Merge commit 'a1bcc76e6036e78f25cbb7323c145056cfca9d93'

* commit 'a1bcc76e6036e78f25cbb7323c145056cfca9d93': (21 commits)
  cmdutils: fix a memleak when specifying an option twice.
  x86: mpegvideo: more sensible names for optimization file and init function
  x86: mpegvideoenc: Split optimizations off into a separate file
  dnxhdenc: x86: more sensible names for optimization file and init function
  svq1/svq3: Move common code out of SVQ1 decoder-specific file
  dirac: add Comments and references to the standard
  lavr: x86: optimized 6-channel flt to fltp conversion
  lavr: x86: optimized 2-channel flt to fltp conversion
  lavr: x86: optimized 6-channel flt to s16p conversion
  lavr: x86: optimized 2-channel flt to s16p conversion
  lavr: x86: optimized 6-channel s16 to fltp conversion
  lavr: x86: optimized 2-channel s16 to fltp conversion
  lavr: x86: optimized 6-channel s16 to s16p conversion
  lavr: x86: optimized 2-channel s16 to s16p conversion
  lavr: x86: optimized 2-channel fltp to flt conversion
  lavr: x86: optimized 6-channel fltp to s16 conversion
  lavr: x86: optimized 2-channel fltp to s16 conversion
  lavr: x86: optimized 6-channel s16p to flt conversion
  lavr: x86: optimized 2-channel s16p to flt conversion
  lavr: x86: optimized 6-channel s16p to s16 conversion
  ...

Conflicts:
	libavcodec/dirac.c
	libavcodec/mpegvideo.h
	libavcodec/x86/Makefile

Merged-by: Michael Niedermayer <michaelni@gmx.at>

19 files changed, 1397 insertions, 180 deletions

diff --git a/cmdutils.c b/cmdutils.c
index ed856301d7..f6f258a9d9 100644
--- a/cmdutils.c
+++ b/cmdutils.c
@@ -301,6 +301,7 @@ int parse_option(void *optctx, const char *opt, const char *arg,
     if (po->flags & OPT_STRING) {
         char *str;
         str = av_strdup(arg);
+//         av_freep(dst);
         *(char **)dst = str;
     } else if (po->flags & OPT_BOOL) {
         *(int *)dst = bool_val;
diff --git a/libavcodec/Makefile b/libavcodec/Makefile
index 450086d29a..c64175d66f 100644
--- a/libavcodec/Makefile
+++ b/libavcodec/Makefile
@@ -405,14 +405,13 @@ OBJS-$(CONFIG_SUBRIP_ENCODER)          += srtenc.o ass_split.o
 OBJS-$(CONFIG_SUBVIEWER_DECODER)       += subviewerdec.o ass.o
 OBJS-$(CONFIG_SUNRAST_DECODER)         += sunrast.o
 OBJS-$(CONFIG_SUNRAST_ENCODER)         += sunrastenc.o
-OBJS-$(CONFIG_SVQ1_DECODER)            += svq1dec.o svq1.o h263.o
+OBJS-$(CONFIG_SVQ1_DECODER)            += svq1dec.o svq1.o svq13.o h263.o
 OBJS-$(CONFIG_SVQ1_ENCODER)            += svq1enc.o svq1.o    \
                                           h263.o ituh263enc.o
-OBJS-$(CONFIG_SVQ3_DECODER)            += h264.o svq3.o                       \
+OBJS-$(CONFIG_SVQ3_DECODER)            += svq3.o svq13.o h263.o h264.o        \
                                           h264_loopfilter.o h264_direct.o     \
                                           h264_sei.o h264_ps.o h264_refs.o    \
-                                          h264_cavlc.o h264_cabac.o cabac.o   \
-                                          svq1dec.o svq1.o h263.o
+                                          h264_cavlc.o h264_cabac.o cabac.o
 OBJS-$(CONFIG_TARGA_DECODER)           += targa.o
 OBJS-$(CONFIG_TARGA_ENCODER)           += targaenc.o rle.o
 OBJS-$(CONFIG_THEORA_DECODER)          += xiph.o
diff --git a/libavcodec/dirac.c b/libavcodec/dirac.c
index 3aa65e3c4d..bab3980382 100644
--- a/libavcodec/dirac.c
+++ b/libavcodec/dirac.c
@@ -57,10 +57,7 @@ static const dirac_source_params dirac_source_parameters_defaults[] = {
     { 7680, 4320, 1, 0, 1, 6,  1, 3840, 2160, 0, 0, 3, 3 },
 };
 
-/**
- * Dirac Specification ->
- * Table 10.4 - Available preset pixel aspect ratio values
- */
+/* [DIRAC_STD] Table 10.4 - Available preset pixel aspect ratio values */
 static const AVRational dirac_preset_aspect_ratios[] = {
     {1, 1},
     {10, 11},
@@ -70,19 +67,16 @@ static const AVRational dirac_preset_aspect_ratios[] = {
     {4, 3},
 };
 
-/**
- * Dirac Specification ->
- * Values 9,10 of 10.3.5 Frame Rate. Table 10.3 Available preset frame rate values
+/* [DIRAC_STD] Values 9,10 of 10.3.5 Frame Rate.
+ * Table 10.3 Available preset frame rate values
  */
 static const AVRational dirac_frame_rate[] = {
     {15000, 1001},
     {25, 2},
 };
 
-/**
- * Dirac Specification ->
- * This should be equivalent to Table 10.5 Available signal range presets
- */
+/* [DIRAC_STD] This should be equivalent to Table 10.5 Available signal
+ * range presets */
 static const struct {
     uint8_t             bitdepth;
     enum AVColorRange   color_range;
@@ -111,19 +105,14 @@ static const struct {
     { AVCOL_PRI_BT709,     AVCOL_SPC_BT709,   AVCOL_TRC_UNSPECIFIED /* DCinema */ },
 };
 
-/**
- * Dirac Specification ->
- * Table 10.2 Supported chroma sampling formats + Luma Offset
- */
+/* [DIRAC_STD] Table 10.2 Supported chroma sampling formats + luma Offset */
 static const enum PixelFormat dirac_pix_fmt[2][3] = {
     { PIX_FMT_YUV444P,  PIX_FMT_YUV422P,  PIX_FMT_YUV420P  },
     { PIX_FMT_YUVJ444P, PIX_FMT_YUVJ422P, PIX_FMT_YUVJ420P },
 };
 
-/**
- * Dirac Specification ->
- * 10.3 Parse Source Parameters. source_parameters(base_video_format)
- */
+/* [DIRAC_STD] 10.3 Parse Source Parameters.
+ * source_parameters(base_video_format) */
 static int parse_source_parameters(AVCodecContext *avctx, GetBitContext *gb,
                                    dirac_source_params *source)
 {
@@ -132,15 +121,18 @@ static int parse_source_parameters(AVCodecContext *avctx, GetBitContext *gb,
     int idx;
 
     /* [DIRAC_STD] 10.3.2 Frame size. frame_size(video_params) */
-    if (get_bits1(gb)) {                         /* [DIRAC_STD] custom_dimensions_flag */
-        source->width  = svq3_get_ue_golomb(gb); /* [DIRAC_STD] FRAME_WIDTH            */
-        source->height = svq3_get_ue_golomb(gb); /* [DIRAC_STD] FRAME_HEIGHT           */
+    /* [DIRAC_STD] custom_dimensions_flag */
+    if (get_bits1(gb)) {
+        source->width  = svq3_get_ue_golomb(gb); /* [DIRAC_STD] FRAME_WIDTH  */
+        source->height = svq3_get_ue_golomb(gb); /* [DIRAC_STD] FRAME_HEIGHT */
     }
 
     /* [DIRAC_STD] 10.3.3 Chroma Sampling Format.
-       chroma_sampling_format(video_params) */
-    if (get_bits1(gb))                                  /* [DIRAC_STD] custom_chroma_format_flag */
-        source->chroma_format = svq3_get_ue_golomb(gb); /*[DIRAC_STD] CHROMA_FORMAT_INDEX        */
+     *  chroma_sampling_format(video_params) */
+    /* [DIRAC_STD] custom_chroma_format_flag */
+    if (get_bits1(gb))
+        /* [DIRAC_STD] CHROMA_FORMAT_INDEX */
+        source->chroma_format = svq3_get_ue_golomb(gb);
     if (source->chroma_format > 2U) {
         av_log(avctx, AV_LOG_ERROR, "Unknown chroma format %d\n",
                source->chroma_format);
@@ -148,8 +140,10 @@ static int parse_source_parameters(AVCodecContext *avctx, GetBitContext *gb,
     }
 
     /* [DIRAC_STD] 10.3.4 Scan Format. scan_format(video_params) */
-    if (get_bits1(gb))                               /* [DIRAC_STD] custom_scan_format_flag */
-        source->interlaced = svq3_get_ue_golomb(gb); /* [DIRAC_STD] SOURCE_SAMPLING         */
+    /* [DIRAC_STD] custom_scan_format_flag */
+    if (get_bits1(gb))
+        /* [DIRAC_STD] SOURCE_SAMPLING */
+        source->interlaced = svq3_get_ue_golomb(gb);
     if (source->interlaced > 1U)
         return AVERROR_INVALIDDATA;
 
@@ -160,23 +154,29 @@ static int parse_source_parameters(AVCodecContext *avctx, GetBitContext *gb,
         if (source->frame_rate_index > 10U)
             return AVERROR_INVALIDDATA;
 
-        if (!source->frame_rate_index){
-            frame_rate.num = svq3_get_ue_golomb(gb); /* [DIRAC_STD] FRAME_RATE_NUMER */
-            frame_rate.den = svq3_get_ue_golomb(gb); /* [DIRAC_STD] FRAME_RATE_DENOM */
+        if (!source->frame_rate_index) {
+            /* [DIRAC_STD] FRAME_RATE_NUMER */
+            frame_rate.num = svq3_get_ue_golomb(gb);
+            /* [DIRAC_STD] FRAME_RATE_DENOM */
+            frame_rate.den = svq3_get_ue_golomb(gb);
         }
     }
-    if (source->frame_rate_index > 0) { /* [DIRAC_STD] preset_frame_rate(video_params,index) */
+    /* [DIRAC_STD] preset_frame_rate(video_params, index) */
+    if (source->frame_rate_index > 0) {
         if (source->frame_rate_index <= 8)
             frame_rate = avpriv_frame_rate_tab[source->frame_rate_index];  /* [DIRAC_STD] Table 10.3 values 1-8  */
         else
-            frame_rate = dirac_frame_rate[source->frame_rate_index-9]; /* [DIRAC_STD] Table 10.3 values 9-10 */
+             /* [DIRAC_STD] Table 10.3 values 9-10 */
+            frame_rate = dirac_frame_rate[source->frame_rate_index-9];
     }
     av_reduce(&avctx->time_base.num, &avctx->time_base.den,
               frame_rate.den, frame_rate.num, 1<<30);
 
-    /* [DIRAC_STD] 10.3.6 Pixel Aspect Ratio. pixel_aspect_ratio(video_params) */
+    /* [DIRAC_STD] 10.3.6 Pixel Aspect Ratio.
+     * pixel_aspect_ratio(video_params) */
     if (get_bits1(gb)) { /* [DIRAC_STD] custom_pixel_aspect_ratio_flag */
-        source->aspect_ratio_index = svq3_get_ue_golomb(gb); /* [DIRAC_STD] index */
+         /* [DIRAC_STD] index */
+        source->aspect_ratio_index = svq3_get_ue_golomb(gb);
 
         if (source->aspect_ratio_index > 6U)
             return AVERROR_INVALIDDATA;
@@ -186,22 +186,30 @@ static int parse_source_parameters(AVCodecContext *avctx, GetBitContext *gb,
             avctx->sample_aspect_ratio.den = svq3_get_ue_golomb(gb);
         }
     }
-    if (source->aspect_ratio_index > 0) /* [DIRAC_STD] Take value from Table 10.4 Available preset pixel aspect ratio values */
+    /* [DIRAC_STD] Take value from Table 10.4 Available preset pixel
+     *  aspect ratio values */
+    if (source->aspect_ratio_index > 0)
         avctx->sample_aspect_ratio =
             dirac_preset_aspect_ratios[source->aspect_ratio_index-1];
 
     /* [DIRAC_STD] 10.3.7 Clean area. clean_area(video_params) */
     if (get_bits1(gb)) { /* [DIRAC_STD] custom_clean_area_flag */
-        source->clean_width        = svq3_get_ue_golomb(gb); /* [DIRAC_STD] CLEAN_WIDTH        */
-        source->clean_height       = svq3_get_ue_golomb(gb); /* [DIRAC_STD] CLEAN_HEIGHT       */
-        source->clean_left_offset  = svq3_get_ue_golomb(gb); /* [DIRAC_STD] CLEAN_LEFT_OFFSET  */
-        source->clean_right_offset = svq3_get_ue_golomb(gb); /* [DIRAC_STD] CLEAN_RIGHT_OFFSET */
+        /* [DIRAC_STD] CLEAN_WIDTH */
+        source->clean_width        = svq3_get_ue_golomb(gb);
+         /* [DIRAC_STD] CLEAN_HEIGHT */
+        source->clean_height       = svq3_get_ue_golomb(gb);
+         /* [DIRAC_STD] CLEAN_LEFT_OFFSET */
+        source->clean_left_offset  = svq3_get_ue_golomb(gb);
+        /* [DIRAC_STD] CLEAN_RIGHT_OFFSET */
+        source->clean_right_offset = svq3_get_ue_golomb(gb);
     }
 
