lavr: add x86-optimized functions for mixing 2 to 1 s16p with float coeffs

libavresample/x86/audio_mix.asm

@@ -21,6 +21,7 @@
 
 %include "x86inc.asm"
 %include "x86util.asm"
+%include "util.asm"
 
 SECTION_TEXT
 
@@ -64,3 +65,47 @@ MIX_2_TO_1_FLTP_FLT
 INIT_YMM avx
 MIX_2_TO_1_FLTP_FLT
 %endif
+
+;-----------------------------------------------------------------------------
+; void ff_mix_2_to_1_s16p_flt(int16_t **src, float **matrix, int len,
+;                             int out_ch, int in_ch);
+;-----------------------------------------------------------------------------
+
+%macro MIX_2_TO_1_S16P_FLT 0
+cglobal mix_2_to_1_s16p_flt, 3,4,6, src, matrix, len, src1
+    mov       src1q, [srcq+gprsize]
+    mov        srcq, [srcq]
+    sub       src1q, srcq
+    mov     matrixq, [matrixq  ]
+    VBROADCASTSS m4, [matrixq  ]
+    VBROADCASTSS m5, [matrixq+4]
+    ALIGN 16
+.loop:
+    mova         m0, [srcq      ]
+    mova         m2, [srcq+src1q]
+    S16_TO_S32_SX 0, 1
+    S16_TO_S32_SX 2, 3
+    cvtdq2ps     m0, m0
+    cvtdq2ps     m1, m1
+    cvtdq2ps     m2, m2
+    cvtdq2ps     m3, m3
+    mulps        m0, m4
+    mulps        m1, m4
+    mulps        m2, m5
+    mulps        m3, m5
+    addps        m0, m2
+    addps        m1, m3
+    cvtps2dq     m0, m0
+    cvtps2dq     m1, m1
+    packssdw     m0, m1
+    mova     [srcq], m0
+    add        srcq, mmsize
+    sub        lend, mmsize/2
+    jg .loop
+    REP_RET
+%endmacro
+
+INIT_XMM sse2
+MIX_2_TO_1_S16P_FLT
+INIT_XMM sse4
+MIX_2_TO_1_S16P_FLT

libavresample/x86/audio_mix_init.c

@@ -27,6 +27,11 @@ extern void ff_mix_2_to_1_fltp_flt_sse(float **src, float **matrix, int len,
 extern void ff_mix_2_to_1_fltp_flt_avx(float **src, float **matrix, int len,
                                        int out_ch, int in_ch);
 
+extern void ff_mix_2_to_1_s16p_flt_sse2(int16_t **src, float **matrix, int len,
+                                        int out_ch, int in_ch);
+extern void ff_mix_2_to_1_s16p_flt_sse4(int16_t **src, float **matrix, int len,
+                                        int out_ch, int in_ch);
+
 av_cold void ff_audio_mix_init_x86(AudioMix *am)
 {
 #if HAVE_YASM
@@ -36,6 +41,14 @@ av_cold void ff_audio_mix_init_x86(AudioMix *am)
         ff_audio_mix_set_func(am, AV_SAMPLE_FMT_FLTP, AV_MIX_COEFF_TYPE_FLT,
                               2, 1, 16, 8, "SSE", ff_mix_2_to_1_fltp_flt_sse);
     }
+    if (mm_flags & AV_CPU_FLAG_SSE2 && HAVE_SSE) {
+        ff_audio_mix_set_func(am, AV_SAMPLE_FMT_S16P, AV_MIX_COEFF_TYPE_FLT,
+                              2, 1, 16, 8, "SSE2", ff_mix_2_to_1_s16p_flt_sse2);
+    }
+    if (mm_flags & AV_CPU_FLAG_SSE4 && HAVE_SSE) {
+        ff_audio_mix_set_func(am, AV_SAMPLE_FMT_S16P, AV_MIX_COEFF_TYPE_FLT,
+                              2, 1, 16, 8, "SSE4", ff_mix_2_to_1_s16p_flt_sse4);
+    }
     if (mm_flags & AV_CPU_FLAG_AVX && HAVE_AVX) {
         ff_audio_mix_set_func(am, AV_SAMPLE_FMT_FLTP, AV_MIX_COEFF_TYPE_FLT,
                               2, 1, 32, 16, "AVX", ff_mix_2_to_1_fltp_flt_avx);

libavresample/x86/util.asm

@@ -0,0 +1,34 @@
+;******************************************************************************
+;* x86 utility macros for libavresample
+;* Copyright (c) 2012 Justin Ruggles <justin.ruggles@gmail.com>
+;*
+;* This file is part of Libav.
+;*
+;* Libav is free software; you can redistribute it and/or
+;* modify it under the terms of the GNU Lesser General Public
+;* License as published by the Free Software Foundation; either
+;* version 2.1 of the License, or (at your option) any later version.
+;*
+;* Libav is distributed in the hope that it will be useful,
+;* but WITHOUT ANY WARRANTY; without even the implied warranty of
+;* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+;* Lesser General Public License for more details.
+;*
+;* You should have received a copy of the GNU Lesser General Public
+;* License along with Libav; if not, write to the Free Software
+;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+;******************************************************************************
+
+%macro S16_TO_S32_SX 2 ; src/low dst, high dst
+%if cpuflag(sse4)
+    pmovsxwd     m%2, m%1
+    psrldq       m%1, 8
+    pmovsxwd     m%1, m%1
+    SWAP %1, %2
+%else
+    punpckhwd    m%2, m%1
+    punpcklwd    m%1, m%1
+    psrad        m%2, 16
+    psrad        m%1, 16
+%endif
+%endmacro