date:20170407

[FFmpeg-cvslog] Merge commit '239d02eff3ffe9f7d40caa21dde50fb4a0e94c24'

2017-04-07 Thread Clément Bœsch

ffmpeg | branch: master | Clément Bœsch  | Fri Apr  7 10:07:43 
2017 +0200| [92e532c18fee1d1ab2b8ed1ffe95696176df68ab] | committer: Clément 
Bœsch

Merge commit '239d02eff3ffe9f7d40caa21dde50fb4a0e94c24'

* commit '239d02eff3ffe9f7d40caa21dde50fb4a0e94c24':
  avisynth: Cast to the right type when loading avisynth library functions

This commit is a noop, we are using dlsym() instead of GetProcAdress().

Merged-by: Clément Bœsch 

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=92e532c18fee1d1ab2b8ed1ffe95696176df68ab
---



___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog

[FFmpeg-cvslog] avisynth: Cast to the right type when loading avisynth library functions

2017-04-07 Thread Diego Biurrun

ffmpeg | branch: master | Diego Biurrun  | Thu Nov 24 
12:46:30 2016 +0100| [239d02eff3ffe9f7d40caa21dde50fb4a0e94c24] | committer: 
Diego Biurrun

avisynth: Cast to the right type when loading avisynth library functions

Fixes a number of related warnings.

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=239d02eff3ffe9f7d40caa21dde50fb4a0e94c24
---

 libavformat/avisynth.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/libavformat/avisynth.c b/libavformat/avisynth.c
index fd5f323..d3b3ec4 100644
--- a/libavformat/avisynth.c
+++ b/libavformat/avisynth.c
@@ -127,7 +127,8 @@ static av_cold int avisynth_load_library(void)
 return AVERROR_UNKNOWN;
 
 #define LOAD_AVS_FUNC(name, continue_on_fail)  \
-avs_library.name = GetProcAddress(avs_library.library, #name); \
+avs_library.name = (name ## _func) \
+   GetProcAddress(avs_library.library, #name); \
 if (!continue_on_fail && !avs_library.name)\
 goto fail;
 

___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog

[FFmpeg-cvslog] Merge commit '3ee5f25d37315b27f0e2d47aa69fc445545ee2e3'

2017-04-07 Thread Clément Bœsch

ffmpeg | branch: master | Clément Bœsch  | Fri Apr  7 10:08:53 
2017 +0200| [e7326e2980a079432bf3f5556fdab38e92982fd1] | committer: Clément 
Bœsch

Merge commit '3ee5f25d37315b27f0e2d47aa69fc445545ee2e3'

* commit '3ee5f25d37315b27f0e2d47aa69fc445545ee2e3':
  dxva2: Adjust printf length modifiers where appropriate

Merged-by: Clément Bœsch 

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=e7326e2980a079432bf3f5556fdab38e92982fd1
---

 libavcodec/dxva2.c | 10 +-
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/libavcodec/dxva2.c b/libavcodec/dxva2.c
index c782f1c..0036b5b 100644
--- a/libavcodec/dxva2.c
+++ b/libavcodec/dxva2.c
@@ -83,7 +83,7 @@ int ff_dxva2_commit_buffer(AVCodecContext *avctx,
 &dxva_data, &dxva_size);
 #endif
 if (FAILED(hr)) {
-av_log(avctx, AV_LOG_ERROR, "Failed to get a buffer for %u: 0x%lx\n",
+av_log(avctx, AV_LOG_ERROR, "Failed to get a buffer for %u: 0x%x\n",
type, hr);
 return -1;
 }
@@ -125,7 +125,7 @@ int ff_dxva2_commit_buffer(AVCodecContext *avctx,
 #endif
 if (FAILED(hr)) {
 av_log(avctx, AV_LOG_ERROR,
-   "Failed to release buffer type %u: 0x%lx\n",
+   "Failed to release buffer type %u: 0x%x\n",
type, hr);
 result = -1;
 }
@@ -179,7 +179,7 @@ int ff_dxva2_common_end_frame(AVCodecContext *avctx, 
AVFrame *frame,
 } while(1);
 
 if (FAILED(hr)) {
-av_log(avctx, AV_LOG_ERROR, "Failed to begin frame: 0x%lx\n", hr);
+av_log(avctx, AV_LOG_ERROR, "Failed to begin frame: 0x%x\n", hr);
 #if CONFIG_D3D11VA
 if (avctx->pix_fmt == AV_PIX_FMT_D3D11VA_VLD)
 if (D3D11VA_CONTEXT(ctx)->context_mutex != INVALID_HANDLE_VALUE)
@@ -278,7 +278,7 @@ int ff_dxva2_common_end_frame(AVCodecContext *avctx, 
AVFrame *frame,
 }
 #endif
 if (FAILED(hr)) {
-av_log(avctx, AV_LOG_ERROR, "Failed to execute: 0x%lx\n", hr);
+av_log(avctx, AV_LOG_ERROR, "Failed to execute: 0x%x\n", hr);
 result = -1;
 }
 
@@ -295,7 +295,7 @@ end:
 hr = IDirectXVideoDecoder_EndFrame(DXVA2_CONTEXT(ctx)->decoder, NULL);
 #endif
 if (FAILED(hr)) {
-av_log(avctx, AV_LOG_ERROR, "Failed to end frame: 0x%lx\n", hr);
+av_log(avctx, AV_LOG_ERROR, "Failed to end frame: 0x%x\n", hr);
 result = -1;
 }
 


==


___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog

[FFmpeg-cvslog] dxva2: Adjust printf length modifiers where appropriate

2017-04-07 Thread Diego Biurrun

ffmpeg | branch: master | Diego Biurrun  | Thu Nov 24 
12:46:31 2016 +0100| [3ee5f25d37315b27f0e2d47aa69fc445545ee2e3] | committer: 
Diego Biurrun

dxva2: Adjust printf length modifiers where appropriate

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=3ee5f25d37315b27f0e2d47aa69fc445545ee2e3
---

 libavcodec/dxva2.c | 10 +-
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/libavcodec/dxva2.c b/libavcodec/dxva2.c
index 9fedd03..6063993 100644
--- a/libavcodec/dxva2.c
+++ b/libavcodec/dxva2.c
@@ -83,7 +83,7 @@ int ff_dxva2_commit_buffer(AVCodecContext *avctx,
 &dxva_data, &dxva_size);
 #endif
 if (FAILED(hr)) {
-av_log(avctx, AV_LOG_ERROR, "Failed to get a buffer for %u: 0x%lx\n",
+av_log(avctx, AV_LOG_ERROR, "Failed to get a buffer for %u: 0x%x\n",
type, hr);
 return -1;
 }
@@ -125,7 +125,7 @@ int ff_dxva2_commit_buffer(AVCodecContext *avctx,
 #endif
 if (FAILED(hr)) {
 av_log(avctx, AV_LOG_ERROR,
-   "Failed to release buffer type %u: 0x%lx\n",
+   "Failed to release buffer type %u: 0x%x\n",
type, hr);
 result = -1;
 }
@@ -179,7 +179,7 @@ int ff_dxva2_common_end_frame(AVCodecContext *avctx, 
AVFrame *frame,
 } while(1);
 
 if (FAILED(hr)) {
-av_log(avctx, AV_LOG_ERROR, "Failed to begin frame: 0x%lx\n", hr);
+av_log(avctx, AV_LOG_ERROR, "Failed to begin frame: 0x%x\n", hr);
 #if CONFIG_D3D11VA
 if (avctx->pix_fmt == AV_PIX_FMT_D3D11VA_VLD)
 if (D3D11VA_CONTEXT(ctx)->context_mutex != INVALID_HANDLE_VALUE)
@@ -278,7 +278,7 @@ int ff_dxva2_common_end_frame(AVCodecContext *avctx, 
AVFrame *frame,
 }
 #endif
 if (FAILED(hr)) {
-av_log(avctx, AV_LOG_ERROR, "Failed to execute: 0x%lx\n", hr);
+av_log(avctx, AV_LOG_ERROR, "Failed to execute: 0x%x\n", hr);
 result = -1;
 }
 