-    /*[DIRAC_STD] 10.3.8 Signal range. signal_range(video_params)
-      WARNING: Some adaptation seemed to be done using the AVCOL_RANGE_MPEG/JPEG values */
-    if (get_bits1(gb)) {                                    /*[DIRAC_STD] custom_signal_range_flag */
-        source->pixel_range_index = svq3_get_ue_golomb(gb); /*[DIRAC_STD] index */
+    /* [DIRAC_STD] 10.3.8 Signal range. signal_range(video_params)
+     * WARNING: Some adaptation seems to be done using the
+     * AVCOL_RANGE_MPEG/JPEG values */
+    if (get_bits1(gb)) { /* [DIRAC_STD] custom_signal_range_flag */
+        /* [DIRAC_STD] index */
+        source->pixel_range_index = svq3_get_ue_golomb(gb);
 
         if (source->pixel_range_index > 4U)
             return AVERROR_INVALIDDATA;
@@ -210,13 +218,14 @@ static int parse_source_parameters(AVCodecContext *avctx, GetBitContext *gb,
         if (!source->pixel_range_index) {
             luma_offset = svq3_get_ue_golomb(gb);
             luma_depth  = av_log2(svq3_get_ue_golomb(gb))+1;
-            svq3_get_ue_golomb(gb); /* chroma offset @Jordi: Why are these two ignored? */
+            svq3_get_ue_golomb(gb); /* chroma offset    */
             svq3_get_ue_golomb(gb); /* chroma excursion */
-
             avctx->color_range = luma_offset ? AVCOL_RANGE_MPEG : AVCOL_RANGE_JPEG;
         }
     }
-    if (source->pixel_range_index > 0) { /*[DIRAC_STD] Take values from Table 10.5 Available signal range presets */
+    /* [DIRAC_STD] Table 10.5
+     * Available signal range presets <--> pixel_range_presets */
+    if (source->pixel_range_index > 0) {
         idx                = source->pixel_range_index-1;
         luma_depth         = pixel_range_presets[idx].bitdepth;
         avctx->color_range = pixel_range_presets[idx].color_range;
@@ -229,7 +238,8 @@ static int parse_source_parameters(AVCodecContext *avctx, GetBitContext *gb,
 
     /* [DIRAC_STD] 10.3.9 Colour specification. colour_spec(video_params) */
     if (get_bits1(gb)) { /* [DIRAC_STD] custom_colour_spec_flag */
-        idx = source->color_spec_index = svq3_get_ue_golomb(gb); /* [DIRAC_STD] index */
+         /* [DIRAC_STD] index */
+        idx = source->color_spec_index = svq3_get_ue_golomb(gb);
 
         if (source->color_spec_index > 4U)
             return AVERROR_INVALIDDATA;
@@ -239,13 +249,13 @@ static int parse_source_parameters(AVCodecContext *avctx, GetBitContext *gb,
         avctx->color_trc       = dirac_color_presets[idx].color_trc;
 
         if (!source->color_spec_index) {
-            /* [DIRAC_STD] 10.3.9.1 Color primaries */
+            /* [DIRAC_STD] 10.3.9.1 Colour primaries */
             if (get_bits1(gb)) {
                 idx = svq3_get_ue_golomb(gb);
                 if (idx < 3U)
                     avctx->color_primaries = dirac_primaries[idx];
             }
-            /* [DIRAC_STD] 10.3.9.2 Color matrix */
+            /* [DIRAC_STD] 10.3.9.2 Colour matrix */
             if (get_bits1(gb)) {
                 idx = svq3_get_ue_golomb(gb);
                 if (!idx)
@@ -267,10 +277,7 @@ static int parse_source_parameters(AVCodecContext *avctx, GetBitContext *gb,
     return 0;
 }
 
-/**
- * Dirac Specification ->
- *  10. Sequence Header. sequence_header()
- */
+/* [DIRAC_STD] 10. Sequence Header. sequence_header() */
 int avpriv_dirac_parse_sequence_header(AVCodecContext *avctx, GetBitContext *gb,
                                    dirac_source_params *source)
 {
@@ -284,7 +291,7 @@ int avpriv_dirac_parse_sequence_header(AVCodecContext *avctx, GetBitContext *gb,
     avctx->profile = svq3_get_ue_golomb(gb);
     avctx->level   = svq3_get_ue_golomb(gb);
     /* [DIRAC_SPEC] sequence_header() -> base_video_format as defined in
-       10.2 Base Video Format, table 10.1 Dirac predefined video formats */
+     * 10.2 Base Video Format, table 10.1 Dirac predefined video formats */
     video_format   = svq3_get_ue_golomb(gb);
 
     if (version_major < 2)
@@ -298,7 +305,8 @@ int avpriv_dirac_parse_sequence_header(AVCodecContext *avctx, GetBitContext *gb,
     /* Fill in defaults for the source parameters. */
     *source = dirac_source_parameters_defaults[video_format];
 
-    // Override the defaults.
+    /* [DIRAC_STD] 10.3 Source Parameters
+     * Override the defaults. */
     if (ret = parse_source_parameters(avctx, gb, source))
         return ret;
 
@@ -307,8 +315,8 @@ int avpriv_dirac_parse_sequence_header(AVCodecContext *avctx, GetBitContext *gb,
 
     avcodec_set_dimensions(avctx, source->width, source->height);
 
-    /*[DIRAC_STD] picture_coding_mode shall be 0 for fields and 1 for frames
-      currently only used to signal field coding */
+    /* [DIRAC_STD] picture_coding_mode shall be 0 for fields and 1 for frames
+     * currently only used to signal field coding */
     picture_coding_mode = svq3_get_ue_golomb(gb);
     if (picture_coding_mode != 0) {
         av_log(avctx, AV_LOG_ERROR, "Unsupported picture coding mode %d",
diff --git a/libavcodec/dnxhdenc.c b/libavcodec/dnxhdenc.c
index 97013f6a94..24767fd1a0 100644
--- a/libavcodec/dnxhdenc.c
+++ b/libavcodec/dnxhdenc.c
@@ -294,9 +294,8 @@ static int dnxhd_encode_init(AVCodecContext *avctx)
        ctx->block_width_l2 = 3;
     }
 
-#if HAVE_MMX
-    ff_dnxhd_init_mmx(ctx);
-#endif
+    if (ARCH_X86)
+        ff_dnxhdenc_init_x86(ctx);
 
     ctx->m.mb_height = (avctx->height + 15) / 16;
     ctx->m.mb_width  = (avctx->width  + 15) / 16;
diff --git a/libavcodec/dnxhdenc.h b/libavcodec/dnxhdenc.h
index 640bbd3995..e57047d5dc 100644
--- a/libavcodec/dnxhdenc.h
+++ b/libavcodec/dnxhdenc.h
@@ -93,6 +93,6 @@ typedef struct DNXHDEncContext {
     void (*get_pixels_8x4_sym)(DCTELEM * /*align 16*/, const uint8_t *, int);
 } DNXHDEncContext;
 
-void ff_dnxhd_init_mmx(DNXHDEncContext *ctx);
+void ff_dnxhdenc_init_x86(DNXHDEncContext *ctx);
 
 #endif /* AVCODEC_DNXHDENC_H */
diff --git a/libavcodec/mpegvideo.c b/libavcodec/mpegvideo.c
index fc05b5f1d1..77108cf834 100644
--- a/libavcodec/mpegvideo.c
+++ b/libavcodec/mpegvideo.c
@@ -185,8 +185,8 @@ av_cold int ff_dct_common_init(MpegEncContext *s)
         s->dct_unquantize_mpeg2_intra = dct_unquantize_mpeg2_intra_bitexact;
     s->dct_unquantize_mpeg2_inter = dct_unquantize_mpeg2_inter_c;
 
-#if HAVE_MMX
-    ff_MPV_common_init_mmx(s);
+#if ARCH_X86
+    ff_MPV_common_init_x86(s);
 #elif ARCH_ALPHA
     ff_MPV_common_init_axp(s);
 #elif HAVE_MMI
diff --git a/libavcodec/mpegvideo.h b/libavcodec/mpegvideo.h
index 3c6f13a2cd..dd1079b0e7 100644
--- a/libavcodec/mpegvideo.h
+++ b/libavcodec/mpegvideo.h
@@ -765,7 +765,8 @@ int ff_MPV_encode_init(AVCodecContext *avctx);
 int ff_MPV_encode_end(AVCodecContext *avctx);
 int ff_MPV_encode_picture(AVCodecContext *avctx, AVPacket *pkt,
                           AVFrame *frame, int *got_packet);
-void ff_MPV_common_init_mmx(MpegEncContext *s);
+void ff_MPV_encode_init_x86(MpegEncContext *s);
+void ff_MPV_common_init_x86(MpegEncContext *s);
 void ff_MPV_common_init_axp(MpegEncContext *s);
 void ff_MPV_common_init_mmi(MpegEncContext *s);
 void ff_MPV_common_init_arm(MpegEncContext *s);
diff --git a/libavcodec/mpegvideo_enc.c b/libavcodec/mpegvideo_enc.c
index 05ae1b6b2d..6ec860984a 100644
--- a/libavcodec/mpegvideo_enc.c
+++ b/libavcodec/mpegvideo_enc.c
@@ -832,6 +832,9 @@ av_cold int ff_MPV_encode_init(AVCodecContext *avctx)
     if (ff_MPV_common_init(s) < 0)
         return -1;
 
+    if (ARCH_X86)
+        ff_MPV_encode_init_x86(s);
+
     if (!s->dct_quantize)
         s->dct_quantize = ff_dct_quantize_c;
     if (!s->denoise_dct)
diff --git a/libavcodec/svq13.c b/libavcodec/svq13.c
new file mode 100644
index 0000000000..65a79025fe
--- /dev/null
+++ b/libavcodec/svq13.c
@@ -0,0 +1,68 @@
+/*
+ * SVQ1/SVQ3 decoder common code
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg 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.
+ *
+ * FFmpeg 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 FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include <stdint.h>
+
+#include "svq1.h"
+
+static const uint16_t checksum_table[256] = {
+  0x0000, 0x1021, 0x2042, 0x3063, 0x4084, 0x50A5, 0x60C6, 0x70E7,
+  0x8108, 0x9129, 0xA14A, 0xB16B, 0xC18C, 0xD1AD, 0xE1CE, 0xF1EF,
+  0x1231, 0x0210, 0x3273, 0x2252, 0x52B5, 0x4294, 0x72F7, 0x62D6,
+  0x9339, 0x8318, 0xB37B, 0xA35A, 0xD3BD, 0xC39C, 0xF3FF, 0xE3DE,
+  0x2462, 0x3443, 0x0420, 0x1401, 0x64E6, 0x74C7, 0x44A4, 0x5485,
+  0xA56A, 0xB54B, 0x8528, 0x9509, 0xE5EE, 0xF5CF, 0xC5AC, 0xD58D,
+  0x3653, 0x2672, 0x1611, 0x0630, 0x76D7, 0x66F6, 0x5695, 0x46B4,
+  0xB75B, 0xA77A, 0x9719, 0x8738, 0xF7DF, 0xE7FE, 0xD79D, 0xC7BC,
+  0x48C4, 0x58E5, 0x6886, 0x78A7, 0x0840, 0x1861, 0x2802, 0x3823,
+  0xC9CC, 0xD9ED, 0xE98E, 0xF9AF, 0x8948, 0x9969, 0xA90A, 0xB92B,
+  0x5AF5, 0x4AD4, 0x7AB7, 0x6A96, 0x1A71, 0x0A50, 0x3A33, 0x2A12,
+  0xDBFD, 0xCBDC, 0xFBBF, 0xEB9E, 0x9B79, 0x8B58, 0xBB3B, 0xAB1A,
+  0x6CA6, 0x7C87, 0x4CE4, 0x5CC5, 0x2C22, 0x3C03, 0x0C60, 0x1C41,
+  0xEDAE, 0xFD8F, 0xCDEC, 0xDDCD, 0xAD2A, 0xBD0B, 0x8D68, 0x9D49,
+  0x7E97, 0x6EB6, 0x5ED5, 0x4EF4, 0x3E13, 0x2E32, 0x1E51, 0x0E70,
+  0xFF9F, 0xEFBE, 0xDFDD, 0xCFFC, 0xBF1B, 0xAF3A, 0x9F59, 0x8F78,
+  0x9188, 0x81A9, 0xB1CA, 0xA1EB, 0xD10C, 0xC12D, 0xF14E, 0xE16F,
+  0x1080, 0x00A1, 0x30C2, 0x20E3, 0x5004, 0x4025, 0x7046, 0x6067,
+  0x83B9, 0x9398, 0xA3FB, 0xB3DA, 0xC33D, 0xD31C, 0xE37F, 0xF35E,
+  0x02B1, 0x1290, 0x22F3, 0x32D2, 0x4235, 0x5214, 0x6277, 0x7256,
+  0xB5EA, 0xA5CB, 0x95A8, 0x8589, 0xF56E, 0xE54F, 0xD52C, 0xC50D,
+  0x34E2, 0x24C3, 0x14A0, 0x0481, 0x7466, 0x6447, 0x5424, 0x4405,
+  0xA7DB, 0xB7FA, 0x8799, 0x97B8, 0xE75F, 0xF77E, 0xC71D, 0xD73C,
+  0x26D3, 0x36F2, 0x0691, 0x16B0, 0x6657, 0x7676, 0x4615, 0x5634,
+  0xD94C, 0xC96D, 0xF90E, 0xE92F, 0x99C8, 0x89E9, 0xB98A, 0xA9AB,
+  0x5844, 0x4865, 0x7806, 0x6827, 0x18C0, 0x08E1, 0x3882, 0x28A3,
+  0xCB7D, 0xDB5C, 0xEB3F, 0xFB1E, 0x8BF9, 0x9BD8, 0xABBB, 0xBB9A,
+  0x4A75, 0x5A54, 0x6A37, 0x7A16, 0x0AF1, 0x1AD0, 0x2AB3, 0x3A92,
+  0xFD2E, 0xED0F, 0xDD6C, 0xCD4D, 0xBDAA, 0xAD8B, 0x9DE8, 0x8DC9,
+  0x7C26, 0x6C07, 0x5C64, 0x4C45, 0x3CA2, 0x2C83, 0x1CE0, 0x0CC1,
+  0xEF1F, 0xFF3E, 0xCF5D, 0xDF7C, 0xAF9B, 0xBFBA, 0x8FD9, 0x9FF8,
+  0x6E17, 0x7E36, 0x4E55, 0x5E74, 0x2E93, 0x3EB2, 0x0ED1, 0x1EF0
+};
+
+uint16_t ff_svq1_packet_checksum (const uint8_t *data, const int length, int value) {
+  int i;
+
+  for (i=0; i < length; i++) {
+    value = checksum_table[data[i] ^ (value >> 8)] ^ ((value & 0xFF) << 8);
+  }
+
+  return value;
+}
diff --git a/libavcodec/svq1dec.c b/libavcodec/svq1dec.c
index 3b8dc1b852..ba240ac74d 100644
--- a/libavcodec/svq1dec.c
+++ b/libavcodec/svq1dec.c
@@ -58,41 +58,6 @@ typedef struct svq1_pmv_s {
   int           y;
 } svq1_pmv;
 
