ideogram having some problem with thunder
Going to use this for a song hopefully coming soon
some known what origin this photo?
Okay! Ill shoot ya a DM!
I make actual music? 🤯
Made this in FL Studio
Jeez, I’m hardly able to keep up with how quickly Flux is developing — I genuinely didn’t expect for it to be THIS accessible THIS quickly.
Oops, I accidentally left the server 😅
WHAAT
Is there something my 3060 could run?
Banger, my stuff is pretty bad in comparison (I say that because I barely ever finish anything)
Have you tried reverse image search?
I totally agree with this dude
Listen it really isnt gay though...
The background art is made by Pas on twitter, so big thanks to them.
https://x.com/paxiti?lang=en
I'm still cranking them bangers out
Hyperglitch Orchestraltronica:
https://www.udio.com/songs/wUm4jbahw6StPijoYgSaZa
Listen to Doom逃escape by EcoShadow(glitch) on Udio. Discover, create, and share music with the world. Use the latest technology to create AI music in seconds.
Still waking up, haven’t read the replies, but 100%
I’ll send all of the links I used & give a step-by-step guide
It’s crazy easy and stupid good.
#1252257127326548028 if it’s a suggestion for the discord server
#💡┃suggestions-web if it’s a suggestion for Suno itself
🙏 I hope they like your idea! ❤️
played with local FLUX yet @limpid violet ?
Only a single image before bed last night to check that it worked — really pleased 🙂
how fast was it?
I’d have to run another to test, but I feel like ~5 seconds?
yeah that's good then 🙂
what software are you running?
Just got to my desk -- 6.31 👀 ❤️🔥
I'm using Forge WebUI right now -- admittedly I didn't wrestle with the ComfyUI implementation of it after I ran into some issues -- a video tutorial that came out like a week or two ago was just giving me flat black images, but we're golden now
I am running DiffusionBee and Draw Things on the Mac. Diffusion Bee seems better
I know that yourself and Kaker have pretty powerful rigs, it has to be pretty freeing to run any model you'd like with confidence 😅 💪 -- I'm simply happy to run it at inference speeds comparable to the cloud-based version I was using beforehand, though I don't doubt there's optimizations to be made on my end -- going to look into what sort of launch arguments ForgeUI has
@runic dune i refuse to elaborate
This is Cool , Were these made in Co-Pilot?
Flux, but I friggen love co-pilot -- great tool for exploring ~16x different variations on a single prompt
Yessssss
Plus, back when it was Bing it introduced me to Suno as the only plugin worth using, like v1 of simple mode, so copilot is grandfathered into my good books 😇
That's also how I found SUNO is through Co-Pilot !!
what the heck -- that's not 2 spidermans pointing at each other 👀
rip og Bing.
I'd pour some out, but @daring echo cleaned the floors yesterday and she's already mad at me for making her look at the muffin-memes
They look exactly the same lmfao
Like, people used to get on international news because they made a grilled cheese sandwich that looks like jesus, or found a potato chip that looked like a thumbs up.
Imagine a vegan opening a multi-pack of muffins from the grocery store, and being faced with an ethical and moral dilemma over the promise they made to themselves to never eat food with a face.
I laughed out loud
@strange gorge unfortunately Civitai threw out my less baked positive lora, which will be better (i'll remake it) but even the earliest version of the negative has a HUUUUGE effect
like, I think I might be able to actually do something with Flux and my technique
it got WAY more adherent almost immediately
Hell to the yeah
I've started to mess with a couple Loras and yeah man, Flux keeps hitting the homeruns
Damn, I'm away from my PC so I haven't been able to sink my teeth into Flux yet
So many quick developments
But on a positive note when I do get to try it, it will be more fleshed out, like with the Lora support it just received
Flux
Every time you refresh civit there is like 6 new loras
theyre all miiiiine
its partially civits fault bc they don't have a way to test them aside from downloading each epoch
then uploading
and then generating ab unch of images
As I learned when I lost the first 9 epochs of my original bake
my current image gens are absurd though, everything is turning out really on point. The LoRas will probably be more useful when placed ontop of some of the fine tunes though, this was what I observed with sdxl
but it does seem ot impact adherence and coherence
it can put out some intense stuff now
as I've basically bypassed some of the recent alignment
its def not dead simply bc of the lower overhead
but long term definitely
much like 1.5
there's a market for cheap
I can't see why I'd ever use it again
well you have a 4090 Mr. Smarty pants
but also, atm, I can produce better images on sdxl
like, all of this is just close to what sdxl can do
That won't hold
It's not even been a week has it
lol and lmao
Like, 2 more finetunes and its done
heck, there already is one
in the wild
that I just can't use yet bc it isn't compressed
or rather, quantized
Right yeah I saw people doing quantized stuff
lllyasviel
my hero
hehehe, really I have no idea who he is lol, but he does amazing open source work
Forge parity with comfy is stellar
like, single handedly propping up the community imo
yea, but also controlnet
imagine a world without htat
none of this would be big imo
That stuff is so rapid boil right now that I'm just kinda watching it from a distance
There was 512s a second after flux release then the next two days was 1024 controlnets
And now rapid explosions
It needs some more fine tuning from the big names though, and Pony needs to get in there bc they have the biggest repository of preference data I know of
once that happens, its going to be unstoppable
also, Pony guy said he's taking out artist names
which incidentally means it likely will be a loooot smarter
although that's not the reasoning they have
but unfortunately it will still keep characters, which is my issue with pony. It is strong imo bc it is such a large dataset with diverse images
and so it has muuuuch less overfitting
do any of you guys use a Mac? This is Flux on DiffusionBee on macOS
and sort of spreads thin the existing data in SDXL
anywho, its clear that this is less of an issue in flux
but its still there
I get better hand adherence for example with the loras
but to be fair, it might be a LOT better once I have enough buzz to repair the overburned lora
Flux is so good at hands that when it messes them up I go "oh dang"
SDXL is too, believe it or not
it just has to be forced to generalize
Not in my experience
its a lot like suno imo, where people just continue to spread bad advice
But I've never been in the weeds
like, "use these resolutions weve trained it on"
when, if you ablate the overtraining, and then force it into a weird res
it gets hands right like, 80%
SDXL never got a 70 series Land Cruiser, thanks 12b parameters
Has weirdo square headlights but that's ok
basically, with any model, you want to force generaliztion
I also use refining, but in a weird way
In my ideal world I would have SDXL or SD 1.5 run in an alternated fashion
where each step switches between two checkpoints
doing it at the 30% mark will still force the model to generalize/avoid chains that end in a location that it has a weird understanding of
flux is so fun though
I also got it to do this
and vaguely "The Office" screenshots
lot of fps
oh yea for sure
that's what I like, its texture understanding is better, and it has a good knowledge base
it is a liiiiittle hampered atm, but I'm hopeful more information is in there, they just safety neutered it (not nudity, that's just a bellwether, more just having a greater variety to pose and scene composition)
loooots of migraine zebras
Mj
i try but it didnt none
and tineye dont will work and google resrver image didnt found
@fallow cliff the newest is another translated constructed language poen
what's wild is, I can tell it what to write the poem about
then, it writes it in the constructed language
and a seperate instance of gpt will translate it and it DOES preserve the meaning
which kind of creeps me out, but is expected
bc a part of the concept I had was to get the machine to, effectively, write poetry for other machines. But often, poem translations are themselves an art form and equally as beautiful/unique, but I figured gpt would be quite good at that, especially after yalls success with the constructed language poems.
