5100 seconds of summer in Tokyo.

Otomatk* was born from interest in sound composition and distraction from MaxMSP. Despite a lifetime of writing code, MaxMSP still confounds me for some reason, and after another recent attempt at coming to grips with MaxMSP, I became frustrated again and veered in another direction, writing a code library to spawn sound from the silicon chips in my computer. The otomatk codebase consists of several modules that run from the command-line in nodeJS. Everything was written from scratch, from the sound design, to oscillators and timing and sequencing and quantization, to percussive elements utilizing FM synthesis and Karplus Strong algorithms, to delays and filters, to track handling, mixing and wav file output, just because it was fun. I began posting some snippets, which were accompanied by video generated using ffmpeg (also launched from the otomatk library). I came up with a reasonable system for cataloguing the findings / compositions in a git repository, as well as utilizing ffmpeg to generate image thumbnails for reference, since the exploration of the sound in tandem with the video output made generating interesting visuals part of the composition / code writing process.

Part of the interest in building the library was to enable me to create aleatoric compositions utilizing code. Reused code-blocks are refactored into separate functions and modules for use in future compositions, and the use of the Mersenne Twister algorithm allows for creating repeatable aleatoric compositions that can be regenerated later with various variables tweaked, such as bpm and bitrate, video size, etc.

Some of the compositions that were interesting visually were also very meditative and drone-like, and so I wanted to pursue various permutations of one composition. According to my feeble attempt at organization, each separate composition is based on a single code file that describes how the composition is created, with a timestamp. The code also allows me to generate multiple versions of the composition utilizing multiple random seeds passed to Mersenne Twister, which can be "watermarked" in the file name with the bpm, the random seed, and the duration specified from the command line. This album features a set of permutations based on one composition, i.e. generated from the same piece of code. Utilizing a bash for-loop, hundreds of different 4 minute segments were generated using hundreds of different random seeds, from which I selected a set of them to include on this collection, and regenerated them at the desired track length for the album. Each permutation has its own random seed which in turn gives each its own key signature and randomized sequences and parameters for panning, delay and timing. All of them were generated at 60bpm. Originating from the same code source, all the pieces sound related, but differences in key signature and timing give them all a slightly different feel, some darker, some lighter, some calm, some filled with tension. Some are self-consistent, and some slowly evolve and take unexpected turns.

In order to impart some sense of time and context, two field recordings were made in west Tokyo (調布市) to capture the sounds of summertime along the Nogawa River (野川), one at rest and one in motion. These field recordings provide background ambience to all the pieces sequentially, so listening to the album from beginning to end allow one to re-experience the time captured in the recordings. One field recording was made in motion, walking, covering a few miles along the Nogawa River, while the other recording was made at rest, in one spot of the river near Nogawa Park. The two recordings are played simultaneously.

This album is dedicated to the memory of and composed in my conception of the spirit of Ryuichi Sakamoto, with other obvious influences being Time Machines by Coil, Thursday Afternoon by Brian Eno, Soliloquy for Lilith by Steven Stapleton (Nurse with Wound), and the ambient collaborations of Carsten Nicolai and Ryuichi Sakamoto. The inspiration for the in motion field recording came from Scanner's (Robin Rimbaud) track Ascent, which I originally mistook to contain a recording of someone walking through fresh snow.

*otomatk is a play on several different words. Obviously, "automatic" because the sound pieces are generated automatically by code. The prefix "oto" in English means "of or pertaining to the ear", but "oto" in Japanese (音) means "sound", so automatic ear or automatic sound, perhaps. oto-ma in Japanese could also be written as 音間 or roughly translated "sound space/interval", as in the space between the sounds.