-static const uint16_t checksum_table[256] = {
-  0x0000, 0x1021, 0x2042, 0x3063, 0x4084, 0x50A5, 0x60C6, 0x70E7,
-  0x8108, 0x9129, 0xA14A, 0xB16B, 0xC18C, 0xD1AD, 0xE1CE, 0xF1EF,
-  0x1231, 0x0210, 0x3273, 0x2252, 0x52B5, 0x4294, 0x72F7, 0x62D6,
-  0x9339, 0x8318, 0xB37B, 0xA35A, 0xD3BD, 0xC39C, 0xF3FF, 0xE3DE,
-  0x2462, 0x3443, 0x0420, 0x1401, 0x64E6, 0x74C7, 0x44A4, 0x5485,
-  0xA56A, 0xB54B, 0x8528, 0x9509, 0xE5EE, 0xF5CF, 0xC5AC, 0xD58D,
-  0x3653, 0x2672, 0x1611, 0x0630, 0x76D7, 0x66F6, 0x5695, 0x46B4,
-  0xB75B, 0xA77A, 0x9719, 0x8738, 0xF7DF, 0xE7FE, 0xD79D, 0xC7BC,
-  0x48C4, 0x58E5, 0x6886, 0x78A7, 0x0840, 0x1861, 0x2802, 0x3823,
-  0xC9CC, 0xD9ED, 0xE98E, 0xF9AF, 0x8948, 0x9969, 0xA90A, 0xB92B,
-  0x5AF5, 0x4AD4, 0x7AB7, 0x6A96, 0x1A71, 0x0A50, 0x3A33, 0x2A12,
-  0xDBFD, 0xCBDC, 0xFBBF, 0xEB9E, 0x9B79, 0x8B58, 0xBB3B, 0xAB1A,
-  0x6CA6, 0x7C87, 0x4CE4, 0x5CC5, 0x2C22, 0x3C03, 0x0C60, 0x1C41,
-  0xEDAE, 0xFD8F, 0xCDEC, 0xDDCD, 0xAD2A, 0xBD0B, 0x8D68, 0x9D49,
-  0x7E97, 0x6EB6, 0x5ED5, 0x4EF4, 0x3E13, 0x2E32, 0x1E51, 0x0E70,
-  0xFF9F, 0xEFBE, 0xDFDD, 0xCFFC, 0xBF1B, 0xAF3A, 0x9F59, 0x8F78,
-  0x9188, 0x81A9, 0xB1CA, 0xA1EB, 0xD10C, 0xC12D, 0xF14E, 0xE16F,
-  0x1080, 0x00A1, 0x30C2, 0x20E3, 0x5004, 0x4025, 0x7046, 0x6067,
-  0x83B9, 0x9398, 0xA3FB, 0xB3DA, 0xC33D, 0xD31C, 0xE37F, 0xF35E,
-  0x02B1, 0x1290, 0x22F3, 0x32D2, 0x4235, 0x5214, 0x6277, 0x7256,
-  0xB5EA, 0xA5CB, 0x95A8, 0x8589, 0xF56E, 0xE54F, 0xD52C, 0xC50D,
-  0x34E2, 0x24C3, 0x14A0, 0x0481, 0x7466, 0x6447, 0x5424, 0x4405,
-  0xA7DB, 0xB7FA, 0x8799, 0x97B8, 0xE75F, 0xF77E, 0xC71D, 0xD73C,
-  0x26D3, 0x36F2, 0x0691, 0x16B0, 0x6657, 0x7676, 0x4615, 0x5634,
-  0xD94C, 0xC96D, 0xF90E, 0xE92F, 0x99C8, 0x89E9, 0xB98A, 0xA9AB,
-  0x5844, 0x4865, 0x7806, 0x6827, 0x18C0, 0x08E1, 0x3882, 0x28A3,
-  0xCB7D, 0xDB5C, 0xEB3F, 0xFB1E, 0x8BF9, 0x9BD8, 0xABBB, 0xBB9A,
-  0x4A75, 0x5A54, 0x6A37, 0x7A16, 0x0AF1, 0x1AD0, 0x2AB3, 0x3A92,
-  0xFD2E, 0xED0F, 0xDD6C, 0xCD4D, 0xBDAA, 0xAD8B, 0x9DE8, 0x8DC9,
-  0x7C26, 0x6C07, 0x5C64, 0x4C45, 0x3CA2, 0x2C83, 0x1CE0, 0x0CC1,
-  0xEF1F, 0xFF3E, 0xCF5D, 0xDF7C, 0xAF9B, 0xBFBA, 0x8FD9, 0x9FF8,
-  0x6E17, 0x7E36, 0x4E55, 0x5E74, 0x2E93, 0x3EB2, 0x0ED1, 0x1EF0
-};
-
 static const uint8_t string_table[256] = {
   0x00, 0xD5, 0x7F, 0xAA, 0xFE, 0x2B, 0x81, 0x54,
   0x29, 0xFC, 0x56, 0x83, 0xD7, 0x02, 0xA8, 0x7D,
@@ -524,16 +489,6 @@ static int svq1_decode_delta_block (MpegEncContext *s, GetBitContext *bitbuf,
   return result;
 }
 
-uint16_t ff_svq1_packet_checksum (const uint8_t *data, const int length, int value) {
-  int i;
-
-  for (i=0; i < length; i++) {
-    value = checksum_table[data[i] ^ (value >> 8)] ^ ((value & 0xFF) << 8);
-  }
-
-  return value;
-}
-
 static void svq1_parse_string (GetBitContext *bitbuf, uint8_t *out) {
   uint8_t seed;
   int     i;
diff --git a/libavcodec/x86/Makefile b/libavcodec/x86/Makefile
index 73a6a6032c..f90f12e877 100644
--- a/libavcodec/x86/Makefile
+++ b/libavcodec/x86/Makefile
@@ -14,7 +14,7 @@ MMX-OBJS                               += x86/dsputil_mmx.o             \
 MMX-OBJS-$(CONFIG_AAC_DECODER)         += x86/sbrdsp_init.o
 MMX-OBJS-$(CONFIG_AC3DSP)              += x86/ac3dsp_init.o
 MMX-OBJS-$(CONFIG_CAVS_DECODER)        += x86/cavsdsp_mmx.o
-MMX-OBJS-$(CONFIG_DNXHD_ENCODER)       += x86/dnxhd_mmx.o
+MMX-OBJS-$(CONFIG_DNXHD_ENCODER)       += x86/dnxhdenc.o
 MMX-OBJS-$(CONFIG_DWT)                 += x86/snowdsp_mmx.o \
                                           x86/dwt.o
 MMX-OBJS-$(CONFIG_ENCODERS)            += x86/dsputilenc_mmx.o
@@ -24,7 +24,8 @@ MMX-OBJS-$(CONFIG_H264DSP)             += x86/h264dsp_init.o
 MMX-OBJS-$(CONFIG_H264PRED)            += x86/h264_intrapred_init.o
 MMX-OBJS-$(CONFIG_LPC)                 += x86/lpc_mmx.o
 MMX-OBJS-$(CONFIG_MPEGAUDIODSP)        += x86/mpegaudiodec_mmx.o
-MMX-OBJS-$(CONFIG_MPEGVIDEO)           += x86/mpegvideo_mmx.o
+MMX-OBJS-$(CONFIG_MPEGVIDEO)           += x86/mpegvideo.o
+MMX-OBJS-$(CONFIG_MPEGVIDEOENC)        += x86/mpegvideoenc.o
 MMX-OBJS-$(CONFIG_PNG_DECODER)         += x86/pngdsp_init.o
 MMX-OBJS-$(CONFIG_PRORES_DECODER)      += x86/proresdsp_init.o
 MMX-OBJS-$(CONFIG_PRORES_LGPL_DECODER) += x86/proresdsp_init.o
diff --git a/libavcodec/x86/dnxhd_mmx.c b/libavcodec/x86/dnxhdenc.c
index ccd1575545..8b0c2ad225 100644
--- a/libavcodec/x86/dnxhd_mmx.c
+++ b/libavcodec/x86/dnxhdenc.c
@@ -54,7 +54,7 @@ static void get_pixels_8x4_sym_sse2(DCTELEM *block, const uint8_t *pixels, int l
 
 #endif /* HAVE_INLINE_ASM */
 
-void ff_dnxhd_init_mmx(DNXHDEncContext *ctx)
+void ff_dnxhdenc_init_x86(DNXHDEncContext *ctx)
 {
 #if HAVE_INLINE_ASM
     if (av_get_cpu_flags() & AV_CPU_FLAG_SSE2) {
diff --git a/libavcodec/x86/mpegvideo_mmx.c b/libavcodec/x86/mpegvideo.c
index f3927f0eae..2d475abe01 100644
--- a/libavcodec/x86/mpegvideo_mmx.c
+++ b/libavcodec/x86/mpegvideo.c
@@ -1,7 +1,4 @@
 /*
- * The simplest mpeg encoder (well, it was the simplest!)
- * Copyright (c) 2000,2001 Fabrice Bellard
- *
  * Optimized for ia32 CPUs by Nick Kurshev <nickols_k@mail.ru>
  * h263, mpeg1, mpeg2 dequantizer & draw_edges by Michael Niedermayer <michaelni@gmx.at>
  *
@@ -31,9 +28,6 @@
 
 #if HAVE_INLINE_ASM
 
-extern uint16_t ff_inv_zigzag_direct16[64];
-
-
 static void dct_unquantize_h263_intra_mmx(MpegEncContext *s,
                                   DCTELEM *block, int n, int qscale)
 {
@@ -588,56 +582,14 @@ static void  denoise_dct_sse2(MpegEncContext *s, DCTELEM *block){
     );
 }
 
-#if HAVE_SSSE3
-#define HAVE_SSSE3_BAK
-#endif
-#undef HAVE_SSSE3
-#define HAVE_SSSE3 0
-
-#undef HAVE_SSE2
-#undef HAVE_MMXEXT
-#define HAVE_SSE2 0
-#define HAVE_MMXEXT 0
-#define RENAME(a) a ## _MMX
-#define RENAMEl(a) a ## _mmx
-#include "mpegvideo_mmx_template.c"
-
-#undef HAVE_MMXEXT
-#define HAVE_MMXEXT 1
-#undef RENAME
-#undef RENAMEl
-#define RENAME(a) a ## _MMX2
-#define RENAMEl(a) a ## _mmx2
-#include "mpegvideo_mmx_template.c"
-
-#undef HAVE_SSE2
-#define HAVE_SSE2 1
-#undef RENAME
-#undef RENAMEl
-#define RENAME(a) a ## _SSE2
-#define RENAMEl(a) a ## _sse2
-#include "mpegvideo_mmx_template.c"
-
-#ifdef HAVE_SSSE3_BAK
-#undef HAVE_SSSE3
-#define HAVE_SSSE3 1
-#undef RENAME
-#undef RENAMEl
-#define RENAME(a) a ## _SSSE3
-#define RENAMEl(a) a ## _sse2
-#include "mpegvideo_mmx_template.c"
-#endif
-
 #endif /* HAVE_INLINE_ASM */
 
