Re: phoneme recognition and stimulus length (Jont Allen )


Subject: Re: phoneme recognition and stimulus length
From:    Jont Allen  <jba(at)RESEARCH.ATT.COM>
Date:    Mon, 19 Mar 2001 10:50:12 -0500

Dear Laszlo, I dont know if this is what you are getting at, but have you looked at th= e paper by Sadaoki Furui "On the role of spectral transitions for speech pe= rception." JASA, Oct. 1986, page 1016+ Jont T=F3th L=E1szl=F3 wrote: > Dear list, > > I wonder how human performance in phoneme recognition depends on the > stimulus length. I'm thinking of an experiment something like "name the > last phoneme you hear", and the length of the sound excerpt played woul= d > vary between, say, 10ms and a couple of seconds (with the end always fi= t > to the end of a phoneme). Since I'm interested in > the importancy of long-term adaptation, nonsense worlds should be used = to > exclude linguistic help. Does anyone know about such or similar > experiments? > > P.S.: My special interest is whether information earlier than the "magi= c" > 250ms is important, and how much (I would like to draw some conclusions > for automatic speech recognition). > > Thanks, > Laszlo Toth > Hungarian Academy of Sciences * "Heavy Metal is the > Research Group on Artificial Intelligence * QUINTessence of musi= c" > e-mail: tothl(at)inf.u-szeged.hu * > http://www.inf.u-szeged.hu/~tothl * /Anonymous/ -- Jont B. Allen AT&T Labs-Research, Shannon Laboratory, E161 180 Park Ave., Florham Park NJ, 07932-0971 973/360-8545voice, x7111fax, http://www.research.att.com/~jba


This message came from the mail archive
http://www.auditory.org/postings/2001/
maintained by:
DAn Ellis <dpwe@ee.columbia.edu>
Electrical Engineering Dept., Columbia University