Pd in ALSA?
ritsch at iem.mhsg.ac.at
Tue Jun 20 09:25:04 CEST 2000
> If you are using a current ALSA release things are pretty well broken. I
> gave up trying to track the changes until things settle down with the ALSA
> development (which will probably be anonther month at least). I had
> another imporoved version that did mmap mode and some other things but
> almost as soon as I got it working things became 'interesting' with the
> ALSA releases. To their credit, though, the ALSA developers are making
> some good changes it seems. The last ALSA I know it works with is 0.5.6 -
> if this doesn't work let me know.
Thats true, since I tried to use them there is each month a major
change which breaks the upwards compatibility. I think this will go on
until kernel 2.5.x when they want ALSA in the kernel, then they have
to stop (hopfully a ALSA 1.x), its also true inside the driver
development. I hope that at that time ALSA has reached the desired
quality for realtime performance since most changes nowadays are to
improve Multiprocessor and Realtime performance.
Low level drivers do not make any conversion anymore, so are straight
forward, and everything until resampling with different rates goes
into the library.
So maybe for good realtime performance we should use lowlevel because
we dont know what ALSA-lib will do with the the samples.
The question is should I make the native RME-drivers also for kernel
2.4.x (since its broken for that versions) or should we use ALSA else
There is a need for the 2.4.0 kernel for Athlon-Boards with VIA
Chipsets to get best performance also with 4xAGP.
More information about the Pd-list