-void ff_MPV_common_init_mmx(MpegEncContext *s)
+void ff_MPV_common_init_x86(MpegEncContext *s)
 {
 #if HAVE_INLINE_ASM
     int mm_flags = av_get_cpu_flags();
 
     if (mm_flags & AV_CPU_FLAG_MMX) {
-        const int dct_algo = s->avctx->dct_algo;
-
         s->dct_unquantize_h263_intra = dct_unquantize_h263_intra_mmx;
         s->dct_unquantize_h263_inter = dct_unquantize_h263_inter_mmx;
         s->dct_unquantize_mpeg1_intra = dct_unquantize_mpeg1_intra_mmx;
@@ -651,21 +603,6 @@ void ff_MPV_common_init_mmx(MpegEncContext *s)
         } else {
                 s->denoise_dct= denoise_dct_mmx;
         }
-
-        if(dct_algo==FF_DCT_AUTO || dct_algo==FF_DCT_MMX){
-#if HAVE_SSSE3
-            if(mm_flags & AV_CPU_FLAG_SSSE3){
-                s->dct_quantize= dct_quantize_SSSE3;
-            } else
-#endif
-            if(mm_flags & AV_CPU_FLAG_SSE2){
-                s->dct_quantize= dct_quantize_SSE2;
-            } else if (mm_flags & AV_CPU_FLAG_MMXEXT) {
-                s->dct_quantize= dct_quantize_MMX2;
-            } else {
-                s->dct_quantize= dct_quantize_MMX;
-            }
-        }
     }
 #endif /* HAVE_INLINE_ASM */
 }
diff --git a/libavcodec/x86/mpegvideoenc.c b/libavcodec/x86/mpegvideoenc.c
new file mode 100644
index 0000000000..2d190f979f
--- /dev/null
+++ b/libavcodec/x86/mpegvideoenc.c
@@ -0,0 +1,96 @@
+/*
+ * The simplest mpeg encoder (well, it was the simplest!)
+ * Copyright (c) 2000,2001 Fabrice Bellard
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg 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.
+ *
+ * FFmpeg 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 FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "libavutil/cpu.h"
+#include "libavutil/x86/asm.h"
+#include "libavcodec/avcodec.h"
+#include "libavcodec/dsputil.h"
+#include "libavcodec/mpegvideo.h"
+#include "dsputil_mmx.h"
+
+#if HAVE_INLINE_ASM
+
+extern uint16_t ff_inv_zigzag_direct16[64];
+
+#if HAVE_SSSE3
+#define HAVE_SSSE3_BAK
+#endif
+#undef HAVE_SSSE3
+#define HAVE_SSSE3 0
+
+#undef HAVE_SSE2
+#undef HAVE_MMXEXT
+#define HAVE_SSE2 0
+#define HAVE_MMXEXT 0
+#define RENAME(a) a ## _MMX
+#define RENAMEl(a) a ## _mmx
+#include "mpegvideoenc_template.c"
+
+#undef HAVE_MMXEXT
+#define HAVE_MMXEXT 1
+#undef RENAME
+#undef RENAMEl
+#define RENAME(a) a ## _MMX2
+#define RENAMEl(a) a ## _mmx2
+#include "mpegvideoenc_template.c"
+
+#undef HAVE_SSE2
+#define HAVE_SSE2 1
+#undef RENAME
+#undef RENAMEl
+#define RENAME(a) a ## _SSE2
+#define RENAMEl(a) a ## _sse2
+#include "mpegvideoenc_template.c"
+
+#ifdef HAVE_SSSE3_BAK
+#undef HAVE_SSSE3
+#define HAVE_SSSE3 1
+#undef RENAME
+#undef RENAMEl
+#define RENAME(a) a ## _SSSE3
+#define RENAMEl(a) a ## _sse2
+#include "mpegvideoenc_template.c"
+#endif
+
+#endif /* HAVE_INLINE_ASM */
+
+void ff_MPV_encode_init_x86(MpegEncContext *s)
+{
+#if HAVE_INLINE_ASM
+    int mm_flags = av_get_cpu_flags();
+    const int dct_algo = s->avctx->dct_algo;
+
+    if (dct_algo == FF_DCT_AUTO || dct_algo == FF_DCT_MMX) {
+#if HAVE_SSSE3
+        if (mm_flags & AV_CPU_FLAG_SSSE3) {
+            s->dct_quantize = dct_quantize_SSSE3;
+        } else
+#endif
+        if (mm_flags & AV_CPU_FLAG_SSE2) {
+            s->dct_quantize = dct_quantize_SSE2;
+        } else if (mm_flags & AV_CPU_FLAG_MMXEXT) {
+            s->dct_quantize = dct_quantize_MMX2;
+        } else {
+            s->dct_quantize = dct_quantize_MMX;
+        }
+    }
+#endif /* HAVE_INLINE_ASM */
+}
diff --git a/libavcodec/x86/mpegvideo_mmx_template.c b/libavcodec/x86/mpegvideoenc_template.c
index d538a39919..d538a39919 100644
--- a/libavcodec/x86/mpegvideo_mmx_template.c
+++ b/libavcodec/x86/mpegvideoenc_template.c
diff --git a/libavresample/x86/audio_convert.asm b/libavresample/x86/audio_convert.asm
index 244c4d1b08..3db64d2f9b 100644
--- a/libavresample/x86/audio_convert.asm
+++ b/libavresample/x86/audio_convert.asm
@@ -30,6 +30,11 @@ pf_s32_inv_scale: times 8 dd 0x30000000
 pf_s32_scale:     times 8 dd 0x4f000000
 pf_s16_inv_scale: times 4 dd 0x38000000
 pf_s16_scale:     times 4 dd 0x47000000
+pb_shuf_unpack_even:      db -1, -1,  0,  1, -1, -1,  2,  3, -1, -1,  8,  9, -1, -1, 10, 11
+pb_shuf_unpack_odd:       db -1, -1,  4,  5, -1, -1,  6,  7, -1, -1, 12, 13, -1, -1, 14, 15
+pb_interleave_words: SHUFFLE_MASK_W  0,  4,  1,  5,  2,  6,  3,  7
+pb_deinterleave_words: SHUFFLE_MASK_W  0,  2,  4,  6,  1,  3,  5,  7
+pw_zero_even:     times 4 dw 0x0000, 0xffff
 
