From 7745346292c580019e4d163aba9a1d9647df1e15 Mon Sep 17 00:00:00 2001 From: Wim Taymans Date: Tue, 12 Jul 2022 10:34:13 +0200 Subject: [PATCH] audioconvert: add sse2 s16 dither functions --- spa/plugins/audioconvert/fmt-ops-sse2.c | 120 ++++++++++++++++++++++++ spa/plugins/audioconvert/fmt-ops.c | 6 ++ spa/plugins/audioconvert/fmt-ops.h | 2 + 3 files changed, 128 insertions(+) diff --git a/spa/plugins/audioconvert/fmt-ops-sse2.c b/spa/plugins/audioconvert/fmt-ops-sse2.c index 6f7464472..7c713daa9 100644 --- a/spa/plugins/audioconvert/fmt-ops-sse2.c +++ b/spa/plugins/audioconvert/fmt-ops-sse2.c @@ -1237,6 +1237,126 @@ conv_f32d_to_s16_sse2(struct convert *conv, void * SPA_RESTRICT dst[], const voi conv_f32d_to_s16_1s_sse2(conv, &d[i], &src[i], n_channels, n_samples); } +static void +conv_f32d_to_s16_1s_dither_sse2(struct convert *conv, void * SPA_RESTRICT dst, const void * SPA_RESTRICT src, + uint32_t n_channels, uint32_t n_samples) +{ + const float *s0 = src; + int16_t *d = dst; + float *dither = SPA_PTR_ALIGN(conv->dither, 16, float); + uint32_t n, unrolled; + __m128 in[2]; + __m128i out[2]; + __m128 int_scale = _mm_set1_ps(S16_SCALE); + __m128 int_max = _mm_set1_ps(S16_MAX); + __m128 int_min = _mm_set1_ps(S16_MIN); + + if (SPA_IS_ALIGNED(s0, 16)) + unrolled = n_samples & ~7; + else + unrolled = 0; + + for(n = 0; n < unrolled; n += 8) { + in[0] = _mm_mul_ps(_mm_load_ps(&s0[n]), int_scale); + in[1] = _mm_mul_ps(_mm_load_ps(&s0[n+4]), int_scale); + in[0] = _mm_add_ps(in[0], _mm_load_ps(&dither[n])); + in[1] = _mm_add_ps(in[1], _mm_load_ps(&dither[n+4])); + out[0] = _mm_cvttps_epi32(in[0]); + out[1] = _mm_cvttps_epi32(in[1]); + out[0] = _mm_packs_epi32(out[0], out[1]); + + d[0*n_channels] = _mm_extract_epi16(out[0], 0); + d[1*n_channels] = _mm_extract_epi16(out[0], 1); + d[2*n_channels] = _mm_extract_epi16(out[0], 2); + d[3*n_channels] = _mm_extract_epi16(out[0], 3); + d[4*n_channels] = _mm_extract_epi16(out[0], 4); + d[5*n_channels] = _mm_extract_epi16(out[0], 5); + d[6*n_channels] = _mm_extract_epi16(out[0], 6); + d[7*n_channels] = _mm_extract_epi16(out[0], 7); + d += 8*n_channels; + } + for(; n < n_samples; n++) { + in[0] = _mm_mul_ss(_mm_load_ss(&s0[n]), int_scale); + in[0] = _mm_add_ss(in[0], _mm_load_ss(&dither[n])); + in[0] = _MM_CLAMP_SS(in[0], int_min, int_max); + *d = _mm_cvttss_si32(in[0]); + d += n_channels; + } +} + +void +conv_f32d_to_s16_dither_sse2(struct convert *conv, void * SPA_RESTRICT dst[], const void * SPA_RESTRICT src[], + uint32_t n_samples) +{ + int16_t *d = dst[0]; + uint32_t i, k, chunk, n_channels = conv->n_channels; + + update_dither_sse2(conv, SPA_MIN(n_samples, conv->dither_size)); + + for(i = 0; i < n_channels; i++) { + const float *s = src[i]; + for(k = 0; k < n_samples; k += chunk) { + chunk = SPA_MIN(n_samples - k, conv->dither_size); + conv_f32d_to_s16_1s_dither_sse2(conv, &d[i + k*n_channels], &s[k], n_channels, chunk); + } + } +} + +static void +conv_f32_to_s16_1_dither_sse2(struct convert *conv, void * SPA_RESTRICT dst, const void * SPA_RESTRICT src, + uint32_t n_samples) +{ + const float *s = src; + int16_t *d = dst; + float *dither = SPA_PTR_ALIGN(conv->dither, 16, float); + uint32_t n, unrolled; + __m128 