@@ -295,7 +295,7 @@ end:
 hr = IDirectXVideoDecoder_EndFrame(DXVA2_CONTEXT(ctx)->decoder, NULL);
 #endif
 if (FAILED(hr)) {
-av_log(avctx, AV_LOG_ERROR, "Failed to end frame: 0x%lx\n", hr);
+av_log(avctx, AV_LOG_ERROR, "Failed to end frame: 0x%x\n", hr);
 result = -1;
 }
 

___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog

[FFmpeg-cvslog] mjpegdec: Check return values of functions that may fail

2017-04-07 Thread Diego Biurrun

ffmpeg | branch: master | Diego Biurrun  | Wed May 11 
08:59:17 2016 +0200| [212c6a1d70df011b6f2a2aa02f7677503287bd00] | committer: 
Diego Biurrun

mjpegdec: Check return values of functions that may fail

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=212c6a1d70df011b6f2a2aa02f7677503287bd00
---

 libavcodec/mjpegdec.c | 70 ++-
 1 file changed, 47 insertions(+), 23 deletions(-)

diff --git a/libavcodec/mjpegdec.c b/libavcodec/mjpegdec.c
index a4a6c63..f6e9038 100644
--- a/libavcodec/mjpegdec.c
+++ b/libavcodec/mjpegdec.c
@@ -68,25 +68,42 @@ static int build_vlc(VLC *vlc, const uint8_t *bits_table,
   huff_code, 2, 2, huff_sym, 2, 2, use_static);
 }
 
-static void build_basic_mjpeg_vlc(MJpegDecodeContext *s)
+static int build_basic_mjpeg_vlc(MJpegDecodeContext *s)
 {
-build_vlc(&s->vlcs[0][0], avpriv_mjpeg_bits_dc_luminance,
-  avpriv_mjpeg_val_dc, 12, 0, 0);
-build_vlc(&s->vlcs[0][1], avpriv_mjpeg_bits_dc_chrominance,
-  avpriv_mjpeg_val_dc, 12, 0, 0);
-build_vlc(&s->vlcs[1][0], avpriv_mjpeg_bits_ac_luminance,
-  avpriv_mjpeg_val_ac_luminance, 251, 0, 1);
-build_vlc(&s->vlcs[1][1], avpriv_mjpeg_bits_ac_chrominance,
-  avpriv_mjpeg_val_ac_chrominance, 251, 0, 1);
-build_vlc(&s->vlcs[2][0], avpriv_mjpeg_bits_ac_luminance,
-  avpriv_mjpeg_val_ac_luminance, 251, 0, 0);
-build_vlc(&s->vlcs[2][1], avpriv_mjpeg_bits_ac_chrominance,
-  avpriv_mjpeg_val_ac_chrominance, 251, 0, 0);
+int ret;
+
+if ((ret = build_vlc(&s->vlcs[0][0], avpriv_mjpeg_bits_dc_luminance,
+ avpriv_mjpeg_val_dc, 12, 0, 0)) < 0)
+return ret;
+
+if ((ret = build_vlc(&s->vlcs[0][1], avpriv_mjpeg_bits_dc_chrominance,
+ avpriv_mjpeg_val_dc, 12, 0, 0)) < 0)
+return ret;
+
+if ((ret = build_vlc(&s->vlcs[1][0], avpriv_mjpeg_bits_ac_luminance,
+ avpriv_mjpeg_val_ac_luminance, 251, 0, 1)) < 0)
+return ret;
+
+if ((ret = build_vlc(&s->vlcs[1][1], avpriv_mjpeg_bits_ac_chrominance,
+ avpriv_mjpeg_val_ac_chrominance, 251, 0, 1)) < 0)
+return ret;
+
+if ((ret = build_vlc(&s->vlcs[2][0], avpriv_mjpeg_bits_ac_luminance,
+ avpriv_mjpeg_val_ac_luminance, 251, 0, 0)) < 0)
+return ret;
+
+if ((ret = build_vlc(&s->vlcs[2][1], avpriv_mjpeg_bits_ac_chrominance,
+ avpriv_mjpeg_val_ac_chrominance, 251, 0, 0)) < 0)
+return ret;
+
+
+return 0;
 }
 
 av_cold int ff_mjpeg_decode_init(AVCodecContext *avctx)
 {
 MJpegDecodeContext *s = avctx->priv_data;
+int ret;
 
 if (!s->picture_ptr) {
 s->picture = av_frame_alloc();
@@ -109,12 +126,13 @@ av_cold int ff_mjpeg_decode_init(AVCodecContext *avctx)
 avctx->chroma_sample_location = AVCHROMA_LOC_CENTER;
 avctx->colorspace = AVCOL_SPC_BT470BG;
 
-build_basic_mjpeg_vlc(s);
+if ((ret = build_basic_mjpeg_vlc(s)) < 0)
+return ret;
 
 if (s->extern_huff) {
-int ret;
 av_log(avctx, AV_LOG_INFO, "mjpeg: using external huffman table\n");
-init_get_bits(&s->gb, avctx->extradata, avctx->extradata_size * 8);
+if ((ret = init_get_bits(&s->gb, avctx->extradata, 
avctx->extradata_size * 8)) < 0)
+return ret;
 if ((ret = ff_mjpeg_decode_dht(s))) {
 av_log(avctx, AV_LOG_ERROR,
"mjpeg: error using external huffman table\n");
@@ -248,7 +266,7 @@ int ff_mjpeg_decode_sof(MJpegDecodeContext *s)
 height= s->height;
 
 av_log(s->avctx, AV_LOG_DEBUG, "sof0: picture: %dx%d\n", width, height);
-if (av_image_check_size(width, height, 0, s->avctx))
+if (av_image_check_size(width, height, 0, s->avctx) < 0)
 return AVERROR_INVALIDDATA;
 
 nb_components = get_bits(&s->gb, 8);
@@ -686,6 +704,9 @@ static int ljpeg_decode_rgb_scan(MJpegDecodeContext *s, int 
predictor,
 
 av_fast_malloc(&s->ljpeg_buffer, &s->ljpeg_buffer_size,
(unsigned)s->mb_width * 4 * sizeof(s->ljpeg_buffer[0][0]));
+if (!s->ljpeg_buffer)
+return AVERROR(ENOMEM);
+
 buffer = s->ljpeg_buffer;
 
 for (i = 0; i < 3; i++)
@@ -1502,14 +1523,15 @@ int ff_mjpeg_decode_frame(AVCodecContext *avctx, void 
*data, int *got_frame,
 av_log(avctx, AV_LOG_DEBUG, "startcode: %X\n", start_code);
 
 /* process markers */
-if (start_code >= 0xd0 && start_code <= 0xd7)
+if (start_code >= 0xd0 && start_code <= 0xd7) {
 av_log(avctx, AV_LOG_DEBUG,
"restart marker: %d\n", start_code & 0x0f);
 /* APP fields */
-else if (start_code >= APP0 && start_code <= APP15)
-mjpeg_decode_app(s);
+} else if (start_code >= APP0 && start_code <= APP15) {
+if ((ret = mjpeg_decode_app(s)) < 0)
+

[FFmpeg-cvslog] Merge commit '212c6a1d70df011b6f2a2aa02f7677503287bd00'

2017-04-07 Thread Clément Bœsch

ffmpeg | branch: master | Clément Bœsch  | Fri Apr  7 10:17:53 
2017 +0200| [9c7ee37490d21350ab1a2e07069284daf1943e52] | committer: Clément 
Bœsch

Merge commit '212c6a1d70df011b6f2a2aa02f7677503287bd00'

* commit '212c6a1d70df011b6f2a2aa02f7677503287bd00':
  mjpegdec: Check return values of functions that may fail

Merged-by: Clément Bœsch 

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=9c7ee37490d21350ab1a2e07069284daf1943e52
---

 libavcodec/mjpegdec.c | 70 +++
 1 file changed, 48 insertions(+), 22 deletions(-)

diff --git a/libavcodec/mjpegdec.c b/libavcodec/mjpegdec.c
index 5470b66..cef3b38 100644
--- a/libavcodec/mjpegdec.c
+++ b/libavcodec/mjpegdec.c
@@ -71,20 +71,36 @@ static int build_vlc(VLC *vlc, const uint8_t *bits_table,
   huff_code, 2, 2, huff_sym, 2, 2, use_static);
 }
 