 SECTION_TEXT
 
@@ -223,6 +228,512 @@ INIT_YMM avx
 CONV_FLT_TO_S32
 %endif
 
+;------------------------------------------------------------------------------
+; void ff_conv_s16p_to_s16_2ch(int16_t *dst, int16_t *const *src, int len,
+;                              int channels);
+;------------------------------------------------------------------------------
+
+%macro CONV_S16P_TO_S16_2CH 0
+cglobal conv_s16p_to_s16_2ch, 3,4,5, dst, src0, len, src1
+    mov       src1q, [src0q+gprsize]
+    mov       src0q, [src0q        ]
+    lea        lenq, [2*lend]
+    add       src0q, lenq
+    add       src1q, lenq
+    lea        dstq, [dstq+2*lenq]
+    neg        lenq
+.loop
+    mova         m0, [src0q+lenq       ]
+    mova         m1, [src1q+lenq       ]
+    mova         m2, [src0q+lenq+mmsize]
+    mova         m3, [src1q+lenq+mmsize]
+    SBUTTERFLY2  wd, 0, 1, 4
+    SBUTTERFLY2  wd, 2, 3, 4
+    mova  [dstq+2*lenq+0*mmsize], m0
+    mova  [dstq+2*lenq+1*mmsize], m1
+    mova  [dstq+2*lenq+2*mmsize], m2
+    mova  [dstq+2*lenq+3*mmsize], m3
+    add        lenq, 2*mmsize
+    jl .loop
+    REP_RET
+%endmacro
+
+INIT_XMM sse2
+CONV_S16P_TO_S16_2CH
+%if HAVE_AVX
+INIT_XMM avx
+CONV_S16P_TO_S16_2CH
+%endif
+
+;------------------------------------------------------------------------------
+; void ff_conv_s16p_to_s16_6ch(int16_t *dst, int16_t *const *src, int len,
+;                              int channels);
+;------------------------------------------------------------------------------
+
+;------------------------------------------------------------------------------
+; NOTE: In the 6-channel functions, len could be used as an index on x86-64
+;       instead of just a counter, which would avoid incrementing the
+;       pointers, but the extra complexity and amount of code is not worth
+;       the small gain. On x86-32 there are not enough registers to use len
+;       as an index without keeping two of the pointers on the stack and
+;       loading them in each iteration.
+;------------------------------------------------------------------------------
+
+%macro CONV_S16P_TO_S16_6CH 0
+%if ARCH_X86_64
+cglobal conv_s16p_to_s16_6ch, 3,8,7, dst, src0, len, src1, src2, src3, src4, src5
+%else
+cglobal conv_s16p_to_s16_6ch, 2,7,7, dst, src0, src1, src2, src3, src4, src5
+%define lend dword r2m
+%endif
+    mov      src1q, [src0q+1*gprsize]
+    mov      src2q, [src0q+2*gprsize]
+    mov      src3q, [src0q+3*gprsize]
+    mov      src4q, [src0q+4*gprsize]
+    mov      src5q, [src0q+5*gprsize]
+    mov      src0q, [src0q]
+    sub      src1q, src0q
+    sub      src2q, src0q
+    sub      src3q, src0q
+    sub      src4q, src0q
+    sub      src5q, src0q
+.loop:
+%if cpuflag(sse2slow)
+    movq        m0, [src0q      ]   ; m0 =  0,  6, 12, 18,  x,  x,  x,  x
+    movq        m1, [src0q+src1q]   ; m1 =  1,  7, 13, 19,  x,  x,  x,  x
+    movq        m2, [src0q+src2q]   ; m2 =  2,  8, 14, 20,  x,  x,  x,  x
+    movq        m3, [src0q+src3q]   ; m3 =  3,  9, 15, 21,  x,  x,  x,  x
+    movq        m4, [src0q+src4q]   ; m4 =  4, 10, 16, 22,  x,  x,  x,  x
+    movq        m5, [src0q+src5q]   ; m5 =  5, 11, 17, 23,  x,  x,  x,  x
+                                    ; unpack words:
+    punpcklwd   m0, m1              ; m0 =  0,  1,  6,  7, 12, 13, 18, 19
+    punpcklwd   m2, m3              ; m2 =  4,  5, 10, 11, 16, 17, 22, 23
+    punpcklwd   m4, m5              ; m4 =  2,  3,  8,  9, 14, 15, 20, 21
+                                    ; blend dwords
+    shufps      m1, m0, m2, q2020   ; m1 =  0,  1, 12, 13,  2,  3, 14, 15
+    shufps      m0, m4, q2031       ; m0 =  6,  7, 18, 19,  4,  5, 16, 17
+    shufps      m2, m4, q3131       ; m2 =  8,  9, 20, 21, 10, 11, 22, 23
+                                    ; shuffle dwords
+    pshufd      m0, m0, q1302       ; m0 =  4,  5,  6,  7, 16, 17, 18, 19
+    pshufd      m1, m1, q3120       ; m1 =  0,  1,  2,  3, 12, 13, 14, 15
+    pshufd      m2, m2, q3120       ; m2 =  8,  9, 10, 11, 20, 21, 22, 23
+    movq   [dstq+0*mmsize/2], m1
+    movq   [dstq+1*mmsize/2], m0
+    movq   [dstq+2*mmsize/2], m2
+    movhps [dstq+3*mmsize/2], m1
+    movhps [dstq+4*mmsize/2], m0
+    movhps [dstq+5*mmsize/2], m2
+    add      src0q, mmsize/2
+    add       dstq, mmsize*3
+    sub       lend, mmsize/4
+%else
+    mova        m0, [src0q      ]   ; m0 =  0,  6, 12, 18, 24, 30, 36, 42
+    mova        m1, [src0q+src1q]   ; m1 =  1,  7, 13, 19, 25, 31, 37, 43
+    mova        m2, [src0q+src2q]   ; m2 =  2,  8, 14, 20, 26, 32, 38, 44
+    mova        m3, [src0q+src3q]   ; m3 =  3,  9, 15, 21, 27, 33, 39, 45
+    mova        m4, [src0q+src4q]   ; m4 =  4, 10, 16, 22, 28, 34, 40, 46
+    mova        m5, [src0q+src5q]   ; m5 =  5, 11, 17, 23, 29, 35, 41, 47
+                                    ; unpack words:
+    SBUTTERFLY2 wd, 0, 1, 6         ; m0 =  0,  1,  6,  7, 12, 13, 18, 19
+                                    ; m1 = 24, 25, 30, 31, 36, 37, 42, 43
+    SBUTTERFLY2 wd, 2, 3, 6         ; m2 =  2,  3,  8,  9, 14, 15, 20, 21
+                                    ; m3 = 26, 27, 32, 33, 38, 39, 44, 45
+    SBUTTERFLY2 wd, 4, 5, 6         ; m4 =  4,  5, 10, 11, 16, 17, 22, 23
+                                    ; m5 = 28, 29, 34, 35, 40, 41, 46, 47
+                                    ; blend dwords
+    shufps      m6, m0, m2, q2020   ; m6 =  0,  1, 12, 13,  2,  3, 14, 15
+    shufps      m0, m4, q2031       ; m0 =  6,  7, 18, 19,  4,  5, 16, 17
+    shufps      m2, m4, q3131       ; m2 =  8,  9, 20, 21, 10, 11, 22, 23
+    SWAP 4,6                        ; m4 =  0,  1, 12, 13,  2,  3, 14, 15
+    shufps      m6, m1, m3, q2020   ; m6 = 24, 25, 36, 37, 26, 27, 38, 39
+    shufps      m1, m5, q2031       ; m1 = 30, 31, 42, 43, 28, 29, 40, 41
+    shufps      m3, m5, q3131       ; m3 = 32, 33, 44, 45, 34, 35, 46, 47
+    SWAP 5,6                        ; m5 = 24, 25, 36, 37, 26, 27, 38, 39
+                                    ; shuffle dwords
+    pshufd      m0, m0, q1302       ; m0 =  4,  5,  6,  7, 16, 17, 18, 19
+    pshufd      m2, m2, q3120       ; m2 =  8,  9, 10, 11, 20, 21, 22, 23
+    pshufd      m4, m4, q3120       ; m4 =  0,  1,  2,  3, 12, 13, 14, 15
+    pshufd      m1, m1, q1302       ; m1 = 28, 29, 30, 31, 40, 41, 42, 43
+    pshufd      m3, m3, q3120       ; m3 = 32, 33, 34, 35, 44, 45, 46, 47
+    pshufd      m5, m5, q3120       ; m5 = 24, 25, 26, 27, 36, 37, 38, 39
+                                    ; shuffle qwords
+    punpcklqdq  m6, m4, m0          ; m6 =  0,  1,  2,  3,  4,  5,  6,  7
+    punpckhqdq  m0, m2              ; m0 = 16, 17, 18, 19, 20, 21, 22, 23
+    shufps      m2, m4, q3210       ; m2 =  8,  9, 10, 11, 12, 13, 14, 15
+    SWAP 4,6                        ; m4 =  0,  1,  2,  3,  4,  5,  6,  7
+    punpcklqdq  m6, m5, m1          ; m6 = 24, 25, 26, 27, 28, 29, 30, 31
+    punpckhqdq  m1, m3              ; m1 = 40, 41, 42, 43, 44, 45, 46, 47
+    shufps      m3, m5, q3210       ; m3 = 32, 33, 34, 35, 36, 37, 38, 39
+    SWAP 5,6                        ; m5 = 24, 25, 26, 27, 28, 29, 30, 31
+    mova   [dstq+0*mmsize], m4
+    mova   [dstq+1*mmsize], m2
+    mova   [dstq+2*mmsize], m0
+    mova   [dstq+3*mmsize], m5
+    mova   [dstq+4*mmsize], m3
+    mova   [dstq+5*mmsize], m1
+    add      src0q, mmsize
+    add       dstq, mmsize*6
+    sub       lend, mmsize/2
+%endif
+    jg .loop
+    REP_RET
+%endmacro
+
+INIT_XMM sse2
+CONV_S16P_TO_S16_6CH
+INIT_XMM sse2slow
+CONV_S16P_TO_S16_6CH
+%if HAVE_AVX
+INIT_XMM avx
+CONV_S16P_TO_S16_6CH
+%endif
+
+;------------------------------------------------------------------------------
+; void ff_conv_s16p_to_flt_2ch(float *dst, int16_t *const *src, int len,
+;                              int channels);
+;------------------------------------------------------------------------------
+
+%macro CONV_S16P_TO_FLT_2CH 0
+cglobal conv_s16p_to_flt_2ch, 3,4,6, dst, src0, len, src1
+    lea       lenq, [2*lend]
+    mov      src1q, [src0q+gprsize]
+    mov      src0q, [src0q        ]
+    lea       dstq, [dstq+4*lenq]
+    add      src0q, lenq
+    add      src1q, lenq
+    neg       lenq
+    mova        m5, [pf_s32_inv_scale]
+.loop:
+    mova        m2, [src0q+lenq]    ; m2 =  0,  2,  4,  6,  8, 10, 12, 14
+    mova        m4, [src1q+lenq]    ; m4 =  1,  3,  5,  7,  9, 11, 13, 15
+    SBUTTERFLY2 wd, 2, 4, 3         ; m2 =  0,  1,  2,  3,  4,  5,  6,  7
+                                    ; m4 =  8,  9, 10, 11, 12, 13, 14, 15
+    pxor        m3, m3
+    punpcklwd   m0, m3, m2          ; m0 =      0,      1,      2,      3
+    punpckhwd   m1, m3, m2          ; m1 =      4,      5,      6,      7
+    punpcklwd   m2, m3, m4          ; m2 =      8,      9,     10,     11
+    punpckhwd   m3, m4              ; m3 =     12,     13,     14,     15
+    cvtdq2ps    m0, m0
+    cvtdq2ps    m1, m1
+    cvtdq2ps    m2, m2
+    cvtdq2ps    m3, m3
+    mulps       m0, m5
+    mulps       m1, m5
+    mulps       m2, m5
+    mulps       m3, m5
+    mova  [dstq+4*lenq         ], m0
+    mova  [dstq+4*lenq+  mmsize], m1
+    mova  [dstq+4*lenq+2*mmsize], m2
+    mova  [dstq+4*lenq+3*mmsize], m3
+    add       lenq, mmsize
+    jl .loop
+    REP_RET
+%endmacro
+
+INIT_XMM sse2
+CONV_S16P_TO_FLT_2CH
+%if HAVE_AVX
+INIT_XMM avx
+CONV_S16P_TO_FLT_2CH
+%endif
+
+;------------------------------------------------------------------------------
+; void ff_conv_s16p_to_flt_6ch(float *dst, int16_t *const *src, int len,
+;                              int channels);
+;------------------------------------------------------------------------------
+
+%macro CONV_S16P_TO_FLT_6CH 0
+%if ARCH_X86_64
+cglobal conv_s16p_to_flt_6ch, 3,8,8, dst, src, len, src1, src2, src3, src4, src5
+%else
+cglobal conv_s16p_to_flt_6ch, 2,7,8, dst, src, src1, src2, src3, src4, src5
+%define lend dword r2m
+%endif
+    mov     src1q, [srcq+1*gprsize]
+    mov     src2q, [srcq+2*gprsize]
+    mov     src3q, [srcq+3*gprsize]
+    mov     src4q, [srcq+4*gprsize]
+    mov     src5q, [srcq+5*gprsize]
+    mov      srcq, [srcq]
+    sub     src1q, srcq
+    sub     src2q, srcq
+    sub     src3q, srcq
+    sub     src4q, srcq
+    sub     src5q, srcq
+    mova       m7, [pf_s32_inv_scale]
+%if cpuflag(ssse3)
+    %define unpack_even m6
+    mova       m6, [pb_shuf_unpack_even]
+%if ARCH_X86_64
+    %define unpack_odd m8
+    mova       m8, [pb_shuf_unpack_odd]
+%else
+    %define unpack_odd [pb_shuf_unpack_odd]
+%endif
+%endif
+.loop:
+    movq       m0, [srcq      ]  ; m0 =  0,  6, 12, 18,  x,  x,  x,  x
+    movq       m1, [srcq+src1q]  ; m1 =  1,  7, 13, 19,  x,  x,  x,  x
+    movq       m2, [srcq+src2q]  ; m2 =  2,  8, 14, 20,  x,  x,  x,  x
+    movq       m3, [srcq+src3q]  ; m3 =  3,  9, 15, 21,  x,  x,  x,  x
+    movq       m4, [srcq+src4q]  ; m4 =  4, 10, 16, 22,  x,  x,  x,  x
+    movq       m5, [srcq+src5q]  ; m5 =  5, 11, 17, 23,  x,  x,  x,  x
+                                 ; unpack words:
+    punpcklwd  m0, m1            ; m0 =  0,  1,  6,  7, 12, 13, 18, 19
+    punpcklwd  m2, m3            ; m2 =  2,  3,  8,  9, 14, 15, 20, 21
+    punpcklwd  m4, m5            ; m4 =  4,  5, 10, 11, 16, 17, 22, 23
+                                 ; blend dwords
+    shufps     m1, m4, m0, q3120 ; m1 =  4,  5, 16, 17,  6,  7, 18, 19
+    shufps         m0, m2, q2020 ; m0 =  0,  1, 12, 13,  2,  3, 14, 15
+    shufps         m2, m4, q3131 ; m2 =  8,  9, 20, 21, 10, 11, 22, 23
+%if cpuflag(ssse3)
+    pshufb     m3, m0, unpack_odd   ; m3 =  12,     13,     14,     15
+    pshufb         m0, unpack_even  ; m0 =   0,      1,      2,      3
+    pshufb     m4, m1, unpack_odd   ; m4 =  16,     17,     18,     19
+    pshufb         m1, unpack_even  ; m1 =   4,      5,      6,      7
+    pshufb     m5, m2, unpack_odd   ; m5 =  20,     21,     22,     23
+    pshufb         m2, unpack_even  ; m2 =   8,      9,     10,     11
+%else
+                                 ; shuffle dwords
+    pshufd     m0, m0, q3120     ; m0 =  0,  1,  2,  3, 12, 13, 14, 15
+    pshufd     m1, m1, q3120     ; m1 =  4,  5,  6,  7, 16, 17, 18, 19
+    pshufd     m2, m2, q3120     ; m2 =  8,  9, 10, 11, 20, 21, 22, 23
+    pxor       m6, m6            ; convert s16 in m0-m2 to s32 in m0-m5
+    punpcklwd  m3, m6, m0        ; m3 =      0,      1,      2,      3
