It's interesting that in the early days of mechanical instruments like the orchestrelle and pianola in the early part of the 20th century, people were buying them because they wanted to vicariously play with “feeling, but if you look at it from the aspect of what is actually happening cognitively when you play music and play it with “feeling”, is different from a machine activating mirror neurons. This is happening today with people generating music with AI: sometimes it generates emotional-sounding music, but you weren't “instrumental” in that emotion--you're just feeling that emotion. (“With no artist in sight, an unexpected art emerged with unforeseen effects.”) When you're actually playing music, it is, in fact, mechanical--even for people who’ve been playing for a long time. Feeling is mostly a fleeting illusion, at least as it relates to playing music. The lasting impact of music on emotions is that if there's some kind of encoding involved--which can occur from the coupling of an experience in a place while listening to music--and the experience of looking at images along with the music, such that when you re-listen to the music or look at the images again it reactivates that experience as a form of nostalgia–”bringing you back”. And typically those kinds of emotions are more powerful than the ones you'd experience when you are playing music. So these feelings that we have for music are the result of mechanical processes, primarily recordings. Many times we’re moved just as intensely by things that work as a beautiful machine, like an expensive sports car. If that’s thrilling in itself, isn’t that a satisfying emotion as well? And if you’re really into cars, you’re not looking for anything else emotionally, and will probably find art and music less interesting in evoking it.
In a Lester Bangs interview with Brian Eno, he said that what people call unemotional simply doesn't have one single overriding emotion to it, and the things he likes best are the ones that are ambiguous on the emotional level. I feel this as well when I'm working with random elements in some kind of AI context, and produces a result that causes some kind of an emotion, but it's not a standard generic emotion that crowds of people would react to. It's an emotion that I'm personally feeling is unique.
***
Postscripts:
"The first piano rolls were edited for playing with only 58 or 65 notes, though rolls using the full 88 notes of the piano keyboard soon became the standard. also, significantly, the punched holes could regulate the duration of a note but could not duplicate the Myriad variations in touch that would bring feeling, to the music."
[Again, it's all about recordings, which are mechanical illusions. they're made possible by time shifting the creative process over months or even years, such that when we finally listen to the final product it appears as one contiguous whole as was performed live. There are many things you can do in a recording that can't be reproduced live and the prime example of that are the Beatles when they stop performing live and just became recording musicians. When we generate music with AI it’s that same kind of a illusory process where we're making things that really can't be easily reproduced. The music that I've generated sounds good as a production, but when you reverse engineer the recordings, there's not much music there and you'd have to mechanically practice playing it so that it would sound as listenable.]
[The way I look at AI authorship is that if you made or designed some kind of a seed that drives the whole work, then you were at least a co-writer. But who are the other writers, either writing the actual music, who are the people that performed on it, who are the people that recorded it, what recording studio was it done at, and so on. What you have to realize about the realm of AI-generated art and music, is that it's all fungible. It's like going down to the grain elevator and getting some grain, working with it, and then packaging it and marketing it as your own. For now, everybody has free access to that grain elevator, and no one is stopping them from taking the grain that they want. But that's probably going to change in the future. Not only are there going to be paywalls, but there will be disclaimers and contracts that might have legal teeth. Who knows, everything is becoming so deregulated now, that it's just the Wild West everywhere. I like the riposte by the Obvious coders: it didn't matter to them whether people liked it or not but the point was is that no one was indifferent, where they would just glance at it and say nothing, or completely ignore it and would have zero views].
