home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.ai.neural-nets
- Path: sparky!uunet!think.com!rpi!zaphod.mps.ohio-state.edu!cis.ohio-state.edu!gecko.cis.ohio-state.edu!pardo
- From: pardo@gecko.cis.ohio-state.edu (Bryan Pardo)
- Subject: Seeking NN package for speech rec.
- Message-ID: <1992Nov10.212043.25993@cis.ohio-state.edu>
- Sender: pardo@cis.ohio-state.edu
- Organization: The Ohio State University Dept. of Computer and Info. Science
- Date: Tue, 10 Nov 1992 21:20:43 GMT
- Lines: 25
-
-
- Hi all,
-
- I am working on a project to test two different kinds of signal processors
- to see which would be better for a speech-recognition project we are working
- on. In order to have a (more or less) unbiased comparison we have determined
- that we would like to train a NN to differentiate between pairs of spoken
- letters (such as p vs d) to see which signal processor seems to give us the
- best results when searching for place of articulation information.
-
- Does anyone out there know of an existing package which could be usefull for
- such a project. It would be great if it had the following characteristics...
-
- 1) Works on a sun with x-11
- 2) ftp able
- 3) Low price (the cheaper the better)
- 4) able to recognize sequential patterns of input (ie give us a response on
- the basis input from a series of timeslices covering the length of the
- utterance of a letter)
-
- Many thanks for any help provided.
-
- --Bryan Pardo
- pardo@cis.ohio-state.edu
-
-