-static void build_basic_mjpeg_vlc(MJpegDecodeContext *s)
+static int build_basic_mjpeg_vlc(MJpegDecodeContext *s)
 {
-build_vlc(&s->vlcs[0][0], avpriv_mjpeg_bits_dc_luminance,
-  avpriv_mjpeg_val_dc, 12, 0, 0);
-build_vlc(&s->vlcs[0][1], avpriv_mjpeg_bits_dc_chrominance,
-  avpriv_mjpeg_val_dc, 12, 0, 0);
-build_vlc(&s->vlcs[1][0], avpriv_mjpeg_bits_ac_luminance,
-  avpriv_mjpeg_val_ac_luminance, 251, 0, 1);
-build_vlc(&s->vlcs[1][1], avpriv_mjpeg_bits_ac_chrominance,
-  avpriv_mjpeg_val_ac_chrominance, 251, 0, 1);
-build_vlc(&s->vlcs[2][0], avpriv_mjpeg_bits_ac_luminance,
-  avpriv_mjpeg_val_ac_luminance, 251, 0, 0);
-build_vlc(&s->vlcs[2][1], avpriv_mjpeg_bits_ac_chrominance,
-  avpriv_mjpeg_val_ac_chrominance, 251, 0, 0);
+int ret;
+
+if ((ret = build_vlc(&s->vlcs[0][0], avpriv_mjpeg_bits_dc_luminance,
+ avpriv_mjpeg_val_dc, 12, 0, 0)) < 0)
+return ret;
+
+if ((ret = build_vlc(&s->vlcs[0][1], avpriv_mjpeg_bits_dc_chrominance,
+ avpriv_mjpeg_val_dc, 12, 0, 0)) < 0)
+return ret;
+
+if ((ret = build_vlc(&s->vlcs[1][0], avpriv_mjpeg_bits_ac_luminance,
+ avpriv_mjpeg_val_ac_luminance, 251, 0, 1)) < 0)
+return ret;
+
+if ((ret = build_vlc(&s->vlcs[1][1], avpriv_mjpeg_bits_ac_chrominance,
+ avpriv_mjpeg_val_ac_chrominance, 251, 0, 1)) < 0)
+return ret;
+
+if ((ret = build_vlc(&s->vlcs[2][0], avpriv_mjpeg_bits_ac_luminance,
+ avpriv_mjpeg_val_ac_luminance, 251, 0, 0)) < 0)
+return ret;
+
+if ((ret = build_vlc(&s->vlcs[2][1], avpriv_mjpeg_bits_ac_chrominance,
+ avpriv_mjpeg_val_ac_chrominance, 251, 0, 0)) < 0)
+return ret;
+
+
+return 0;
 }
 
 static void parse_avid(MJpegDecodeContext *s, uint8_t *buf, int len)
@@ -110,6 +126,7 @@ static void init_idct(AVCodecContext *avctx)
 av_cold int ff_mjpeg_decode_init(AVCodecContext *avctx)
 {
 MJpegDecodeContext *s = avctx->priv_data;
+int ret;
 
 if (!s->picture_ptr) {
 s->picture = av_frame_alloc();
@@ -131,11 +148,13 @@ av_cold int ff_mjpeg_decode_init(AVCodecContext *avctx)
 avctx->chroma_sample_location = AVCHROMA_LOC_CENTER;
 avctx->colorspace = AVCOL_SPC_BT470BG;
 
-build_basic_mjpeg_vlc(s);
+if ((ret = build_basic_mjpeg_vlc(s)) < 0)
+return ret;
 
 if (s->extern_huff) {
 av_log(avctx, AV_LOG_INFO, "using external huffman table\n");
-init_get_bits(&s->gb, avctx->extradata, avctx->extradata_size * 8);
+if ((ret = init_get_bits(&s->gb, avctx->extradata, 
avctx->extradata_size * 8)) < 0)
+return ret;
 if (ff_mjpeg_decode_dht(s)) {
 av_log(avctx, AV_LOG_ERROR,
"error using external huffman table, switching back to 
internal\n");
@@ -307,7 +326,7 @@ int ff_mjpeg_decode_sof(MJpegDecodeContext *s)
 height= s->height;
 
 av_log(s->avctx, AV_LOG_DEBUG, "sof0: picture: %dx%d\n", width, height);
-if (av_image_check_size(width, height, 0, s->avctx))
+if (av_image_check_size(width, height, 0, s->avctx) < 0)
 return AVERROR_INVALIDDATA;
 
 nb_components = get_bits(&s->gb, 8);
@@ -969,6 +988,9 @@ static int ljpeg_decode_rgb_scan(MJpegDecodeContext *s, int 
nb_components, int p
 
 av_fast_malloc(&s->ljpeg_buffer, &s->ljpeg_buffer_size,
(unsigned)s->mb_width * 4 * sizeof(s->ljpeg_buffer[0][0]));
+if (!s->ljpeg_buffer)
+return AVERROR(ENOMEM);
+
 buffer = s->ljpeg_buffer;
 
 for (i = 0; i < 4; i++)
@@ -2116,19 +2138,22 @@ int ff_mjpeg_decode_frame(AVCodecContext *avctx, void 
*data, int *got_frame,
 av_log(avctx, AV_LOG_DEBUG, "startcode: %X\n", start_code);
 
 /* process markers */
-if (start_code >= 0xd0 && start_code <= 0xd7)
+if (start_code >= 0xd0 && start_code <= 0xd7) {
 av_log(avctx, AV_LOG_DEBUG,
"restart marker: %d\n", star

[FFmpeg-cvslog] avcodec/h264: Check weight values to be within the specs limits.

2017-04-07 Thread Michael Niedermayer

ffmpeg | branch: master | Michael Niedermayer  | Tue 
Mar 21 23:46:14 2017 +0100| [08117a40157464f8a9dcc2df393fa5fe299c1e98] | 
committer: Michael Niedermayer

avcodec/h264: Check weight values to be within the specs limits.

Fixes: integer overflows
Fixes: 911/clusterfuzz-testcase-5415105606975488

Found-by: continuous fuzzing process 
https://github.com/google/oss-fuzz/tree/master/targets/ffmpeg
Reviewed-by: "Ronald S. Bultje" 
Signed-off-by: Michael Niedermayer 

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=08117a40157464f8a9dcc2df393fa5fe299c1e98
---

 libavcodec/h264_parse.c | 9 +
 libavcodec/h264_slice.c | 7 +--
 2 files changed, 14 insertions(+), 2 deletions(-)

diff --git a/libavcodec/h264_parse.c b/libavcodec/h264_parse.c
index 0c87319..ea202e7 100644
--- a/libavcodec/h264_parse.c
+++ b/libavcodec/h264_parse.c
@@ -59,6 +59,9 @@ int ff_h264_pred_weight_table(GetBitContext *gb, const SPS 
*sps,
 if (luma_weight_flag) {
 pwt->luma_weight[i][list][0] = get_se_golomb(gb);
 pwt->luma_weight[i][list][1] = get_se_golomb(gb);
+if ((int8_t)pwt->luma_weight[i][list][0] != 
pwt->luma_weight[i][list][0] ||
+(int8_t)pwt->luma_weight[i][list][1] != 
pwt->luma_weight[i][list][1])
+goto out_range_weight;
 if (pwt->luma_weight[i][list][0] != luma_def ||
 pwt->luma_weight[i][list][1] != 0) {
 pwt->use_weight = 1;
@@ -76,6 +79,9 @@ int ff_h264_pred_weight_table(GetBitContext *gb, const SPS 
*sps,
 for (j = 0; j < 2; j++) {
 pwt->chroma_weight[i][list][j][0] = get_se_golomb(gb);
 pwt->chroma_weight[i][list][j][1] = get_se_golomb(gb);
+if ((int8_t)pwt->chroma_weight[i][list][j][0] != 
pwt->chroma_weight[i][list][j][0] ||
+(int8_t)pwt->chroma_weight[i][list][j][1] != 
pwt->chroma_weight[i][list][j][1])
+goto out_range_weight;
 if (pwt->chroma_weight[i][list][j][0] != chroma_def ||
 pwt->chroma_weight[i][list][j][1] != 0) {
 pwt->use_weight_chroma= 1;
@@ -104,6 +110,9 @@ int ff_h264_pred_weight_table(GetBitContext *gb, const SPS 
*sps,
 }
 pwt->use_weight = pwt->use_weight || pwt->use_weight_chroma;
 return 0;
+out_range_weight:
+avpriv_request_sample(logctx, "Out of range weight\n");
+return AVERROR_INVALIDDATA;
 }
 