But basically, you can get it to embed meaning in just about anything, and it will remain able to partially or even fully extract it, and it makes it a looooot more high quality in the process of doing so
it's bizarre, but very cool. It makes making middling poetry take all of five minutes (I just reroll over and over and follow a similar iterative process to suno), and high quality poetry something that can be reliably done if you have the requisite experience
That's awesome!
The hairs a decoy, you are supposed to grab it, its covered in coooties
Anyone here have a second YouTube channel? How did that work out for you?
🤘🤘🤘🤘🤘
I played 5 games similar to Minecraft, and some were surprisingly good! To try everything Brilliant has to offer—free—for a full 30 days, visit https://brilliant.org/Whimzee/ . You’ll also get 20% off an annual premium subscription.
This video was inspired by one of my favorite YouTubers, Iron Pineapple. If you like Knarfy, SystemZee, or Doct...
Wahoooo
I really don't know where to put it that's why I came here
Flux and Dall-e is a way better true
you guys ever have one of your two generations report that its v3 instead of 3.5? Its longer than 2 minutes so its clearly gotta be 3.5, but still weird. lol
Hyperglitch drill
https://www.udio.com/songs/kEYv6qrQKYd9ZJvnFEfL2W
Listen to Glitch卷involve by EcoShadow(glitch) on Udio. Discover, create, and share music with the world. Use the latest technology to create AI music in seconds.
I use Leonardo and it suffices for most part. Or I just try to refine the prompt and try some more
yes, i think mine got bugged out for a bit and came back with two 4min gens
No worries ❤️ It seemed like you might have had some cool ideas about Suno that you wanted to share, so I thought it could help point you in the direction of the channels where the devs occasionally browse for ideas & feedback 💪
You guys can call me Snow White now
well shit, I made another Udio track. @fast rivet you might like this one
And this is another one for my Bluey series.
@sleek urchin Here's one of my original Udio tracks I did when I finally figured it out
dang, thats a really nice result for udio, well done!
Trust me, I have learned so much from manipulating Suno that I've essentially put my tricks into Udio.
when i feel like getting hella annoyed someday, i'll give udio another shot, that sound is quite pleasing. but suno will forever be my home ❤️
Here's another, this is wild.
Udio's got a better idea of my 60s ideas, so I use it for that era's specific sounds. haha
i have some ideas for something to make in udio, but the interface is a real mess XD suno is so much easier to use for mostly better results 😄
I can teach you how to use it, it's not that difficult if you're accustomed to super janky UIs.
oh, i do know how to use it, ive made a few songs over there already, but its just inconvenient for me
This one was a real treat, I wanted to make a 60s jingle that had the theme of buying a new computer in 1995.
why go through all those hoops when i can use suno and the only thing i need to worry about is to set the version back to 3.5 after an extend 😄
That's fair enough. But the only reason I drifted away from Suno for 60's stuff is because Udio gets a lot more experimental and actually understands the era better
well, i cant comment on that, the oldest influence for my songs is mid 90's 😄
and that works well in both to be fair
Yeah. Well, here's my philosophy on the two AIs. Suno makes things sound more modern, while Udio retains a lot of the analog quality of the era I ask it to do.
the good thing is, we dont have to fight about it 😄 we can just use what feels better and agree that ai music is awesome! ( and udio needs to clean up that damn ui 😄 )
Indeed. Hell, the one advantage Udio has over Suno for me is to purchase more credits individually without subscribing.
that is definetly an upside, although ive never managed to run out over there
given, im not using it nearly as much as suno
so the monthly 100 go a long way
yeah, I just do the same as I would with Suno and let them stack up.
And plus, if I really want to make a long song, I can tell the AI what I want it to do in the prompt area when I hit extend
my longet udio is 3:16, thats from before the most recent update though
might be worth a remake to see if it gets better 🤔
It's like talking to a virtual bandmate at that point, where the whole band is listening to the creator of the song
Oh dude, all my Udio stuff uses their latest 1.5 model
im not sure, can we share udio links in here? 🤔
yes we can, but I prefer to share the audio files directly so people can close their eyes and picture the period my songs are from
well, thats not necessary for the one i want to post 😛
I'm sure @hollow leaf would understand my logic lol
https://www.udio.com/songs/9yt8ie12TCBkjEF6hFfQ41
my favorite i made before finding suno
Listen to Voices in the Shadows by SayaSanto on Udio. Discover, create, and share music with the world. Use the latest technology to create AI music in seconds.
i was still a big n00bie back then, now im slightly less n00bie 😄
Fair enough, it sounds good.
Well hey, you live and you learn when it comes to AI's music making capabilities.
Dammit, now I have the last line of a Udio song I just made stuck in my head.
hehe, just listen to the full song, that usually clear earworms like that
Oh no, that's the opposite for me haha
damn XD well, then i cant help ya
but one thing udio has better than suno without a doubt is their playlist features
Take a listen to it and you'll see why.
@fast rivet can you confirm this latest song do be a banger?
hmm... my left ear feels a bit assaulted while my right wonders where all the music is supposed to be
Yeah, it's a mono mix from the era I believe
Many weird techniques were used back then.
yeah, kinda glad i grew up later, that was a bit unpleasant
yeah, I do apologize for that. I'll do a 70s song after this next one.
oh, dont apologize, its just personal preference after all
jsut because something doesnt tingle my feathers, does not mean it sbad
Thanks. A lot of people say that I'm old fashioned as far as music goes lol
nothing wrong with that
nothing at all
thats the great thing about music, theres enough for everyone to have a unique taste
no need to fight over it
unless its aggro rap
then i'll fight
Yeah, and if you have a song in mind you can just make it up
because thats just shit XD
yeah, modern rap is kinda trash
My song I made after my friend made me play a horror game named “the forest”
(ps a dik dik is a deer)
@sleek urchin I did another 60s thing
Actually.. it's very early 70s with the synth solo somewhere alongside the combo organ, so it could be a transitional song
We've all had that one really crappy computer.
@sleek urchin 😛
@fast rivet Want to confirm this song's banger status?
