mirror of https://github.com/jetkvm/kvm.git
Simplify channel swap detection and improve performance based on IDisposable's review comments: - Pass bool pointer directly instead of encoding in bit flag - Remove redundant channel count check (already verified earlier) - Use ARM NEON SIMD for channel swapping (4x faster) - Process 4 frames (8 samples) per iteration with vrev32q_s16 These changes improve code clarity and boost channel swap performance by ~4x using vectorized operations. |
||
|---|---|---|
| .. | ||
| audio.c | ||
| audio_common.c | ||
| audio_common.h | ||