CAVI (2021)

a coadaptive audiovisual instrument using mixture-density recurrent neural networks (MDRNNs) to generate control signals in real-time based on muscle and motion data of the performer's actions. The generated control signals automate the live sound processing based on layered time-based effects modules.



Playing in the "Air" (2020)

a long short-term memory (LSTM) artificial recurrent neural network (RNN) model that is trained with raw electromyographic (EMG) signals as the input and predicts the sound energy envelope (RMS) for controlling an audio program using physics-based synthesis. (paper)



RAW (2019)

a muscle-based collaborative body–machine instrument that relies on `sculpting' the sonification of raw bioelectric music (EMG) signals. The instrument is built around two Myo armbands located on the forearms of the performer. These are used to investigate muscle contraction, which is again used for the sonic interaction design. RAW explores a chaotic behavior in control and automatized ensemble interaction.



Vrengt (2018)

a body–machine instrument for an active co-performance between a dancer and a musician. It uses sonification as a tool for systematically exploring the dancer's bodily expressions. The exploration is based on a ``spatiotemporal matrix,'' focusing on sonic microinteraction. Two Myo armbands are used for capturing muscle activity of the arm and leg of the dancer, together with a wireless headset microphone capturing the sound of breathing.



Biostomp (2017)

a new musical interface that relies on mechanomyography (MMG) as a biocontrol mechanism in live performance situations. Designed in the form of a stompbox, Biostomp translates a performer’s muscle movements into control signals. A custom MMG sensor captures acoustic signals of muscle tissue oscillations resulting from contractions. An analog circuit amplifies and filters these signals, and a micro-controller translates the results into pulses. These pulses are used to activate a stepper motor mechanism, which is mounted on parameter knobs on effects pedals. The primary goal in creating Biostomp is to offer a robust, inexpensive, and easy-to-operate platform for integrating biological signals into both traditional and contemporary music performance practices without requiring an intermediary computer software.



LDS (2016)

a light-dependent noise box with a step sequencer. 




Ellix (2016)

a modular wearable interface using mechanical muscle sensing (MMG), integrated motion units (IMUs), infrared (IR) sensors, flex sensors, an heart rate (HR) sensor + joysticks & switches 




spectacleX (2015)

a pair of glasses equipped with an integrated motion unit (IMU) and infrared (IR) sensors.




modGuit_ (2015)

an augmented guitar using an integrated motion unit (IMU), force-sensing resistors (FSRs), and infrared (IR) sensors.




dat-I (2015)

a wearable multimodal sensing system to monitor bioelectric muscle activity (EMG), galvanic skin response (GSR), blood pressure, and three-dimensional acceleration and orientation.



NoiseWig~ (2014)

a "punk" style wig equipped with integrated motion unit (IMU) sensors.


kaybolm[ ]. (2014)

a generative sonification of the dataset of the "Enforced Disappearance" cases by Hafıza Merkezi. The piece is developed in Max. It aims for a sound-mapping of the geographic distribution of people who were "lost" under custody in Turkey between 1980 and 2004.