Peter Meerwald 3a3c4eb462 resampler: Improve s16<-->s32 conversion, use s16 work format if input or output is s16 ... Problem: s16 to s32 conversion is performed as s16->float->s32 (via work format float) for resamplers TRIVIAL, COPY, PEAKS. Precision and efficiency suffers: e.g. 0x9fff results in 0x9ffe4001 (instead of 0x9fff0000) and there are two sample format conversions instead of one conversion. Solution: If input or output format is s16, then choose the work format to be s16 as well. If remapping is to be performed, we could stick to work format float32ne for precision reseans. This is debateable. Signed-off-by: Peter Meerwald <p.meerwald@bct-electronic.com>		2013-02-01 10:10:30 +02:00
..
daemon	cpulimit: Explicitly ignore pa_read() return value.	2012-12-19 12:31:50 +02:00
modules	reserve: Fix leaking NameLost signals after release+acquire	2013-01-31 14:04:45 +02:00
pulse	pulse: Initialize pa_operation objects to all-zero.	2013-01-13 03:15:23 +02:00
pulsecore	resampler: Improve s16<-->s32 conversion, use s16 work format if input or output is s16	2013-02-01 10:10:30 +02:00
tests	tests: Fix sconv sample correctness in cpu-test	2013-02-01 10:06:03 +02:00
utils	pactl: Add a command for setting the default sink/source.	2013-02-01 07:16:20 +02:00
.gitignore	.gitignore: Add cpu-test.	2012-09-07 17:22:10 +03:00
depmod.py	remove remaining $	2008-06-18 22:38:32 +02:00
Makefile.am	role-ducking: Apply a ducking effect based on streams roles	2012-12-19 12:31:48 +02:00
map-file	pulse: Add pa_operation_set_state_callback() API	2013-01-12 17:55:07 +02:00