Pianist And Coder Dan Tepfer Composes Music With The Help Of Artificial Intelligence NPR's Ailsa Chang talks with Dan Tepfer, pianist, composer and coder, about how he uses artificial intelligence and virtual reality to bring his music to life.
NPR logo

Pianist And Coder Dan Tepfer Composes Music With The Help Of Artificial Intelligence

  • Download
  • <iframe src="https://www.npr.org/player/embed/721172186/721172187" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript
Pianist And Coder Dan Tepfer Composes Music With The Help Of Artificial Intelligence

Pianist And Coder Dan Tepfer Composes Music With The Help Of Artificial Intelligence

  • Download
  • <iframe src="https://www.npr.org/player/embed/721172186/721172187" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript

AILSA CHANG, HOST:

When musician Dan Tepfer was a kid, he taught himself to code on an early Macintosh computer that his dad brought home one day.

DAN TEPFER: What I love about programming is it's so powerful. You can be a little kid, you know, with very little power over the world, and you can tell this very powerful machine to do whatever you want it to do. And it will do it for you, you know, flawlessly over and over again.

CHANG: That little kid grew up to be a world-renowned pianist. And now he's coded another machine - his piano.

(SOUNDBITE OF MUSIC)

CHANG: What you're hearing is not just Tepfer playing his piano. It's also his piano playing along with him. Let me explain. He wrote a series of algorithms that direct his specially modified piano to respond in certain ways. When he plays, the piano plays back according to the algorithms rules. It's like a duet with phantom hands or, to put it another way...

I mean, in a way, you've sort of designed this imaginary friend that can play piano with you whenever you want. There's...

TEPFER: I love that.

CHANG: (Laughter).

TEPFER: You know, I'm an only child.

CHANG: (Laughter).

TEPFER: And no one has ever put it that way, but I think you might be getting somewhere (laughter).

CHANG: Tepfer's new album explores the fusion of human and machine - a dynamic he says has actually been at the heart of music for centuries.

TEPFER: Let's go way back. If you look at, like, medieval composers like Ockeghem, baroque classical composers like Bach, their music lives at the intersection of the algorithmic and the spiritual - equal parts rules and equal parts intuition and, you know, for lack of a better word, spirituality.

(SOUNDBITE OF MUSIC)

TEPFER: So, you know, I spent a lot of time with that idea. And about five years ago, I sat down at one of these Yamaha disklaviers. And I realized - it was a kind of light bulb went off in my head that because the input for the algorithms is what I played, which, by necessity, has this organic quality because I'm human (laughter)...

CHANG: It's an extension of your aesthetic.

TEPFER: Exactly. It's actually going to be right at that intersection of that emotional quality and also the algorithmic.

(SOUNDBITE OF MUSIC)

CHANG: I want to hear how this works, so let's just start with something small. I want to hear how the piano reacts to you. So what happens when you play just one note?

TEPFER: Well, you know, it really depends. They're actually 11 different algorithms that I use. There's a track on the record called "Inversion." And in that one, I set up an axis of symmetry in the middle of the piano. And then everything that I play on one side of the axis gets mirrored onto the other side.

CHANG: Wow.

TEPFER: So maybe - I don't know. I'll play "Mary Had A Little Lamb" or something (laughter).

CHANG: That's pretty simple.

TEPFER: All right. Check this out. So...

(SOUNDBITE OF MUSIC)

TEPFER: OK.

CHANG: Wow.

TEPFER: So I'm actually playing only one of those lines. I'm playing the upper line right now. And what you're hearing in the bass is the response of the piano.

CHANG: I love it. Can you show us, like - stay with "Mary Had A Little Lamb," but get more complex gradually. And explain it as you're going so we can hear how the computer's affecting the melody, the music.

TEPFER: OK. So I'm going to start with the computer responding almost immediately to what I play.

(SOUNDBITE OF MUSIC)

TEPFER: Now I'm going to make it respond a little later. So if I do this, you'll see...

(SOUNDBITE OF NOTES PLAYING)

CHANG: Right.

TEPFER: It responds...

CHANG: Yup.

TEPFER: ...With a little delay. And suddenly, we have rhythm, which gets me really excited.

(SOUNDBITE OF MUSIC)

CHANG: That's amazing. It sounds like Mary is drinking a lot of wine and having a great time.

(LAUGHTER)

TEPFER: This is what I love about this the most is I'm imposing a rule on myself. But what you perceive and, actually, what I perceive too as the musician is an emotional result.

CHANG: When you know every keystroke your duet partner will play, does it take away, at some level, from the joy that can come from just improvising during a duet with another human from being surprised once in a while by your human partner?

TEPFER: Well, you know, I really love playing with a human partner. And being surprised is so dear to my heart. But these algorithms that I've written, they absolutely still surprise me because I don't think we humans really have the right kind of brainpower to anticipate exactly how these things are going to happen. You know, if I take an algorithm I wrote like the demonic march - I like this one I call the demonic march...

CHANG: Let's hear it.

TEPFER: ...Because it sounds that way to me. You know, I'll do something like...

(SOUNDBITE OF MUSIC)

TEPFER: All right.

CHANG: (Laughter).

TEPFER: So I've done that before, and I know what that sounds like. But what if I do something like - I don't know. I'm just going to try to do something random.

(SOUNDBITE OF MUSIC)

TEPFER: (Laughter) I literally just surprised myself. I mean...

CHANG: (Laughter).

TEPFER: There's stuff even rhythmically that I wouldn't expect. And, in fact, you know, the first time I sat down and played with these algorithms as I was programming them, there would just be this blissful moment of discovery. But once that initial surprise wore off, I was having trouble really getting that same feeling that we're exploring. And it suddenly dawned on me a couple of years in. I have to be listening to what the piano's doing in exactly the same way as I am listening to a human that I'm playing with. And then from that day on, I didn't have any trouble anymore.

(SOUNDBITE OF MUSIC)

CHANG: Do you think this method of creating music alongside artificial intelligence will become more and more common? Or do you think it'll always be kind of this novelty?

TEPFER: You know, the history of music is defined, in many ways, by progress and technology. You know, the organ was one of the most complex pieces of tech around when it was developed. And I just can't imagine humans shying away from the possibilities that it brings. And I think over time, the allure of novelty wears off. And then we're left with something much deeper, which is simply how these new tools can open up new artistic horizons to creators.

(SOUNDBITE OF DAN TEPFER'S "CANON AT THE OCTAVE/METRICMOD")

CHANG: Dan Tepfer's new album is called "Natural Machines."

Thank you very much for joining us today.

TEPFER: Thank you so much.

(SOUNDBITE OF DAN TEPFER'S "CANON AT THE OCTAVE/METRICMOD")

Copyright © 2019 NPR. All rights reserved. Visit our website terms of use and permissions pages at www.npr.org for further information.

NPR transcripts are created on a rush deadline by Verb8tm, Inc., an NPR contractor, and produced using a proprietary transcription process developed with NPR. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.