While thematic melodies are motivic. There are non-motivic melodies. Good examples are transitions where the melody section is often just playing scale runs arpeggios. A common popular example is the guitar solo as well
Music theory isn't rules. Its a model. The same way a cannon ball shot up into the air doesn't follow an exact parabolic curve, but it helps to model it as one. Reality has wind, air pressure is not constant, maybe the ball is spinning one way or another. Reality never consistently follows the model, but it is a good and useful estimate of reality.
The main problem is that while you can emulate any stationary signal in time with sine waves, you have to change their amplitude over time to actually replicate a realistic piano sound.
That’s not a problem for a Fourier transform. The resulting sum will just add up to zero at the start and end of the signal. Yes, it may and will diverge from zero outside the given range, but we’re only interested the nonzero part of the signal anyway.
Huh? I mean, its the same problem with FFTing a chirp signal. If you have non-stationary frequencies over the sample, then the FFT can't replicate the signal.
Have you mastered making gorgeous sounding diatonic/functional chord progressions? (I'll bet $1000 the answer is no!) I've had this conversation with people trying to write great sounding melodies / solos, and in either case the real important thing is: can you make something sound nice just using simple tools? Can you make a great sounding melody with just say 5 or 4 notes? If you can't, you won't be able to make a great sounding melody with all of the notes. The same applies to chord progressions - if you can't make a great sounding simple diatonic chord progression, no amount of chords will help you. Also, having an extremely complicated harmony makes having a nice sounding melody less & less likely, and melody is king in 99% of the music that you listen to.
Its actually quite common in music from the romantic period. Legit, the darth vader and darth sideous chord progressions (or atleast a major part of them) are i - iii and i - biii over and over. (Can really call them prolongations in a way) Not to mention the minor plagal cadence or the I - III - IV - iv loop used in radioheads creep.
There are a few approaches, the other ones are chromatic voice leading techniques (probably schenkarian or something similar). That and Neo-Riemannian theory or NRT. This is found in the book Audacious Euphony.
Because music isn't representational like a lot of visual art is, it would likely be very difficult to accurately describe what you might hear in your head. Especially for non-musicians
In ideal world, that may be true. But in real world, there is issues and dangers. One of biggest issues I personally afraid of is stagnation in post-AI visual arts.
I don't think AI GANs are necessarily just averaging. I mean, its an interesting way to put it, but the math isn't simple like that. It might make sense for certain things like linear regression where central tendency plays a much more crucial role. But, we aren't doing linear regression. AI art is not unimodal and it isn't just trying to denoise all the data samples. Neural networks are really complex
To be fair, I think the bigger limitation, especially going into the future is going to be more about the limitations of humans than AI generators. As humans, we like to think we are creative, but we are absolute clique machines. Our brains are wired to find patterns, explicitly or implicitly and follow them. The real question is when does/will the latent space be larger than any possible human permutation? What value will there be in an infinite latent space?
I would argue that its an incomplete analogy. While classification (aka vocabulary/trying to box things) is a part of music theory. Arguably, a lot more of music theory also has to do with models and structures too.
Kind of apples to oranges isn't it? I mean, if I drew a clown nose on your face without consent and you don't like it. I wouldn't compare that to rape, even if they have this shared element.
I find this argument silly because the designation of what is art and isn't art has a long history of sexism, racism, etc. I mean, just look at the differences between arts and crafts. What is considered art is often more of an exclusionary process versus an inclusive one.
Personally, I just see AI art as just a different medium. Yes, it is a visual medium and yes, it can estimate representative subject matter. But its also way more constrained than you realize (its really bad at rendering unless its a super vivid idea). Plus you wouldn't say that digital art is out to kill and replace traditional painting either, its a goofy narrative. Same for photography and phone cameras or whatever dumb bs. Will there be an impact? Yes. But its not a doomsday event. Not to mention that there is more to art than drawing or painting. There is more to art than representative art. There is more to art than aestheticism.