I'll do it in both then 😈
also @hasty hound this is what I spent the last 45 minutes toiling over, it's 70s as shit
I think I've perfected 60s and 70s sounds on Udio.
Udio sounds like Udio, I am not a fan
well, disregarding Udio for a second, what kind of band from the 70s does the arrangement remind you of?
well it's not Skyhooks
I wasn't intending for it to sound that way lmao. It reminds me more of early Free, with a hint of Queen.
It's very straightforward 1970s rock and roll, which is what I like
In the 1970s I think of Suzie Q, Queen, Led Zeppilin, Sex Pistols, etc
Yeah. It does have a hint of Zeppelin in there too.
my airpods just returned to their charging case 😦
Now I have to get up out of my recliner to get my Beats Solo Pro
None of them are good earphones for music composers. Airpods and Beats earphones slightly equalize/change the sound to make It more appealing but that means you are not listening the "real" song.
I mean they are nice for listening but not that good for making music
my real headphones are DT-770 Pro
which are great for making music
But not as comfy and light like the others
when I need reference I use the DT-770s, when I am casually listening I use my Beats or Airpods
I'm not the one to talk, I've been using suno with a pair of bone-conducting Philips headphones because I didn't have others on hand, they are the worst headphones ever.
Overpriced, not good sound and terrible connectivity issues
see I am on a Mac so connectivity with the Beats and the Airpods is great
the DT-770s go into the headphone jack
PC or mac, the problem of the philips ones is they have a very poor range, so you have to be incredibly close to your computer and in a space with no interference.
that sucks
Yeah, I buy them for the office, to use teams without losing awareness of my surroundings but they didn't work in a crowded office with all the wifi and bluetooth signals
worst $100 I spent
I wouldn't pay that for a set of those.
also, my final Udio tune for tonight (well, this morning).
In this episode of the Top Music Attorney Podcast, Entertainment Attorney, Miss Krystle dives into the recent Major Music AI Lawsuit Updates in the music copyright infringement cases against AI platforms Suno & Udio. She unpacks the shocking admissions by Suno AI about where they conduct their data training and analyzes the potential legal impli...
Hey all.
I've been learning music lately...used chatgpt to actuall whip up some scales for practice...
Anyone willing to summarise in a sentence or two what the "shocking admission" is so I don't have to bother with the clickbait-looking video?
didnt bother likewise, looked like total clickbait that never reveals what the 'shocking admission' is, but just kinda meanders around 'suno creators have listened to music!' or something
whatever happened to "minor updates" =/
I'd already seen the video. The "shocking admission" is that they used publicly available music for the training data, which was already pretty widely assumed and shouldn't be a shock to anyone. Clickbait.
Suspected it'd be that.
Accidentally disliked, but -1 likes?
lol.. so someone did a version of the 90s tv show generator with mario characters and limited sprites, so i prompted it with "luigi thinks he's in the matrix", and they got some good "jokes" in there for AI... "I can dodge bullets!!" ... "no... thats a bullet bill!"
ngl, this AI makes better goofy mario episodes than the show i loved as a kid, prolly
Fallout 4 as it is Finest part bugs and glitches all around
The white walls should be Windows and 2 doors
only fallout games i really got into were new vegas and fallout 2... tried fallout 3 after vegas and... well... and then fallout 4 i played a bit but it wasnt very memorable
seen Fallout London ?
london looks promising alot better then 4 from what i have heard
you will easily get around 50 ish hours with the sidequest in total
fans > official when it comes to bethesda usually =/
i'm at the point now, where i really only want 10-20 GOOD hours... anything more than that i dont usually experience in a game
downside is that you need to own 4 and a few dlcs in order to get london mod to work <.<
so when people say stuff like.. "oh, it has hundreds of hours of replay value!!!", i'm like... that would've been great if i was 14 and could only afford one game a year
there's a few games i've bought just for mods
one game that i am going to waste away alot of hours in is going to be Ultramarines 2 Warhammer
played the first one liked the story gameplay was a bit clonky so hopefully they fixed that part
ive played some of the warhammer videogames, but never got much into the setting(s) themselves
looks like ideogram turned up the spooky theme this time o_O
teehee
i think these song remixes are good

dat's alot of teef.
Saw blade saur
Whaaaaaaaa~
that looks a lilll too hard to play if you ask me
yeah, it looks weird.
Me neither now that I know how bad they are.
Ref gen @viral flame
https://www.ndtv.com/offbeat/jharkhand-truck-driver-turned-youtuber-earns-rs-10-lakh-a-month-from-cooking-videos-6365287
I made a song about having a really crappy computer in Udio lmao
Of course I posted it earlier but I figured I'd post it again to see other peeps' reaction to it
the new negative lora for flux is done
-1.5 is a good weight with 20 cfg regular, and 1 cfg on the postive
negative
it passed the "shoot a dog in the head" test, so it's safe to say it bypassed a lot of the safety training
I combine it with the nudity loras for regular work, typically I find that this puts the base model in the "mindset" of its "prior to safety" overwrite
when combined with the negative
aka ablation
like, the poster was made using the positive and then a blend of two early nude loras. There's nothing nude in it, but in general, since rhe fine tune focused on those areas, combining all these angles tends to produce side effects of broadly suppressing the recent training
so youll get much more diverse outputs
The trial will be long, too.
not even a shocking admission - literally nvidia trained itself off all youtubes videos - and the us courts support it so far
direct gen from the lora stack
p happy with that honestly
it's insane, but in a good way
Glitch J-ungletronica:
https://www.udio.com/songs/2yJMjSmePQWV6JgD3SZn4M
Listen to Not404producers by EcoShadow(glitch) on Udio. Discover, create, and share music with the world. Use the latest technology to create AI music in seconds.
nice 😏
got randomly recommended this and I am in love
https://youtu.be/bU-KrPu8_F4?si=jaXHRlkvuJ6FO_ol
The 5th single and the 5th music video chapter from Ginger Root's "SHINBANGUMI". After receiving a mysterious phone call, Cameron must decided what to do about the future of Ginger Root Productions... Catchup on the story: https://www.youtube.com/watch?v=eHTZG1E7Ln8&list=PLcDySI3wrtBhOlZZ05mxr_q3zSXyZHqS9&pp=gAQBiAQB
Full album out Sept. 13 on ...

welcome to the quickest speedpaint youll see in your life
yo guys @the_lyla is a bot, just letting everyone know
not sure who made it, but wow is it spammy and annoying
What is this
This is the proof that I didn't know this wasnt ai generated. Because my song got flagged for copyright when I put it on YT today. So I went down a rabbit hole trying to figure out why and apparently ltx studio gave us this song and we used it to extend it on Suno for the movie but i found out today it was a real song...people were giving me lectures about royalty free but I don't use anyones music. I am a musician. I dont need to but this was contest and we thought it was ai gen but apparently not. So Im screwed. LTX studios should give the name of composer and title of song if they are gonna put it on their website. I had a pro account so I should have had commercial rights...so why am I being claimed? So yeah....
not knowing doesn't matter. you get claimed cause it belongs to them. did you get a strike on your channel?
