h264: DSP'ize MBAFF loopfilter.

[ffmpeg.git] / libavcodec / h264dsp.c

diff --git a/libavcodec/h264dsp.c b/libavcodec/h264dsp.c
index b0821caa23d9d91992c1d33d936e75cf9d33d4c1..1ef6a26db4187baf0a9703c88ce64cd7948b4398 100644 (file)
--- a/libavcodec/h264dsp.c
+++ b/libavcodec/h264dsp.c
@@ -2,25 +2,25 @@
  * H.26L/H.264/AVC/JVT/14496-10/... encoder/decoder
  * Copyright (c) 2003-2010 Michael Niedermayer <michaelni@gmx.at>
  *
  * H.26L/H.264/AVC/JVT/14496-10/... encoder/decoder
  * Copyright (c) 2003-2010 Michael Niedermayer <michaelni@gmx.at>
  *

- * This file is part of FFmpeg.
+ * This file is part of Libav.

  *
  *

- * FFmpeg is free software; you can redistribute it and/or
+ * Libav is free software; you can redistribute it and/or

  * modify it under the terms of the GNU Lesser General Public
  * License as published by the Free Software Foundation; either
  * version 2.1 of the License, or (at your option) any later version.
  *
  * 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,
+ * Libav is distributed in the hope that it will be useful,

  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  * Lesser General Public License for more details.
  *
  * You should have received a copy of the GNU Lesser General Public
  * 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
+ * License along with Libav; if not, write to the Free Software

  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  */
 
 /**
  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  */
 
 /**

- * @file libavcodec/h264dsp.c
+ * @file

  * H.264 / AVC / MPEG4 part10 DSP functions.
  * @author Michael Niedermayer <michaelni@gmx.at>
  */
  * H.264 / AVC / MPEG4 part10 DSP functions.
  * @author Michael Niedermayer <michaelni@gmx.at>
  */


@@ -99,15 +99,15 @@ H264_WEIGHT(2,2)
 #undef op_scale2
 #undef H264_WEIGHT
 
 #undef op_scale2
 #undef H264_WEIGHT
 

-static av_always_inline av_flatten void h264_loop_filter_luma_c(uint8_t *pix, int xstride, int ystride, int alpha, int beta, int8_t *tc0)
+static av_always_inline av_flatten void h264_loop_filter_luma_c(uint8_t *pix, int xstride, int ystride, int inner_iters, int alpha, int beta, int8_t *tc0)

 {
     int i, d;
     for( i = 0; i < 4; i++ ) {
         if( tc0[i] < 0 ) {
 {
     int i, d;
     for( i = 0; i < 4; i++ ) {
         if( tc0[i] < 0 ) {

-            pix += 4*ystride;
+            pix += inner_iters*ystride;

             continue;
         }
             continue;
         }

-        for( d = 0; d < 4; d++ ) {
+        for( d = 0; d < inner_iters; d++ ) {

             const int p0 = pix[-1*xstride];
             const int p1 = pix[-2*xstride];
             const int p2 = pix[-3*xstride];
             const int p0 = pix[-1*xstride];
             const int p1 = pix[-2*xstride];
             const int p2 = pix[-3*xstride];


@@ -143,17 +143,21 @@ static av_always_inline av_flatten void h264_loop_filter_luma_c(uint8_t *pix, in
 }
 static void h264_v_loop_filter_luma_c(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0)
 {
 }
 static void h264_v_loop_filter_luma_c(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0)
 {

-    h264_loop_filter_luma_c(pix, stride, 1, alpha, beta, tc0);
+    h264_loop_filter_luma_c(pix, stride, 1, 4, alpha, beta, tc0);

 }
 static void h264_h_loop_filter_luma_c(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0)
 {
 }
 static void h264_h_loop_filter_luma_c(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0)
 {

-    h264_loop_filter_luma_c(pix, 1, stride, alpha, beta, tc0);
+    h264_loop_filter_luma_c(pix, 1, stride, 4, alpha, beta, tc0);
+}
+static void h264_h_loop_filter_luma_mbaff_c(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0)
+{
+    h264_loop_filter_luma_c(pix, 1, stride, 2, alpha, beta, tc0);

 }
 
 }
 

-static av_always_inline av_flatten void h264_loop_filter_luma_intra_c(uint8_t *pix, int xstride, int ystride, int alpha, int beta)
+static av_always_inline av_flatten void h264_loop_filter_luma_intra_c(uint8_t *pix, int xstride, int ystride, int inner_iters, int alpha, int beta)

 {
     int d;
 {
     int d;