Let's be honest - the Ai image generation created a whole army of people who came out of nowhere, learned how to type git pull or paid $30 for midjourney sub and then became self proclaimed Ai artist and descended on artstation, deviantart and you name it.
Idk bro, if a drawing noobie wants to call themselves an artist, despite that its their first "proper" doodle. I might not like that they're not skilled like a professional, but to give them crap for it? Yeesh.
You can argue for it since its only a single data point, but it does raise the question of how different visual art communities see AI art. Especially AI arts impact, but also how it might benefit them. I'm sure the pottery people aren't nearly as threatened as much as digital concept artists
Claus Ogermann was one of the most prolific and studied jazz arrangers and came out of a purely strict classical, contrapuntal background, so stop spreading misinformation.
Honestly, its such a stupid mistake to get all elitist about historical formal devices. Especially just harmonic devices. There is more to music than harmony, there is more to art than form (in the art sense, not the music sense).
I'll put it like this. The more theory you know the more you like composers and players who are crazy with theory and the more simple things sound dull.
Honestly no, I just listen to music for fun and also for my brain with different playlists. Also it helps me appreciate the complex structures even in simple music. Nothing is truely simple
I love analyzing classical form by William Caplin (the classroom edition). Its a book on the structure of musical ideas. Great for getting an intuitive grasp on melodic, harmonic, and various other musical structures from the small to the large scale.
Honestly both ways is fine. They both are modelling harmony and they both have their limitations and tradeoffs. Just got to fit the model for the right job
Can't tell if this is april fools post or you don't know what your talking about.
While thematic melodies are motivic. There are non-motivic melodies. Good examples are transitions where the melody section is often just playing scale runs arpeggios. A common popular example is the guitar solo as well
I didn't know those were still considered melodies! Thank you!
I mean, you can play around with the concept of melody. Its better explained in analyzing classical form by William Caplin (the classroom edition)
Maybe find a book on the semiotics of music or musical semiotics.
Most times, other systems handle chromaticism better, whether its just a chromatic passing motion or using neo-Riemannian theory
What are these like? They don’t exist in my country but I’ve seen people talk very well about them.
Very sweet and they last a long time. But they are hard as rocks
Music theory isn't rules. Its a model. The same way a cannon ball shot up into the air doesn't follow an exact parabolic curve, but it helps to model it as one. Reality has wind, air pressure is not constant, maybe the ball is spinning one way or another. Reality never consistently follows the model, but it is a good and useful estimate of reality.
The main problem is that while you can emulate any stationary signal in time with sine waves, you have to change their amplitude over time to actually replicate a realistic piano sound.
That’s not a problem for a Fourier transform. The resulting sum will just add up to zero at the start and end of the signal. Yes, it may and will diverge from zero outside the given range, but we’re only interested the nonzero part of the signal anyway.
Huh? I mean, its the same problem with FFTing a chirp signal. If you have non-stationary frequencies over the sample, then the FFT can't replicate the signal.
Have you mastered making gorgeous sounding diatonic/functional chord progressions? (I'll bet $1000 the answer is no!) I've had this conversation with people trying to write great sounding melodies / solos, and in either case the real important thing is: can you make something sound nice just using simple tools? Can you make a great sounding melody with just say 5 or 4 notes? If you can't, you won't be able to make a great sounding melody with all of the notes. The same applies to chord progressions - if you can't make a great sounding simple diatonic chord progression, no amount of chords will help you. Also, having an extremely complicated harmony makes having a nice sounding melody less & less likely, and melody is king in 99% of the music that you listen to.
Its actually quite common in music from the romantic period. Legit, the darth vader and darth sideous chord progressions (or atleast a major part of them) are i - iii and i - biii over and over. (Can really call them prolongations in a way) Not to mention the minor plagal cadence or the I - III - IV - iv loop used in radioheads creep.
There are a few approaches, the other ones are chromatic voice leading techniques (probably schenkarian or something similar). That and Neo-Riemannian theory or NRT. This is found in the book Audacious Euphony.
analyzing classical form by william caplin, the classroom edition
https://www.youtube.com/@FransAbsil
Because music isn't representational like a lot of visual art is, it would likely be very difficult to accurately describe what you might hear in your head. Especially for non-musicians
There's a lot to unpack here.