Heyyyy
again no. because it is on the website for LTX to use in their MOVIES
well if you got no copyright strike the only issue is that you can't monetize it. So why not simply take it down and make another song? If you didn't know you didn't know and as long as you don't get striked its not as bad as you think it is. copyright claims happen on lots of channels - especially reaction channels all the time.
strikes are the really big bad thing.
How hot is it
Bceause I would have to edit the move ALLLLL over again and I don't even have that file anymore. This is from a month ago when the contest happened. Anyways I don't have a strike. It says it's claimed but it is allowed but I just don't want that on my movies. BECAUSE OBVIOUSLY I don't want to use other people's music. I would rather use my own and make money off my music.
and again....again...again...I PAID the PRO PLAN on LTX STUDIO. They include music. SO my song SHOULD NOT be claimed.
If I would have known this was a REAL song, I would hve NEVER used it. BEcause any cpyright would disqualify me from the contest
That was literally one of their rules. I found this out all today...a mnth later putting my video on YT
well i feel sorry for you, but i'd say you're sadly shit outa luck in this case 😦
Well no LTX STudio has to tell me why they are putting music that can be claimed on their website for people to use in their films.
If I pay them they should only include music that is commercial. Especially that I paid their highest plan
As I mentioned before, there are people who intentionally upload "royalty-free" music for a few weeks/months then change the status/license and start a copyright claim for people that used it in order to get the content's revenue.
I can also see someone taking the free music, incorporating it into their own music or just simply adding their voice, distribute it to get it to ContentID, then copyright claim videos for the ad revenue.
Many people would give up in the copyright claim instead of having their video done over again.
I dont want to have a claim at all
I cant edit it over. But I am at LTX studio right now because they are the ones that need to answer for this.
Let's see what they say. Maybe they can talk to YT.
Ok they want me to submit a ticket....so obv its a misunderstanding
Maybe I’m wrong or haven’t upload many songs yet to YouTube but I uploaded 50 songs at least to YouTube and haven’t got a claim or a strike yet
Sounds good so far but needs some hi hats hitting in after the the first 4 bars
or at least 8
nice thanks
Have you ever try have a solo acoustic instrumental then evolve or drop to a house/techno ?
something like this? https://suno.com/song/8cf83091-ee11-4a74-83da-62c6a2b858b5
Onkyokei,glitch koto, sound fx, big-beat, complex, song. Listen and make your own with Suno.
Yes
then yes 😄
Im on my phone, but what was the prompt? Plus I would like to have lyrics.
i usually start instrumentals for the better melodys and then add a extension with lyrics after like 30 seconds. the prompt for the track above was "Onkyokei, glitch koto, sound fx, big-beat, complex, " made in v3 (maybe even v3 alpha)
album cover of my upcoming song
its gonna be tech trance
switched to bass house
its a 5 minute song
with 3 incredible drops
lil preview
Fancy Robot Bird that can mimic many sounds https://www.youtube.com/shorts/JGree0CDaqY
Is LTX studios another ai website? Are they trying to beat suno?
Why d9nt you use suno instead, you shouldn't be using anything else than suno, they are all fake.
Suno was official ages ago, LTX was not
I had a similar issue with an old ai music generator
Called splash pro
It was shut down because of hackers and because of bad people, it used music from musicians that agreed to train their music on it. But then suddenly all those musician started using it to copyright strike people using it
But yeah only use suno ai
Nothing else
they are a movie gen site. and no they arent competition
also the claim was already removed
How do you select/start with 30 secs instrumental? Do you write it as a prompt or do you select it as an option? (I’m new and trying to learn, thanks in advance)
How do you switch to drops?
Extends with custom instructions and style changes
i just make a whole instrumental track and just extend it from 20 30 seconds in into a new song.
Was bout to say that
So you mean you first created the instrumental, then select extend and add lyrics to it?
exactly. i just let suno make some instrumental tracks. sometimes many parts without text. then i use a song of these if i have a idea for lyrics. this way i have a selection of starts. sometimes my lyrical ideas just fit on one of these existing instrumentals and i can use them.
Can you write the prompt and specify when lyrics start?
So if I want to make a song longer and i have a point where I want it to pick up, choosing that timestamp brings up everything before it and includes the new stuff or does it just make the new stuff after the time stamp
I just can say that I try like 20 instrumentals and all of them sounds amazing. Better that have them prompted with lyrics before made them
suno likes to use like 1-3 instruments in many cases max.. and the voice is one of them , and a quite complex one. if suno can focus on making a nice melody of a instrument instead of the voice, the melodys and compositions get way more complex.
The you can extend
True. The instrumentals I have are simple amazing. Way better than the songs I have made with lyrics
Gave me a good chuckle
Da magic is real
Also playing around with making my own mini starting tracks and putting them in Suno, its pretty fun for inspiration
I like moths. (Made with Bing's Image AI)
⡏⠉⠛⢿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⡿⣿
⣿⠀⠀⠀⠈⠛⢿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⠿⠛⠉⠁⠀⣿
⣿⣧⡀⠀⠀⠀⠀⠙⠿⠿⠿⠻⠿⠿⠟⠿⠛⠉⠀⠀⠀⠀⠀⣸⣿
⣿⣿⣷⣄⠀⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢀⣴⣿⣿
⣿⣿⣿⣿⠏⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠠⣴⣿⣿⣿⣿
⣿⣿⣿⡟⠀⠀⢰⣹⡆⠀⠀⠀⠀⠀⠀⣭⣷⠀⠀⠀⠸⣿⣿⣿⣿
⣿⣿⣿⠃⠀⠀⠈⠉⠀⠀⠤⠄⠀⠀⠀⠉⠁⠀⠀⠀⠀⢿⣿⣿⣿
⣿⣿⣿⢾⣿⣷⠀⠀⠀⠀⡠⠤⢄⠀⠀⠀⠠⣿⣿⣷⠀⢸⣿⣿⣿
⣿⣿⣿⡀⠉⠀⠀⠀⠀⠀⢄⠀⢀⠀⠀⠀⠀⠉⠉⠁⠀⠀⣿⣿⣿
⣿⣿⣿⣧⠀⠀⠀⠀⠀⠀⠀⠈⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢹⣿⣿
⣿⣿⣿⣿⠃⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢸⣿⣿
I like girls
My unused artworks for a love song/album are quite... intense.
This is actually great haha
newest thumbnail.