+    punpckhwd  m4, m6, m0        ; m4 =     12,     13,     14,     15
+    punpcklwd  m0, m6, m1        ; m0 =      4,      5,      6,      7
+    punpckhwd  m5, m6, m1        ; m5 =     16,     17,     18,     19
+    punpcklwd  m1, m6, m2        ; m1 =      8,      9,     10,     11
+    punpckhwd      m6, m2        ; m6 =     20,     21,     22,     23
+    SWAP 6,2,1,0,3,4,5           ; swap registers 3,0,1,4,5,6 to 0,1,2,3,4,5
+%endif
+    cvtdq2ps   m0, m0            ; convert s32 to float
+    cvtdq2ps   m1, m1
+    cvtdq2ps   m2, m2
+    cvtdq2ps   m3, m3
+    cvtdq2ps   m4, m4
+    cvtdq2ps   m5, m5
+    mulps      m0, m7            ; scale float from s32 range to [-1.0,1.0]
+    mulps      m1, m7
+    mulps      m2, m7
+    mulps      m3, m7
+    mulps      m4, m7
+    mulps      m5, m7
+    mova  [dstq         ], m0
+    mova  [dstq+  mmsize], m1
+    mova  [dstq+2*mmsize], m2
+    mova  [dstq+3*mmsize], m3
+    mova  [dstq+4*mmsize], m4
+    mova  [dstq+5*mmsize], m5
+    add      srcq, mmsize/2
+    add      dstq, mmsize*6
+    sub      lend, mmsize/4
+    jg .loop
+    REP_RET
+%endmacro
+
+INIT_XMM sse2
+CONV_S16P_TO_FLT_6CH
+INIT_XMM ssse3
+CONV_S16P_TO_FLT_6CH
+%if HAVE_AVX
+INIT_XMM avx
+CONV_S16P_TO_FLT_6CH
+%endif
+
+;------------------------------------------------------------------------------
+; void ff_conv_fltp_to_s16_2ch(int16_t *dst, float *const *src, int len,
+;                              int channels);
+;------------------------------------------------------------------------------
+
+%macro CONV_FLTP_TO_S16_2CH 0
+cglobal conv_fltp_to_s16_2ch, 3,4,3, dst, src0, len, src1
+    lea      lenq, [4*lend]
+    mov     src1q, [src0q+gprsize]
+    mov     src0q, [src0q        ]
+    add      dstq, lenq
+    add     src0q, lenq
+    add     src1q, lenq
+    neg      lenq
+    mova       m2, [pf_s16_scale]
+%if cpuflag(ssse3)
+    mova       m3, [pb_interleave_words]
+%endif
+.loop:
+    mulps      m0, m2, [src0q+lenq] ; m0 =    0,    2,    4,    6
+    mulps      m1, m2, [src1q+lenq] ; m1 =    1,    3,    5,    7
+    cvtps2dq   m0, m0
+    cvtps2dq   m1, m1
+%if cpuflag(ssse3)
+    packssdw   m0, m1               ; m0 = 0, 2, 4, 6, 1, 3, 5, 7
+    pshufb     m0, m3               ; m0 = 0, 1, 2, 3, 4, 5, 6, 7
+%else
+    packssdw   m0, m0               ; m0 = 0, 2, 4, 6, x, x, x, x
+    packssdw   m1, m1               ; m1 = 1, 3, 5, 7, x, x, x, x
+    punpcklwd  m0, m1               ; m0 = 0, 1, 2, 3, 4, 5, 6, 7
+%endif
+    mova  [dstq+lenq], m0
+    add      lenq, mmsize
+    jl .loop
+    REP_RET
+%endmacro
+
+INIT_XMM sse2
+CONV_FLTP_TO_S16_2CH
+INIT_XMM ssse3
+CONV_FLTP_TO_S16_2CH
+
+;------------------------------------------------------------------------------
+; void ff_conv_fltp_to_s16_6ch(int16_t *dst, float *const *src, int len,
+;                              int channels);
+;------------------------------------------------------------------------------
+
+%macro CONV_FLTP_TO_S16_6CH 0
+%if ARCH_X86_64
+cglobal conv_fltp_to_s16_6ch, 3,8,7, dst, src, len, src1, src2, src3, src4, src5
+%else
+cglobal conv_fltp_to_s16_6ch, 2,7,7, dst, src, src1, src2, src3, src4, src5
+%define lend dword r2m
+%endif
+    mov        src1q, [srcq+1*gprsize]
+    mov        src2q, [srcq+2*gprsize]
+    mov        src3q, [srcq+3*gprsize]
+    mov        src4q, [srcq+4*gprsize]
+    mov        src5q, [srcq+5*gprsize]
+    mov         srcq, [srcq]
+    sub        src1q, srcq
+    sub        src2q, srcq
+    sub        src3q, srcq
+    sub        src4q, srcq
+    sub        src5q, srcq
+    movaps      xmm6, [pf_s16_scale]
+.loop:
+%if cpuflag(sse2)
+    mulps         m0, m6, [srcq      ]
+    mulps         m1, m6, [srcq+src1q]
+    mulps         m2, m6, [srcq+src2q]
+    mulps         m3, m6, [srcq+src3q]
+    mulps         m4, m6, [srcq+src4q]
+    mulps         m5, m6, [srcq+src5q]
+    cvtps2dq      m0, m0
+    cvtps2dq      m1, m1
+    cvtps2dq      m2, m2
+    cvtps2dq      m3, m3
+    cvtps2dq      m4, m4
+    cvtps2dq      m5, m5
+    packssdw      m0, m3            ; m0 =  0,  6, 12, 18,  3,  9, 15, 21
+    packssdw      m1, m4            ; m1 =  1,  7, 13, 19,  4, 10, 16, 22
+    packssdw      m2, m5            ; m2 =  2,  8, 14, 20,  5, 11, 17, 23
+                                    ; unpack words:
+    movhlps       m3, m0            ; m3 =  3,  9, 15, 21,  x,  x,  x,  x
+    punpcklwd     m0, m1            ; m0 =  0,  1,  6,  7, 12, 13, 18, 19
+    punpckhwd     m1, m2            ; m1 =  4,  5, 10, 11, 16, 17, 22, 23
+    punpcklwd     m2, m3            ; m2 =  2,  3,  8,  9, 14, 15, 20, 21
+                                    ; blend dwords:
+    shufps        m3, m0, m2, q2020 ; m3 =  0,  1, 12, 13,  2,  3, 14, 15
+    shufps        m0, m1, q2031     ; m0 =  6,  7, 18, 19,  4,  5, 16, 17
+    shufps        m2, m1, q3131     ; m2 =  8,  9, 20, 21, 10, 11, 22, 23
+                                    ; shuffle dwords:
+    shufps        m1, m2, m3, q3120 ; m1 =  8,  9, 10, 11, 12, 13, 14, 15
+    shufps        m3, m0,     q0220 ; m3 =  0,  1,  2,  3,  4,  5,  6,  7
+    shufps        m0, m2,     q3113 ; m0 = 16, 17, 18, 19, 20, 21, 22, 23
+    mova  [dstq+0*mmsize], m3
+    mova  [dstq+1*mmsize], m1
+    mova  [dstq+2*mmsize], m0
+%else ; sse
+    movlps      xmm0, [srcq      ]
+    movlps      xmm1, [srcq+src1q]
+    movlps      xmm2, [srcq+src2q]
+    movlps      xmm3, [srcq+src3q]
+    movlps      xmm4, [srcq+src4q]
+    movlps      xmm5, [srcq+src5q]
+    mulps       xmm0, xmm6
+    mulps       xmm1, xmm6
+    mulps       xmm2, xmm6
+    mulps       xmm3, xmm6
+    mulps       xmm4, xmm6
+    mulps       xmm5, xmm6
+    cvtps2pi     mm0, xmm0
+    cvtps2pi     mm1, xmm1
+    cvtps2pi     mm2, xmm2
+    cvtps2pi     mm3, xmm3
+    cvtps2pi     mm4, xmm4
+    cvtps2pi     mm5, xmm5
+    packssdw     mm0, mm3           ; m0 =  0,  6,  3,  9
+    packssdw     mm1, mm4           ; m1 =  1,  7,  4, 10
+    packssdw     mm2, mm5           ; m2 =  2,  8,  5, 11
+                                    ; unpack words
+    pshufw       mm3, mm0, q1032    ; m3 =  3,  9,  0,  6
+    punpcklwd    mm0, mm1           ; m0 =  0,  1,  6,  7
+    punpckhwd    mm1, mm2           ; m1 =  4,  5, 10, 11
+    punpcklwd    mm2, mm3           ; m2 =  2,  3,  8,  9
+                                    ; unpack dwords
+    pshufw       mm3, mm0, q1032    ; m3 =  6,  7,  0,  1
+    punpckldq    mm0, mm2           ; m0 =  0,  1,  2,  3 (final)
+    punpckhdq    mm2, mm1           ; m2 =  8,  9, 10, 11 (final)
+    punpckldq    mm1, mm3           ; m1 =  4,  5,  6,  7 (final)
+    mova  [dstq+0*mmsize], mm0
+    mova  [dstq+1*mmsize], mm1
+    mova  [dstq+2*mmsize], mm2
+%endif
+    add       srcq, mmsize
+    add       dstq, mmsize*3
+    sub       lend, mmsize/4
+    jg .loop
+%if mmsize == 8
+    emms
+    RET
+%else
+    REP_RET
+%endif
+%endmacro
+
+INIT_MMX sse
+CONV_FLTP_TO_S16_6CH
+INIT_XMM sse2
+CONV_FLTP_TO_S16_6CH
+%if HAVE_AVX
+INIT_XMM avx
+CONV_FLTP_TO_S16_6CH
+%endif
+
+;------------------------------------------------------------------------------
+; void ff_conv_fltp_to_flt_2ch(float *dst, float *const *src, int len,
+;                              int channels);
+;------------------------------------------------------------------------------
+
+%macro CONV_FLTP_TO_FLT_2CH 0
+cglobal conv_fltp_to_flt_2ch, 3,4,5, dst, src0, len, src1
+    mov  src1q, [src0q+gprsize]
+    mov  src0q, [src0q]
+    lea   lenq, [4*lend]
+    add  src0q, lenq
+    add  src1q, lenq
+    lea   dstq, [dstq+2*lenq]
+    neg   lenq
+.loop
+    mova    m0, [src0q+lenq       ]
+    mova    m1, [src1q+lenq       ]
+    mova    m2, [src0q+lenq+mmsize]
+    mova    m3, [src1q+lenq+mmsize]
+    SBUTTERFLYPS 0, 1, 4
+    SBUTTERFLYPS 2, 3, 4
+    mova  [dstq+2*lenq+0*mmsize], m0
+    mova  [dstq+2*lenq+1*mmsize], m1
+    mova  [dstq+2*lenq+2*mmsize], m2
+    mova  [dstq+2*lenq+3*mmsize], m3
+    add   lenq, 2*mmsize
+    jl .loop
+    REP_RET
+%endmacro
+
+INIT_XMM sse
+CONV_FLTP_TO_FLT_2CH
+%if HAVE_AVX
+INIT_XMM avx
+CONV_FLTP_TO_FLT_2CH
+%endif
+
 ;-----------------------------------------------------------------------------
 ; void ff_conv_fltp_to_flt_6ch(float *dst, float *const *src, int len,
 ;                              int channels);
@@ -303,3 +814,449 @@ CONV_FLTP_TO_FLT_6CH
 INIT_XMM avx
 CONV_FLTP_TO_FLT_6CH
 %endif
+
+;------------------------------------------------------------------------------
+; void ff_conv_s16_to_s16p_2ch(int16_t *const *dst, int16_t *src, int len,
+;                              int channels);
+;------------------------------------------------------------------------------
+
+%macro CONV_S16_TO_S16P_2CH 0
+cglobal conv_s16_to_s16p_2ch, 3,4,4, dst0, src, len, dst1
+    lea       lenq, [2*lend]
+    mov      dst1q, [dst0q+gprsize]
+    mov      dst0q, [dst0q        ]
+    lea       srcq, [srcq+2*lenq]
+    add      dst0q, lenq
+    add      dst1q, lenq
+    neg       lenq
+%if cpuflag(ssse3)
+    mova        m3, [pb_deinterleave_words]
+%endif
+.loop:
+    mova        m0, [srcq+2*lenq       ]  ; m0 =  0,  1,  2,  3,  4,  5,  6,  7
+    mova        m1, [srcq+2*lenq+mmsize]  ; m1 =  8,  9, 10, 11, 12, 13, 14, 15
+%if cpuflag(ssse3)
+    pshufb      m0, m3                    ; m0 =  0,  2,  4,  6,  1,  3,  5,  7
+    pshufb      m1, m3                    ; m1 =  8, 10, 12, 14,  9, 11, 13, 15
+    SBUTTERFLY2 qdq, 0, 1, 2              ; m0 =  0,  2,  4,  6,  8, 10, 12, 14
+                                          ; m1 =  1,  3,  5,  7,  9, 11, 13, 15
+%else ; sse2
+    pshuflw     m0, m0, q3120             ; m0 =  0,  2,  1,  3,  4,  5,  6,  7
+    pshufhw     m0, m0, q3120             ; m0 =  0,  2,  1,  3,  4,  6,  5,  7
+    pshuflw     m1, m1, q3120             ; m1 =  8, 10,  9, 11, 12, 13, 14, 15
+    pshufhw     m1, m1, q3120             ; m1 =  8, 10,  9, 11, 12, 14, 13, 15
+    DEINT2_PS    0, 1, 2                  ; m0 =  0,  2,  4,  6,  8, 10, 12, 14
+                                          ; m1 =  1,  3,  5,  7,  9, 11, 13, 15
+%endif
+    mova  [dst0q+lenq], m0
+    mova  [dst1q+lenq], m1
+    add       lenq, mmsize
+    jl .loop
+    REP_RET
+%endmacro
+
+INIT_XMM sse2
+CONV_S16_TO_S16P_2CH
+INIT_XMM ssse3
+CONV_S16_TO_S16P_2CH
+%if HAVE_AVX
+INIT_XMM avx
+CONV_S16_TO_S16P_2CH
+%endif
+
+;------------------------------------------------------------------------------
+; void ff_conv_s16_to_s16p_6ch(int16_t *const *dst, int16_t *src, int len,
+;                              int channels);
+;------------------------------------------------------------------------------
+
+%macro CONV_S16_TO_S16P_6CH 0
+%if ARCH_X86_64
+cglobal conv_s16_to_s16p_6ch, 3,8,5, dst, src, len, dst1, dst2, dst3, dst4, dst5
+%else
+cglobal conv_s16_to_s16p_6ch, 2,7,5, dst, src, dst1, dst2, dst3, dst4, dst5
+%define lend dword r2m
+%endif
+    mov     dst1q, [dstq+  gprsize]
+    mov     dst2q, [dstq+2*gprsize]
+    mov     dst3q, [dstq+3*gprsize]
+    mov     dst4q, [dstq+4*gprsize]
+    mov     dst5q, [dstq+5*gprsize]
+    mov      dstq, [dstq          ]
+    sub     dst1q, dstq
+    sub     dst2q, dstq
+    sub     dst3q, dstq
+    sub     dst4q, dstq
+    sub     dst5q, dstq
+.loop:
+    mova       m0, [srcq+0*mmsize]      ; m0 =  0,  1,  2,  3,  4,  5,  6,  7
+    mova       m3, [srcq+1*mmsize]      ; m3 =  8,  9, 10, 11, 12, 13, 14, 15
+    mova       m2, [srcq+2*mmsize]      ; m2 = 16, 17, 18, 19, 20, 21, 22, 23
+    PALIGNR    m1, m3, m0, 12, m4       ; m1 =  6,  7,  8,  9, 10, 11,  x,  x
+    shufps     m3, m2, q1032            ; m3 = 12, 13, 14, 15, 16, 17, 18, 19
+    psrldq     m2, 4                    ; m2 = 18, 19, 20, 21, 22, 23,  x,  x
+    SBUTTERFLY2 wd, 0, 1, 4             ; m0 =  0,  6,  1,  7,  2,  8,  3,  9
+                                        ; m1 =  4, 10,  5, 11,  x,  x,  x,  x
+    SBUTTERFLY2 wd, 3, 2, 4             ; m3 = 12, 18, 13, 19, 14, 20, 15, 21
+                                        ; m2 = 16, 22, 17, 23,  x,  x,  x,  x
+    SBUTTERFLY2 dq, 0, 3, 4             ; m0 =  0,  6, 12, 18,  1,  7, 13, 19
+                                        ; m3 =  2,  8, 14, 20,  3,  9, 15, 21
+    punpckldq  m1, m2                   ; m1 =  4, 10, 16, 22,  5, 11, 17, 23
+    movq    [dstq      ], m0
+    movhps  [dstq+dst1q], m0
