https://www.youtube.com/watch?v=JSswFa8Ta1k
Def. prefer Suno to Mureka here. But that's not really important, it will get much better in a very exponencial fashion. Just 2 years ago (or so) I thought cracking vocals and acoustic sounding music will be quite hard for AI. But it's doing it.
Yes there are artifacts and it sounds like a lowres MP3, but that's just a matter of resolution IMO. The crazy part is how hard it is to find vocalists this good IRL. The other crazy part is the pop stuff that you get with a click of a button is often (dare I say) better than a majority of mediocre human made pop. Which is more (!) repetitive and doesn't flow as good, unless you have really talented musicians. Which you often don't get as it's a hobby thing for most aspiring musicians and they just never get really good in their craft.
So in a few years, I wouldn't be shocked that a huge portion of pop, game and film music will be just dominated with AI creations.
From the creator point of view, I think AI generated orchestras, instruments and vocalists will creep into the productions. Esp. AI generating an orchestra seems like a better way to fake it than with the massive multi-sampled libraries that are used today. However good these tools are for faking some performances, the window of faking it well is quite narrow, leading to kinda same-ish results. One would say "genre defining limitations". AI generated orchestras won't have these limitations, they just have to get much more hi-ress than what Suno is doing today.
HurvajsRumcajs / Hurvajs Rumcajs
npub1mw…pd09j
2025-04-04 02:04:43
Author Public Key
npub1mwkpvd2snl2tlxtmtgvrj2vr5mwf2nmuynua4a0jmzgtkvwt93rq8pd09jPublished at
2025-04-04 02:04:43Event JSON
{
"id": "0a0e7a6ab8483a40bad9f132bc1d856d80cb27b5643e81fda60523803dc51ad7",
"pubkey": "dbac1635509fd4bf997b5a18392983a6dc954f7c24f9daf5f2d890bb31cb2c46",
"created_at": 1743732283,
"kind": 1,
"tags": [
[
"r",
"https://www.youtube.com/watch?v=JSswFa8Ta1k"
]
],
"content": "https://www.youtube.com/watch?v=JSswFa8Ta1k\n\nDef. prefer Suno to Mureka here. But that's not really important, it will get much better in a very exponencial fashion. Just 2 years ago (or so) I thought cracking vocals and acoustic sounding music will be quite hard for AI. But it's doing it.\n\nYes there are artifacts and it sounds like a lowres MP3, but that's just a matter of resolution IMO. The crazy part is how hard it is to find vocalists this good IRL. The other crazy part is the pop stuff that you get with a click of a button is often (dare I say) better than a majority of mediocre human made pop. Which is more (!) repetitive and doesn't flow as good, unless you have really talented musicians. Which you often don't get as it's a hobby thing for most aspiring musicians and they just never get really good in their craft.\n\nSo in a few years, I wouldn't be shocked that a huge portion of pop, game and film music will be just dominated with AI creations.\n\nFrom the creator point of view, I think AI generated orchestras, instruments and vocalists will creep into the productions. Esp. AI generating an orchestra seems like a better way to fake it than with the massive multi-sampled libraries that are used today. However good these tools are for faking some performances, the window of faking it well is quite narrow, leading to kinda same-ish results. One would say \"genre defining limitations\". AI generated orchestras won't have these limitations, they just have to get much more hi-ress than what Suno is doing today.",
"sig": "0c6116e92629972e0864bb6ba1d66397d3321eb5336f66d111fe4886af10d6a810b9954664fed8f617ebd7f0b3caef050bd65def74a085cf14964346fe8110e5"
}