 /**
diff --git a/libavcodec/h264_slice.c b/libavcodec/h264_slice.c
index d4d31cc..34bd6cf 100644
--- a/libavcodec/h264_slice.c
+++ b/libavcodec/h264_slice.c
@@ -1780,9 +1780,12 @@ static int h264_slice_header_parse(const H264Context *h, 
H264SliceContext *sl,
 }
 if ((pps->weighted_pred && sl->slice_type_nos == AV_PICTURE_TYPE_P) ||
 (pps->weighted_bipred_idc == 1 &&
- sl->slice_type_nos == AV_PICTURE_TYPE_B))
-ff_h264_pred_weight_table(&sl->gb, sps, sl->ref_count,
+ sl->slice_type_nos == AV_PICTURE_TYPE_B)) {
+ret = ff_h264_pred_weight_table(&sl->gb, sps, sl->ref_count,
   sl->slice_type_nos, &sl->pwt, h->avctx);
+if (ret < 0)
+return ret;
+}
 
 sl->explicit_ref_marking = 0;
 if (nal->ref_idc) {

___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog

[FFmpeg-cvslog] avcodec/tta: Don't try to read more than MIN_CACHE_BITS bits

2017-04-07 Thread James Almer

ffmpeg | branch: master | James Almer  | Fri Apr  7 12:48:31 
2017 -0300| [7c1566fec39492815eda0dc1fdfd18f8bf7ca635] | committer: James Almer

avcodec/tta: Don't try to read more than MIN_CACHE_BITS bits

This fixes assertion failures introduced in 4fbb56acbe.

Reviewed-by: michaelni
Reviewed-by: durandal_1707

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=7c1566fec39492815eda0dc1fdfd18f8bf7ca635
---

 libavcodec/tta.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/libavcodec/tta.c b/libavcodec/tta.c
index 118b9f7..8f097b3 100644
--- a/libavcodec/tta.c
+++ b/libavcodec/tta.c
@@ -285,7 +285,7 @@ static int tta_decode_frame(AVCodecContext *avctx, void 
*data,
 }
 
 if (k) {
-if (k >= 32 || unary > INT32_MAX >> k) {
+if (k > MIN_CACHE_BITS || unary > INT32_MAX >> k) {
 ret = AVERROR_INVALIDDATA;
 goto error;
 }

___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog

[FFmpeg-cvslog] lavc/mjpegdec: allow failure while decoding APP

2017-04-07 Thread Clément Bœsch

ffmpeg | branch: master | Clément Bœsch  | Fri Apr  7 
17:50:44 2017 +0200| [8d94d9798a6903fc65ceb8e5af843cb6a793e045] | committer: 
Clément Bœsch

lavc/mjpegdec: allow failure while decoding APP

Fix decoding frame.jpg from ticket #267

Regression since 9c7ee3749 / 212c6a1d7

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=8d94d9798a6903fc65ceb8e5af843cb6a793e045
---

 libavcodec/mjpegdec.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/libavcodec/mjpegdec.c b/libavcodec/mjpegdec.c
index cef3b38..98cbd41 100644
--- a/libavcodec/mjpegdec.c
+++ b/libavcodec/mjpegdec.c
@@ -2144,7 +2144,8 @@ int ff_mjpeg_decode_frame(AVCodecContext *avctx, void 
*data, int *got_frame,
 /* APP fields */
 } else if (start_code >= APP0 && start_code <= APP15) {
 if ((ret = mjpeg_decode_app(s)) < 0)
-return ret;
+av_log(avctx, AV_LOG_ERROR, "unable to decode APP fields: 
%s\n",
+   av_err2str(ret));
 /* Comment */
 } else if (start_code == COM) {
 ret = mjpeg_decode_com(s);

___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog

[FFmpeg-cvslog] h264: don't re-call ff_h264_direct_ref_list_init() w/ frame-mt.

2017-04-07 Thread Ronald S. Bultje

ffmpeg | branch: master | Ronald S. Bultje  | Thu Apr  6 
11:47:03 2017 -0400| [7f05c5cea04112471d8147487aa3b44141922d09] | committer: 
Ronald S. Bultje

h264: don't re-call ff_h264_direct_ref_list_init() w/ frame-mt.

I'm hoping that this will address the remaining tsan fate-h264 issues:

WARNING: ThreadSanitizer: data race (pid=24478)
  Read of size 8 at 0x7dbc0001c828 by main thread (mutexes: write M3243):
#0 ff_h264_ref_picture src/libavcodec/h264_picture.c:107 
(ffmpeg+0x013b78d8)
[..]
  Previous write of size 1 at 0x7dbc0001c82e by thread T2 (mutexes: write 
M3245):
#0 ff_h264_direct_ref_list_init src/libavcodec/h264_direct.c:137 
(ffmpeg+0x01382c93)

But I'm not sure because I haven't been able to reproduce locally.

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=7f05c5cea04112471d8147487aa3b44141922d09
---

 libavcodec/h264_slice.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/libavcodec/h264_slice.c b/libavcodec/h264_slice.c
index 34bd6cf..acf6a73 100644
--- a/libavcodec/h264_slice.c
+++ b/libavcodec/h264_slice.c
@@ -1891,7 +1891,8 @@ static int h264_slice_init(H264Context *h, 
H264SliceContext *sl,
 
 if (sl->slice_type_nos == AV_PICTURE_TYPE_B && !sl->direct_spatial_mv_pred)
 ff_h264_direct_dist_scale_factor(h, sl);
-ff_h264_direct_ref_list_init(h, sl);
+if (!h->setup_finished)
+ff_h264_direct_ref_list_init(h, sl);
 
 if (h->avctx->skip_loop_filter >= AVDISCARD_ALL ||
 (h->avctx->skip_loop_filter >= AVDISCARD_NONKEY &&

___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog

[FFmpeg-cvslog] pthread_frame: make accesses to debug field be protected by owner lock.

2017-04-07 Thread Ronald S. Bultje

ffmpeg | branch: master | Ronald S. Bultje  | Thu Apr  6 
13:58:59 2017 -0400| [2e664b9c1e73c80aab91070c1eb7676f04bdd12d] | committer: 
Ronald S. Bultje

pthread_frame: make accesses to debug field be protected by owner lock.

The av_log() is done outside the lock, but this way the accesses to the
field (reads and writes) are always protected by a mutex. The av_log()
is not run inside the lock context because it may involve user callbacks
and doing that in performance-sensitive code is probably not a good idea.

This should fix occasional tsan warnings when running fate-h264, like:

WARNING: ThreadSanitizer: data race (pid=10916)
  Write of size 4 at 0x7d64000174fc by main thread (mutexes: write M2313):
#0 update_context_from_user src/libavcodec/pthread_frame.c:335 
(ffmpeg+0x00df7b06)
[..]
  Previous read of size 4 at 0x7d64000174fc by thread T1 (mutexes: write M2311):
#0 ff_thread_await_progress src/libavcodec/pthread_frame.c:592 
(ffmpeg+0x00df8b3e)

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=2e664b9c1e73c80aab91070c1eb7676f04bdd12d
---

 libavcodec/pthread_frame.c | 6 ++
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/libavcodec/pthread_frame.c b/libavcodec/pthread_frame.c
index aaf576d..46c6292 100644
--- a/libavcodec/pthread_frame.c
+++ b/libavcodec/pthread_frame.c
@@ -566,12 +566,11 @@ void ff_thread_report_progress(ThreadFrame *f, int n, int 
field)
 
 p = f->owner[field]->internal->thread_ctx;
 
+pthread_mutex_lock(&p->progress_mutex);
 if (f->owner[field]->debug&FF_DEBUG_THREADS)
 av_log(f->owner[field], AV_LOG_DEBUG,
"%p finished %d field %d\n", progress, n, field);
 
