|
Every single thing I've listened to in this thread sounds like dogshit to me.
|
# ? Mar 25, 2024 21:20 |
|
|
# ? Apr 27, 2024 09:34 |
|
Music is 99% dogshit regardless of who makes it.
|
# ? Mar 26, 2024 01:01 |
|
SCheeseman posted:Music is 99% dogshit regardless of who makes it. No one is making this so that means it's good?
|
# ? Mar 26, 2024 03:07 |
|
massive spider posted:Every single thing I've listened to in this thread sounds like poo poo pant to me.
|
# ? Mar 26, 2024 03:12 |
|
Nobody Interesting posted:No one is making this so that means it's good? Since 99% of all music is dogshit, logically that percentage would carry over to anything generated from a dataset based on all music.
|
# ? Mar 26, 2024 03:19 |
|
SCheeseman posted:Since 99% of all music is dogshit, logically that percentage would carry over to anything generated from a dataset based on all music. This is lazy cynicism as a shield. The “akchully everything sucks” thing. Nope lots of things are good. And in music there’s plenty of acts who even if they aren’t getting into any hall of fames are at least serviceable or have something notable about them. This is just an unpleasant listening experience. Also the logic above shows the issue with trying to generate music from AI. A human who had listened to all music ever might hopefully generate some good taste. An awareness of what cliche is and also an awareness of how to avoid it. An AI just learns more efficiently to cliche. I get the appeal of it as a shitposting toy but the people going “wow this is like radio ready for the general public” or saying they’re gonna use it to score their next project wtf are you doing lol.
|
# ? Mar 26, 2024 07:10 |
|
Your original post was lazy, I was just reciprocating that. Not that I don't think most music isn't bad, though I also didn't say there wasn't good music, 1% of tens of millions is still a lot. I'm less interested in the raw output of the model, sometimes you can make it do something weird and hosed up and cool but it usually leans bland instead. At this point I use it to generate samples, picking the most interesting of the generations, which get broken apart into stems and remixed in a DAW, smashing the audio together much like when I was screwing around with fruityloops as a teen. I think I know what maybe a quarter of the options in these VST plugins do? I can hear all the problems, the audio is full of artifacts, but it's more flexible than a bunch of wav loops or MIDI and the vocal performances it can generate are better than anything I could possibly perform or afford. So it has some novelty. Who knows what the hell the public will accept. They accepted Crazy Frog Axel F. SCheeseman fucked around with this message at 08:39 on Mar 26, 2024 |
# ? Mar 26, 2024 07:27 |
|
It’s not more flexible than midi because midi is pure note information so it can trigger anything up to and including physical light rigs. And if you used wav loops maybe the limitations might force you to come up with something actually interesting. By lazy cynicism I mean “Music is mostly bad anyway” is one of those populist cliches that people will nod along with because it’s not cool to admit you like things. But followed to its conclusion it’s just justifying a race to the bottom-Everything is poo poo anyway so why not make poo poo. The thing is, if someone posts music for critique I’ll generally try and at least be constructive, because yknow theres a human there and there’s effort and heart that went into that even when it kinda blows. With this though- why even try and be polite? It’s soulless slop and there wasn’t anything put into it. Youre simultaneously sort of backhandedly trying to justify it while also being unable to ignore that you know it’s slop.
