Muhammad Hafiz Wan Rosli
PhD, March 2017. Committee members: Curtis Roads (Chair), Clarence Barlow, JoAnn Kuchera-Morin, Andres Cabrera, Matthew Wright.
The process of segmenting a sound signal into small grains (less than 100 ms), and reassembling them into a new time order is known as granulation. Articulation of the grains’ spatial characteristics can be achieved through various techniques, allowing one to choreograph the position and movement of individual grains. However, this spatial information is generally arbitrarily generated during the synthesis stage.
Dr. Wan Rosli's PhD thesis introduces a novel theory and technique called "Spatiotemporal Granulation". Through the use of spatially encoded signals, the proposed technique segments spatial and temporal information producing grains that are localized in both space and time.
Ambisonics is a technology that represents full-sphere spatial sound through the use of Spherical Harmonics. Using the encoded spatial information as part of the analysis-synthesis process enables the extraction of grains by spatial position, in addition to temporal position. Examples of potential effects and transformations include manipulation of spatial characteristics, spatial cross synthesis, and spatial gating. Angkasa is a software tool that was written to demonstrate the viability of this technique.