+    movq    [dstq+dst2q], m3
+    movhps  [dstq+dst3q], m3
+    movq    [dstq+dst4q], m1
+    movhps  [dstq+dst5q], m1
+    add      srcq, mmsize*3
+    add      dstq, mmsize/2
+    sub      lend, mmsize/4
+    jg .loop
+    REP_RET
+%endmacro
+
+%define PALIGNR PALIGNR_MMX
+INIT_XMM sse2
+CONV_S16_TO_S16P_6CH
+%define PALIGNR PALIGNR_SSSE3
+INIT_XMM ssse3
+CONV_S16_TO_S16P_6CH
+%if HAVE_AVX
+INIT_XMM avx
+CONV_S16_TO_S16P_6CH
+%endif
+
+;------------------------------------------------------------------------------
+; void ff_conv_s16_to_fltp_2ch(float *const *dst, int16_t *src, int len,
+;                              int channels);
+;------------------------------------------------------------------------------
+
+%macro CONV_S16_TO_FLTP_2CH 0
+cglobal conv_s16_to_fltp_2ch, 3,4,5, dst0, src, len, dst1
+    lea       lenq, [4*lend]
+    mov      dst1q, [dst0q+gprsize]
+    mov      dst0q, [dst0q        ]
+    add       srcq, lenq
+    add      dst0q, lenq
+    add      dst1q, lenq
+    neg       lenq
+    mova        m3, [pf_s32_inv_scale]
+    mova        m4, [pw_zero_even]
+.loop:
+    mova        m1, [srcq+lenq]
+    pslld       m0, m1, 16
+    pand        m1, m4
+    cvtdq2ps    m0, m0
+    cvtdq2ps    m1, m1
+    mulps       m0, m0, m3
+    mulps       m1, m1, m3
+    mova  [dst0q+lenq], m0
+    mova  [dst1q+lenq], m1
+    add       lenq, mmsize
+    jl .loop
+    REP_RET
+%endmacro
+
+INIT_XMM sse2
+CONV_S16_TO_FLTP_2CH
+%if HAVE_AVX
+INIT_XMM avx
+CONV_S16_TO_FLTP_2CH
+%endif
+
+;------------------------------------------------------------------------------
+; void ff_conv_s16_to_fltp_6ch(float *const *dst, int16_t *src, int len,
+;                              int channels);
+;------------------------------------------------------------------------------
+
+%macro CONV_S16_TO_FLTP_6CH 0
+%if ARCH_X86_64
+cglobal conv_s16_to_fltp_6ch, 3,8,7, dst, src, len, dst1, dst2, dst3, dst4, dst5
+%else
+cglobal conv_s16_to_fltp_6ch, 2,7,7, dst, src, dst1, dst2, dst3, dst4, dst5
+%define lend dword r2m
+%endif
+    mov     dst1q, [dstq+  gprsize]
+    mov     dst2q, [dstq+2*gprsize]
+    mov     dst3q, [dstq+3*gprsize]
+    mov     dst4q, [dstq+4*gprsize]
+    mov     dst5q, [dstq+5*gprsize]
+    mov      dstq, [dstq          ]
+    sub     dst1q, dstq
+    sub     dst2q, dstq
+    sub     dst3q, dstq
+    sub     dst4q, dstq
+    sub     dst5q, dstq
+    mova       m6, [pf_s16_inv_scale]
+.loop:
+    mova       m0, [srcq+0*mmsize]  ; m0 =  0,  1,  2,  3,  4,  5,  6,  7
+    mova       m3, [srcq+1*mmsize]  ; m3 =  8,  9, 10, 11, 12, 13, 14, 15
+    mova       m2, [srcq+2*mmsize]  ; m2 = 16, 17, 18, 19, 20, 21, 22, 23
+    PALIGNR    m1, m3, m0, 12, m4   ; m1 =  6,  7,  8,  9, 10, 11,  x,  x
+    shufps     m3, m2, q1032        ; m3 = 12, 13, 14, 15, 16, 17, 18, 19
+    psrldq     m2, 4                ; m2 = 18, 19, 20, 21, 22, 23,  x,  x
+    SBUTTERFLY2 wd, 0, 1, 4         ; m0 =  0,  6,  1,  7,  2,  8,  3,  9
+                                    ; m1 =  4, 10,  5, 11,  x,  x,  x,  x
+    SBUTTERFLY2 wd, 3, 2, 4         ; m3 = 12, 18, 13, 19, 14, 20, 15, 21
+                                    ; m2 = 16, 22, 17, 23,  x,  x,  x,  x
+    SBUTTERFLY2 dq, 0, 3, 4         ; m0 =  0,  6, 12, 18,  1,  7, 13, 19
+                                    ; m3 =  2,  8, 14, 20,  3,  9, 15, 21
+    punpckldq  m1, m2               ; m1 =  4, 10, 16, 22,  5, 11, 17, 23
+    S16_TO_S32_SX 0, 2              ; m0 =      0,      6,     12,     18
+                                    ; m2 =      1,      7,     13,     19
+    S16_TO_S32_SX 3, 4              ; m3 =      2,      8,     14,     20
+                                    ; m4 =      3,      9,     15,     21
+    S16_TO_S32_SX 1, 5              ; m1 =      4,     10,     16,     22
+                                    ; m5 =      5,     11,     17,     23
+    SWAP 1,2,3,4
+    cvtdq2ps   m0, m0
+    cvtdq2ps   m1, m1
+    cvtdq2ps   m2, m2
+    cvtdq2ps   m3, m3
+    cvtdq2ps   m4, m4
+    cvtdq2ps   m5, m5
+    mulps      m0, m6
+    mulps      m1, m6
+    mulps      m2, m6
+    mulps      m3, m6
+    mulps      m4, m6
+    mulps      m5, m6
+    mova  [dstq      ], m0
+    mova  [dstq+dst1q], m1
+    mova  [dstq+dst2q], m2
+    mova  [dstq+dst3q], m3
+    mova  [dstq+dst4q], m4
+    mova  [dstq+dst5q], m5
+    add      srcq, mmsize*3
+    add      dstq, mmsize
+    sub      lend, mmsize/4
+    jg .loop
+    REP_RET
+%endmacro
+
+%define PALIGNR PALIGNR_MMX
+INIT_XMM sse2
+CONV_S16_TO_FLTP_6CH
+%define PALIGNR PALIGNR_SSSE3
+INIT_XMM ssse3
+CONV_S16_TO_FLTP_6CH
+INIT_XMM sse4
+CONV_S16_TO_FLTP_6CH
+%if HAVE_AVX
+INIT_XMM avx
+CONV_S16_TO_FLTP_6CH
+%endif
+
+;------------------------------------------------------------------------------
+; void ff_conv_flt_to_s16p_2ch(int16_t *const *dst, float *src, int len,
+;                              int channels);
+;------------------------------------------------------------------------------
+
+%macro CONV_FLT_TO_S16P_2CH 0
+cglobal conv_flt_to_s16p_2ch, 3,4,6, dst0, src, len, dst1
+    lea       lenq, [2*lend]
+    mov      dst1q, [dst0q+gprsize]
+    mov      dst0q, [dst0q        ]
+    lea       srcq, [srcq+4*lenq]
+    add      dst0q, lenq
+    add      dst1q, lenq
+    neg       lenq
+    mova        m5, [pf_s16_scale]
+.loop:
+    mova       m0, [srcq+4*lenq         ]
+    mova       m1, [srcq+4*lenq+  mmsize]
+    mova       m2, [srcq+4*lenq+2*mmsize]
+    mova       m3, [srcq+4*lenq+3*mmsize]
+    DEINT2_PS   0, 1, 4
+    DEINT2_PS   2, 3, 4
+    mulps      m0, m0, m5
+    mulps      m1, m1, m5
+    mulps      m2, m2, m5
+    mulps      m3, m3, m5
+    cvtps2dq   m0, m0
+    cvtps2dq   m1, m1
+    cvtps2dq   m2, m2
+    cvtps2dq   m3, m3
+    packssdw   m0, m2
+    packssdw   m1, m3
+    mova  [dst0q+lenq], m0
+    mova  [dst1q+lenq], m1
+    add      lenq, mmsize
+    jl .loop
+    REP_RET
+%endmacro
+
+INIT_XMM sse2
+CONV_FLT_TO_S16P_2CH
+%if HAVE_AVX
+INIT_XMM avx
+CONV_FLT_TO_S16P_2CH
+%endif
+
+;------------------------------------------------------------------------------
+; void ff_conv_flt_to_s16p_6ch(int16_t *const *dst, float *src, int len,
+;                              int channels);
+;------------------------------------------------------------------------------
+
+%macro CONV_FLT_TO_S16P_6CH 0
+%if ARCH_X86_64
+cglobal conv_flt_to_s16p_6ch, 3,8,7, dst, src, len, dst1, dst2, dst3, dst4, dst5
+%else
+cglobal conv_flt_to_s16p_6ch, 2,7,7, dst, src, dst1, dst2, dst3, dst4, dst5
+%define lend dword r2m
+%endif
+    mov     dst1q, [dstq+  gprsize]
+    mov     dst2q, [dstq+2*gprsize]
+    mov     dst3q, [dstq+3*gprsize]
+    mov     dst4q, [dstq+4*gprsize]
+    mov     dst5q, [dstq+5*gprsize]
+    mov      dstq, [dstq          ]
+    sub     dst1q, dstq
+    sub     dst2q, dstq
+    sub     dst3q, dstq
+    sub     dst4q, dstq
+    sub     dst5q, dstq
+    mova       m6, [pf_s16_scale]
+.loop:
+    mulps      m0, m6, [srcq+0*mmsize]
+    mulps      m3, m6, [srcq+1*mmsize]
+    mulps      m1, m6, [srcq+2*mmsize]
+    mulps      m4, m6, [srcq+3*mmsize]
+    mulps      m2, m6, [srcq+4*mmsize]
+    mulps      m5, m6, [srcq+5*mmsize]
+    cvtps2dq   m0, m0
+    cvtps2dq   m1, m1
+    cvtps2dq   m2, m2
+    cvtps2dq   m3, m3
+    cvtps2dq   m4, m4
+    cvtps2dq   m5, m5
+    packssdw   m0, m3               ; m0 =  0,  1,  2,  3,  4,  5,  6,  7
+    packssdw   m1, m4               ; m1 =  8,  9, 10, 11, 12, 13, 14, 15
+    packssdw   m2, m5               ; m2 = 16, 17, 18, 19, 20, 21, 22, 23
+    PALIGNR    m3, m1, m0, 12, m4   ; m3 =  6,  7,  8,  9, 10, 11,  x,  x
+    shufps     m1, m2, q1032        ; m1 = 12, 13, 14, 15, 16, 17, 18, 19
+    psrldq     m2, 4                ; m2 = 18, 19, 20, 21, 22, 23,  x,  x
+    SBUTTERFLY2 wd, 0, 3, 4         ; m0 =  0,  6,  1,  7,  2,  8,  3,  9
+                                    ; m3 =  4, 10,  5, 11,  x,  x,  x,  x
+    SBUTTERFLY2 wd, 1, 2, 4         ; m1 = 12, 18, 13, 19, 14, 20, 15, 21
+                                    ; m2 = 16, 22, 17, 23,  x,  x,  x,  x
+    SBUTTERFLY2 dq, 0, 1, 4         ; m0 =  0,  6, 12, 18,  1,  7, 13, 19
+                                    ; m1 =  2,  8, 14, 20,  3,  9, 15, 21
+    punpckldq  m3, m2               ; m3 =  4, 10, 16, 22,  5, 11, 17, 23
+    movq    [dstq      ], m0
+    movhps  [dstq+dst1q], m0
+    movq    [dstq+dst2q], m1
+    movhps  [dstq+dst3q], m1
+    movq    [dstq+dst4q], m3
+    movhps  [dstq+dst5q], m3
+    add      srcq, mmsize*6
+    add      dstq, mmsize/2
+    sub      lend, mmsize/4
+    jg .loop
+    REP_RET
+%endmacro
+
+%define PALIGNR PALIGNR_MMX
+INIT_XMM sse2
+CONV_FLT_TO_S16P_6CH
+%define PALIGNR PALIGNR_SSSE3
+INIT_XMM ssse3
+CONV_FLT_TO_S16P_6CH
+%if HAVE_AVX
+INIT_XMM avx
+CONV_FLT_TO_S16P_6CH
+%endif
+
+;------------------------------------------------------------------------------
+; void ff_conv_flt_to_fltp_2ch(float *const *dst, float *src, int len,
+;                              int channels);
+;------------------------------------------------------------------------------
+
+%macro CONV_FLT_TO_FLTP_2CH 0
+cglobal conv_flt_to_fltp_2ch, 3,4,3, dst0, src, len, dst1
+    lea    lenq, [4*lend]
+    mov   dst1q, [dst0q+gprsize]
+    mov   dst0q, [dst0q        ]
+    lea    srcq, [srcq+2*lenq]
+    add   dst0q, lenq
+    add   dst1q, lenq
+    neg    lenq
+.loop:
+    mova     m0, [srcq+2*lenq       ]
+    mova     m1, [srcq+2*lenq+mmsize]
+    DEINT2_PS 0, 1, 2
+    mova  [dst0q+lenq], m0
+    mova  [dst1q+lenq], m1
+    add    lenq, mmsize
+    jl .loop
+    REP_RET
+%endmacro
+
+INIT_XMM sse
+CONV_FLT_TO_FLTP_2CH
+%if HAVE_AVX
+INIT_XMM avx
+CONV_FLT_TO_FLTP_2CH
+%endif
+
+;------------------------------------------------------------------------------
+; void ff_conv_flt_to_fltp_6ch(float *const *dst, float *src, int len,
+;                              int channels);
+;------------------------------------------------------------------------------
+
+%macro CONV_FLT_TO_FLTP_6CH 0
+%if ARCH_X86_64
+cglobal conv_flt_to_fltp_6ch, 3,8,7, dst, src, len, dst1, dst2, dst3, dst4, dst5
+%else
+cglobal conv_flt_to_fltp_6ch, 2,7,7, dst, src, dst1, dst2, dst3, dst4, dst5
+%define lend dword r2m
+%endif
+    mov     dst1q, [dstq+  gprsize]
+    mov     dst2q, [dstq+2*gprsize]
+    mov     dst3q, [dstq+3*gprsize]
+    mov     dst4q, [dstq+4*gprsize]
+    mov     dst5q, [dstq+5*gprsize]
+    mov      dstq, [dstq          ]
+    sub     dst1q, dstq
+    sub     dst2q, dstq
+    sub     dst3q, dstq
+    sub     dst4q, dstq
+    sub     dst5q, dstq
+.loop:
+    mova       m0, [srcq+0*mmsize]  ; m0 =  0,  1,  2,  3
+    mova       m1, [srcq+1*mmsize]  ; m1 =  4,  5,  6,  7
+    mova       m2, [srcq+2*mmsize]  ; m2 =  8,  9, 10, 11
+    mova       m3, [srcq+3*mmsize]  ; m3 = 12, 13, 14, 15
+    mova       m4, [srcq+4*mmsize]  ; m4 = 16, 17, 18, 19
+    mova       m5, [srcq+5*mmsize]  ; m5 = 20, 21, 22, 23
+
+    SBUTTERFLY2 dq, 0, 3, 6         ; m0 =  0, 12,  1, 13
+                                    ; m3 =  2, 14,  3, 15
+    SBUTTERFLY2 dq, 1, 4, 6         ; m1 =  4, 16,  5, 17
+                                    ; m4 =  6, 18,  7, 19
+    SBUTTERFLY2 dq, 2, 5, 6         ; m2 =  8, 20,  9, 21
+                                    ; m5 = 10, 22, 11, 23
+    SBUTTERFLY2 dq, 0, 4, 6         ; m0 =  0,  6, 12, 18
+                                    ; m4 =  1,  7, 13, 19
+    SBUTTERFLY2 dq, 3, 2, 6         ; m3 =  2,  8, 14, 20
+                                    ; m2 =  3,  9, 15, 21
+    SBUTTERFLY2 dq, 1, 5, 6         ; m1 =  4, 10, 16, 22
+                                    ; m5 =  5, 11, 17, 23
+    mova [dstq      ], m0
+    mova [dstq+dst1q], m4
+    mova [dstq+dst2q], m3
+    mova [dstq+dst3q], m2
+    mova [dstq+dst4q], m1
+    mova [dstq+dst5q], m5
+    add      srcq, mmsize*6
+    add      dstq, mmsize
+    sub      lend, mmsize/4
+    jg .loop
+    REP_RET
+%endmacro
+
+INIT_XMM sse2
+CONV_FLT_TO_FLTP_6CH
+%if HAVE_AVX
+INIT_XMM avx
+CONV_FLT_TO_FLTP_6CH
+%endif
diff --git a/libavresample/x86/audio_convert_init.c b/libavresample/x86/audio_convert_init.c
index 2de49709c2..a3589c6957 100644
--- a/libavresample/x86/audio_convert_init.c
+++ b/libavresample/x86/audio_convert_init.c
@@ -22,6 +22,8 @@
 #include "libavutil/cpu.h"
 #include "libavresample/audio_convert.h"
 