Analyzing classical form by william caplin. The classroom edition if you like your sanity.
Think about it a little bit 🧐
In ideal world, that may be true. But in real world, there is issues and dangers. One of biggest issues I personally afraid of is stagnation in post-AI visual arts.
I don't think AI GANs are necessarily just averaging. I mean, its an interesting way to put it, but the math isn't simple like that. It might make sense for certain things like linear regression where central tendency plays a much more crucial role. But, we aren't doing linear regression. AI art is not unimodal and it isn't just trying to denoise all the data samples. Neural networks are really complex
I mean,
To be fair, I think the bigger limitation, especially going into the future is going to be more about the limitations of humans than AI generators. As humans, we like to think we are creative, but we are absolute clique machines. Our brains are wired to find patterns, explicitly or implicitly and follow them. The real question is when does/will the latent space be larger than any possible human permutation? What value will there be in an infinite latent space?
AIs are smart, and they do cheat. But like any security problem, its a matter of what things you allow it to do.
I would argue that its an incomplete analogy. While classification (aka vocabulary/trying to box things) is a part of music theory. Arguably, a lot more of music theory also has to do with models and structures too.
Kind of apples to oranges isn't it? I mean, if I drew a clown nose on your face without consent and you don't like it. I wouldn't compare that to rape, even if they have this shared element.
The truth is that AI art is not art, nor is it considered art of any kind as it normalizes theft.
I find this argument silly because the designation of what is art and isn't art has a long history of sexism, racism, etc. I mean, just look at the differences between arts and crafts. What is considered art is often more of an exclusionary process versus an inclusive one.
It's still not art of any kind if it endorses theft and the active movement of replacing artists everywhere and makes art strictly for AI.
Personally, I just see AI art as just a different medium. Yes, it is a visual medium and yes, it can estimate representative subject matter. But its also way more constrained than you realize (its really bad at rendering unless its a super vivid idea). Plus you wouldn't say that digital art is out to kill and replace traditional painting either, its a goofy narrative. Same for photography and phone cameras or whatever dumb bs. Will there be an impact? Yes. But its not a doomsday event. Not to mention that there is more to art than drawing or painting. There is more to art than representative art. There is more to art than aestheticism.
Clickbait for any art form is everywhere. Its the limitation of youtube as being more of an entertainment platform than an educational one.
Let's be honest - the Ai image generation created a whole army of people who came out of nowhere, learned how to type git pull or paid $30 for midjourney sub and then became self proclaimed Ai artist and descended on artstation, deviantart and you name it.
Idk bro, if a drawing noobie wants to call themselves an artist, despite that its their first "proper" doodle. I might not like that they're not skilled like a professional, but to give them crap for it? Yeesh.
You can argue for it since its only a single data point, but it does raise the question of how different visual art communities see AI art. Especially AI arts impact, but also how it might benefit them. I'm sure the pottery people aren't nearly as threatened as much as digital concept artists
Claus Ogermann was one of the most prolific and studied jazz arrangers and came out of a purely strict classical, contrapuntal background, so stop spreading misinformation.
Honestly, its such a stupid mistake to get all elitist about historical formal devices. Especially just harmonic devices. There is more to music than harmony, there is more to art than form (in the art sense, not the music sense).
I'll put it like this. The more theory you know the more you like composers and players who are crazy with theory and the more simple things sound dull.
Honestly no, I just listen to music for fun and also for my brain with different playlists. Also it helps me appreciate the complex structures even in simple music. Nothing is truely simple
I love analyzing classical form by William Caplin (the classroom edition). Its a book on the structure of musical ideas. Great for getting an intuitive grasp on melodic, harmonic, and various other musical structures from the small to the large scale.
Absolutely,
Honestly both ways is fine. They both are modelling harmony and they both have their limitations and tradeoffs. Just got to fit the model for the right job