Newsgroups: comp.speech
Path: pavo.csi.cam.ac.uk!doc.ic.ac.uk!daresbury!mrccrc!news.dcs.warwick.ac.uk!warwick!uknet!pipex!howland.reston.ans.net!news.intercon.com!uhog.mit.edu!nntp.club.cc.cmu.edu!honeydew.srv.cs.cmu.edu!mkant
From: mkant+@cs.cmu.edu (Mark Kantrowitz)
Subject: Re: Network Audio - is this possible?
Message-ID: <CKo8Gt.MB1.3@cs.cmu.edu>
Keywords: network audio
Sender: news@cs.cmu.edu (Usenet News System)
Nntp-Posting-Host: glinda.oz.cs.cmu.edu
Organization: School of Computer Science, Carnegie Mellon
References: <1994Feb3.110321.43848@ucl.ac.uk>
Date: Thu, 3 Feb 1994 23:12:29 GMT
Lines: 27

In article <1994Feb3.110321.43848@ucl.ac.uk>,
Mr Andrew Simpson <uclyams@ucl.ac.uk> wrote:
>Is it possible to use the audio device of one machine from another
>machine on the same network?  I'm interested in playing out
>audio files stored on a Sun SPARC station on a variety of 
>different platforms including other SPARC stations, a 486 machine
>running Linux, and a DOS/Windows platform.   Is it possible to
>do this within the context of X-Windows - i.e. to somehow use the
>communication channel between Client and Server to carry the 
>audio signal?

Yes, it is possible. I did it for my software-only speech synthesizer.
The synthesizer is implemented in highly optimized common lisp, and
runs in real time on a HP Snake. I have a SPARC on my desk, and wanted
to play the audio out of that, so I wrote a socket-based interface.
From the Lisp end, it looks just like any other stream. A simple
server on the Sparc end takes the bits and writes them to /dev/audio.
Sound quality was indistinguishable from writing to file and catting
the file to /dev/audio, even when the synthesizer was run in
incremental unbuffered mode (e.g., lots of little writes, instead of a
block write).

--mark 

P.S. For the curious, the Common Lisp implementation was faster
than the equivalent optimized C code. Numeric support in CMU Common
Lisp is outstanding.
