Merge remote-tracking branch 'qatar/master'

* qatar/master:
  dnxhddec: avoid a branch in 10-bit decode_dct_block()
  H.264: Add optimizations to predict x86 assembly.
  riff: Add mpgv MPEG-2 fourcc
  add Flash Screen Video 2 decoder

Conflicts:
	configure
	doc/general.texi
	libavcodec/Makefile
	libavcodec/allcodecs.c
	libavcodec/version.h
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Changelog

@@ -6,6 +6,7 @@ version next:
 - openal input device added
 - boxblur filter added
 - BWF muxer
+- Flash Screen Video 2 decoder
 
 
 version 0.8:

configure

@@ -1280,6 +1280,7 @@ flac_encoder_select="golomb lpc"
 flashsv_decoder_select="zlib"
 flashsv_encoder_select="zlib"
 flashsv2_encoder_select="zlib"
+flashsv2_decoder_select="zlib"
 flv_decoder_select="h263_decoder"
 flv_encoder_select="h263_encoder"
 fraps_decoder_select="huffman"

doc/general.texi

@@ -401,7 +401,7 @@ following image formats are supported:
     @tab experimental lossless codec (fourcc: FFV1)
 @item Flash Screen Video v1  @tab  X  @tab  X
     @tab fourcc: FSV1
-@item Flash Screen Video v2  @tab  X
+@item Flash Screen Video v2  @tab  X  @tab  X
 @item Flash Video (FLV)      @tab  X  @tab  X
     @tab Sorenson H.263 used in Flash
 @item Fraps                  @tab     @tab  X

libavcodec/Makefile

@@ -153,6 +153,7 @@ OBJS-$(CONFIG_FLAC_ENCODER)            += flacenc.o flacdata.o flac.o vorbis_dat
 OBJS-$(CONFIG_FLASHSV_DECODER)         += flashsv.o
 OBJS-$(CONFIG_FLASHSV_ENCODER)         += flashsvenc.o
 OBJS-$(CONFIG_FLASHSV2_ENCODER)        += flashsv2enc.o
+OBJS-$(CONFIG_FLASHSV2_DECODER)        += flashsv.o
 OBJS-$(CONFIG_FLIC_DECODER)            += flicvideo.o
 OBJS-$(CONFIG_FOURXM_DECODER)          += 4xm.o
 OBJS-$(CONFIG_FRAPS_DECODER)           += fraps.o

libavcodec/allcodecs.c

@@ -109,7 +109,7 @@ void avcodec_register_all(void)
     REGISTER_ENCDEC  (FFV1, ffv1);
     REGISTER_ENCDEC  (FFVHUFF, ffvhuff);
     REGISTER_ENCDEC  (FLASHSV, flashsv);
-    REGISTER_ENCODER (FLASHSV2, flashsv2);
+    REGISTER_ENCDEC (FLASHSV2, flashsv2);
     REGISTER_DECODER (FLIC, flic);
     REGISTER_ENCDEC  (FLV, flv);
     REGISTER_DECODER (FOURXM, fourxm);

libavcodec/dnxhddec.c

@@ -246,7 +246,7 @@ static av_always_inline void dnxhd_decode_dct_block(DNXHDContext *ctx,
         //av_log(ctx->avctx, AV_LOG_DEBUG, "j %d\n", j);
         //av_log(ctx->avctx, AV_LOG_DEBUG, "level %d, weight %d\n", level, weight_matrix[i]);
         level = (2*level+1) * qscale * weight_matrix[i];
-        if (weight_matrix[i] != level_bias)
+        if (level_bias < 32 || weight_matrix[i] != level_bias)
             level += level_bias;
         level >>= level_shift;
 

libavcodec/version.h

@@ -21,7 +21,7 @@
 #define AVCODEC_VERSION_H
 
 #define LIBAVCODEC_VERSION_MAJOR 53
-#define LIBAVCODEC_VERSION_MINOR  8
+#define LIBAVCODEC_VERSION_MINOR  9
 #define LIBAVCODEC_VERSION_MICRO  0
 