-pthread_mutex_lock(&p->progress_mutex);
-
 atomic_store_explicit(&progress[field], n, memory_order_release);
 
 pthread_cond_broadcast(&p->progress_cond);
@@ -589,11 +588,10 @@ void ff_thread_await_progress(ThreadFrame *f, int n, int 
field)
 
 p = f->owner[field]->internal->thread_ctx;
 
+pthread_mutex_lock(&p->progress_mutex);
 if (f->owner[field]->debug&FF_DEBUG_THREADS)
 av_log(f->owner[field], AV_LOG_DEBUG,
"thread awaiting %d field %d from %p\n", n, field, progress);
-
-pthread_mutex_lock(&p->progress_mutex);
 while (atomic_load_explicit(&progress[field], memory_order_relaxed) < n)
 pthread_cond_wait(&p->progress_cond, &p->progress_mutex);
 pthread_mutex_unlock(&p->progress_mutex);

___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog

[FFmpeg-cvslog] avcodec/utvideodec: add support for gradient prediction

2017-04-07 Thread Paul B Mahol

ffmpeg | branch: master | Paul B Mahol  | Fri Apr  7 20:09:22 
2017 +0200| [faa94a576f5f3de10fc7016e0d94229faa1c2159] | committer: Paul B Mahol

avcodec/utvideodec: add support for gradient prediction

Signed-off-by: Paul B Mahol 

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=faa94a576f5f3de10fc7016e0d94229faa1c2159
---

 libavcodec/utvideodec.c | 230 ++--
 1 file changed, 225 insertions(+), 5 deletions(-)

diff --git a/libavcodec/utvideodec.c b/libavcodec/utvideodec.c
index 7c65d77..a49cae8 100644
--- a/libavcodec/utvideodec.c
+++ b/libavcodec/utvideodec.c
@@ -602,6 +602,189 @@ static void restore_median_packed_il(uint8_t *src, int 
step, ptrdiff_t stride,
 }
 }
 
+static void restore_gradient_planar(UtvideoContext *c, uint8_t *src, ptrdiff_t 
stride,
+int width, int height, int slices, int 
rmode)
+{
+int i, j, slice;
+int A, B, C;
+uint8_t *bsrc;
+int slice_start, slice_height;
+const int cmask = ~rmode;
+
+for (slice = 0; slice < slices; slice++) {
+slice_start  = ((slice * height) / slices) & cmask;
+slice_height = slice + 1) * height) / slices) & cmask) -
+   slice_start;
+
+if (!slice_height)
+continue;
+bsrc = src + slice_start * stride;
+
+// first line - left neighbour prediction
+bsrc[0] += 0x80;
+c->llviddsp.add_left_pred(bsrc, bsrc, width, 0);
+bsrc += stride;
+if (slice_height <= 1)
+continue;
+for (j = 1; j < slice_height; j++) {
+// second line - first element has top prediction, the rest uses 
gradient
+bsrc[0] = (bsrc[0] + bsrc[-stride]) & 0xFF;
+for (i = 1; i < width; i++) {
+A = bsrc[i - stride];
+B = bsrc[i - (stride + 1)];
+C = bsrc[i - 1];
+bsrc[i] = (A - B + C + bsrc[i]) & 0xFF;
+}
+bsrc += stride;
+}
+}
+}
+
+static void restore_gradient_planar_il(UtvideoContext *c, uint8_t *src, 
ptrdiff_t stride,
+  int width, int height, int slices, int 
rmode)
+{
+int i, j, slice;
+int A, B, C;
+uint8_t *bsrc;
+int slice_start, slice_height;
+const int cmask   = ~(rmode ? 3 : 1);
+const ptrdiff_t stride2 = stride << 1;
+
+for (slice = 0; slice < slices; slice++) {
+slice_start= ((slice * height) / slices) & cmask;
+slice_height   = slice + 1) * height) / slices) & cmask) -
+ slice_start;
+slice_height >>= 1;
+if (!slice_height)
+continue;
+
+bsrc = src + slice_start * stride;
+
+// first line - left neighbour prediction
+bsrc[0] += 0x80;
+A = c->llviddsp.add_left_pred(bsrc, bsrc, width, 0);
+c->llviddsp.add_left_pred(bsrc + stride, bsrc + stride, width, A);
+bsrc += stride2;
+if (slice_height <= 1)
+continue;
+for (j = 1; j < slice_height; j++) {
+// second line - first element has top prediction, the rest uses 
gradient
+bsrc[0] = (bsrc[0] + bsrc[-stride2]) & 0xFF;
+for (i = 1; i < width; i++) {
+A = bsrc[i - stride2];
+B = bsrc[i - (stride2 + 1)];
+C = bsrc[i - 1];
+bsrc[i] = (A - B + C + bsrc[i]) & 0xFF;
+}
+for (i = 0; i < width; i++) {
+A = bsrc[i - stride];
+B = bsrc[i - (1 + stride)];
+C = bsrc[i - 1 + stride];
+bsrc[i + stride] = (A - B + C + bsrc[i + stride]) & 0xFF;
+}
+bsrc += stride2;
+}
+}
+}
+
+static void restore_gradient_packed(uint8_t *src, int step, ptrdiff_t stride,
+int width, int height, int slices, int 
rmode)
+{
+int i, j, slice;
+int A, B, C;
+uint8_t *bsrc;
+int slice_start, slice_height;
+const int cmask = ~rmode;
+
+for (slice = 0; slice < slices; slice++) {
+slice_start  = ((slice * height) / slices) & cmask;
+slice_height = slice + 1) * height) / slices) & cmask) -
+   slice_start;
+
+if (!slice_height)
+continue;
+bsrc = src + slice_start * stride;
+
+// first line - left neighbour prediction
+bsrc[0] += 0x80;
+A = bsrc[0];
+for (i = step; i < width * step; i += step) {
+bsrc[i] += A;
+A= bsrc[i];
+}
+bsrc += stride;
+if (slice_height <= 1)
+continue;
+for (j = 1; j < slice_height; j++) {
+// second line - first element has top prediction, the rest uses 
gradient
+C= bsrc[-stride];
+bsrc[0] += C;
+for (i = step; i < width * step; i += step) {
+A = bsrc

[FFmpeg-cvslog] ffmpeg: use av_stream_new_side_data()

2017-04-07 Thread James Almer

ffmpeg | branch: master | James Almer  | Mon Dec 26 18:33:06 
2016 -0300| [9f102653fd723005f26c6e8c7525fec585631a72] | committer: James Almer

ffmpeg: use av_stream_new_side_data()

Signed-off-by: James Almer 

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=9f102653fd723005f26c6e8c7525fec585631a72
---

 ffmpeg.c | 17 -
 1 file changed, 4 insertions(+), 13 deletions(-)

diff --git a/ffmpeg.c b/ffmpeg.c
index 444b963..55467c0 100644
--- a/ffmpeg.c
+++ b/ffmpeg.c
@@ -3062,23 +3062,14 @@ static int init_output_stream_streamcopy(OutputStream 
*ost)
 ost->st->disposition = ist->st->disposition;
 
 if (ist->st->nb_side_data) {
-ost->st->side_data = av_realloc_array(NULL, ist->st->nb_side_data,
-  sizeof(*ist->st->side_data));
-if (!ost->st->side_data)
-return AVERROR(ENOMEM);
-
-ost->st->nb_side_data = 0;
 for (i = 0; i < ist->st->nb_side_data; i++) {
 const AVPacketSideData *sd_src = &ist->st->side_data[i];
-AVPacketSideData *sd_dst = 
&ost->st->side_data[ost->st->nb_side_data];
+uint8_t *dst_data;
 
-sd_dst->data = av_malloc(sd_src->size);
-if (!sd_dst->data)
+dst_data = av_stream_new_side_data(ost->st, sd_src->type, 
sd_src->size);
+if (!dst_data)
 return AVERROR(ENOMEM);
-memcpy(sd_dst->data, sd_src->data, sd_src->size);
-sd_dst->size = sd_src->size;
-sd_dst->type = sd_src->type;
-ost->st->nb_side_data++;
+memcpy(dst_data, sd_src->data, sd_src->size);
 }
 }
 

___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog

[FFmpeg-cvslog] lavf: use the new bitstream filter for extracting extradata

2017-04-07 Thread James Almer

ffmpeg | branch: master | James Almer  | Fri Mar 24 18:42:19 
2017 -0300| [b8f26779d615dfb466e90627323b1a4e40639f76] | committer: James Almer

lavf: use the new bitstream filter for extracting extradata

This merges commits 8e2ea691351c5079cdab245ff7bfa5c0f3e3bfe4 and
096a8effa3f8f3455292c958c3ed07e798def7bd by Anton Khirnov, with the
following change:

- extract_extradata_check() is added to know if the codec is supported
by the bsf before trying to initialize it. This behaviour is similar to
the old AVCodecParser.split checks.

