• TalkBass has been independent since 1998. Add your voice.
    Create a free account to reply to discussions, view embedded media, and browse with fewer display ads.
    Join freeLog in
    Want zero display ads or expanded classifieds tools? Compare plans.

Ambient/Post-Rock/Textural bass playing thread

It isn’t my favorite genre. But I did listen to a lot of the AI generated stuff tonight, and I probably liked it more than some of the man made stuff. I assume the “same key” issue is related to the open string usage in djent.

In all the fun about djent, I forgot to ask what else caught your attention. Sometime next year I have to write something on AI-generated music for work.
 
25800587[/URL], member: 37282"]In all the fun about djent, I forgot to ask what else caught your attention. Sometime next year I have to write something on AI-generated music for work.

I think I’d have passed it by and dismissed it if it wasn’t for the fact I knew it was AI. I’m not a fan of the genre, but I can appreciate it. From what I heard, the AI riffs were “real enough” to fool me into thinking it could’ve been composed by human players. I couldn’t really discern verse/chorus/bridge sections, but the riffs did seem to work together, and there were breakdowns that seemed to fit. All the big picture elements were there for me; syncopated/polyrhythmic feeling sections, with djent tones, and big keyboard/pad breaks. The instruments seemed coordinated in a way that felt right. The sampled guitar sound helped make it all convincing enough for me to drive by and think “that checks out.”
 
  • Like
Reactions: dukeorock and JES
I think I’d have passed it by and dismissed it if it wasn’t for the fact I knew it was AI. I’m not a fan of the genre, but I can appreciate it. From what I heard, the AI riffs were “real enough” to fool me into thinking it could’ve been composed by human players. I couldn’t really discern verse/chorus/bridge sections, but the riffs did seem to work together, and there were breakdowns that seemed to fit. All the big picture elements were there for me; syncopated/polyrhythmic feeling sections, with djent tones, and big keyboard/pad breaks. The instruments seemed coordinated in a way that felt right. The sampled guitar sound helped make it all convincing enough for me to drive by and think “that checks out.”
Djent is soooooo snapped to the grid and heavily edited no wonder an AI can pull it off.
Ever hear real drums in Djent?? Triggers and resample ftw.

My 2 cents.
 
Yeah, but any other year and I'd call that "damning with faint praise" :)

Hah. Yeah. I listened to some more of them last night. It took me right back to 2004, and I liked a lot of what I heard. I’ll be spinning more today at work. I might even buy an album (whatever that means in 2021).

They seem to have gotten some attention for being the first “major” metal band with a black female guitarist, which is cool, and wild considering how long metal has been around. She’s certainly a strong player regardless of the diversity factor. The songs I heard didn’t have a ton of solos/leads, but I liked the ones I found.

My only real gripe is that the lyrics and subjects aren’t up my alley. I’m an early-onset geezer and this stuff has a young target audience. I get it, but my days of feeling like a social outcast at school are behind me.
 
I think I’d have passed it by and dismissed it if it wasn’t for the fact I knew it was AI. I’m not a fan of the genre, but I can appreciate it. From what I heard, the AI riffs were “real enough” to fool me into thinking it could’ve been composed by human players. I couldn’t really discern verse/chorus/bridge sections, but the riffs did seem to work together, and there were breakdowns that seemed to fit. All the big picture elements were there for me; syncopated/polyrhythmic feeling sections, with djent tones, and big keyboard/pad breaks. The instruments seemed coordinated in a way that felt right. The sampled guitar sound helped make it all convincing enough for me to drive by and think “that checks out.”

thanks. That makes sense. In your travels, did you hear other AI generated music that either was interesting or really not?
 
animals as leaders is sorta fusion djent, i guess. i like them.

Prior to ^^^this post the only thing I'd ever heard from Animals As Leaders is one tune "Monomyth" (which I'll say is one of the more gorgeous music videos I've seen in perhaps decades), but for these past couple days I've been checking out more of their stuff on youtube.

