Subject: Re: Lossy data compression and Tinnitus From: Thomas G Brennan <g_brennantg(at)TITAN.SFASU.EDU> Date: Sat, 4 Jan 2003 04:22:05 -0600I missed the original post on this or would have commented sooner. Being a speech user for computers and a lot of other devices I'm fairly familiar with current speech systems as well as systems used over the past 20 or so years. I also am involved both with blind clients and a number of internet lists composed of blind computer users and have come across no user complaints concerning auditory problems with digital speech. If for the moment we set aside the probability that there is at least some auditory involvement with dyslexia, we should then consider which speech system is being used. Treditionally lpc was used for the production of speech because it took little memory compaired to digital speech. In recent years dedicated speech synthesizers have been replaced by the use of sound cards. Some of these systems use digital speech and some still rely on lpc. As sound card use has become more popular with drivers such as DEC32 access blind users are finding that they prefer to go back to dedicated speech synthesizers. In part this is because of the poor responsiveness of sound cards and in part it is because of poorer speech quality. I believe that the research group at the American Foundation for the Blind has begun looking at speech synthesis systems rather than only providing reports on screen access programs. Listening to compressed or synthesized speech is a bit tiring and it has been my experience that speech produced with a sound card is particularly tiring because of background sound associated with the speech. Although the speech produced by dedicated synthesizers may not be as clear or human sounding as that producable by sound cards in many cases, I find it easier to listen to for extended periods. It also often helps if you use a larger speaker than that provided by synthesizers. Plugging your sound card into a stereo may make a world of difference in the ease with which you can listen to speech even though mistakes made by the synthesizer's chip programming will become more apparent. Tom Tom Brennan, CCC-A/SLP, RHD web page http://titan.sfasu.edu/~g_brennantg/sonicpage.html