[Gllug] Linux Sound APIs

Chris Bell chrisbell at overview.demon.co.uk
Sat Apr 9 22:12:58 UTC 2005


On Sat 09 Apr, Peter Childs wrote:
> 

> 
> At the Low level we have OSS and ALSA and ALSA seams to be the prefered 
> way, However neither offer a standard api for playing and recording noise.
> 
   There is not so much a standard method of storing sounds, more a whole
load of different methods, with original digital recordings made to
different data depths at different sampling rates, with or without analogue
compression before digital conversion, which can either be stored as huge
amounts of raw data or compressed using a variety of methods.
   The human ear is extremely sensitive to minute amounts of waveform
distortion and differential frequency sensitive delays, yet relatively
insensitive to huge loudness variations, which makes the hardware design
extremely difficult. The data can be compressed following a logarithmic law,
or by describing the shape of the waveform, doing a Fourier transformation,
or a combination of different methods, but the result may depend on the time
and processing power available (does it have to be compressed in real
time?).
   Some methods will produce a better quality result, but with a huge price
tag. I know someone who got rid of his Hi-Fi kit which cost 50,000 UK pounds
to make room for his new kit which cost about 10 times as much because he
was not happy with the quality. At that price it is possible to hear the
difference in quality between two apparently identical connecting cables.
   Add to that the various patent problems associated with different
methods, agreed standards being ignored by certain near-monopoly
organisations, and you end up with the current mess.




-- 
Chris Bell

-- 
Gllug mailing list  -  Gllug at gllug.org.uk
http://lists.gllug.org.uk/mailman/listinfo/gllug




More information about the GLLUG mailing list