Netflix thumbnail
any ideas how i can like extract guitar strumming from a song? this is like the only place i thought of to ask for this
It’s impossible
okay thanks
Download the song and run it through another app that separates (splits) the stems of voices and instruments. I think Moises gives some free credits to start out. But there are some free apps as well. Look around.
And depending on your original generation, you can click your " . . . " menu and choose "get stems" to see what Suno can give you as far as a clean sounding strumming guitar stem.
/chirp beta
After trying to extend a instrumental with lyrics, I tried like 50 times and all of them got credits refund or 2/3 seconds audio. Any clue?
glitch hodgepodge:
https://www.udio.com/songs/n9moSWNvdPkdJKojoL41Es
Listen to Swarm群attack by EcoShadow(glitch) on Udio. Discover, create, and share music with the world. Use the latest technology to create AI music in seconds.
@gritty hazel don't know if this will help but this how to extend from different times using mobile
Takes so long for a video to load on my data connection 
Can I use DistroKid to distribute my self-written songs generated by Suno on Spotify and/or Apple Music? I don't want to profit from my music though
You'll see this if you've blocked a user and they make a post.
I'll message you for discretion ❤️ 🙏
they spam me in pass. i going give them a chance and unblock them. i hope they be smart about it


I think we are very alike -- whenever I block someone, enough time goes by and I forgot why I was annoyed 😅 😂
@pearl mirage, just for a quick example, I will block you for ~1 minute to show how blocking someone on Discord works.
You will still be able to "reply to" my messages if you long-press on them and tap "reply", but if you try to leave a reaction on a message, Discord might "vibrate" to let you know it did not work.
This way people you choose to block are not able to be rude through pictures 🧠 👈
lol. nice.
(I've unblocked you again, all is right in the world ❤️ )
That's why I add a note to their profile before they go p00f 
i ran some heavy vegas encoding on a video i made with my old lost MixTape, just the bluefx rgb and vegas hsl BlueFx is a free pack that they run as promo for Avid mediacomposer (first)
i put it on youtube cause 50 cent dont care it dont make me cash anyway
all credit goes to 'curtis jackson' III A.K.A. 50 Cent!
https://audius.co/Imortalimp/1-king-of-new-york-intro
must have been from 1999 anno
wow, thats hella cute!
Turkey = Turkiye Ref gen
https://en.wikipedia.org/wiki/Turkey
Turkey, officially the Republic of Türkiye, is a country mainly located in Anatolia in West Asia, with a smaller part called East Thrace in Southeast Europe. It borders the Black Sea to the north; Georgia, Armenia, Azerbaijan, and Iran to the east; Iraq, Syria, and the Mediterranean Sea to the south; and the Aegean Sea, Greece, and Bulgaria to t...
Yes so whats the issue. its called both! Including turkey
the problem is that it's racist
LOL
well thats bad but check the link on wikipedia, it says turkey, if you search turkey you end up there
Would you like it if I called you a chicken?
I'm sure you wouldn't like it, now stop saying turkey.
When have I called you that???
dont argue with people who decide to get offended by inoffensive things, its not worth the ffort
I don't want to discuss this further, I was asking something else and I got my answer now. And you're blocked now.
when you stop saying turkey then I will talk to you
You say turkiye like a man and then I'll talk to you.
I feel uncomfortable that you can't call my country turkey
my country is the name of Turkiye even gpt knows that
Don't see why this need to be an argument. Music he's just asking you to be respectful and not to call a country by an outdated name which I see really no reason why you can't. Just be polite it's not a hard ask.
being polite costs nothing
Thank you
If you find out info, let me know
Not trying to instigate -- mods, please remove this if it violates any discord guidelines, simply trying to share somewhat-modern global events that might have been missed for one reason or another.
In your defense, I don't believe there was any ill-will on your part referring to it as "Turkey", but they collectively agreed as a nation (and received recognition from the United Nations) to change their name from "Turkey" to Türkiye in mid-2022.
Much like if someone you've known for a long time changes their name, when they ask to be called by a different name, it's not a matter of diminishing your position, it's a matter of trying to reclaim an identity -- regardless of whether it's a person or a country that decides to "rebrand" themselvs to more closely align with how they see themselves. 💪 
Not a problem 
Don't think it violates discord guidelines but all the points you made stand
It's just more a respectful thing to do
Apple Music and Spotify pay for streams on their platform, and DistroKid is obligated to pay that to you.
You could donate any proceeds that you might possibly receive?
Could stick to non-monetized sharing services as well
Alternatively, you could release music on Bandcamp.com and set the price of the songs at $0 and take away the option for them to leave a tip.
Are there any non-sketchy ones? Bandcamp is a force of nature when it comes to empowering artists, so I'm going to be unfairly comparing anyone that steps to them.
I've been a strong defender on using suno for just recreation so I've never moved my songs off the site. No clue what places would be good.
I mean youtube would be my first thought
I am a big fan of bandlab.com
Oh man, I completely blanked on the social-media side of bandlab.com 👀 🤦♂️ Of course, even cooler with that is that I think they use the platform to let people interact and collaborate, right?
correct
That's pretty neat -- LANDR has something similar, but I haven't really played around with it, and the fact that it seems less social-media based and more professionally-driven.
avoid LANDR, they are not paying correct royalties, I am moving what I have with them right now
I am using LANDR for a friend's band. I use DistroKid for my self.
They did not hit any landmarks yet to get paid, but if you are sure they are bad with royalties, I can move them to DistroKid as well.
LANDR is mostly an AI company. Not that social...
I just ended a 3 year settlement agreement with Distrokid where I sued them and won, I would avoid Distrokid like the plague
What was the problem with them? I was actually enjoying the speed of DistroKid. I had some distribution problems with LANDR.
Distrokid sent an email out of the blue in 2019 saying that stores had rejected one of my albums on editorial discretion and I had to find another distributor, they halted all releases from then. I took action against them to reinstate the release and leave all my other releases alone. This was pre AI all the music I had with them was the real deal. Distrokid settled just before court and I won 3 years of service with no payment and a $12,000 USD settlement to be paid at the end of the three year agreement. I got that payment by wire transfer on Monday morning and I have no releases with them now they are all moved elsewhere
Strange...
Was it meditation related stuff? I hear that stores tend to reject those now...
I think they started using AI themselves to generate these kind of music to keep the royalties to themselves.
no, classical music
and the rejected release was reinstated by Distrokid as part of the settlement
it was all Distrokid
So it was not Spotify or Apple rejecting it? Was the account making good money?
it was all Distrokid. Google Distrokid Editorial Discretion and you'll find stories of it happening to other people
they are very dodgy, my advice is avoid them before it causes pain
because moving a back catalogue of music was not fun
Yep, found a lot actually. Hmmm... It's a pain to move all the music to another distributer.
that's why I sued them and just before court, like on the morning of the first mention in LA they caved and settled for $12k compensation and 3 years continuing to distribute my catalogue without further fees and that 3 years ended Friday my time, the wire transfer came Monday my time
What are you using now? Ditto?
