Technical data:

Sound and visual installation using computational agents to decide the whole behaviour. Year : 2021

The set of recordings made from the opening of the fully functional website https://deas.sonoscop.net is the base material that initially feeds the nuclear application of the piece, responsible for its real-time appearance, which I call Syringa, integrates a reworking of NettyMcNetFace, it is made up of 16 computational agents and behaves, during a randomly chosen amount of time, as follows:

1. Each of its agents chooses a recording from the database according to the following criteria:

1.1 If no other agent has made a choice, choose a recording at random from all those available.

1.2 If other agents have made a choice, according to a probability greater than 70%, choose a random recording from the set of semantically similar recordings those chosen by other agents or, according to the complementary probability, never greater than 30% Therefore, choose a recording at random among all those available. The semantic markers emerge from the analysis of the texts carried out by isolating.

2. Each agent applies a sequence of transformations in the recording chosen according to a criterion that depend on the semantic field assigned to each recording. The idea pursues the association of fixed sequences of the treatments mentioned above (transposition, juxtaposition, overlap, convolution, fragmentation, granulation, filtering and spatialization) to the semantic markers. Agents are programmed in order to respect a series of perceptual limits according to the empirically obtained information in the programming process. They extract the musicality of the voices in a similar way to the musical work of Cocuprenant de Quaresme.

3. The musical results of the processing carried out by each agent are broadcast to the elastic sound space of installation during the installation time, and to the IP space, the Internet, to give ideas to eventual remote improvisers and to be registered, to become part of the database.

4. Parallel to the sound processes, the agents also choose, according to an equally semantic criterion, words of which the appearances written in their original alphabets will also be treated and, later, virtually projected onto 3D objects, which, in turn, will form part of the images projected by four projectors on the architectural elements of the space finally assigned to the piece by the Mixtur festival and which is still to be decided. Likewise, this production will be broadcast in streaming and its behavior will be determined by changing sound characteristics, such as aspects of the color, frequency, rhythmic density, etc. of voices, thanks to the use of tools long known as sound descriptors [27]

5. Elapsed the chosen time, go to 1.

D’Eco a Siringa

Composition of a musical and visual flow in continuous transformation, without formal resolving elements throughout the installation time, whatever its duration, dynamically built, to starting from a database accessible in the IP space and initially made up of materials obtained by reading words from world women's literature, spoken by diverse female voices in as many languages as possible and constantly and dynamically maintained. At present there are sound samples from 60 languages. The process began with languages spoken in Barcelona. Currently, any language is welcome (https://deas.sonoscop.net). Images are dynamically constructed from the graphic treatment of texts in the various languages and alphabets, chosen according to spectral characteristics of sounds and procedures selected at variable time intervals. Its instantaneous behavior is reactive in real time to the general dynamics of sounds. Despite the continuous change in the sound and visual flow that is generated as a consequence of ethical-aesthetic reasons dictated by the need to construct works avoiding the traditional musical forms marked by patriarchy, I consider the problems of the distillation of the essence and the maintenance of the identity of the work of art, which, in this case, is achieved through the “intelligent” manipulation of the timbral colorations of voices and the balance of the images. To this end, recordings of readings and the appearance of texts are subjected to various forms of signal processing; in particular, transposition, juxtaposition, superposition, convolution, fragmentation, granulation, filtering, and spatialization.