I definitely like them, but I also feel the same way about them that I do about a lot of contemporary "prog" ...which is to say A) I might like them more if 80s hair metal had never happened; and B) Pat, I'd Like To Buy A Chord Change. Not sure if I could sit through an entire album or concert if everything is so harmonically monolithic.

otoh, a quadruple bill of Animals As Leaders with Battles, 05ric, and Don Caballero sounds like it could be a seriously fun night.
 
does it take more time to program the ai to "write" a piece or to actually write a piece?

Well, at the 1:00 mark he says there's no AI or Machine Learning being used. He's got some other "generative" techniques which don't fall into those categories.

If the goal is to "write a song" using any means available, then for most musicians the obvious answer is that it is faster to write. However, I'd wager that most musicians couldn't produce 10 hours of content with the same speed as the generative method, even factoring in the programming time.

Of course, the purpose of this was not to generate a song. It was to display that songs could be generated, with musical and stylistic results. I like that it demands examination of the "songwriting" process in order to achieve the end product in a specific genre.

If you are generating gaming music for a paycheck, this tool could be quite handy. It creates an atmosphere with minimal effort required to produce endless variations that stay in the lane.
 
Well, at the 1:00 mark he says there's no AI or Machine Learning being used. He's got some other "generative" techniques which don't fall into those categories.

If the goal is to "write a song" using any means available, then for most musicians the obvious answer is that it is faster to write. However, I'd wager that most musicians couldn't produce 10 hours of content with the same speed as the generative method, even factoring in the programming time.

Of course, the purpose of this was not to generate a song. It was to display that songs could be generated, with musical and stylistic results. I like that it demands examination of the "songwriting" process in order to achieve the end product in a specific genre.

If you are generating gaming music for a paycheck, this tool could be quite handy. It creates an atmosphere with minimal effort required to produce endless variations that stay in the lane.


there's a lot in here to unpack, mostly cause it raises more questions for me.


the operative word here seems to be "generate," not "create" . . . .

perhaps that's a small distinction, but for me it seems huge. one thing is about "content" and the other, well, is about something else.


the concept of quickly generating fairly faceless "musical content" for anything these days feels to me like it's further devaluing music *as an art*—it just becomes more wallpaper. seems to me that it creates a situation where "music" means almost nothing to both the person "generating" it and the person "consuming" it (though whether it's consuming it or just having it there is a difference i need to thing about). maybe it's "curating an experience" without really *experiencing* it . . .


so, i've made statements. i'll try to generate the questions:

"is content generation the same thing as creation?"

"does this activity create a devaluation of music?"
 
Great questions @steubig. I really want to watch the video. Just been slammed.

But I can't resist. This is all IMO of course. I'm not ready to die on any hills for these hot takes yet.

"is content generation the same thing as creation?"

No. However, using automated processes as part of composition or improvisation is the same thing as creation. So it's about the relationship to the process (artistic, economic, conceptual).

"does this activity create a devaluation of music?"

In the hands of artists, no. In the hands of startups and blowhards, yes. Tech giants are both yes and no because of their big basic research divisions that sometimes funnel money to artists.

I had a meeting yesterday with a couple people at one of the big 4 tech companies who are doing work on music generation. A lapsed composer turned into a computer scientist. It was striking to me how interested in and concerned about musicians they are--though to be fair, the people doing the research are not in charge of the business direction. One of them also told some stories about a streaming company (being vague here, forgive me -- I wouldn't if we weren't talking in public) where some researchers up and quit when management got more interested in content than music. Which of course didn't change anything about what the company was doing.
 
Last edited:
to be fair, record companies have always been interested in selling content. same with movies. i read a series of interviews with novelist steve erickson (not the fantasy guy) who basically says that publishing houses are going this way too and suggests that young writers go to self-publishing. (he's a guy who's often lauded as one of the u.s.'s great writers, but can't get published half the time.)

serialism is sort of about "generating" material, but it also gets into making decisions about how to use that material.

i'm currently working on a 5-piece set that uses a number of 8-note sequences that i manipulate in real time and do some serialistic operations, but still gets me spending more time on the manipulation, etc., than the sequence itself. all of which is to say that i'm not a stranger to that idea.

tbh, i think streaming is totally devaluing music. you can hear whatever, whenever, with very little effort.

i think the blowhards almost always win, btw.
 
Last edited: