Oregon Graduate Institute of Science and Technology has created a public resource for experimenters. Download trial software for speech recognizers.

A voice processing toolbox for Matlab, very useful for prototyping algorithms and has routines to do most of the basic useful tasks in speech.

A library in C++ for developing Open Source speech and machine learning applications. The code is under the Gnu Public License.

Provides a range of information on speech technology, including speech synthesis, speech recognition, speech coding, and related material. Pages look a little old (1997) but a very wide range of FAQ here. Algorithm for building simple recognizers.

A detailed paper describing speech recognition technology.

Pictures of early voice synthesizors hosted by UCSC - site seems always under construction.

Has advanced software algorithms for speech compression, enhancement, recognition and speaker identification (i.e., voice recognition).

Main research directions are speech recognition/understanding, methodology of speech dialogue systems, applied systems, hidden Markov algorithms.

Article describing the rapid-deployment Text-to-Speech processes and development timeframe used in the DIPLOMAT System.

Graduate student thesis which may give ideas for difficult environments. Mostly in German, but there is also an English page (click flag). Approach develops a little bit of intelligence between speech recognition and text to speech engine.