The ICSI Speech FAQ is the latest attempt to provide some documentation for speech research theory and practice here at ICSI. The vision is to create an evolving, continually updated directory of issues and solutions associated with speech recognition research.
Answers to common questions are prepared by group members by filling in a template file (which is actually XML, for no particularly compelling reason). Answer files have an arbitrary file name, but with an extension of ".xml". They are put in the "anssrc" subdirectory of the main speech FAQ directory, /usr/local/www/Speech/faq/ .These source files are then automatically translated into HTML and an index file is built at the same time by running "make" in the main directory. A second plain-text source file, ICSI_SPEECH_FAQ gives hints for the ordering of the questions within the index page. For a contributed answer to be listed against one of the queries in this list, the TITLE field of the answer XML file must exactly match the question in the hints file. Multiple answers to the same question are supported.
Please feel free to add questions to the hints file (which is accessible through RCS), or to write new answers and put them in the anssrc directory.
Back to ICSI Speech Group Home