|
# ? Mar 26, 2024 09:50 |
|
massive spider posted:It’s not more flexible than midi because midi is pure note information so it can trigger anything up to and including physical light rigs. And if you used wav loops maybe the limitations might force you to come up with something actually interesting. I'm not sure what MIDI's use as part of live performances has to do with pre-recorded music production. I meant it's flexibility when used for instrumentation. There's some intent in the creation of music through Suno driven by the input of lyrics and metatags that control the overall structure of the song. Most of the rest is still providing style prompts, rolling the dice and curating the song chunk by chunk. Or generating bits and pieces, stripping them into stems and rebuilding the song more traditionally which allows for more control over the final work. So I wouldn't say it's all slop, at least if some effort is put into it. 99% slop. I wouldn't bother to ask you to critique anything here because you clearly have a chip on your shoulder. I don't really blame you, generative AI is being used to gently caress over workers. But that's not what anyone in this thread is doing, so cool off. SCheeseman fucked around with this message at 11:39 on Mar 26, 2024 |
# ? Mar 26, 2024 11:32 |
|
You know I suppose I do have a chip on my shoulder because music is supposed to generate some emotion or discussion and in the absence of this creating anything worth saying about it the best I can muster is the emotion to reject it. In the face of something this creatively dead might as well go full Alex Jones: “This is a human! This is what it looks like!”.
massive spider fucked around with this message at 11:53 on Mar 26, 2024 |
# ? Mar 26, 2024 11:49 |
|
It's making you angry, that's an emotion. I wrote the lyrics for a song about my cat and how he's always sitting down in places where people will trip over him, my nieces enjoy it, it's not high art or top 100 material but I enjoy it for what it is, it's basically a poem that a computer made music for (and then I took the music and changed it a bit so it flowed better). Absence of emotion? Slop? Maybe to you, but it isn't for you. If you're comparing yourself to Alex Jones, maybe that's an indication that you're treading too far into reactionary territory. You can reject AI generated music the same way anyone with taste rejects modern american country music, by ignoring it.
|
# ? Mar 26, 2024 12:10 |
|
https://cdn1.suno.ai/64898042-8eee-4679-b072-4347668155d9.mp4
|
# ? Mar 26, 2024 12:55 |
|
A lot of the Suno community are annoyed that the V3 update is actually has inferior sound quality to the V3alpha that ran beforehand. It seems like that's the case to me as well, so you might want to hold off a bit before splurging too many credits.
|
# ? Mar 26, 2024 14:03 |
|
SCheeseman posted:It's making you angry, that's an emotion. Are you the generative AI behind Suno? Astrochicken fucked around with this message at 14:12 on Mar 26, 2024 |
# ? Mar 26, 2024 14:05 |
|
Astrochicken posted:Are you the generative AI behind Suno? Responses like this are more derivative than anything I wrote, though I guess any LLM dataset wouldn't include the thousands of examples of people making this joke yet.
|
# ? Mar 26, 2024 14:35 |
|
how is this different from any other millennial music?
|
# ? Mar 26, 2024 15:13 |
|
ChubbyChecker posted:how is this different from any other millennial music? It is free from any pending or possible sexual impropriety.
|
# ? Mar 26, 2024 16:53 |
|
I tried to listen to this for the sake of a good roast, made it about 10 seconds in of a robot burping before i switched off. Jesus I know im old man yelling at clouds here but it reminded me of how back in the day people used to do dramatic readings of posts they hated, or full fledged diss tracks, or even just like jokes? Now we have dead eyed iPad kids so sapped of vital life force chatgpt writes their burns for them https://voca.ro/1mqKs3zkiboR
|
# ? Mar 27, 2024 15:40 |
|
massive spider posted:piss piss whine whine So, with that in mind, could I possibly ask you to gently caress off, you insufferable sack of misery? Anyway, in other news, did you guys know that there isn't an actual separate species of animal called a "black panther"? They're just leopards with a genetic chromatic aberration, kind of like a reverse-albino. Heh, makes you wonder what they'd think if they were capable of understanding that concept. I think it would go a little something like this!
|
# ? Mar 27, 2024 16:10 |
|
lyrics copyright massive spider https://voca.ro/1i7wmMC1VyBW
|
# ? Mar 28, 2024 08:43 |
|
Interview with the Dream Walker, creepy dialogue https://voca.ro/1okRhp0K5YYv It would be cool do designate Voice 1 and Voice 2 or something like that.