-    for( d = 0; d < 16; d++ ) {
+    for( d = 0; d < 4 * inner_iters; d++ ) {

         const int p2 = pix[-3*xstride];
         const int p1 = pix[-2*xstride];
         const int p0 = pix[-1*xstride];
         const int p2 = pix[-3*xstride];
         const int p1 = pix[-2*xstride];
         const int p0 = pix[-1*xstride];


@@ -200,23 +204,27 @@ static av_always_inline av_flatten void h264_loop_filter_luma_intra_c(uint8_t *p
 }
 static void h264_v_loop_filter_luma_intra_c(uint8_t *pix, int stride, int alpha, int beta)
 {
 }
 static void h264_v_loop_filter_luma_intra_c(uint8_t *pix, int stride, int alpha, int beta)
 {

-    h264_loop_filter_luma_intra_c(pix, stride, 1, alpha, beta);
+    h264_loop_filter_luma_intra_c(pix, stride, 1, 4, alpha, beta);

 }
 static void h264_h_loop_filter_luma_intra_c(uint8_t *pix, int stride, int alpha, int beta)
 {
 }
 static void h264_h_loop_filter_luma_intra_c(uint8_t *pix, int stride, int alpha, int beta)
 {

-    h264_loop_filter_luma_intra_c(pix, 1, stride, alpha, beta);
+    h264_loop_filter_luma_intra_c(pix, 1, stride, 4, alpha, beta);
+}
+static void h264_h_loop_filter_luma_mbaff_intra_c(uint8_t *pix, int stride, int alpha, int beta)
+{
+    h264_loop_filter_luma_intra_c(pix, 1, stride, 2, alpha, beta);

 }
 
 }
 

-static av_always_inline av_flatten void h264_loop_filter_chroma_c(uint8_t *pix, int xstride, int ystride, int alpha, int beta, int8_t *tc0)
+static av_always_inline av_flatten void h264_loop_filter_chroma_c(uint8_t *pix, int xstride, int ystride, int inner_iters, int alpha, int beta, int8_t *tc0)

 {
     int i, d;
     for( i = 0; i < 4; i++ ) {
         const int tc = tc0[i];
         if( tc <= 0 ) {
 {
     int i, d;
     for( i = 0; i < 4; i++ ) {
         const int tc = tc0[i];
         if( tc <= 0 ) {

-            pix += 2*ystride;
+            pix += inner_iters*ystride;

             continue;
         }
             continue;
         }

-        for( d = 0; d < 2; d++ ) {
+        for( d = 0; d < inner_iters; d++ ) {

             const int p0 = pix[-1*xstride];
             const int p1 = pix[-2*xstride];
             const int q0 = pix[0];
             const int p0 = pix[-1*xstride];
             const int p1 = pix[-2*xstride];
             const int q0 = pix[0];


@@ -237,17 +245,21 @@ static av_always_inline av_flatten void h264_loop_filter_chroma_c(uint8_t *pix,
 }
 static void h264_v_loop_filter_chroma_c(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0)
 {
 }
 static void h264_v_loop_filter_chroma_c(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0)
 {

-    h264_loop_filter_chroma_c(pix, stride, 1, alpha, beta, tc0);
+    h264_loop_filter_chroma_c(pix, stride, 1, 2, alpha, beta, tc0);

 }
 static void h264_h_loop_filter_chroma_c(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0)
 {
 }
 static void h264_h_loop_filter_chroma_c(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0)
 {

-    h264_loop_filter_chroma_c(pix, 1, stride, alpha, beta, tc0);
+    h264_loop_filter_chroma_c(pix, 1, stride, 2, alpha, beta, tc0);
+}
+static void h264_h_loop_filter_chroma_mbaff_c(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0)
+{
+    h264_loop_filter_chroma_c(pix, 1, stride, 1, alpha, beta, tc0);

 }
 
 }
 

-static av_always_inline av_flatten void h264_loop_filter_chroma_intra_c(uint8_t *pix, int xstride, int ystride, int alpha, int beta)
+static av_always_inline av_flatten void h264_loop_filter_chroma_intra_c(uint8_t *pix, int xstride, int ystride, int inner_iters, int alpha, int beta)

 {
     int d;
 {
     int d;

-    for( d = 0; d < 8; d++ ) {
+    for( d = 0; d < 4 * inner_iters; d++ ) {

         const int p0 = pix[-1*xstride];
         const int p1 = pix[-2*xstride];
         const int q0 = pix[0];
         const int p0 = pix[-1*xstride];
         const int p1 = pix[-2*xstride];
         const int q0 = pix[0];


