This seminar is based in the article "A Tutorial on EEG Signal-processing: techniques for mental-state recognition in Brain-computer Interfaces" by Fabien Lotte, a chapter of the book "Guide to Brain-computer Music Interfacing" edited by Eduardo R. Miranda and Julien Castet.
The main stages in the development of Brain-computer Interfaces through Electroencephalography will be presented, with the description of the most used techniques for Feature Extraction an Feature Classification..
In this talk we will present many research projects developed from the collaboration between Computer Music and professionals from other areas like Musica, Engineering, Physics, and Medicine. The discussion will focus on the contributions that the Computer Music area provided to the researches and researchers, and we will discuss the interaction between the professionals that had to deal with different technical languages, methods, and scientific approaches in front of the problems to be solved.
The works discussed during this talk had the presenter as a member, and they comprise: musical installation using mobile devices and webservice; the use of mobile sensors during musical performances; a concert taking advantage of cloud computing and web audio; difficult airways evaluation from formant analysis of vowels pronounced by patients; the disgnostic of sleep apnea with mobile devices and sensors; and some works developed with the HardwareLivreUSP group including analog synthesizers and MIDI controllers connected to floppy disks. Some of these researches are in progress, but they already provided many interesting experiences that will be presented during this talk.
Within the research line in Creative Processes focused on music composition and computer-aided orchestration, this work describes an investigation into the instrumental combination process using a computing environment using audio descriptors in PureData (PD) and the PDescriptors library to analyze the sound characteristics of a database of audio files with various musical instruments and extended techniques.
The research relies mainly on three fundamental aspects: a) the study of audio descriptors as an analysis tool for computer-assisted musical orchestration; b) the analysis of works and procedures developed by composers who served as reference for the development of writing models; c) the creation of compositional studies, in which the orchestration hypotheses will be tested. From the study of audio descriptors as an analytical tool, item a, the next step focused on the generation of a universe of instrumental combinations described by their sonic characteristics. This procedure helps the decision making related to the orchestration and expanded techniques. Regarding item b, we used the audio descriptors to analyze the sound characteristics of the following works: the First Movement of the "Quattro Pezzi per Orchestra" by Giacinto Scelsi; three solo pieces "Guero", "Pression" and "Dal Niente - Interieur III", by Helmut Lachenmann; and "Reflections" by Jônatas Manzolli. As concerns the item c, we present a portfolio whose orchestral compositional planning was based on the dialogue between the results obtained from studies of audio descriptors with the analysis of the works of reference composers.
It is expected that the whole process of research reported here will contribute to developing knowledge in the field of computer-assisted musical orchestration, establishing a more refined blend of sound diversity as a method of combination of musical instruments and extended techniques.