A voice sample generated by AI using SampleRNN, which learnt 6 hour recording of Tomomi Adachi's voice performance.

Voice generated by AI  1(SampleRNN)

Voice generated by AI 2 (SampleRNN)

Voice generated by AI 3 (SampleRNN)

Voice generated by AI 4 (SampleRNN)

A voice sample generated by AI using FFTNet, which learnt 6 hour recording of Tomomi Adachi's voice performance.

Voice generated by AI 5 ( FTTNet)

This is a duo improvisation by Tomomibot (left channel) and Tomomibot (right channel).
Tomomibot learnt Adachi's improvisation. In this recording, two Tomomibots improvise together with two different improvisation models and the same sample sets which consist of AI generated voices and Tomomi's prerecorded voices. Recorded as a live public performance on July 25 2018.

Duo Improvisation Tomomibot & Tomomibot

This is a duo improvisation by Tomomibot (left channel) and Tomomi Adachi (right channel).
Tomomibot learnt Adachi's improvisation. In this recording Tomomibot is uttering exclusively voices generated by Neural Networks (SampleRNN and FTTNet).
Recorded on July 24 2018.

Improvisation Tomomibot & Tomomi Adachi

This is a duo improvisation by Tomomibot (left channel) and Tomomi Adachi (right channel).
Tomomibot learnt Adachi's improvisation. In this recording, the Tomomibot improvises with Tomomi Adachi, is uttering a sample set which consists of AI generated voices and Tomomi's prerecorded voices. Recorded on July 24 2018.

In this recording, Two Tomomibot drives two orchestra sound sets reacting Adachi's voice.
Recorded in Akiyoshidai Summer Contemporary Music Festival on August 20th, 2018.

Tomomibot drives 2 orchestras

These sounds were generated by WaveNet, which learnt Tomomi Adachi's voice performances and cat meows equally.

Tomomi's voice + Cat meow generated by AI (WaveNet)

Voices from AI in Experimental Improvisation is a project by Tomomi Adachi, Andreas Dzialocha and Marcello Lussana. We built an AI named Tomomibot (sometimes it is called Tomomi Adachi II) which improvises like Tomomi Adachi.

Our next lecture/workshop/performance is 
April 18th 2019, 20:00 at KuLe (Auguststraße 10, 10117 Berlin)

You find source codes for this project at here
https://github.com/adzialocha/tomomibot
The project is funded by Musikfonds e.V. in 2019 and the Initiative Neue Musik Berlin e.V. in 2018.

This project received the Award of Distinction from Ars electronica 2019.

Tomomibot - AI Improvisation

Popular tracks