ietf-822
[Top] [All Lists]

Re: New content types

1993-02-11 06:20:12
On 10 Feb 93 22:16:00 GMT, Ed Hall <edhall(_at_)rand(_dot_)org> said:

        Ed> In article 
<9302102116(_dot_)AA12266(_at_)cbfsb(_dot_)cb(_dot_)att(_dot_)com>
        Ed> mark(_at_)cbfsb(_dot_)cb(_dot_)att(_dot_)com writes:

        Mark> . . . .
        Mark> By the way, what is the core Audio format?  I get the uneasy
        Mark> feeling it's "whatever Sun's /dev/audio understands" which
        Mark> makes me very nervous.

        Ed> Well, no, 8-bit u-law is actually a national standard for telephone-
        Ed> quality audio. (Europe uses a similar but not idential A-law 
encoding).
        Ed> The fact that Suns support this format isn't a coincidence--it is a
        Ed> result of Sun using a digital telephone interface chip for their 
audio
        Ed> port.  But the standard preceded /dev/audio, not the other way 
around.

Indeed.

        Ed> I think it's a good choice, irrespective of Sun's support for it.
        Ed> Better than 8-bit linear (with its limited dynamic range),
        Ed> thriftier than 16-bit linear, and not CPU-consumptive, like more
        Ed> highly compressed formats.

I can't say I agree here. Looking at recent history, CPU performance
increases MUCH faster than available bandwidth, especially among
connected LANs. And while ATM promises greater speed soon, it will be
a while before all the ethernet machines go away; ATM will initially
be deployed as a wide-area technology.

Most mail traffic is across local nets, at least in the Environments
I have seen. Since having a machine that can play back sound usually
means having a CPU on your desktop, it is only neighborly to trade
your personal CPU for reduced network (public) impact.

Don't get me wrong; U-LAW is a pretty good choice (for americans!) of
a standard, but suboptimal. A compressed type should have been the
standard  (4-bit, g.721  ADPCM?) and should be registered pretty soon.
Changing the base audio type, I think, would constitute a rather too
substantive change to the spec at this late date.

-Rens
  __   ___  ,    __  
 /__) /__  /| / (    |  J. Laurens Troost    Lehman Brothers Technical Services
/  \ (___ / |/ __)   |  *Opinions expressed herein are mine. Mine, I tell you!*
-------------------------------------------------------------------------------
INET: rens(_at_)shearson(_dot_)com        VOICE: (212) 464-3705        FAX: 
(212) 464-2040

<Prev in Thread] Current Thread [Next in Thread>
  • Re: New content types, Rens Troost <=