On February 19 I play a concert with pianist and improviser mastermind Mattias Risberg. Also joining will be the amazing Anna Lindal and Eva Lindal on violin. The concert is set at the small and cozy Omnikvariatet in Uppsala that continues to be one of the few places for new music in Uppsala. This was the first time a played with the Benjolin made by Derek Holzer in a special version he made for me with a small preamp built in.

This is a simple video that I shot in order to document what I did in this improvised performance. Unfortunately the camera only picked up a part of the modules. The sound is just a mobile phone recording but it gives a rough approximation of the performance.

On October 16 I'm playing at Håkan Lidbo's lab Rumtiden on Närkesgatan 6 in Stockholm. I will again play a set on the Dataton 3000 modules discussed that I've played on numerous times in the last few years (see Norberg and Forum Artistic Research ).

One thing I will try this time is to play with a kind of feedback in one of the oscillators. What I'm doing is actually merely modulating the signal with itself which creates strange and unexpected effects, and sounds that I can't get to otherwise. Each oscillator can be modulated by one of the other oscillators, but connecting the output of the first oscillator to the modulation input of the same channel is not possible without physically connecting the output to the input.

y

A PhD student at KTH that I am supervising did as his master's project an annotated human database with a thousand samples. A number of composers and researchers submitted sounds an annotated them as can be seen from the huggingface database version of the dataset.

Together with my collegue Rikard Lindell in the IRESAP project we set out to create an machine learning algorithm that would be able to classify sounds according to Pierre Sceaffer's.¹ The first step was to map the sounds of the dataset to Schaeffer's typology in the picture to the right. Table 1 below is a reduced version of this mapping excercise that we set up where for an N category sound, the sample must have been annotated both with Harmonic sound and Flat sustain. The preliminary results of this was quite convincing but I will return to the actual model that I created from this set in a later post.

Here's a simple shell script to extract data from the downloaded database according to two dimensions of the table above. In this case it is Impulse and Iteration, of N' and N'' in Schaeffer's typology.

#!/bin/bash
dataset="../../dataset/SCHAEFFER/"
keyA=Impulse
keyB=Iteration
base="../classification/training_data/"
dirA="$base""$keyA"
dirB="$base""$keyB"

echo "Clearing directories...\n\n"
rm -rf $dirB/*.wav
rm -rf $dirA/*.wav

echo "Copying files from dataset...\n\n"
for d in "$dataset"*/; do
    for j in "$d"*.json; do
        if test -f "$j"
        then
            if [[ "$(cat "$j" | jq '.object.labels.sustain')" == "\""$keyB"\"" ]]; then
                if [ ! -d "$dirB" ]; then
                    mkdir "$dirB"
                fi
                cp "$d""`cat "$j" | jq -r '.object.filename'`" "$dirB"
            fi
            if [[ "$(cat "$j" | jq '.object.labels."pulse-typology"')" == "\""$keyA"\"" ]]; then
                if [ ! -d "$dirA" ]; then
                    mkdir "$dirA"
                fi
                cp "$d""`cat "$j" | jq -r '.object.filename'`" "$dirA"
            fi
        fi
    done
done

echo "Done!\n"

This paper is submitted and accepted for publication in the Journal of the Audio Engineering Society

This paper os co-written by Jan Schacher, University of the Arts, Helsinki, Sibelius Academy and make an attempt to discuss the experiential dimension of spatial listening in an attempt develop new methods, both for that activity, as well as for working with large speaker setups.

The paper The Imperfect Copy: Role Playing Reenactments of Historical Electronic Sound Instruments which was presented at the NIME conference and which is co-written with Derek Holzer (part of his highly recommended PhD compilation thesis was considered "Highly Commended" at the conference in Canberra, Australia in June. The paper is available here.

I'm performing my piece The other side of black noise, an improvised performance using mainly sound sources from modules of the Dataton 3000 series instruments dating from the early 1970's. This is related to thoughts developed starting a few years ago and presented at Vintage Materialities at the Orpheus Institute.

See also the performance at norberg last year and at the Frequenz festival before that.

The premiere of my piece Imaginary Associations was premiered at the Art Exhibition Center in Växjö on March 8. This is a work in progress and I'm looking forward to working with Thuy further on this piece.

A tuner for the scordatura for a new piece for the wonderful Dan Tranh player Nguyen Thanh Thuy with whom I've worked with for decades in the group The Six Tones

Listen to the piece here