3 commits

Author	SHA1	Message	Date
Wim Taymans	8fe83e5304	audiomixer: optimize avx mixer some more ... Add avx mixer to test and benchmark Rework and unroll the avx mixer some more. The SSE one is 10 times faster than the C one, The AVX is 20 times faster. The SSE2 function is 5 times faster than the C one.	2022-07-10 23:13:24 +02:00
Wim Taymans	abd6806d71	tests: add sse and sse2 versions to tests	2022-07-09 20:11:46 +02:00
Wim Taymans	371b5a1836	audiomixer: rewrite the mixer functions ... Let the mixer functions accumulate the intermediate results into a larger size variable and then clamp to the final precission. This avoids distortions because of intermediate clamping. Although the access pattern of the reads are no longer sequential, the writes are sequential and we don't need to read intermediate values. Together with the avoided clamping this is probably faster overall. Add a unit test for the various cases.	2022-07-09 18:11:13 +02:00