-Written by Arthur Chan

Developers of Sphinx Here we list all Sphinx's developes who have
contributed to Sphinx.  Since, in the early stage of Sphinx's
development we didn't use RCS and CVS.  This version only include the
contributions which can be derived from the source code.  The
description may not be accurate enough.  Please kindly correct us if
there is any descripencies of this file.

There are also countless contributers for Sphinx. It is possibly
impossible to list them in a single file.  Do protest to us if you
want your name to be somewhere here.  We are more than willing to do
it. 

-Arthur Chan.  20040825

Kai Fu Lee - The one who proposed Sphinx. 

Hsiao-wen Hon - According to Dr. Kai-Fu Lee's "Sphinx: a speaker
independent speech recognition system", Hsiao-wen is the implementer
of Sphinx I. 

Fil Alleva - 
Robert Brennan -

Mei-Yuh Hwang - The proposer of the concept of senones. 

Xue Dong Huang- Manager of Sphinx 2.  A lot of improvement were
introduced from Sphinx 1 to Sphinx 2.

Ravishankar Mosur - Implementer of Sphinx 2(fbs-6,7,8) and Sphinx 3
(both flat lexicon decoder and tree lexicon decoder.  Many fast
algorithms in S2 and S3 were designed and implemented by him. 

Eric Thayer - Architect of SphinxTrain, he unified a lot of toolkits
of CMU's researchers and create the backbone of the current
SphinxTrain.

Rita Singh - Proposer and Implementer of automatic question
generation.  She also implemented the first version of live mode
decoder and write a comprehensive on-line tutorial for SphinxTrain,
sphinx2 and sphinx3.  Many researchers were benefited by her version
of modified code. 

Bhiksha Raj - Implementer of silence deletion in the alignment in
s3align and a version of SAT in SphinxTrain.  Main designer of Sphinx
4.

Ricky Houghton - He is the author of the backbone of the perl scripts
for SCHMM training.  He also fixed a lot of memory leaking problems in
Sphinx 3.

Mike Seltzer - He is the author of the current version of wave2feat. 

Sam-Joo Doh - He is the author of a speaker adaptation package. The
current version of adaptation routine only used small portion of its
package.

Kevin Lenzo - Packaged Sphinx 2 and developed several versions of
Sphinx2 to make it has open-source quality.  He also started to put
the code into Sourceforge. 

Alan W Black - Packaged SphinxTrain's script and code. He makes it work
for speech synthesis.  

Evandro Gouvea - Package Sphinx3.0 and Sphinx 3.x, maintainer of
Sphinx X including the code, websites, machine queues and networks.

Developers from Sun, including Willie Walker, Paul Lamere and Philip
Kwok.  They first ported s3.3 to s3j, a java version of s3.3.  Later,
they proposed to design and implement Sphinx 4.  For Sphinx 4's
contributors, please consult Sphinx 4's documentation. 

Jahanzeb Sherwan - Implementer of phoneme lookahead algorithm in
Sphinx 3.x

Yitao Sun - Designer and Implementer of Sphinx 3.x live mode APIs. 

David Huggins-Daines - Implementer of unsupervised speaker adaptation
routine.

Scott Silliman - Tester of libutil library.  Help us to fence out a
lot of bugs.

Arthur Chan - The author of this document. :-) Before I left CMU,
let's make this description empty. :-)

Special Thanks

Carl Quillen - He submitted a patch for allowing three s3.0 tools to
s3.x.  This allows developers of Sphinx to incorporate other tools
very quickly to s3.x.  He also submitted a patch of mean
transformation.  

Joseph Donohue - He gave us a lot of invaluable advices on code
integrity.  We took most of it and was benefited a lot. 

Ziad Al Bawab - He implemented the current end-pointing routine. 

Rong Zhang - He implemented the C-version of confidence annotation
routines. 

Michael Pust, Jia Pu and Keith Herold - They give us invaluable
suggestions and advices on how to improve the code quality.  We used
their suggestions to adopt the use of doxygen. 

Thanks

Rich Stern - Experts in biaurnal speech processing and he
manages CMU's robust group.

Alex Rudnicky - The coordinator of Hephaestus.  Gave a lot of valuable
suggestions to Sphinx and various open source speech projects of CMU. 

Jack Mostow - The principal investigator of project LISTEN.  A lot of
features of Sphinx are motivated by project LISTEN. From example,
FSM's version of Sphinx 2.

Jim Baker - Interaction with him benefits a lot of young researchers
and develoeprs in CMU's speech group. 

Raj Reddy - For his vision in speech recognition and his good sense of
choosing the name "Sphinx". 

We would also like to thank to all users who have kindly feedbacked to
us and make Sphinx's better. 

To make this list complete, we may need to put all the names of CMU's
graduates and faculties here. 


Others we'd like to thank

Dr Takuji Nishimura and Dr Makoto Matsumoto. - Thanks to their effort
in open sourcing a Mersene Twister prime-based random generator, we
now have an extremely good portable random generator.