From: pv AT cs DOT montana DOT edu (Paul Peavyhouse)
Newsgroups: comp.os.msdos.djgpp
Subject: How to write a network "voice phone" program?
Date: Mon, 24 Mar 1997 18:49:22 GMT
Organization: Montana State University
Lines: 29
Message-ID: <5h7b0e$slj@netra.montana.edu>
NNTP-Posting-Host: esus.cs.montana.edu
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp

        This question is not necessarily a djgpp question, but I'd like to see 
how djgpp can handle this idea.  I'm thinking more along the lines of 
Microsoft's Direct sound, but even as low-latency as they say their sound card 
routines are, I have my doubts as to how well it can be implemented even in 
DirectSound.
        The idea is to write a barebones, no bells and whistles network "voice 
phone" program.  The idea is MINIMAL bandwidth and MINIMAL features.  Voice 
quality need not be anything better than you get on a typical CB or Walkie 
Talkie.
        So, my question is, does anyone know of any good libraries already 
wriiten that allow line/mic input to a buffer (which I can then write the code 
to transmit that buffer to another computer on a TCP/IP network) and then the 
ability to echo that buffer to a lineout/speaker?  This is a fairly 
simple/common concept, so I thought I would check the global library before I 
go about writing it myself.  The trick though is that  don't know much about 
ANALOG input.  Reading/writing digital WAVs and MODs is one thing, but to 
process real-time analog input is over my head.  Anyone have any suggestions 
on the SIMPLEST way to do this?  Should I do this in DJGPP (w/ ASM), or ignore 
gnu and just use DirectSound?

                                                        PV


______________________________________________________________________________

 Paul Peavyhouse                          
 http://www.cs.montana.edu/~pv           
 email: pv AT cs DOT montana DOT edu                
______________________________________________________________________________