 #define LIBAVCODEC_VERSION_INT  AV_VERSION_INT(LIBAVCODEC_VERSION_MAJOR, \

libavcodec/x86/h264_intrapred.asm

@@ -2611,12 +2611,11 @@ cglobal pred4x4_down_left_mmxext, 3,3
     punpckldq m1, [r1]
     movq      m2, m1
     movq      m3, m1
-    movq      m4, m1
     psllq     m1, 8
     pxor      m2, m1
     psrlq     m2, 8
-    pxor      m3, m2
-    PRED4x4_LOWPASS m0, m1, m3, m4, m5
+    pxor      m2, m3
+    PRED4x4_LOWPASS m0, m1, m2, m3, m4
     lea       r1, [r0+r2*2]
     psrlq     m0, 8
     movd      [r0+r2*1], m0

libavcodec/x86/h264_intrapred_init.c

@@ -45,7 +45,6 @@ void ff_pred8x8_ ## TYPE ## _ ## DEPTH ## _ ## OPT (uint8_t *src, int stride);
 
 PRED8x8(dc, 10, mmxext)
 PRED8x8(dc, 10, sse2)
-PRED8x8(top_dc, 10, mmxext)
 PRED8x8(top_dc, 10, sse2)
 PRED8x8(plane, 10, sse2)
 PRED8x8(vertical, 10, sse2)
@@ -55,23 +54,28 @@ PRED8x8(horizontal, 10, sse2)
 void ff_pred8x8l_ ## TYPE ## _ ## DEPTH ## _ ## OPT (uint8_t *src, int has_topleft, int has_topright, int stride);
 
 PRED8x8L(dc, 10, sse2)
-PRED8x8L(dc, 10, ssse3)
+PRED8x8L(dc, 10, avx)
 PRED8x8L(128_dc, 10, mmxext)
 PRED8x8L(128_dc, 10, sse2)
 PRED8x8L(top_dc, 10, sse2)
-PRED8x8L(top_dc, 10, ssse3)
+PRED8x8L(top_dc, 10, avx)
 PRED8x8L(vertical, 10, sse2)
-PRED8x8L(vertical, 10, ssse3)
+PRED8x8L(vertical, 10, avx)
 PRED8x8L(horizontal, 10, sse2)
 PRED8x8L(horizontal, 10, ssse3)
+PRED8x8L(horizontal, 10, avx)
 PRED8x8L(down_left, 10, sse2)
 PRED8x8L(down_left, 10, ssse3)
+PRED8x8L(down_left, 10, avx)
 PRED8x8L(down_right, 10, sse2)
 PRED8x8L(down_right, 10, ssse3)
+PRED8x8L(down_right, 10, avx)
 PRED8x8L(vertical_right, 10, sse2)
 PRED8x8L(vertical_right, 10, ssse3)
+PRED8x8L(vertical_right, 10, avx)
 PRED8x8L(horizontal_up, 10, sse2)
 PRED8x8L(horizontal_up, 10, ssse3)
+PRED8x8L(horizontal_up, 10, avx)
 
 #define PRED16x16(TYPE, DEPTH, OPT)\
 void ff_pred16x16_ ## TYPE ## _ ## DEPTH ## _ ## OPT (uint8_t *src, int stride);
@@ -298,7 +302,6 @@ void ff_h264_pred_init_x86(H264PredContext *h, int codec_id, const int bit_depth
             h->pred4x4[HOR_UP_PRED         ] = ff_pred4x4_horizontal_up_10_mmxext;
 
             h->pred8x8[DC_PRED8x8          ] = ff_pred8x8_dc_10_mmxext;
-            h->pred8x8[TOP_DC_PRED8x8      ] = ff_pred8x8_top_dc_10_mmxext;
 
             h->pred8x8l[DC_128_PRED        ] = ff_pred8x8l_128_dc_10_mmxext;
 