|
# ? Mar 29, 2024 06:30 |
|
https://twitter.com/RomanIsntOnline/status/1774439272956719232
|
# ? Apr 1, 2024 00:46 |
|
I've only had a chance to try a few tracks but Stable Audio just got a big upgrade. "high-quality, full tracks with coherent musical structure up to three minutes long at 44.1 kHz stereo from a single prompt." https://twitter.com/StabilityAI/status/1775501906321793266
|
# ? Apr 3, 2024 14:00 |
|
Some indie/psychedelic/alternative creations mit improved songwriting https://youtube.com/playlist?list=PL-2ZbI2n2IM9W70mK5n3m8mv6StQhqW3o
|
# ? Apr 4, 2024 01:26 |
|
Youtube told me to watch/listen this and I am honestly not entirely sure if it's suno or not https://www.youtube.com/watch?v=qA_wBSqe-lg
|
# ? Apr 7, 2024 01:32 |
|
Nobody Interesting posted:Youtube told me to watch/listen this and I am honestly not entirely sure if it's suno or not https://www.youtube.com/watch?v=qA_wBSqe-lg Everything on this channel is loving hilarious. Theres no way AI could write lyrics that funny, but the music is probably AI
|
# ? Apr 7, 2024 02:01 |
|
V3 vs v2 comparison V3: https://app.suno.ai/song/ec538c25-e23e-411f-a89a-561e6943dbcd/ V2: https://app.suno.ai/song/533c0efb-6a97-4f19-a923-92820d342772/ (started in v2 then used v3) V3 as it is tends to get distorted and it has more "filler" sounds that sound awful rather than clear individual instruments. The premise of the song is unsually better and takes less generations to get something usable
|
# ? Apr 7, 2024 04:00 |
|
Nobody Interesting posted:Youtube told me to watch/listen this and I am honestly not entirely sure if it's suno or not https://www.youtube.com/watch?v=qA_wBSqe-lg Suno definitely generated the audio.
|
# ? Apr 7, 2024 04:23 |
|
Still a work in progress, I'm still in the process of mixing it but I haven't posted anything in a while other than a troll song and I like how this is sounding https://voca.ro/1c4pNnkfgCck Quest of the Best of the Best of the Test (To Rhyme The Word Best With Other Words Expressed)
|
# ? Apr 7, 2024 09:22 |
|
SCheeseman posted:Still a work in progress, I'm still in the process of mixing it but I haven't posted anything in a while other than a troll song and I like how this is sounding I appreciated the Taylor Swift cameo
|
# ? Apr 8, 2024 05:43 |
Hi from discord - which of you was responsible for "cat cat cat cat"? I introduced my dad (72) to this on the weekend, set him up with the free account to noodle with. He's into making AI video so he was happy to be able to make little scores for them (he's just making little cartoons for his grandkids about mice princesses & frog knights, nothing serious). I just use it to make workout music, free five second intros for videos I have to make at work (edu stuff) and to have fun trying to make cursed noise, for no reason other than it's a fun toy and I enjoy playing with it. I'm still not a fan of the way the singing sounds though. I'm hoping more and more control & tools get added so it can make something that has more human 'storytelling' to it - it can make some nice sounds but lacks that intelligent human element, I suck at explaining - things like knowing where to the put the breaks/drops and how to evolve things and so on. Sometimes even with tags it's doing it's own thing, I feel like a lot of the stuff I get that sort of works is mostly luck. The ability to continue from any second helps but nothing beats having the ability to cut/move loops around exactly where they're needed & having more control on how loud parts of the track are etc etc. All stuff to come I expect. In the meantime I'm enjoying being able to knock up crazy genre combinations while working out to make working out suck less (or, I guess, suck in new more interesting ways). edit: i just figured out i can also use it to make pipe-organ/merry-go-round/generic ride theme music for planet coaster Tambaloneus fucked around with this message at 06:59 on Apr 8, 2024 |
|
# ? Apr 8, 2024 06:44 |
|
Tambaloneus posted:I'm hoping more and more control & tools get added so it can make something that has more human 'storytelling' to it - it can make some nice sounds but lacks that intelligent human element, I suck at explaining - things like knowing where to the put the breaks/drops and how to evolve things and so on. Sometimes even with tags it's doing it's own thing, I feel like a lot of the stuff I get that sort of works is mostly luck. The ability to continue from any second helps but nothing beats having the ability to cut/move loops around exactly where they're needed & having more control on how loud parts of the track are etc etc. All stuff to come I expect. In the meantime I'm enjoying being able to knock up crazy genre combinations while working out to make working out suck less (or, I guess, suck in new more interesting ways). My process is I start with an instrumental that sounds good (it's best to go heavy or weird with the prompts, I almost always add the word "experimental) then generate an extension of that to pick the vocals I want. I try not to get more than 3-4 extensions deep since quality starts to drop. I then import those to https://moises.ai/ which lets me split the tracks into separate stems that I export and load into Reaper (though you could use Garage Band or whatever else). From there you can arrange the song largely however you want or even chop samples up and create whole new bits of music from the constituent elements.