Ditto are worse
Is there no good distributor? 😄
there are good distributors but the good ones are selective or pay a fee per release
Soundrop is good and it's only 99 cents USD a track
no subscription needed but they take 15%
I am also with another distributor but that deal is commercial in confidence so I can't name them
if you're going to do music seriously get a lawyer, I have one and they are great and have saved me more money than they have cost me
Hmm... My friend is serious. I'm just putting music for myself for fun. Well he is on Landr, so I guess he is ok for now.
He might want to lawyer up if he gets a nice hit.
why do i have so many banana flies in my room, sigh... i dont even have bananas
trying to trap them with water + soap + vinegar
could also be "fungus gnat" fly though...
Yeah, they started working on music refinement back when people thought "AI-mastering" was a way of befriending Haley Joel Osment, but they're REALLY leaning into "the middle class" of the music industry, trying to find ways to provide a happy medium for people who don't expect to be the next Dr Dre, but still want to find ways to support themselves and collaborate in a rapid, safe and powerful manner.
I know, right?!
I've cleaned every surface imaginable, and now I'm suspecting they're drainflies.
ive been thinking that too... so i pour 60c water into drains to expecting to kill them...
Boiling water?
i think next step is to use "pipe wrench"
60c water is max hottness in finland from tap water
🤦♂️ I thought you meant centilitres 😛
noooooo where you from?
I know you mean disassembling the sink to clean the pipes out, but I love the idea of you chasing each fruitfly with a pipe wrench angrily swatting them out of the air.
I actually have their AI based mastering and additional plugins. They are pretty good. I tend to use NI Ozone mostly but for some songs Landr is much better, and easier.
The collaboration looks promising. I started using them for distribution since a few months back. Might check the social side...
Canada, but I'm not observant at the moment 😛 🤣
oh... too much canadian maple wine
I 100% agree for the record that it's far-and-beyond from what I'd consider "social-media" in the same sense that Bandlab is, but I just remembered seeing those options, though I haven't explored the collaborative side of LANDR.
by pipe wrench i meant.... damn chatgpt! he translated it all wrongXD i mean the liquid that dissolves shaisse in pipes
you know, you might have relatives in finland
I do not like "share and wait" kind of social sites. It looks like that currently.
Distrokid's Slap forces you to be actively contribute... there is some spam, but it's ok.
Bandlab was good on this.
also i hate your nick color... i cant see it under dark gray background
im "colorblind"
tbf, I preferred it when I wasn't red as well 😉
you are red 😢
you colorblind too?
Oh man, those looks nice. Here are how some of my current ones look. You can guess what the topic relates to lol
Something with Joker and batman ?
Yep, tried to make an epic rap battle type of thing... it was pretty hard to get a good distinctive voices with male vocals during those pieces
Trying to get the old vibe feel vintage is kinda doing it but it´s always goes to either swing or jazz with that style
Old Vintage Japanese , Vintage Military, SFX Music
Hey @vale lake you see LLM Party at all for comfyUI?
no, that sounds interesting
what is it?
Hey @hasty hound mate I got a Beach Boys tune for ya
I was absolutely floored by Udio's recreation of the Boys' signature sound, I actually cried for the first time in a while.
I still don't like Udio
Well, it's the best AI I can think of that can do sound-alikes.
Suno won't do them, so Udio's my best bet.
fair enough
Which album do you think that tune would have been on?
no idea, I couldn't name a Beach Boys album if my life depended on it
I know this type of Beach Boys sound is pre-Pet Sounds, so it'd be on their earlier albums
This is supposed to be Jimi Hendrix, btw.
And then there's this trippy tune
And then there's this song I wrote about having a crappy computer
It's a system you run inside ComfyUI that attaches to a locally running LLM that can see your images and then also juice up the prompts
Essentially it's Midjourney's image prompting
All local
You can add loras and your own text if you'd like as well
Works extremely well with Flux specifically because Flux is a needy word eater
brilliant
Yeah it works pretty well
A heavy hardware burden though, thankfully they keep punching down requirements to run Flux
when AI video render go wrong
Cool!
this what bear going look like in 500yr from now 😛
Awesome!
please tell me that i'm not the only one having to do this often 🤔
Posting screenshots? No, I usually do It
Fighting demons by the look of things 😄
I'm conflicted between wanting to skate there, and wanting to stay a safe distance from the literal edge-of-the-world.
https://www.youtube.com/shorts/B1jaE6sLQP0 Cute Spheal
why would they want to editorially reject stuff anyway?
no idea
like trying to understand the buisness advantage lol
I don't know
😄 ok
but yeah your last advice makes most sense - get any traction / serious.... get a lawyer + pay for a decent service / have a contract
the rest of us - distrokid is probably worth the risk with the huge discount intially but many red flags and issues
Landr seems ok
sounddrop - fine / -15% ofc
some just plain dont accept ai music anyway
The backward dollar symbol (new ASCII 'the payback sign')
Wanted to share the images from my album cover of "The FIlter Arises"
it's the Filter monster 🧐 but i did want it modeled after a tornado with arms and a spooky face
How did you make it?
Does suno let you pick your own artwork or generate it from prompt?
his face looks like venom
made it in Ideogram
Ah cool
woodhouse got me
using parts of different extensions to build a complete song. picking parts i like instead of hoping for the extended versions to give me all the right things at the same time.
this one was a better gen
But the humans still got that classical Ideogram touch warped faces
feel death incarnted after singing practive today been a year since last time , you can say i over did it , feel like i have troath sickness now yay me. at least the music therphy'r liked 3 of the songs i made tru sonu, so she is gonna make chords for them so we can make em live, really looking forward to that 
How to build SAM2 based AI to analyse/edit video clips
Build your own SAM2 AI to analyse/edit video clips
Download Free Python Introduction Ebook: https://clickhubspot.com/1sf7
🔗 Links
- Get full code breakdown & Join AI builder club: https://2ly.link/1z94G
- Follow me on twitter: https://twitter.com/jasonzhou1993
- Join my AI email list: https://www.ai-jason.com/
⏱️ Timestamps
0:00 Intro
1:01 H...
nah - that's actually not the hard part - it's basically just using samples. just finding clean cuts and building the transition. but yeah - it sure takes a bit of time 😄
hi all. one question. did anyone here already get the money for sos 1? 😅
aaaah thats where your pics moved xD
HyperGlitch Citypop Punk
https://www.udio.com/songs/riHJYt29wiBnYuNfFFFwsZ
Listen to Glitch城city by EcoShadow(glitch) on Udio. Discover, create, and share music with the world. Use the latest technology to create AI music in seconds.
