Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!usc!apple!agate!shelby!csli!poser
From: poser@csli.Stanford.EDU (Bill Poser)
Newsgroups: comp.dsp
Subject: Re: A simple, practical sound board
Keywords: 29 Meg 8-bit speech
Message-ID: <16262@csli.Stanford.EDU>
Date: 8 Nov 90 02:00:55 GMT
References: <16002@netcom.UUCP> <16150@csli.Stanford.EDU> <16015@netcom.UUCP> <1990Nov7.064409.17295@engin.umich.edu>
Reply-To: poser@csli.stanford.edu (Bill Poser)
Organization: Center for the Study of Language and Information, Stanford U.
Lines: 19

In article <1990Nov7.064409.17295@engin.umich.edu> cerberus@caen.engin.umich.edu (R Eric Bennett) writes:
>
>I don't know what everybody is giving gross figures for but I've used an 8-bit
>codec sampler at 8Khz, and the results are great for speech. The man who asked
>the question is obviously not doing research.
>I was quite impressed, as somewhat of a layman, with NeXT's built in sampling.
>Mac's 8-bit, 22Khz sampling, sounds awful.
>Perhaps it has something to do with the speakers I'm listening to.
>Anyway, my point is that NeXT does 8-bit sampling at 8Khz, meaning 29 Meg/hour.

The estimate I made was 43MB/hour, which will give somewhat better
quality. Sampling at 8Khz gives you telephone quality speech, which is
not too bad but nonetheless clearly distorted. (Try listening to /s/ and/sh/.)
The central point is not whether he can shave 25% off the filesize
by reducing the sampling rate a bit, but the order of magnitude.
His original hope of getting an hour into a  megabyte was off by over
an order of magnitude, and that is the crucial thing to understand.

								Bill