eht@f.word.cs.cmu.edu (Eric Thayer) (02/09/90)
Yes, SPHINX does run on the NeXT with some caveats. 1) Development of new recognition tasks sometimes requires expert knowledge to perform optimally. 2) Running existing recognition tasks optimally requires some care as to mic choice and placement. 3) Documentation is sparse, but improving. We are working on these areas continually, but currently, the casual user can easily be disappointed after he/she gets to the point where he/she wants to build a new vocabulary and grammar. However, we have transferred the technology to a number of major companies that have speech recognition efforts and some have successfully built their own tasks. The tasks that NeXT demos were developed here at CMU. We were confident that they would perform reasonably well given that NeXT's SE's were aware of what mic's would work and after some practice with the systems. -- Eric H. Thayer School of Computer Science, Carnegie Mellon (412) 268-7679 5000 Forbes Ave, Pittsburgh, PA 15213
SLVQC@CUNYVM (Salvatore Saieva) (02/09/90)
Is SPHINX available for ftp anywhere? Sal.