As far as I understand OSS, Alsa, etc. are just drivers to let the sound card work, ie they take care of making a device (/dev/dsp) for the soundcard and let the kernel use it. Then there are soundservers like aRTS and esd that take control over the soundcard, they lock /dev/dsp. For this to work, your programs must know that they have to use the soundserver instead of the soundcard. For example xmms can be compiled with an esd output plugin and mpg123 can also be compiled to use esd. For programs that don't have this option you can use esddsp <name of program>. esddsp takes care of letting the sound go to esd, which takes care of sending it to the sound-card. If you use aRTS, there exists artsdsp that does the same. The advantage is that you can mix sound streams to your soundcard, whereas playing directly to the soundcard from xmms locks the device. I think the arts project also focuses on using effects on the sound, like the software you get with a soundblaster live. another cool thing is that you can use a soundserver to let other computers use your soundcard. (at least, if your network connection between these 2 computers is fast enough). You can let esd and artsd use a tcpip-socket instead of a unix socket and let them accept connections from other hosts. With esddsp you can specify a server and port to use, with artsdsp I don't know how to do it. (maybe not implemented yet?) htbsc, (hehe)
From Ron Cordell to suse-linux-e@suse.com about [SLE] [OT] Sound confusion:
Hi everyone,
Can anyone point me to a _concise_ place to read about sound in Linux? Here is what I am trying to understand:
It seems that there are several different ways to support sound hardware in Linux; among them, OSS, Alsa, etc.
KDE 2.1 seems to want to use Arts. What I don't understand is how all these sound systems work together. Applications want to use other sound systems, for example XMMS can use OSS, or "default" sound system.
Is there some low level set of interfaces that all these sound systems are working with, or are they each trying to manipulate /dev/dsp? When should I use one over the other?
Just what does the Arts sound server do, anyway?
Thanks to all for sharing their understanding and knowledge!
-ronc
-- dieter