It depends on the application. NN are good for limited tasks with small vocabularies but I would say that, for the most demanding tasks (large vocabularies, continuous speech, etc.), HMM currently give the best results. Check Cambridge's system: http://htk.eng.cam.ac.uk/docs/cuhtk.shtml Aldebaro ========================== Aldebaro Klautau UCSD - Ph.D. student http://speech.ucsd.edu/aldebaro On Wed, 14 Aug 2002, mansoor vali wrote: > Hello my dears > I want to knowe state of the arts in continuous speech > recgnition are based on HMM methods or Neural Network > methods ? > thanks a lot > M. Vali

(You need to be a member of speech-recognition -- send a blank email to speech-recognition-subscribe@yahoogroups.com )