Jordan and the jam_bot

Tonight’s performance is the culmination of three years of developing jam_bot at the MIT Media Lab, in Professor Joe Paradiso’s Responsive Environments lab. Our amazing team, Lancelot Blanchard and Perry Naseck, and I have been on this journey together through residencies here at MIT, visits to my home studio, and countless hours working together, sometimes even across continents while I was in the midst of a 19-month Dream Theater tour. This has truly been a labor of love for all of us. We’ve spent three years experimenting, building, tweaking and continually discovering new possibilities.

Tonight isn’t the end of the project. It’s a stop on a journey that’s still unfolding. We’re excited to share where jam_bot is right now and to have all of you here with us for this chapter of the adventure.

Improvisation has always been at the center of my musical life. With jam_bot, we’ve been exploring what happens when an AI system trained on my own musical language becomes another musician in that improvisational conversation.

Nothing you’re about to hear has been composed in advance, with the exception of the Bach chorale near the end of the program. These pieces are happening in the moment. I don’t know what jam_bot is going to play, jam_bot doesn’t know what I’m going to play. Each piece explores a different way for us to listen, respond, lead, follow, and create together.

1. Emergence

jam_bot solo

The system comes alive, and the concert begins before I even reach the keyboard. As I enter, jam_bot is generating a string ensemble piece, creating a kind of processional and welcoming me into our shared musical space.

2. String Theory

jam_bot & Jordan — trading off

jam_bot begins on its own with a world of strings, and then I join in. From there, we begin listening and responding to one another, passing musical ideas back and forth and seeing where they take us.

3. Free Flow

jam_bot & Jordan — trading musical ideas

This is improvisation in its most open form. jam_bot and I trade musical ideas freely, with neither of us knowing where the other is going next. Sometimes I lead. Sometimes I follow. Sometimes one musical idea sends us somewhere neither of us expected.

jam_bot is trained to make intelligent musical decisions and is fairly advanced at this point, but sometimes it makes choices that surprise me. That’s part of what makes playing with it so interesting, fun and challenging.

4. Adagio

jam_bot expressive strings / Jordan piano, guitar & GeoShred

I’ve always loved Barber’s Adagio for Strings, and its emotional and harmonic world was the inspiration for this model.

Here, jam_bot takes responsibility for the harmony, creating the chords in real time. My job is to listen, interact, and influence the harmony while creating the melody. My Strandberg guitar will take a solo and then, GeoShred, the iOS app based on physical modeling that I developed with a team out of CCRMA@Stanford University, will join with a cello solo. 

5. Progressive Current

jam_bot & Jordan — trading musical ideas

A change of sonic world. Using a mixture of mallet percussion and other textures, jam_bot and I return to our musical conversation, trading lines and reacting to each other in real time.

6. Kinection

Jordan, jam_bot, tap dancer & hip hop dancers

Now another human voice enters the improvisation, but this one speaks through rhythm and movement. A dancer joins jam_bot and me, adding another layer to the conversation. We listen and respond to each other in real time as melody, percussion, rhythm, and physical movement all become part of the improvisation.

7. Take it, jam_bot!

Jordan harmony / jam_bot lead

I’ll create the chords and bass while jam_bot takes the lead, improvising melodic shapes over the musical foundation I give it. For this one, I’m laying down the chords and bass. jam_bot listens and, if it “behaves”, will play over my changes like a pro..

8. Role Reversal

jam_bot harmony / Jordan lead

Now we switch. jam_bot becomes my accompanist, providing the musical foundation while I take the lead. After controlling myself for most of the evening, this is where I FINALLY get to shred.

9. Invisible Orchestra

Piano & gesture-controlled orchestra

After all that shredding, it’s time to switch gears and return to the piano where  my hands take on another role. Through gesture. I can control an orchestra of strings and brass while I play, shaping this virtual ensemble with the movement of my hands.

10. From Bach to Now

Jordan, jam_bot & MIT Chamber Chorus. Ryan Turner, conductor

We begin with Bach’s Chorale,  “Komm, süßer Tod.”. 

From there, we move from music that has existed for centuries into music that is being created in the moment.

As I improvise at the piano, jam_bot listens to what I’m playing and creates musical notation in real time. It separates that music into individual vocal parts and sends the newly created sheet music directly to the choir. The singers sight-read what jam_bot has just written and sing it back to me. Then I respond to them. Piano to jam_bot. jam_bot to written music. Written music to choir. Choir back to me.

___________________________________________________________________________

Thank you for joining us tonight and for joining us  on our continuing explorations of creativity, music, and technology.

For me, this journey has always been about staying curious, experimenting, and discovering new ways to make music. Preparing for this concert was an excellent impetus to propel our work forward. I really hope that you leave tonight feeling inspired. 

Jordan Rudess

Please visit the jam_bot installation at the MIT Museum