The FATE reference changes are due to the filtered out NAL units that
the old AVCodecParser.split implementation left alone.
Decoding is unchanged as the functions that parse extradata simply
ignored said unnecessary NAL units.

Signed-off-by: James Almer 

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=b8f26779d615dfb466e90627323b1a4e40639f76
---

 libavformat/internal.h   |   9 ++
 libavformat/utils.c  | 146 ---
 tests/ref/fate/copy-trac2211-avi |   4 +-
 tests/ref/fate/h264_mp4toannexb_ticket2991   |   2 +-
 tests/ref/fate/h264_mp4toannexb_ticket5927   |   2 +-
 tests/ref/fate/h264_mp4toannexb_ticket5927_2 |   2 +-
 tests/ref/fate/segment-mp4-to-ts |   2 +-
 7 files changed, 148 insertions(+), 19 deletions(-)

diff --git a/libavformat/internal.h b/libavformat/internal.h
index 63a1724..c856945 100644
--- a/libavformat/internal.h
+++ b/libavformat/internal.h
@@ -178,6 +178,15 @@ struct AVStreamInternal {
 
 enum AVCodecID orig_codec_id;
 
+/* the context for extracting extradata in find_stream_info()
+ * inited=1/bsf=NULL signals that extracting is not possible (codec not
+ * supported) */
+struct {
+AVBSFContext *bsf;
+AVPacket *pkt;
+int inited;
+} extract_extradata;
+
 /**
  * Whether the internal avctx needs to be updated from codecpar (after a 
late change to codecpar)
  */
diff --git a/libavformat/utils.c b/libavformat/utils.c
index a059046..ba82a76 100644
--- a/libavformat/utils.c
+++ b/libavformat/utils.c
@@ -3370,6 +3370,127 @@ void ff_rfps_calculate(AVFormatContext *ic)
 }
 }
 
+static int extract_extradata_check(AVStream *st)
+{
+const AVBitStreamFilter *f;
+
+f = av_bsf_get_by_name("extract_extradata");
+if (!f)
+return 0;
+
+if (f->codec_ids) {
+const enum AVCodecID *ids;
+for (ids = f->codec_ids; *ids != AV_CODEC_ID_NONE; ids++)
+if (*ids == st->codecpar->codec_id)
+return 1;
+}
+
+return 0;
+}
+
+static int extract_extradata_init(AVStream *st)
+{
+AVStreamInternal *i = st->internal;
+const AVBitStreamFilter *f;
+int ret;
+
+f = av_bsf_get_by_name("extract_extradata");
+if (!f)
+goto finish;
+
+/* check that the codec id is supported */
+ret = extract_extradata_check(st);
+if (!ret)
+goto finish;
+
+i->extract_extradata.pkt = av_packet_alloc();
+if (!i->extract_extradata.pkt)
+return AVERROR(ENOMEM);
+
+ret = av_bsf_alloc(f, &i->extract_extradata.bsf);
+if (ret < 0)
+goto fail;
+
+ret = avcodec_parameters_copy(i->extract_extradata.bsf->par_in,
+  st->codecpar);
+if (ret < 0)
+goto fail;
+
+i->extract_extradata.bsf->time_base_in = st->time_base;
+
+/* if init fails here, we assume extracting extradata is just not
+ * supported for this codec, so we return success */
+ret = av_bsf_init(i->extract_extradata.bsf);
+if (ret < 0) {
+av_bsf_free(&i->extract_extradata.bsf);
+ret = 0;
+}
+
+finish:
+i->extract_extradata.inited = 1;
+
+return 0;
+fail:
+av_bsf_free(&i->extract_extradata.bsf);
+av_packet_free(&i->extract_extradata.pkt);
+return ret;
+}
+
+static int extract_extradata(AVStream *st, AVPacket *pkt)
+{
+AVStreamInternal *i = st->internal;
+AVPacket *pkt_ref;
+int ret;
+
+if (!i->extract_extradata.inited) {
+ret = extract_extradata_init(st);
+if (ret < 0)
+return ret;
+}
+
+if (i->extract_extradata.inited && !i->extract_extradata.bsf)
+return 0;
+
+pkt_ref = i->extract_extradata.pkt;
+ret = av_packet_ref(pkt_ref, pkt);
+if (ret < 0)
+return ret;
+
+ret = av_bsf_send_packet(i->extract_extradata.bsf, pkt_ref);
+if (ret < 0) {
+av_packet_unref(pkt_ref);
+return ret;
+}
+
+while (ret >= 0 && !i->avctx->extradata) {
+int extradata_size;
+uint8_t *extradata;
+
+ret = av_bsf_receive_packet(i->extract_extradata.bsf, pkt_ref);
+if (ret < 0) {
+if (ret != AVERROR(EAGAIN) && ret != AVERROR_EOF)
+return ret;
+continue;
+}
+
+extradata = av_packet_get_side_data(pkt_ref, AV_PKT_DATA_NEW_EXTRADATA,
+&extradata_s

[FFmpeg-cvslog] avcodec/hevc_parse: ignore all non parameter set NAL units in extradata

2017-04-07 Thread James Almer

ffmpeg | branch: master | James Almer  | Sun Apr  2 14:35:03 
2017 -0300| [e7fb6bc32d1bbffbc1a466c2a505b138cf8a7e0e] | committer: James Almer

avcodec/hevc_parse: ignore all non parameter set NAL units in extradata

While they shouldn't be present, they are harmless if they are.

Reviewed-by: Michael Niedermayer 
Signed-off-by: James Almer 

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=e7fb6bc32d1bbffbc1a466c2a505b138cf8a7e0e
---

 libavcodec/hevc_parse.c | 21 ++---
 1 file changed, 2 insertions(+), 19 deletions(-)

diff --git a/libavcodec/hevc_parse.c b/libavcodec/hevc_parse.c
index 6c1138e..3fe6662 100644
--- a/libavcodec/hevc_parse.c
+++ b/libavcodec/hevc_parse.c
@@ -41,25 +41,8 @@ static int hevc_decode_nal_units(const uint8_t *buf, int 
buf_size, HEVCParamSets
 case HEVC_NAL_VPS: ff_hevc_decode_nal_vps(&nal->gb, logctx, ps);
break;
 case HEVC_NAL_SPS: ff_hevc_decode_nal_sps(&nal->gb, logctx, ps, 1); 
break;
 case HEVC_NAL_PPS: ff_hevc_decode_nal_pps(&nal->gb, logctx, ps);
break;
-case HEVC_NAL_TRAIL_R:
-case HEVC_NAL_TRAIL_N:
-case HEVC_NAL_TSA_N:
-case HEVC_NAL_TSA_R:
-case HEVC_NAL_STSA_N:
-case HEVC_NAL_STSA_R:
-case HEVC_NAL_BLA_W_LP:
-case HEVC_NAL_BLA_W_RADL:
-case HEVC_NAL_BLA_N_LP:
-case HEVC_NAL_IDR_W_RADL:
-case HEVC_NAL_IDR_N_LP:
-case HEVC_NAL_CRA_NUT:
-case HEVC_NAL_RADL_N:
-case HEVC_NAL_RADL_R:
-case HEVC_NAL_RASL_N:
-case HEVC_NAL_RASL_R:
-av_log(logctx, AV_LOG_ERROR, "Invalid NAL unit: %d\n", nal->type);
-ret = AVERROR_INVALIDDATA;
-goto done;
+default:
+av_log(logctx, AV_LOG_VERBOSE, "Ignoring NAL type %d in 
extradata\n", nal->type);
 break;
 }
 }

