Our goal for the workshop

The field of audio and speech processing has been long split between two major schools of thought. On one hand there are the "perceptualists", who devise systems based on physiological and psychological observations, and on the other hand there are the "statisticians" who prefer to employ mathematically rigorous methods that usually have little in common with life processes. To further divide the situation there is a further segmentation between the work of the speech, general audio, and the music information communities. Although the original intentions and philosophies of all these schools differ considerably, the problems that they are applied on are quite often the same. As a result we have seen the development of elaborate, yet unusually distinct, systems, which often perform the same tasks.

The purpose of this workshop is to bring together people from all of these fields and examine the differences, agreements, and intersections of their work. We hope to expose participants to different types of thinking when it comes to audio and speech processing, and to initiate cross-pollination between disparate approaches. We are particularly interested to see how hybrid approaches can be conjured and whether this is a good way to advance the state of the art to the next level.

This workshop will take part as a satellite of ICSLP2004. The workshop will consist of four sessions spread over a full day. Each session will address a selection of topics, covering perceptual, the statistical and the hybrid approaches. The topics that we are interested in are wide in range and are centered in the field of audio and speech processing.

For a complete list of the topics to be covered by the workshop, click here.