@@ -344,18 +347,28 @@ void ff_h264_pred_init_x86(H264PredContext *h, int codec_id, const int bit_depth
             h->pred4x4[VERT_RIGHT_PRED     ] = ff_pred4x4_vertical_right_10_ssse3;
             h->pred4x4[HOR_DOWN_PRED       ] = ff_pred4x4_horizontal_down_10_ssse3;
 
-            h->pred8x8l[VERT_PRED           ] = ff_pred8x8l_vertical_10_ssse3;
             h->pred8x8l[HOR_PRED            ] = ff_pred8x8l_horizontal_10_ssse3;
-            h->pred8x8l[DC_PRED             ] = ff_pred8x8l_dc_10_ssse3;
-            h->pred8x8l[TOP_DC_PRED         ] = ff_pred8x8l_top_dc_10_ssse3;
             h->pred8x8l[DIAG_DOWN_LEFT_PRED ] = ff_pred8x8l_down_left_10_ssse3;
+            h->pred8x8l[DIAG_DOWN_RIGHT_PRED] = ff_pred8x8l_down_right_10_ssse3;
+            h->pred8x8l[VERT_RIGHT_PRED     ] = ff_pred8x8l_vertical_right_10_ssse3;
+            h->pred8x8l[HOR_UP_PRED         ] = ff_pred8x8l_horizontal_up_10_ssse3;
         }
 #if HAVE_AVX
         if (mm_flags & AV_CPU_FLAG_AVX) {
             h->pred4x4[DIAG_DOWN_LEFT_PRED ] = ff_pred4x4_down_left_10_avx;
             h->pred4x4[DIAG_DOWN_RIGHT_PRED] = ff_pred4x4_down_right_10_avx;
+            h->pred4x4[VERT_LEFT_PRED      ] = ff_pred4x4_vertical_left_10_avx;
             h->pred4x4[VERT_RIGHT_PRED     ] = ff_pred4x4_vertical_right_10_avx;
             h->pred4x4[HOR_DOWN_PRED       ] = ff_pred4x4_horizontal_down_10_avx;
+
+            h->pred8x8l[VERT_PRED           ] = ff_pred8x8l_vertical_10_avx;
+            h->pred8x8l[HOR_PRED            ] = ff_pred8x8l_horizontal_10_avx;
+            h->pred8x8l[DC_PRED             ] = ff_pred8x8l_dc_10_avx;
+            h->pred8x8l[TOP_DC_PRED         ] = ff_pred8x8l_top_dc_10_avx;
+            h->pred8x8l[DIAG_DOWN_RIGHT_PRED] = ff_pred8x8l_down_right_10_avx;
+            h->pred8x8l[DIAG_DOWN_LEFT_PRED ] = ff_pred8x8l_down_left_10_avx;
+            h->pred8x8l[VERT_RIGHT_PRED     ] = ff_pred8x8l_vertical_right_10_avx;
+            h->pred8x8l[HOR_UP_PRED         ] = ff_pred8x8l_horizontal_up_10_avx;
         }
 #endif /* HAVE_AVX */
     }

libavformat/riff.c

@@ -131,6 +131,7 @@ const AVCodecTag ff_codec_bmp_tags[] = {
     { CODEC_ID_MPEG2VIDEO,   MKTAG('s', 'l', 'i', 'f') },
     { CODEC_ID_MPEG2VIDEO,   MKTAG('E', 'M', '2', 'V') },
     { CODEC_ID_MPEG2VIDEO,   MKTAG('M', '7', '0', '1') }, /* Matrox MPEG2 intra-only */
+    { CODEC_ID_MPEG2VIDEO,   MKTAG('m', 'p', 'g', 'v') },
     { CODEC_ID_MJPEG,        MKTAG('M', 'J', 'P', 'G') },
     { CODEC_ID_MJPEG,        MKTAG('L', 'J', 'P', 'G') },
     { CODEC_ID_MJPEG,        MKTAG('d', 'm', 'b', '1') },