___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog

[FFmpeg-cvslog] doc/libav-merge: remove line about extract_extradata_bsf usage

2017-04-07 Thread James Almer

ffmpeg | branch: master | James Almer  | Fri Apr  7 17:31:13 
2017 -0300| [3f8d7342c32c8b22ab177d62f5c7660db213e342] | committer: James Almer

doc/libav-merge: remove line about extract_extradata_bsf usage

It was addressed in b8f26779d615dfb466e90627323b1a4e40639f76

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=3f8d7342c32c8b22ab177d62f5c7660db213e342
---

 doc/libav-merge.txt | 1 -
 1 file changed, 1 deletion(-)

diff --git a/doc/libav-merge.txt b/doc/libav-merge.txt
index 2040c6a..454f454 100644
--- a/doc/libav-merge.txt
+++ b/doc/libav-merge.txt
@@ -96,7 +96,6 @@ Stuff that didn't reach the codebase:
   - e7078e842 hevcdsp: add x86 SIMD for MC
 - VAAPI VP8 decode hwaccel (currently under review: 
http://ffmpeg.org/pipermail/ffmpeg-devel/2017-February/thread.html#207348)
 - Removal of the custom atomic API (5cc0057f49, see 
http://ffmpeg.org/pipermail/ffmpeg-devel/2017-March/209003.html)
-- Use the new bitstream filter for extracting extradata (8e2ea69135 and 
096a8effa3, see 
https://ffmpeg.org/pipermail/ffmpeg-devel/2017-March/209068.html)
 - Read aac_adtstoasc extradata updates from packet side data on Matroska once 
mov and the bsf in question are fixed (See 13a211e632 and 5ef1959080)
 - new bistream reader (see 
http://ffmpeg.org/pipermail/ffmpeg-devel/2017-April/209609.html)
 

___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog

[FFmpeg-cvslog] vf_pad: center image on padded area if negative x/y

2017-04-07 Thread Ricardo Constantino

ffmpeg | branch: master | Ricardo Constantino  | Mon Apr  3 
19:34:09 2017 +0100| [57c3670896c69714ca1514728edf5ee48b2abcc9] | committer: 
Rostislav Pehlivanov

vf_pad: center image on padded area if negative x/y

or if x/y go beyond padded area.

This is mostly useful when paired with the aspect option.
Defaults aren't changed.

Idea for this was taken from mpv's soon-to-be-removed expand vf.

Reviewed-by: Paul B Mahol 

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=57c3670896c69714ca1514728edf5ee48b2abcc9
---

 doc/filters.texi |  3 +++
 libavfilter/vf_pad.c | 12 +++-
 2 files changed, 10 insertions(+), 5 deletions(-)

diff --git a/doc/filters.texi b/doc/filters.texi
index bc37e66..32720db 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -10430,6 +10430,9 @@ expression, and vice versa.
 
 The default value of @var{x} and @var{y} is 0.
 
+If @var{x} or @var{y} evaluate to a negative number, they'll be changed
+so the input image is centered on the padded area.
+
 @item color
 Specify the color of the padded area. For the syntax of this option,
 check the "Color" section in the ffmpeg-utils manual.
diff --git a/libavfilter/vf_pad.c b/libavfilter/vf_pad.c
index 44a8fec..725dc29 100644
--- a/libavfilter/vf_pad.c
+++ b/libavfilter/vf_pad.c
@@ -173,8 +173,13 @@ static int config_input(AVFilterLink *inlink)
 goto eval_fail;
 s->x = var_values[VAR_X] = res;
 
+if (s->x < 0 || s->x + inlink->w > s->w)
+s->x = var_values[VAR_X] = (s->w - inlink->w) / 2;
+if (s->y < 0 || s->y + inlink->h > s->h)
+s->y = var_values[VAR_Y] = (s->h - inlink->h) / 2;
+
 /* sanity check params */
-if (s->w < 0 || s->h < 0 || s->x < 0 || s->y < 0) {
+if (s->w < 0 || s->h < 0) {
 av_log(ctx, AV_LOG_ERROR, "Negative values are not acceptable.\n");
 return AVERROR(EINVAL);
 }
@@ -192,10 +197,7 @@ static int config_input(AVFilterLink *inlink)
inlink->w, inlink->h, s->w, s->h, s->x, s->y,
s->rgba_color[0], s->rgba_color[1], s->rgba_color[2], 
s->rgba_color[3]);
 
-if (s->x <  0 || s->y <  0  ||
-s->w <= 0 || s->h <= 0  ||
-(unsigned)s->x + (unsigned)inlink->w > s->w ||
-(unsigned)s->y + (unsigned)inlink->h > s->h) {
+if (s->w <= 0 || s->h <= 0) {
 av_log(ctx, AV_LOG_ERROR,
"Input area %d:%d:%d:%d not within the padded area 0:0:%d:%d or 
zero-sized\n",
s->x, s->y, s->x + inlink->w, s->y + inlink->h, s->w, s->h);

___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog

[FFmpeg-cvslog] opusenc: minor style changes

2017-04-07 Thread Rostislav Pehlivanov

ffmpeg | branch: master | Rostislav Pehlivanov  | Sat Apr  
8 00:29:03 2017 +0100| [42ffc67208b18e7048c0deed0ff3b289c870c66d] | committer: 
Rostislav Pehlivanov

opusenc: minor style changes

Signed-off-by: Rostislav Pehlivanov 

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=42ffc67208b18e7048c0deed0ff3b289c870c66d
---

 libavcodec/opusenc.c | 32 ++--
 1 file changed, 14 insertions(+), 18 deletions(-)

diff --git a/libavcodec/opusenc.c b/libavcodec/opusenc.c
index a84e51f..d80ae7a 100644
--- a/libavcodec/opusenc.c
+++ b/libavcodec/opusenc.c
@@ -280,7 +280,7 @@ static int celt_frame_map_norm_bands(OpusEncContext *s, 
CeltFrame *f)
 return !noise;
 }
 
-static void celt_enc_tf(OpusEncContext *s, OpusRangeCoder *rc, CeltFrame *f)
+static void celt_enc_tf(OpusRangeCoder *rc, CeltFrame *f)
 {
 int i, tf_select = 0, diff = 0, tf_changed = 0, tf_select_needed;
 int bits = f->transient ? 2 : 4;
@@ -307,7 +307,7 @@ static void celt_enc_tf(OpusEncContext *s, OpusRangeCoder 
*rc, CeltFrame *f)
 f->tf_change[i] = 
ff_celt_tf_select[f->size][f->transient][tf_select][f->tf_change[i]];
 }
 
-static void celt_bitalloc(OpusEncContext *s, OpusRangeCoder *rc, CeltFrame *f)
+static void ff_celt_enc_bitalloc(OpusRangeCoder *rc, CeltFrame *f)
 {
 int i, j, low, high, total, done, bandbits, remaining, tbits_8ths;
 int skip_startband  = f->start_band;
@@ -675,7 +675,7 @@ static void celt_quant_coarse(OpusEncContext *s, 
OpusRangeCoder *rc, CeltFrame *
 }
 }
 
-static void celt_quant_fine(OpusEncContext *s, OpusRangeCoder *rc, CeltFrame 
*f)
+static void celt_quant_fine(OpusRangeCoder *rc, CeltFrame *f)
 {
 int i, ch;
 for (i = f->start_band; i < f->end_band; i++) {
@@ -712,7 +712,7 @@ static void celt_quant_final(OpusEncContext *s, 
OpusRangeCoder *rc, CeltFrame *f
 }
 }
 
