Path: utzoo!attcan!uunet!midway!ux1.cso.uiuc.edu!brutus.cs.uiuc.edu!zaphod.mps.ohio-state.edu!usc!randvax!news
From: edhall@rand.org (Ed Hall)
Newsgroups: comp.music
Subject: Re: Time change without pitch change, WHY?
Message-ID: <1990Jul16.202855.7640@rand.org>
Date: 16 Jul 90 20:28:55 GMT
References: <31559@cup.portal.com> <2650002@hpsad.HP.COM> <31762@cup.portal.com>
Sender: news@rand.org
Reply-To: edhall@rand.org (Ed Hall)
Organization: The RAND Corporation, Santa Monica, CA
Lines: 24

In article <31762@cup.portal.com> Alvin@cup.portal.com (Alvin Henry White) writes:
>The reason I want time change without pitch change is so that we can achieve
>the modern equivalent of inter-linear language translation on stereo audio
>channels.  Such as, you have a computer look up a clue word in a second
>language, of your choice, thesaurus.  Then you have the computer speech
>function determine how long it will take to synthesize the word in the
>original tongue.  It then determines how long it would take to synthesize the
>clue word in the second language. Next, it applies time change without pitch
>change to the second language speech synthesis. and Voile. You hear the 
>original song in one ear and the clue words, dubbed in the other ear, in time
>to the music.

How are you going to handle languages with a different word order,
like English and Japanese, or Spanish and German?  Word-for-word
translation would at least produce a large shift in meaning, if not
result in gibberish.  It just isn't that easy.  Even if it were, how
are you going to handle cases where the substitute-word has four times
as many (or fewer) syllables as the original?  A 4X speedup would
render it unintelligably fast.  A 4X slowdown would render many
diphthongs inaudible.  The most speedup or slowdown people can handle
(with practice) is a bit over 2X, even with constant pitch.

		-Ed Hall
		edhall@rand.org