in[2]; + __m128i out[2]; + __m128 int_scale = _mm_set1_ps(S16_SCALE); + __m128 int_max = _mm_set1_ps(S16_MAX); + __m128 int_min = _mm_set1_ps(S16_MIN); + + if (SPA_IS_ALIGNED(s, 16)) + unrolled = n_samples & ~7; + else + unrolled = 0; + + for(n = 0; n < unrolled; n += 8) { + in[0] = _mm_mul_ps(_mm_load_ps(&s[n]), int_scale); + in[1] = _mm_mul_ps(_mm_load_ps(&s[n+4]), int_scale); + in[0] = _mm_add_ps(in[0], _mm_load_ps(&dither[n])); + in[1] = _mm_add_ps(in[1], _mm_load_ps(&dither[n+4])); + out[0] = _mm_cvttps_epi32(in[0]); + out[1] = _mm_cvttps_epi32(in[1]); + out[0] = _mm_packs_epi32(out[0], out[1]); + _mm_storeu_si128((__m128i*)(&d[n]), out[0]); + } + for(; n < n_samples; n++) { + in[0] = _mm_mul_ss(_mm_load_ss(&s[n]), int_scale); + in[0] = _mm_add_ss(in[0], _mm_load_ss(&dither[n])); + in[0] = _MM_CLAMP_SS(in[0], int_min, int_max); + d[n] = _mm_cvtss_si32(in[0]); + } +} + +void +conv_f32d_to_s16d_dither_sse2(struct convert *conv, void * SPA_RESTRICT dst[], const void * SPA_RESTRICT src[], + uint32_t n_samples) +{ + uint32_t i, k, chunk, n_channels = conv->n_channels; + + update_dither_sse2(conv, SPA_MIN(n_samples, conv->dither_size)); + + for(i = 0; i < n_channels; i++) { + const float *s = src[i]; + int16_t *d = dst[i]; + for(k = 0; k < n_samples; k += chunk) { + chunk = SPA_MIN(n_samples - k, conv->dither_size); + conv_f32_to_s16_1_dither_sse2(conv, &d[k], &s[k], chunk); + } + } +} + void conv_f32d_to_s16_2_sse2(struct convert *conv, void * SPA_RESTRICT dst[], const void * SPA_RESTRICT src[], uint32_t n_samples) diff --git a/spa/plugins/audioconvert/fmt-ops.c b/spa/plugins/audioconvert/fmt-ops.c index 634415b2d..7f28b5067 100644 --- a/spa/plugins/audioconvert/fmt-ops.c +++ b/spa/plugins/audioconvert/fmt-ops.c @@ -199,6 +199,9 @@ static struct conv_info conv_table[] = MAKE(F32, S16, 0, conv_f32_to_s16_c), MAKE(F32P, S16P, 0, conv_f32d_to_s16d_shaped_c, 0, CONV_SHAPE), +#if defined (HAVE_SSE2) + MAKE(F32P, S16P, 0, conv_f32d_to_s16d_dither_sse2, SPA_CPU_FLAG_SSE2, CONV_DITHER), +#endif MAKE(F32P, S16P, 0, conv_f32d_to_s16d_dither_c, 0, CONV_DITHER), #if defined (HAVE_SSE2) MAKE(F32P, S16P, 0, conv_f32d_to_s16d_sse2, SPA_CPU_FLAG_SSE2), @@ -208,6 +211,9 @@ static struct conv_info conv_table[] = MAKE(F32, S16P, 0, conv_f32_to_s16d_c), MAKE(F32P, S16, 0, conv_f32d_to_s16_shaped_c, 0, CONV_SHAPE), +#if defined (HAVE_SSE2) + MAKE(F32P, S16, 0, conv_f32d_to_s16_dither_sse2, SPA_CPU_FLAG_SSE2, CONV_DITHER), +#endif MAKE(F32P, S16, 0, conv_f32d_to_s16_dither_c, 0, CONV_DITHER), #if defined (HAVE_NEON) MAKE(F32P, S16, 0, conv_f32d_to_s16_neon, SPA_CPU_FLAG_NEON), diff --git a/spa/plugins/audioconvert/fmt-ops.h b/spa/plugins/audioconvert/fmt-ops.h index 5359ee388..c23dc9f07 100644 --- a/spa/plugins/audioconvert/fmt-ops.h +++ b/spa/plugins/audioconvert/fmt-ops.h @@ -407,7 +407,9 @@ DEFINE_FUNCTION(f32d_to_s32_dither, sse2); DEFINE_FUNCTION(f32_to_s16, sse2); DEFINE_FUNCTION(f32d_to_s16_2, sse2); DEFINE_FUNCTION(f32d_to_s16, sse2); +DEFINE_FUNCTION(f32d_to_s16_dither, sse2); DEFINE_FUNCTION(f32d_to_s16d, sse2); +DEFINE_FUNCTION(f32d_to_s16d_dither, sse2); DEFINE_FUNCTION(32_to_32d, sse2); DEFINE_FUNCTION(32s_to_32d, sse2); DEFINE_FUNCTION(32d_to_32, sse2);