algorithmes, images, sons
exposition double de Daniel Maszkowicz et Robert Turner
Forde, du 12.12.19 au 19.12.19
Nous partons de la figure de l’arbre, trace d’une vision commune et d’une expérience primordiale, point de départ de constructions algorithmiques. L’information se crée en décomposant et triant des données; pour mettre en exergue ce procédé nous décomposons et recomposons des sons et des images afin d’obtenir des réalités altérées. Cette exposition est constituée de “l’Arbre Sonore” (DM) – installation hexadécaphonique immersive – et des images de “Pohutukawa” (RT) – accrochage d’abstractions numériques.
Sonic Tree by Daniel Maszkowicz
algorithmic-hexadecaphonic-sound-installation
Sonic Tree is a metallic structure in a room of concrete, contrasting with the organic movements it incarnates. Its virtual neuron cells exchange electric signals, and its metallic cells propagate sound faster than in the air. Just like any tree, it offers a comforting shelter with the sounds of insects and the soft brease in its leaves.
This sound installation features an autonomous audio dedicated programmable processor containing the algorithms for the real-time processing of audio signals. This processor allows controlling sixteen independent channels, each of them driving sixteen small active speakers distributed alongside the branches of the tree. A small piezoelectric contact microphone place at the top of the trunk allows amplifying the sounds produced by the movements of the metallic chains.
An artificial neural network has been implemented with sixteen neurons, each of them corresponding to one audio channel. The signal emitted by each neuron results from a matrix combination which coefficients reflect the sound interactions in the network. The original sound source comes from the contact microphone, and is processed by the neuron network for an effect of an organic movement through the branches of the tree.
Daniel Maszkowicz
Pohutukawa by Robert Turner
To a computer, an image is an array of numbers. In order to save space, images can be compressed using algorithms to create a decomposition into three new arrays. The middle of these is made up of a single diagonal line, where values decrease sliding downwards. An image is compressed – with some loss of information – by neglecting the smaller values near the bottom. By retaining only the first few values, an abstracted version of the original is created.
In the simplest case, only the top-left value is retained, leading to an image in which each horizontal line and each vertical line is assigned a colour. The result depends only on the merging of colours as the lines intersect. This kind of pattern is familiar in weaves, tartans and the like. By keeping several values rather than just the one, an image is built up superimposing images of this simple type.
All works displayed here are based on a single image of a tree – the magnificent pohutukawa of New Zealand – to mirror Daniel’s sonic tree. Built out of the same basic constituents, each image is recomposed in its own unique way. Colour, light, and relative importance of the building blocks have all been adjusted to evoke different realities emerging from a single source.