+/* flat conversions */
+
 extern void ff_conv_s16_to_s32_sse2(int16_t *dst, const int32_t *src, int len);
 
 extern void ff_conv_s16_to_flt_sse2(float *dst, const int16_t *src, int len);
@@ -38,6 +40,49 @@ extern void ff_conv_flt_to_s16_sse2(int16_t *dst, const float *src, int len);
 extern void ff_conv_flt_to_s32_sse2(int32_t *dst, const float *src, int len);
 extern void ff_conv_flt_to_s32_avx (int32_t *dst, const float *src, int len);
 
+/* interleave conversions */
+
+extern void ff_conv_s16p_to_s16_2ch_sse2(int16_t *dst, int16_t *const *src,
+                                         int len, int channels);
+extern void ff_conv_s16p_to_s16_2ch_avx (int16_t *dst, int16_t *const *src,
+                                         int len, int channels);
+
+extern void ff_conv_s16p_to_s16_6ch_sse2(int16_t *dst, int16_t *const *src,
+                                         int len, int channels);
+extern void ff_conv_s16p_to_s16_6ch_sse2slow(int16_t *dst, int16_t *const *src,
+                                             int len, int channels);
+extern void ff_conv_s16p_to_s16_6ch_avx (int16_t *dst, int16_t *const *src,
+                                         int len, int channels);
+
+extern void ff_conv_s16p_to_flt_2ch_sse2(float *dst, int16_t *const *src,
+                                         int len, int channels);
+extern void ff_conv_s16p_to_flt_2ch_avx (float *dst, int16_t *const *src,
+                                         int len, int channels);
+
+extern void ff_conv_s16p_to_flt_6ch_sse2 (float *dst, int16_t *const *src,
+                                          int len, int channels);
+extern void ff_conv_s16p_to_flt_6ch_ssse3(float *dst, int16_t *const *src,
+                                         int len, int channels);
+extern void ff_conv_s16p_to_flt_6ch_avx  (float *dst, int16_t *const *src,
+                                          int len, int channels);
+
+extern void ff_conv_fltp_to_s16_2ch_sse2 (int16_t *dst, float *const *src,
+                                          int len, int channels);
+extern void ff_conv_fltp_to_s16_2ch_ssse3(int16_t *dst, float *const *src,
+                                          int len, int channels);
+
+extern void ff_conv_fltp_to_s16_6ch_sse (int16_t *dst, float *const *src,
+                                         int len, int channels);
+extern void ff_conv_fltp_to_s16_6ch_sse2(int16_t *dst, float *const *src,
+                                         int len, int channels);
+extern void ff_conv_fltp_to_s16_6ch_avx (int16_t *dst, float *const *src,
+                                         int len, int channels);
+
+extern void ff_conv_fltp_to_flt_2ch_sse(float *dst, float *const *src, int len,
+                                        int channels);
+extern void ff_conv_fltp_to_flt_2ch_avx(float *dst, float *const *src, int len,
+                                        int channels);
+
 extern void ff_conv_fltp_to_flt_6ch_mmx (float *dst, float *const *src, int len,
                                          int channels);
 extern void ff_conv_fltp_to_flt_6ch_sse4(float *dst, float *const *src, int len,
@@ -45,6 +90,58 @@ extern void ff_conv_fltp_to_flt_6ch_sse4(float *dst, float *const *src, int len,
 extern void ff_conv_fltp_to_flt_6ch_avx (float *dst, float *const *src, int len,
                                          int channels);
 
+/* deinterleave conversions */
+
+extern void ff_conv_s16_to_s16p_2ch_sse2(int16_t *const *dst, int16_t *src,
+                                         int len, int channels);
+extern void ff_conv_s16_to_s16p_2ch_ssse3(int16_t *const *dst, int16_t *src,
+                                          int len, int channels);
+extern void ff_conv_s16_to_s16p_2ch_avx (int16_t *const *dst, int16_t *src,
+                                         int len, int channels);
+
+extern void ff_conv_s16_to_s16p_6ch_sse2 (int16_t *const *dst, int16_t *src,
+                                          int len, int channels);
+extern void ff_conv_s16_to_s16p_6ch_ssse3(int16_t *const *dst, int16_t *src,
+                                          int len, int channels);
+extern void ff_conv_s16_to_s16p_6ch_avx  (int16_t *const *dst, int16_t *src,
+                                          int len, int channels);
+
+extern void ff_conv_s16_to_fltp_2ch_sse2(float *const *dst, int16_t *src,
+                                         int len, int channels);
+extern void ff_conv_s16_to_fltp_2ch_avx (float *const *dst, int16_t *src,
+                                         int len, int channels);
+
+extern void ff_conv_s16_to_fltp_6ch_sse2 (float *const *dst, int16_t *src,
+                                          int len, int channels);
+extern void ff_conv_s16_to_fltp_6ch_ssse3(float *const *dst, int16_t *src,
+                                          int len, int channels);
+extern void ff_conv_s16_to_fltp_6ch_sse4 (float *const *dst, int16_t *src,
+                                          int len, int channels);
+extern void ff_conv_s16_to_fltp_6ch_avx  (float *const *dst, int16_t *src,
+                                          int len, int channels);
+
+extern void ff_conv_flt_to_s16p_2ch_sse2(int16_t *const *dst, float *src,
+                                         int len, int channels);
+extern void ff_conv_flt_to_s16p_2ch_avx (int16_t *const *dst, float *src,
+                                         int len, int channels);
+
+extern void ff_conv_flt_to_s16p_6ch_sse2 (int16_t *const *dst, float *src,
+                                          int len, int channels);
+extern void ff_conv_flt_to_s16p_6ch_ssse3(int16_t *const *dst, float *src,
+                                          int len, int channels);
+extern void ff_conv_flt_to_s16p_6ch_avx  (int16_t *const *dst, float *src,
+                                          int len, int channels);
+
+extern void ff_conv_flt_to_fltp_2ch_sse(float *const *dst, float *src, int len,
+                                        int channels);
+extern void ff_conv_flt_to_fltp_2ch_avx(float *const *dst, float *src, int len,
+                                        int channels);
+
+extern void ff_conv_flt_to_fltp_6ch_sse2(float *const *dst, float *src, int len,
+                                         int channels);
+extern void ff_conv_flt_to_fltp_6ch_avx (float *const *dst, float *src, int len,
+                                         int channels);
+
 av_cold void ff_audio_convert_init_x86(AudioConvert *ac)
 {
 #if HAVE_YASM
@@ -56,10 +153,25 @@ av_cold void ff_audio_convert_init_x86(AudioConvert *ac)
         ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_FLTP,
                                   6, 1, 4, "MMX", ff_conv_fltp_to_flt_6ch_mmx);
     }
+    if (mm_flags & AV_CPU_FLAG_SSE && HAVE_SSE) {
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_FLTP,
+                                  6, 1, 2, "SSE", ff_conv_fltp_to_s16_6ch_sse);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_FLTP,
+                                  2, 16, 8, "SSE", ff_conv_fltp_to_flt_2ch_sse);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_FLTP, AV_SAMPLE_FMT_FLT,
+                                  2, 16, 4, "SSE", ff_conv_flt_to_fltp_2ch_sse);
+    }
     if (mm_flags & AV_CPU_FLAG_SSE2 && HAVE_SSE) {
         if (!(mm_flags & AV_CPU_FLAG_SSE2SLOW)) {
             ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_S32,
                                       0, 16, 16, "SSE2", ff_conv_s32_to_s16_sse2);
+            ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_S16P,
+                                      6, 16, 8, "SSE2", ff_conv_s16p_to_s16_6ch_sse2);
+            ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_FLTP,
+                                      6, 16, 4, "SSE2", ff_conv_fltp_to_s16_6ch_sse2);
+        } else {
+            ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_S16P,
+                                      6, 1, 4, "SSE2SLOW", ff_conv_s16p_to_s16_6ch_sse2slow);
         }
         ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S32, AV_SAMPLE_FMT_S16,
                                   0, 16, 8, "SSE2", ff_conv_s16_to_s32_sse2);
@@ -71,6 +183,42 @@ av_cold void ff_audio_convert_init_x86(AudioConvert *ac)
                                   0, 16, 16, "SSE2", ff_conv_flt_to_s16_sse2);
         ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S32, AV_SAMPLE_FMT_FLT,
                                   0, 16, 16, "SSE2", ff_conv_flt_to_s32_sse2);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_S16P,
+                                  2, 16, 16, "SSE2", ff_conv_s16p_to_s16_2ch_sse2);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_S16P,
+                                  2, 16, 8, "SSE2", ff_conv_s16p_to_flt_2ch_sse2);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_S16P,
+                                  6, 16, 4, "SSE2", ff_conv_s16p_to_flt_6ch_sse2);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_FLTP,
+                                  2, 16, 4, "SSE2", ff_conv_fltp_to_s16_2ch_sse2);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S16P, AV_SAMPLE_FMT_S16,
+                                  2, 16, 8, "SSE2", ff_conv_s16_to_s16p_2ch_sse2);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S16P, AV_SAMPLE_FMT_S16,
+                                  6, 16, 4, "SSE2", ff_conv_s16_to_s16p_6ch_sse2);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_FLTP, AV_SAMPLE_FMT_S16,
+                                  2, 16, 8, "SSE2", ff_conv_s16_to_fltp_2ch_sse2);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_FLTP, AV_SAMPLE_FMT_S16,
+                                  6, 16, 4, "SSE2", ff_conv_s16_to_fltp_6ch_sse2);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S16P, AV_SAMPLE_FMT_FLT,
+                                  2, 16, 8, "SSE2", ff_conv_flt_to_s16p_2ch_sse2);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S16P, AV_SAMPLE_FMT_FLT,
+                                  6, 16, 4, "SSE2", ff_conv_flt_to_s16p_6ch_sse2);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_FLTP, AV_SAMPLE_FMT_FLT,
+                                  6, 16, 4, "SSE2", ff_conv_flt_to_fltp_6ch_sse2);
+    }
+    if (mm_flags & AV_CPU_FLAG_SSSE3 && HAVE_SSE) {
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_S16P,
+                                  6, 16, 4, "SSSE3", ff_conv_s16p_to_flt_6ch_ssse3);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_FLTP,
+                                  2, 16, 4, "SSSE3", ff_conv_fltp_to_s16_2ch_ssse3);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S16P, AV_SAMPLE_FMT_S16,
+                                  2, 16, 8, "SSSE3", ff_conv_s16_to_s16p_2ch_ssse3);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S16P, AV_SAMPLE_FMT_S16,
+                                  6, 16, 4, "SSSE3", ff_conv_s16_to_s16p_6ch_ssse3);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_FLTP, AV_SAMPLE_FMT_S16,
+                                  6, 16, 4, "SSSE3", ff_conv_s16_to_fltp_6ch_ssse3);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S16P, AV_SAMPLE_FMT_FLT,
+                                  6, 16, 4, "SSSE3", ff_conv_flt_to_s16p_6ch_ssse3);
     }
     if (mm_flags & AV_CPU_FLAG_SSE4 && HAVE_SSE) {
         ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_S16,
@@ -83,8 +231,34 @@ av_cold void ff_audio_convert_init_x86(AudioConvert *ac)
                                   0, 32, 16, "AVX", ff_conv_s32_to_flt_avx);
         ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S32, AV_SAMPLE_FMT_FLT,
                                   0, 32, 32, "AVX", ff_conv_flt_to_s32_avx);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_S16P,
+                                  2, 16, 16, "AVX", ff_conv_s16p_to_s16_2ch_avx);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_S16P,
+                                  6, 16, 8, "AVX", ff_conv_s16p_to_s16_6ch_avx);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_S16P,
+                                  2, 16, 8, "AVX", ff_conv_s16p_to_flt_2ch_avx);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_S16P,
+                                  6, 16, 4, "AVX", ff_conv_s16p_to_flt_6ch_avx);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_FLTP,
+                                  6, 16, 4, "AVX", ff_conv_fltp_to_s16_6ch_avx);
         ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_FLTP,
                                   6, 16, 4, "AVX", ff_conv_fltp_to_flt_6ch_avx);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S16P, AV_SAMPLE_FMT_S16,
+                                  2, 16, 8, "AVX", ff_conv_s16_to_s16p_2ch_avx);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S16P, AV_SAMPLE_FMT_S16,
+                                  6, 16, 4, "AVX", ff_conv_s16_to_s16p_6ch_avx);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_FLTP, AV_SAMPLE_FMT_S16,
+                                  2, 16, 8, "AVX", ff_conv_s16_to_fltp_2ch_avx);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_FLTP, AV_SAMPLE_FMT_S16,
+                                  6, 16, 4, "AVX", ff_conv_s16_to_fltp_6ch_avx);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S16P, AV_SAMPLE_FMT_FLT,
+                                  2, 16, 8, "AVX", ff_conv_flt_to_s16p_2ch_avx);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_S16P, AV_SAMPLE_FMT_FLT,
+                                  6, 16, 4, "AVX", ff_conv_flt_to_s16p_6ch_avx);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_FLTP, AV_SAMPLE_FMT_FLT,
+                                  2, 16, 4, "AVX", ff_conv_flt_to_fltp_2ch_avx);
+        ff_audio_convert_set_func(ac, AV_SAMPLE_FMT_FLTP, AV_SAMPLE_FMT_FLT,
+                                  6, 16, 4, "AVX", ff_conv_flt_to_fltp_6ch_avx);
     }
 #endif
 }
diff --git a/libavresample/x86/util.asm b/libavresample/x86/util.asm
index ca7fde513a..0ce953159c 100644
--- a/libavresample/x86/util.asm
+++ b/libavresample/x86/util.asm
@@ -33,3 +33,9 @@
     psrad        m%1, 16
 %endif
 %endmacro
+
+%macro DEINT2_PS 3 ; src0/even dst, src1/odd dst, temp
+    shufps  m%3, m%1, m%2, q3131
+    shufps       m%1, m%2, q2020
+    SWAP %2,%3
+%endmacro
diff --git a/libavutil/x86/x86util.asm b/libavutil/x86/x86util.asm
index 979ae41577..cf5ea363f6 100644
--- a/libavutil/x86/x86util.asm
+++ b/libavutil/x86/x86util.asm
@@ -637,3 +637,15 @@
         %rotate 1
     %endrep
 %endmacro
+
+%macro PMOVSXWD 2; dst, src
+%if cpuflag(sse4)
+    pmovsxwd     %1, %2
+%else
+    %ifnidn %1, %2
+    mova         %1, %2
+    %endif
+    punpcklwd    %1, %1
+    psrad        %1, 16
+%endif
+%endmacro