From LOJBAN@CUVMB.CC.COLUMBIA.EDU Sat Mar 6 22:58:49 2010 Return-Path: Delivered-To: veion@XIRON.PC.HELSINKI.FI Received: (qmail 13577 invoked from network); 28 Dec 1996 23:58:05 -0000 Received: from segate.sunet.se (192.36.125.6) by xiron.pc.helsinki.fi with SMTP; 28 Dec 1996 23:58:05 -0000 Received: from segate.sunet.se by SEGATE.SUNET.SE (LSMTP for OpenVMS v1.1a) with SMTP id <5.CD2F890A@SEGATE.SUNET.SE>; Sun, 29 Dec 1996 0:58:00 +0100 Date: Sat, 28 Dec 1996 23:53:31 GMT Reply-To: Graeme Dunbar Sender: Lojban list From: Graeme Dunbar Organization: The Robert Gordon University Subject: Voice output on a PC X-To: lojban@cuvmb.cc.columbia.edu To: Veijo Vilva Content-Length: 1537 Lines: 32 Message-ID: Santa kindly left a sound card and I have been playing with some shareware versions of text to speech software (eg PC Wholeware clip and talk). Unfortunately they are all too clever and are designed to cope with the peculiarities of english in the conversion. I understand that lojban pronunciation is a lot more straight forward and regular. Does anyone know of software to drive a "Soundblaster" compatable card that would allow all the lojban letter/sounds to be mapped? I think I am out of my (linguistic) depth here but will there be a problem because such software is designed around english language phonemes which do not match the lojban ones? Do I need software that deals with more fundamental speech components (allophones?) to create a lojban phoneme set? Is each lojban letter a unique phoneme (what about dipthongs?) or do the letter sounds depend on context? Has anyone tried this sort of thing? Best wishes for '97! Regards, Graeme +-------------------------------------------------------------------+ | Graeme Dunbar | | School of Electronic and Electrical Engineering | | The Robert Gordon University | | Schoolhill Tel. +44 1224 262415 | | Aberdeen AB10 1FR Fax +44 1224 262444 | | Scotland U.K. email g.r.a.dunbar@rgu.ac.uk | +-------------------------------------------------------------------+