Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!rice!sun-spots-request
From: soleil!gopstein@rutgers.edu (Rich Gopstein)
Newsgroups: comp.sys.sun
Subject: Re: Sparc Station Audio
Keywords: Miscellaneous
Message-ID: <4855@brazos.Rice.edu>
Date: 8 Feb 90 14:56:59 GMT
Sender: root@rice.edu
Organization: Sun-Spots
Lines: 9
Approved: Sun-Spots@rice.edu
X-Refs:  Original: v9n31
X-Sun-Spots-Digest: Volume 9, Issue 38, message 7

I believe the sample speed is fixed at 8000 samples/second in the
hardware, so you're not going to see that change.  The chip they used can
support u-LAW, a-LAW, and standard binary (probably at 8-bits-per-sample,
though).

There is NO WAY (in general) to convert back from u-law to binary without
losing information.  The binary-to-ulaw conversion process is a
many-to-one mapping.  This is very similar to a floating point roundoff
problem.