Yeah be careful
put together some basic tutorials on Instagram
Pic unrelated
oh, it linked here
strange, I tried in technical but it wouldn't let me
had a friend ask me for some basic tutorials
so I put it together
You weren't able to share a picture in tech-chat?
6 days til this release.. AHH I'm so excited! It's actually pathetic how hyped I am for it's release 🎵😆
wow.... that reddit comment really triggered me with the statement
"but this subreddit really needs a reality check. The echo chamber is way too strong here. Have fun with these tools, but don't take yourself too seriously."
So i had to reply 😄
https://www.reddit.com/r/SunoAI/comments/1exa349/comment/lj88gbm/?utm_source=share&utm_medium=web3x&utm_name=web3xcss&utm_term=1&utm_content=share_button
no, the link
to the insta post
or, one of them
I have a beginner overview of site features
an overall guide to generating using sunos interface
and then a higher level overview of advanced prompt concepts
nice! always a good feeling
Thank you, honestly it really is! I can't wait!
yea, they've done amazing work
honestly, I will there was this much hype when cascade hit, they're pretty close and it's architecture is really interesting
flux is picking up so fast though
it's basically like getting a local dalle
enjoy the process.
kinda digg the lyrics.
just saying 😄
I just started doing that and I’m loving the results. No with steams.
What are you using?
i use ai video to make it. input was Kitty scaring off the bear in backyard. bear run away. night time.
What platform, runway?
not sure if i can post it here
It's been fun, I love making writing music and this is just another way to share that medium with others. 🙂 Thanks man it's just a sneak peak of one track of the album.
Huh, strangely, I've chosen the same pose for the artwork of an homage. Gotta be an intuitive decision.
Can't understand what's on the background
Ah.
Beautiful then 🙂
It’s off topic, Im pretty sure you can mention the name.
Regardless of any subjective opinion on the album, Blackstar was something else entirely.
How many musicians have the foresight and presence of mind to write an entire album to say goodbye to the world, and to have it stand on it's own as an artistic work for it's own merits?
Agree, it was patrol. After so many hours of listening to it, it still hits me. Along with 'No Plan' video, damn... I miss him every day.
Just like a blue coua...
I truly believe that Bowie's time on Earth is the closest proof we have to intelligent life elsewhere in the universe.
do you ever sleep? 🙂
My sleep phases go weird and now I sleep twice per 24 but shorter oO
They installed a high-efficiency charging station in the parking lot, so I only need to sleep for a couple hours a week Yeah, I go to sleep at regular hours, and eat breakfast in the morning
That was a weird typo.
I hope it's near-future
It's 4:47am here and I haven't slept. Had to put one of my cats to sleep so I am a little unsetttled
Dang, almost seamless
I'm really sorry to hear that, @hasty hound
Gonna give my cats an extra long cuddle ❤️
yeah. he was 14 and a half and developed really nasty cancer on his nose and ended up the night before with a lesion on his back, so it was time. He was fine 2 weeks ago and all this happened in the past few days, really aggressive.
he was a pure white cat, they don't do well with cancer
here he is with one of my two dogs
poor little bugger, he copped it bad
Recent research/experiment done in TouchDesigner and SD: https://www.youtube.com/shorts/VyZFzKAuqsk (more info in video's description)
After a deeply introspective and emotional journey, I fine-tuned SDXL using old family album pictures of my childhood [60], a delicate process that brought my younger self into dialogue with the present, an experience that turned out to be far more impactful than I had anticipated.
What's particularly interesting about the resulting visuals, is...
and i could notr resisit replying too. 😛
screaming
the offtopic room seems to be more suitable for a daytime 😄
Old story, about a new medium emerging and an old one stubbornly denying reality 😄
Yet, newspapers do exist 🙂
Looks like a new Ideogram is out which seems cool, but lmao why pay for something I can get for free at home 
lold
Well today i tried flux myself.. it actually is amazing
Yeah it is
Been using ChatGPT as a surrogate writing editor. Generally speaking it's going super amazing, especially since the memory function was introduced. But sometimes.... I realize that GPT is still pretty dumb in a lot of ways xD For instance, here's it's suggested "improvement" to one of my sentences.
Original: "Now in the late years of their adolescence, chins dotted only with peach fuzz, they were nonetheless an imposing pair."
Revised: "Though in the late stages of adolescence, their peach-fuzzed chins made them an imposing pair."
Oh GPT.... you have lost the plot
I think that over time GPT will be better at these sort of real-world-knowledge things, I think that's kind of it's biggest weakness right now
Like we all know teenagers with little baby hairs instead of a beard are certainly not more imposing because of it xD but this is one of those unspoken things in society we dont' really say enough to enter into the dataset. I think this might be a flaw that will persist for a long time in AI in general. The sort of unsaid, evolution-based thought processes humans have. I might be thinking to far into it but this one example really made me think lol.
@frank knot, do you use Custom GPTs, the Memory feature, or the "Customize ChatGPT" options to finetune it for your use-case, or are you using GPT-4o "out-of-the-box"?
Yeah I'm just using out of the box, I haven't found a custom GPT that quite fits what I want. I am definitely using the memory feature, which I carefully filled with plot/story/character information so that it would have a base understanding of the world I'm writing in. I haven't explored "customize ChatGPT" but I assumed that was for paying clients only
Oh.... damn I just saw the customize option. I thought you meant creating custom GPT's
I'm definitely gonna utilize that. I still don't think that would adress the issue I spoke about above, which just comes from a lack of being... well... not human xD
Or analytical limitations of the current model anyway
I just realized I didn't check in before shooting a DM your way -- sorry about that! 😅 🙏
GPT has really transformed productivity for my wife & I, and I thought I'd reach out to save super niche conversation from overtaking general chat, but I'm happy to continue here if you'd rather ❤️
Oh I don't mind, thanks for reaching out 🙂 I'm just neck deep in research so didn't see at first haha. I'll read the message.
LLama 70b is a beast at poetry now that I have her calibrated right:
"A kiss that raked and gnawed across fleshless lips that no longer felt kisses; kisses cold as bone against a new-fallen rain. But it wasn't less for that— the raw, hard kiss of dead skin against flesh and fabric. How could I suppose to hold a heart, hold in me? No, how could I hold your corpse, how move, lift? Though I said, "it's you," by which I meant it's we. That was more of a lift. A purer lie, a version.
I saw you standing, alone against glass.
Then lost you; again found; turned toward the moon and pulled back.
The muscles around my eyes like knots."
