I feel the futuristic relationship between technology and human in Dan Tepfer, a jazz pianist (2)

#BachUpsideDown #1: Here’s Bach’s 1st Goldberg Variation, normal, then upside down! Been meaning to do this for a while. My computer‘s doing a chromatic inversion of what I just played, so the intervals are preserved. You get incredible harmonic movement this way. Whoo! pic.twitter.com/8Ul9jNZjn8

— Dan Tepfer (@tepferdan) May 31, 2019

The above is a quote from his recent tweet. Since the previous article, I've been writing a story that the jazz pianist Dan Tepfer's performance suggests the art of the future.  After playing the first piece of Bach's Goldberg Variations, the program he wrote is auto-playing notes,  each of which has the same strength and timing of the keyboard touch he played, but the scale is upside-down. Every time the original sound goes up by a semitone, the new one moves down the same musical length (= chromatic inversion). As a result, the tonality changes from major to minor.

Why do I pick up these things and say "the futuristic relationship between technology and human"?  So far, when people hear of the usage of computers for music, many people may have imagined so-called computer music or electronic music. On the other hand, Dan's music is using a computer for human augmentation. Those who are watching the video above know that the computer is  automatically playing in the latter half, but those who are just listening to the sound will not notice. His performance is based on a new philosophy. From the viewpoint of technology, his technology is not what stands in the human's way, but expand human skills naturally standing by the human.

As I wrote in the previous blog article, I talked with Dan Tepfer when he visited Tokyo (June 4, 2019). I will write about the conversation with him starting from the question I wanted to ask him very much.

That is, "So far, you have been extending your performance with your own program (= an algorithm you created), but do you have any plans to use deep learning  technology too?"

The meaning of this question might be hard to understand for people who are not familiar with deep learning, so I will explain about it a little bit. Deep learning is the core technology of so-called artificial intelligence (AI). The rapid development of this technology is the major reason that non-negligible number of experts say that  people may lose their jobs because of AI. Although I wrote "your own program (= an algorithm you created)" in the above, in fact this equation holds almost true only for the programs not using deep learning and is not correct for the deep learning programs. In the deep learning, the equation breaks down. This is because deep learning is more like making a black box that models very complex rules and know-how from an extremely large amount of data that cannot be handled by humans, rather than making an algorithm.

In the case of music, the "a large amount of data" mainly means the recorded data played by various past and current players. That is, to use deep learning means “using a large number of rules and know-how that can not be explained logically, which is mainly extracted from the accumulation of various people's performances in the past". That is, using the deep learning is similar to taking in the abstracted collection of unspecified number of player performances.

So, if I say my question in other words, "You have been using the automatic play with a computer within the range the you know the logic of the computer play, but  are you planning to extend it to the range that you cannot predict what kind of sounds the computer plays next? In a sense, it also incorporates the performance of other people in abstract meaning." Actually, it's technically not too difficult compared with what he has been doing, and some are doing experimentally. The point is whether a top-level pianist like him is positive or negative to use such techniques to extend his music.

I asked him this question with such a complicated meaning, but he instantly understood my intention and said, "Yes, that's an important question!". He continued, "Actually, I'm very  interested in it, but currently I'm a bit skeptical." From his words, I thought that his spirit of inquiry would urge him to use it experimentally, but he was wondering if it would be really effective and worth extending his music (I myself have ever given the Goldberg Variations to Google Magenta engine and have generated sounds, and noticed that I will need various ideas and techniques to make them into natural and comfortable music.)

Furthermore, he said a surprising things to me. "Actually, I'm planning to do research in a research institute Ph.D. course about technology and music. 

Here, I noticed that he was aiming at even deeper and more advanced field than I had imagined. I thought, "Dan Tepfer is trying to work on the music of the future much more flexibly than I imagined". When I thought so, I remembered his own home page. He writes on his home page:

if you always play tennis against the same person, you only get better at playing against that person. If you always play with different people, you get better at tennis. What I’m trying to do is clarify my message, independently of style. To get right at the music itself.

He is a jazz pianist, but he is also a pianist who plays difficult classical music pieces freely as in his Bach's Goldberg variation video shown first.  There may be some people who think that classical music pianists and jazz pianists are quite similar except the category of the pieces they play, but actually, they are different not only in the music genre, but in quite basic levels. There are few people who play both without any boundaries. In old days,  classical pianos were played by "musicians" like Mozart, Chopin, and Liszt who composed and played, but many classical music pianists today are "players" since the composers and the players are separated professionally in the modern age. On the other hand, jazz pianists are "composers + players" whose play is mainly ad lib (improvisation), and they naturally play while they are doing the activity that is very close to composition. There are some researches which say the usage of brain is different between classical pianists and jazz pianists. There are some musicians who started from classical pianos and moved to jazz (such as Bill Evans), but I've never heard of the reverse, and there are very few musicians who move around classical and jazz piano field freely. I'm not sure but the situation of other types of musicians will not be much different the piano. Honestly, I like that kind of musicians "who freely goes back and forth between classical music and jazz" or "the ones who freely straddle the genre of music", even without mentioning the names such as Andre Previn and Philip Glass. It is because I feel that they are pursuing the music itself, if I borrow Dan Tapfer's words.

Now, Dan Tepfer is going even beyond the categories of classical music and jazz,  and is freely crossing the music genre of "human and machine"  to master music. This suggests us various things when we think about  the relationship between AI and human beings in the future (to be continued).

<- go to the first article            -> go to the next article