3 May
2007
3 May
'07
9:11 p.m.
Hi all,
What do you think of this patch? I tested this using a stac9205 HDA codec with all sample formats supported by softvol (s32, s16, s24, big and little endian).
The potential drawback I see is that softvol now has to use a intermediate long long multiply and add when computing amplified 32-bit samples. This wasn't an issue with the system I tested on, but possibly could be on slower machines. Also, are their any portability issues with using a long long datatype in alsa-lib? What is the list of embedded architectures that alsa-lib is targeted to build on? I see long long used in many other places in alsa-lib, so I'm hoping this isn't an issue.
Steve