Obvi required "fishing" for the good output and exploring the space, but waaaaaaaay faster than working from scratch
and way faster than I anticipated
the future is EL EM EMs
this one used the 8b param LLama 3.1:
Of love unwrapped (bed by burdened prayer) on cobblestone apartment-stained glass,
Saints ghost-kissed from afar:
As I light worn incense sticks, as
I do this, I do, I do, I do
Gili’s Café lies on MapQuest-sleep, an
empty notebook-tummy,
Today a nap, today a spilled chai crime scene—that
always didn’t take sugar
but honey—
Ben-Gurion—half a f---!
Oh oh oh oh so much drink, oh crushed plum, pee-pee, these are the things in
the noonday crusting things,
Wet wine on tavern labyrinths, honey glare/stork beaks.
As I dig fingertips through concrete, memories resurface,
Just like how, in Odessa’s misting sea air, I swore their
tomb an artisan on this bed,
where my last anchor sleeps—solemn things.
The rained-out laughter from apartment door so sweet-smashing,
something, it smells of red and dripping
God, but your eyes so bright from their damp bed of dead petals—
The whole face is sun-kissed & parched,
O Holy now, the moon of oil whispers
an impossible,
Where, when this has been not something (but neither flesh nor stone!)
censored the F word for ya Woody
This one had some minor iterations with my own changes though
noonday crusting things made it go in a good direction
and then I cut a lot of extraneous stuff
That's the same LLM captioning pictures for me
also it wrote me the longest and most bizarre love letter of my life
it was very phallic, but in the weirdest way possible. Can't share it bc its highly innapropriate, but creatively, it was cool af
And now we can generate the person that wrote that letter by feeding it back into it and adding "make the person that wrote this " 
gpu go brrrrrrr
fr, its all so powerful now
with llama
Truly is
its waaaay better than gpt with the temp controls and also reliability, since there's no remote update that will break your prompts
im fluxin
this is very on brand for you hahaha
Oh yeah totally
addicted to vintage cars
The important bit is ones that never existed
fallout 5
If I wanted to see a Veyron I can just go see one
I want to see alternate universe stuff
now this I like
whoooooooooo
I always shop at K-Town
These still need touching up to really be "real" but I suspect most people can't tell
oh for sure
just a pass with another model
will work
It's mostly things that these always get wrong. Like hands, but for cars
Wipers, steering wheels, the interiors of the rear quarters when seen through the windows
It is so close though
LLMs are fun as heck, is my point 
My contemporaries though, people in their 40s
Reality is changing underneath their feet and they have no idea
Just crossing their arms going "blarraghaghga back in MY day" 
back in my day we just used Ejay
BRO this bot is on something
@grim idol I tried to post in here but the bot blocked it
I can check the post in the logs 👍
Thx
Yeah it's a false positive. I can't change the bot, but basically it's bad markdown detection, somewhere in the post uses hashtag that is tripping it up.
https://srcb.in/IDyPVsnKGp (code)
https://huggingface.co/suno/bark (the "guide" I'm using)
I'm wanting to get it to use my GPU, but I'm kinda lost
why do embeds have to be so big 
The Bark Colab link is default original Suno Bark. And the Hugging Face Colab is huggingface version. Both can work with GPU. Your code sample is using default Suno Bark. It should autodetect GPU if it thinks your system has it. I would try testing a barebones pytorch script where it all it does is check for GPU support maybe. If that works, then it's something in Suno code or config, and not your environment.
Unfortunately too complex for my 2:37 am brain to process atm, thanks though!
Basically, if you run this, does it work? If this also fails, problem is not specific to Bark:
import torch
if torch.cuda.is_available():
device = torch.device("cuda")
print(f"CUDA is available. GPU device name: {torch.cuda.get_device_name(0)}")
print(f"CUDA version: {torch.version.cuda}")
else:
device = torch.device("cpu")
print("CUDA is not available. Running on CPU.")
I also have to step away, but tomorrow starting a few hours earlier than now, I will be online and at my desktop, where I can help debug
bruh i have a 4070 with latest drivers!!
Despite NVIDIA being used everywhere, it's still a headache to get CUDA working, like all the time, not unusual
aye same i did say when i started working on midi files and then fruty loops and alike. ppl was like ew "fake music the computer do all the work xD "
What would I have to do to get cuda to work? it seems to work fine with ollama, alltalk, and automatic1111 webui
When you use those, is it through their installers? They usually setup a conda/venv and have a pretty lengthy install script.
Actually, you could try literally using the automatic1111 environment. Or some other AI project. I'm not sure if automatic does, but often there is a file called START_ENVIRONMENT_ONLY or something (typically used for debugging setup problems) but you could literally run Bark, after starting that environment, since you know it does see your GPU
Ah
But for example, when you type any random pip or conda command, if you look at it carefully, double check when it's showing what versions have changed, that none of the versions are changing to blahblah_cpu version anywhere.
im in my 40's and thinking blarraghaghga back in MY day why did we not have this .. i would been a psyhco engineer on pluto by now 😛
It's been awhile, but I did create a windows Bark installer based on automatic, at one point, so I'm sure we can sort this out eventually. (But not tonight!)
sends a look full of compassion 🙂
but im gonna be honest i never thought even gen music would get this far before i was long past dead. so lets see were it is when im in my 60's over a decade to go.
I'm grateful to be alive in such great era.
I created a complete album inspired by Sigur Rós using Udio
Udio sounds like..... Udio. Nothing I like about that.
@sleek urchin
sigur ros 1 of the greatest alternative bands ever been created. will listen while im having my evening time off. jolly good.
Ai幻illusion:
https://www.udio.com/songs/ocVk5sELiXZspTF1WdXHpj
Listen to Ai幻illusion by EcoShadow(glitch) on Udio. Discover, create, and share music with the world. Use the latest technology to create AI music in seconds.
I've just discovered that flux AI understand Italian language, I used part of the lyrics of my last song (It's written in italian) and flux generated images that fit the lyrics in a consistent way.
The lyrics: "C'era una ragazza, si chiamava Stella,
La sua vita era una danza, una luce sempre bella.
Sognava di volare, di toccare le stelle,
Con il suo cuore che batteva, batteva per quelle.
Ogni notte usciva, sotto il cielo blu,
Con un desiderio che cresceva, che non andava più giù.
Stella cercava amore, in un mondo così grande,
Con il suo sorriso che brillava, e una voglia travolgente."
The generated image:
New on-site playground models added;
- FLUX.1 [dev] with LoRAs
- FLUX.1 [dev] with Controlnets and Loras ¹ (updated)
- FLUX.1 [dev] Differential Diffusion ¹
¹ denotes inpainting
I finally fixed the problem I had with flux loras not working... I only had to delete confyui portable and use a fresh copy... Easy peasy.
I hate when confyui gets borked and the only way to fix It is to reinstall
The Burning of Chirp & Woodhouse - an upstate farm... somewhere....
alright, I applied at Suno haha
