next up previous contents
Next: FPMs Up: Task 5.1: Status Previous: Task 5.1: Status

ICSI

We have tested two common assumptions on multi-band ASR: 1) the objection of the critics of multi-band ASR that it is inherently inferior to a full-band approach because phonetic information is lost due to the division of the frequency space into sub-bands; and 2) the assumption by multi-band ASR researchers that transitions in frequency bands often occur asynchronously (i.e., at different times than the full-band transition).

Addressing these two assumptions is important because they have been posited by many in the ASR community, the first by multi-band critics, and the second by the enthusiasts. If true, the first objection would be a serious weakness of multi-band, and the second would be a considerable advantage. Hence, it is important to study these untested assumptions.



Christophe Ris
1998-11-10