[Apparently, Stravinsky was an early adopter of the player piano and composed an etude for the pianola in 1917. This is sort of like my use of AI music now. I think creative people generally don't push back on new technologies because they're naturally curious people and when there are new tools they always want to try them out to see what they can do with them. What you have to be vigilant about is that you're not spending too much time with the new tools and letting your other skills go to seed. Like all new technologies, lots of people panned it for being annoying. This was one of the complaints in the 1980s when people started using sequencers and MIDI–that it had a monotonous machine like quality they didn't like. Even to this day there are lots of people who think that the 1980s was a period in music in which it started to become too non-human. This goes back to my theory about things either being round or square: In visual art people are naturally attracted to things that are curvilinear and don't like things that are too square or rectilinear because it's not human enough. Also roundness has a certain sensuality to it, almost erotic in some sense. If we were to have sex with a rectangular object it wouldn't feel very good. But for the sake of structure, things always need rectilinearity. Buildings can't be an assemblage of curvilinear forms. A good example is the proscenium at the Pritzker Pavilion in Millennium Park in Chicago. If you look behind it there's a superstructure of rectilinear forms. And it isn't that the rectilinear forms are an aesthetic aspect of it. it's simply to hold the thing up.]
[AI generated music is essentially like a photograph of actual music being played by musicians, as a photograph would be of a painting, it's a reproduction. but in the case of AI generated music it's a unique photograph that isn't representational. it's not representing anything specific in the sense that if I take a photograph of a flower I'm representing it. But with artificial intelligence it's finding thousands of different flowers based upon your text prompt. In music, it's taking a musical sample and mapping a melody based on your text over it. It would actually be interesting to have a camera or app that would do this kind of thing that would take your description and combine it with the photograph that you're taking and make something new out of it. This probably exists already.]
"Ragtime called for the qualities critics valued in classical music: shading, sensitivity, and so on, and also exceptional technical facility and speed, as well as Mastery of musical time. Tasked the music became popular, people, including critics, who appeared to lose sight of the relevance of humanistic values in the playing of ragtime, associating the music mainly with technical and mechanical values. Player pianos deserved part of the blame."
[Once something becomes convenient and mechanized as well as popular, high art Gets pushed off its pedestal. The most popular videos on YouTube ever since the beginning have always been people playing virtuosic things on instruments, which is the thing that people admire most, rather than the composition itself. people aren't really interested in art per se, they're interested in how they themselves can access artistry in the easiest way possible. It makes it look easy, and machines or automation allows them to realize their fantasy about playing like them. People are naturally impressed (especially men) by musical dexterity, and machines allow for an easy way to achieve (vicarious) virtuosity, the same every time, and never a bad performance].
[The more text to speech voices are used, I think people are going to start talking like them. once you're not able to tell a machine generated voice from a real voice, People will naturally mimic them, in the same way that the machines shaped the way we understood music after player pianos became popular. We learn languages by the sound and the music of it, so it stands to reason that the more we start using text to speech that will start to sound like that. This is actually a good thing because people will begin to talk more articulately.]
[With AI music it's sort of a player piano but it's playing other people's music that they've already conceived of and recorded, as opposed to a player piano which is just a way of mechanically playing the keys via a pattern.]
[It's interesting and accurate that the instruments that you play, play you. People have the misconception about musicians that if they suddenly see you in another context in which they normally do in everyday life–and this carries over to the musicians themselves–when they switch instruments they start to feel like a different person, and the sounds that are coming from the instrument–which are no different from sounds coming from other instruments in a musical sense (An A-natural is an A-natural if it's played on acoustic piano or if it's played on a bass. But if people are used to seeing you playing a keyboard and then they see you playing an electric bass, they're taken aback by it and might think that it doesn't suit you or you can't “own” the music because they just don't see you that way. To musicians, an instrument is an instrument. If I play a B flat on a keyboard as a bass sound and play a B-flat on on an electric bass it still will be a B-flat to me and fits with a B -flat minor chord. But the audience sees it entirely differently. They want people to “own” the music as the authentic character that they think you are based upon the instrument that they know you for. So not only are instruments machines and mechanical processes, it's also mechanical in the sense that it's just prosaic and ordinary and to the musician it's neither here nor there.]
[There's only a slim margin of randomness that can come from a machine, so the surprises that you might see are fairly constrained, and in that sense it's really not that creative. You can definitely be more creative than a machine.]
Comments