@@ -265,11 +277,15 @@ static av_always_inline av_flatten void h264_loop_filter_chroma_intra_c(uint8_t
 }
 static void h264_v_loop_filter_chroma_intra_c(uint8_t *pix, int stride, int alpha, int beta)
 {
 }
 static void h264_v_loop_filter_chroma_intra_c(uint8_t *pix, int stride, int alpha, int beta)
 {

-    h264_loop_filter_chroma_intra_c(pix, stride, 1, alpha, beta);
+    h264_loop_filter_chroma_intra_c(pix, stride, 1, 2, alpha, beta);

 }
 static void h264_h_loop_filter_chroma_intra_c(uint8_t *pix, int stride, int alpha, int beta)
 {
 }
 static void h264_h_loop_filter_chroma_intra_c(uint8_t *pix, int stride, int alpha, int beta)
 {

-    h264_loop_filter_chroma_intra_c(pix, 1, stride, alpha, beta);
+    h264_loop_filter_chroma_intra_c(pix, 1, stride, 2, alpha, beta);
+}
+static void h264_h_loop_filter_chroma_mbaff_intra_c(uint8_t *pix, int stride, int alpha, int beta)
+{
+    h264_loop_filter_chroma_intra_c(pix, 1, stride, 1, alpha, beta);

 }
 
 void ff_h264dsp_init(H264DSPContext *c)
 }
 
 void ff_h264dsp_init(H264DSPContext *c)


@@ -282,6 +298,7 @@ void ff_h264dsp_init(H264DSPContext *c)
     c->h264_idct8_add4     = ff_h264_idct8_add4_c;
     c->h264_idct_add8      = ff_h264_idct_add8_c;
     c->h264_idct_add16intra= ff_h264_idct_add16intra_c;
     c->h264_idct8_add4     = ff_h264_idct8_add4_c;
     c->h264_idct_add8      = ff_h264_idct_add8_c;
     c->h264_idct_add16intra= ff_h264_idct_add16intra_c;

+    c->h264_luma_dc_dequant_idct= ff_h264_luma_dc_dequant_idct_c;

 
     c->weight_h264_pixels_tab[0]= weight_h264_pixels16x16_c;
     c->weight_h264_pixels_tab[1]= weight_h264_pixels16x8_c;
 
     c->weight_h264_pixels_tab[0]= weight_h264_pixels16x16_c;
     c->weight_h264_pixels_tab[1]= weight_h264_pixels16x8_c;


@@ -306,12 +323,16 @@ void ff_h264dsp_init(H264DSPContext *c)
 
     c->h264_v_loop_filter_luma= h264_v_loop_filter_luma_c;
     c->h264_h_loop_filter_luma= h264_h_loop_filter_luma_c;
 
     c->h264_v_loop_filter_luma= h264_v_loop_filter_luma_c;
     c->h264_h_loop_filter_luma= h264_h_loop_filter_luma_c;

+    c->h264_h_loop_filter_luma_mbaff= h264_h_loop_filter_luma_mbaff_c;

     c->h264_v_loop_filter_luma_intra= h264_v_loop_filter_luma_intra_c;
     c->h264_h_loop_filter_luma_intra= h264_h_loop_filter_luma_intra_c;
     c->h264_v_loop_filter_luma_intra= h264_v_loop_filter_luma_intra_c;
     c->h264_h_loop_filter_luma_intra= h264_h_loop_filter_luma_intra_c;

+    c->h264_h_loop_filter_luma_mbaff_intra= h264_h_loop_filter_luma_mbaff_intra_c;

     c->h264_v_loop_filter_chroma= h264_v_loop_filter_chroma_c;
     c->h264_h_loop_filter_chroma= h264_h_loop_filter_chroma_c;
     c->h264_v_loop_filter_chroma= h264_v_loop_filter_chroma_c;
     c->h264_h_loop_filter_chroma= h264_h_loop_filter_chroma_c;

+    c->h264_h_loop_filter_chroma_mbaff= h264_h_loop_filter_chroma_mbaff_c;

     c->h264_v_loop_filter_chroma_intra= h264_v_loop_filter_chroma_intra_c;
     c->h264_h_loop_filter_chroma_intra= h264_h_loop_filter_chroma_intra_c;
     c->h264_v_loop_filter_chroma_intra= h264_v_loop_filter_chroma_intra_c;
     c->h264_h_loop_filter_chroma_intra= h264_h_loop_filter_chroma_intra_c;

+    c->h264_h_loop_filter_chroma_mbaff_intra= h264_h_loop_filter_chroma_mbaff_intra_c;

     c->h264_loop_filter_strength= NULL;
 
     if (ARCH_ARM) ff_h264dsp_init_arm(c);
     c->h264_loop_filter_strength= NULL;
 
     if (ARCH_ARM) ff_h264dsp_init_arm(c);