|
# ? Apr 8, 2024 07:31 |
|
Suno is a gently caress. Udio is my love now. https://www.udio.com/ https://i.imgur.com/WhCAQLL.mp4
|
# ? Apr 11, 2024 05:27 |
|
I've been trying to get something Devo-like out of Suno without much success, tried with Udio and managed to get something somewhat decent? I still need another verse or so and an ending, also did some work on the track in post to shorten segments of it and try to punch it up a bit (though there's some clipping). https://voca.ro/1gX8HQI5A2GY "Don't Stand Up" SCheeseman fucked around with this message at 13:09 on Apr 11, 2024 |
# ? Apr 11, 2024 08:59 |
|
Mola Yam posted:Suno is a gently caress. Heck yeh, It does hardcore/punk stuff a little better. Did a corny pop punk song right away. (Loving You is a) Self-Inflicted Life Sentence https://voca.ro/1mpZgTvnuYoc
|
# ? Apr 12, 2024 01:32 |
|
Roman posted:I've only had a chance to try a few tracks but Stable Audio just got a big upgrade. Are they releasing the model/weights behind this, like they did with their image generator? Because I want to see what the internet-at-large does with this, once they're actually allowed to play (...and don't have to pretend to care about copyright)
|
# ? Apr 12, 2024 09:32 |
|
https://i.imgur.com/Mh1IXrA.mp4 https://i.imgur.com/Brn8m6l.mp4
|
# ? Apr 12, 2024 15:04 |
|
Udio is amazing. Been using it all day instead of work. https://www.udio.com/songs/6zJwn71AKHLvfvyoM82Qhc https://www.udio.com/songs/2qP6b5hDE64En9eA6C4y5j Prince Reggie K fucked around with this message at 00:51 on Apr 13, 2024 |
# ? Apr 12, 2024 22:52 |
|
https://voca.ro/195f3p7Y3bXM WIP of trying to get Udio to do late-era Beatles, with my own lyrics. It's hard to get it to tone it down a bit and that guitar break is a bit of a mess. I've been spoiled by Suno's ability to continue from any arbitrary point in a song. It helps cut out junk from extensions that you don't want in there and Udio appears to lack the feature (though being able to generate backwards is really cool) SCheeseman fucked around with this message at 12:26 on Apr 13, 2024 |
# ? Apr 13, 2024 12:24 |
|
|
# ? Apr 27, 2024 09:34 |
|
Found a program called stem roller. It will split up your songs into bass, Drums, others (lead instruments), the whole instrumentals, and vocals https://www.stemroller.com/ Longpig Bard posted:Heck yeh, It does hardcore/punk stuff a little better. here are the vocals https://vocaroo.com/1eCLbgSxncwR
|
# ? Apr 13, 2024 20:34 |