Newsgroups: comp.speech
Path: pavo.csi.cam.ac.uk!warwick!pipex!uunet!zaphod.mps.ohio-state.edu!saimiri.primate.wisc.edu!hp9000.csc.cuhk.hk!cucs5.cs.cuhk.hk!yu030
From: yu030@cs.cuhk.hk (Yu Wai Leung)
Subject: Re: Klatt SYnthesizer (How to work)
Message-ID: <1993Jan12.084959.7515@cucs5.cs.cuhk.hk>
Sender: news@cucs5.cs.cuhk.hk
Organization: Faculty of Engineering, The Chinese U. of Hong Kong
References: <1993Jan8.024840.12894@cucs5.cs.cuhk.hk> <gordon.726507433@spot.Colorado.EDU> <1993Jan11.013217.8139@cucs5.cs.cuhk.hk> <gordon.726766695@spot.Colorado.EDU>
Date: Tue, 12 Jan 1993 08:49:59 GMT
Lines: 42

gordon@spot.Colorado.EDU (GORDON ALLEN R) writes:

>Thanks,  

>The .au file was not written.  Examination of the code in klatt.c revealed that
>the code segment that created the header and wrote the data to the .au file 
>was in a #if FALSE ....  #endif   conditional.   Removal of this conditional,
>then allowed for the creation of the file, which is an audio data file 
>(according to file).  However,  playing the test file hello.par produces sound
>which is unintelligble.  It is identical to playing back the audio file 
>hello.au.  Furthermore, something that wasn't clear in my original note was thatwhile hello.xgr file could be displayed with xplot, the plot really didn't 
>look like speech.  Similarly for hello.dat.

>What am I missing here?  I'll check the short array iwave to see if I can find
>anything else.  However, as I read klatt.c, iwave is filtered through
>audio_s2u to produce plabuf.  The ascii files hello.xgr and hello.dat are
>produced from audio_u2s(plabuf).  Shouldn't this be the same as iwave, then?
>The file hello.au is essentially plabuf with an audio header.
>-- 
>Allen Gordon 		         *If the folly of but one of us was changed to*
>Research Associate         	 *intelligence, and divided amongst a thousand*
>gordon@tramp.colorado.edu   	 *toads, each would be more intelligent than  *
>				 *Aristotle				      *

Hello:

     Looking from "klatt.c", I've discovered that the marco "audio_s2u"
is defined as:

     #define audio_s2u(x) ((x >> 8) ^ (0x80))

     Obviously, this definition is WRONG! You can try to eliminate this
definition and use the "original" macro definition in the include file
<multimedia/ulaw2linear.h> instead.

Cheers!

Yu Wai-Leung
Dept of CS,
Chinese University of Hong Kong
e-mail: yu030@cs.cuhk.hk