-static void celt_quant_bands(OpusEncContext *s, OpusRangeCoder *rc, CeltFrame 
*f)
+static void celt_quant_bands(OpusRangeCoder *rc, CeltFrame *f)
 {
 float lowband_scratch[8 * 22];
 float norm[2 * 8 * 100];
@@ -725,6 +725,7 @@ static void celt_quant_bands(OpusEncContext *s, 
OpusRangeCoder *rc, CeltFrame *f
 int i, j;
 
 for (i = f->start_band; i < f->end_band; i++) {
+uint32_t cm[2] = { (1 << f->blocks) - 1, (1 << f->blocks) - 1 };
 int band_offset = ff_celt_freq_bands[i] << f->size;
 int band_size   = ff_celt_freq_range[i] << f->size;
 float *X = f->block[0].coeffs + band_offset;
@@ -733,8 +734,7 @@ static void celt_quant_bands(OpusEncContext *s, 
OpusRangeCoder *rc, CeltFrame *f
 int consumed = opus_rc_tell_frac(rc);
 float *norm2 = norm + 8 * 100;
 int effective_lowband = -1;
-unsigned int cm[2];
-int b;
+int b = 0;
 
 /* Compute how many bits we want to allocate to this band */
 if (i != f->start_band)
@@ -743,8 +743,7 @@ static void celt_quant_bands(OpusEncContext *s, 
OpusRangeCoder *rc, CeltFrame *f
 if (i <= f->coded_bands - 1) {
 int curr_balance = f->remaining / FFMIN(3, f->coded_bands-i);
 b = av_clip_uintp2(FFMIN(f->remaining2 + 1, f->pulses[i] + 
curr_balance), 14);
-} else
-b = 0;
+}
 
 if (ff_celt_freq_bands[i] - ff_celt_freq_range[i] >= 
ff_celt_freq_bands[f->start_band] &&
 (update_lowband || lowband_offset == 0))
@@ -769,10 +768,7 @@ static void celt_quant_bands(OpusEncContext *s, 
OpusRangeCoder *rc, CeltFrame *f
 cm[0] |= f->block[0].collapse_masks[j];
 cm[1] |= f->block[f->channels - 1].collapse_masks[j];
 }
-} else
-/* Otherwise, we'll be using the LCG to fold, so all blocks will 
(almost
-always) be non-zero.*/
-cm[0] = cm[1] = (1 << f->blocks) - 1;
+}
 
 if (f->dual_stereo && i == f->intensity_stereo) {
 /* Switch off dual stereo to do intensity */
@@ -786,13 +782,13 @@ static void celt_quant_bands(OpusEncContext *s, 
OpusRangeCoder *rc, CeltFrame *f
 effective_lowband != -1 ? norm + 
(effective_lowband << f->size) : NULL, f->size,
 norm + band_offset, 0, 1.0f, 
lowband_scratch, cm[0]);
 
-cm[1] = ff_celt_encode_band(f, rc, i, Y, NULL, band_size, b/2, 
f->blocks,
+cm[1] = ff_celt_encode_band(f, rc, i, Y, NULL, band_size, b / 2, 
f->blocks,
 effective_lowband != -1 ? norm2 + 
(effective_lowband << f->size) : NULL, f->size,
 norm2 + band_offset, 0, 1.0f, 
lowband_scratch, cm[1]);
 } else {
 cm[0] = ff_celt_encode_band(f, rc, i, X, Y, band_size, b, 
f->blocks,
 effective_lowband != -1 ? norm + 
(effective_lowband << f->size) : NULL, f->size,
-norm + band_offset, 0, 1.0

[FFmpeg-cvslog] opusenc: remove unused header entries and simplify normalization

2017-04-07 Thread Rostislav Pehlivanov

ffmpeg | branch: master | Rostislav Pehlivanov  | Mon Apr  
3 13:13:14 2017 +0100| [63744d8afd57851479ed2ff3e9a843c70d1b0031] | committer: 
Rostislav Pehlivanov

opusenc: remove unused header entries and simplify normalization

Signed-off-by: Rostislav Pehlivanov 

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=63744d8afd57851479ed2ff3e9a843c70d1b0031
---

 libavcodec/opus_celt.h |  3 ---
 libavcodec/opusenc.c   | 18 +++---
 2 files changed, 7 insertions(+), 14 deletions(-)

diff --git a/libavcodec/opus_celt.h b/libavcodec/opus_celt.h
index 23c2089..1ed57f3 100644
--- a/libavcodec/opus_celt.h
+++ b/libavcodec/opus_celt.h
@@ -67,9 +67,6 @@ typedef struct CeltBlock {
 
 uint8_t collapse_masks[CELT_MAX_BANDS];
 
-int band_bins[CELT_MAX_BANDS]; /* MDCT bins per band */
-float *band_coeffs[CELT_MAX_BANDS];
-
 /* buffer for mdct output + postfilter */
 DECLARE_ALIGNED(32, float, buf)[2048];
 DECLARE_ALIGNED(32, float, coeffs)[CELT_MAX_FRAME_SIZE];
diff --git a/libavcodec/opusenc.c b/libavcodec/opusenc.c
index cecc8f2..a84e51f 100644
--- a/libavcodec/opusenc.c
+++ b/libavcodec/opusenc.c
@@ -255,24 +255,20 @@ static int celt_frame_map_norm_bands(OpusEncContext *s, 
CeltFrame *f)
 
 for (ch = 0; ch < f->channels; ch++) {
 CeltBlock *block = &f->block[ch];
-float *start = block->coeffs;
 for (i = 0; i < CELT_MAX_BANDS; i++) {
 float ener = 0.0f;
+int band_offset = ff_celt_freq_bands[i] << f->size;
+int band_size   = ff_celt_freq_range[i] << f->size;
+float *coeffs   = &block->coeffs[band_offset];
 
-/* Calculate band bins */
-block->band_bins[i] = ff_celt_freq_range[i] << f->size;
-block->band_coeffs[i] = start;
-start += block->band_bins[i];
-
-/* Normalize band energy */
-for (j = 0; j < block->band_bins[i]; j++)
-ener += block->band_coeffs[i][j]*block->band_coeffs[i][j];
+for (j = 0; j < band_size; j++)
+ener += coeffs[j]*coeffs[j];
 
 block->lin_energy[i] = sqrtf(ener) + FLT_EPSILON;
 ener = 1.0f/block->lin_energy[i];
 
-for (j = 0; j < block->band_bins[i]; j++)
-block->band_coeffs[i][j] *= ener;
+for (j = 0; j < band_size; j++)
+coeffs[j] *= ener;
 
 block->energy[i] = log2f(block->lin_energy[i]) - 
ff_celt_mean_energy[i];
 

___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog

[FFmpeg-cvslog] opus_celt: fix some style issues

2017-04-07 Thread Rostislav Pehlivanov

ffmpeg | branch: master | Rostislav Pehlivanov  | Sun Apr  
2 20:55:11 2017 +0100| [6d0b62a12b46edbb3d141c356a7bcf7856a1e351] | committer: 
Rostislav Pehlivanov

opus_celt: fix some style issues

Signed-off-by: Rostislav Pehlivanov 

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=6d0b62a12b46edbb3d141c356a7bcf7856a1e351
---

 libavcodec/opus_celt.c | 14 +-
 1 file changed, 5 insertions(+), 9 deletions(-)

diff --git a/libavcodec/opus_celt.c b/libavcodec/opus_celt.c
index af3c100..aee8ddc 100644
--- a/libavcodec/opus_celt.c
+++ b/libavcodec/opus_celt.c
@@ -700,6 +700,7 @@ static void celt_decode_bands(CeltFrame *f, OpusRangeCoder 
*rc)
 memset(f->block[1].coeffs, 0, sizeof(f->block[0].coeffs));
 
 for (i = f->start_band; i < f->end_band; i++) {
+uint32_t cm[2] = { (1 << f->blocks) - 1, (1 << f->blocks) - 1 };
 int band_offset = ff_celt_freq_bands[i] << f->size;
 int band_size   = ff_celt_freq_range[i] << f->size;
 float *X = f->block[0].coeffs + band_offset;
@@ -708,8 +709,7 @@ static void celt_decode_bands(CeltFrame *f, OpusRangeCoder 
*rc)
 int consumed = opus_rc_tell_frac(rc);
 float *norm2 = norm + 8 * 100;
 int effective_lowband = -1;
-unsigned int cm[2];
-int b;
+int b = 0;
 
 /* Compute how many bits we want to allocate to this band */
 if (i != f->start_band)
@@ -718,15 +718,14 @@ static void celt_decode_bands(CeltFrame *f, 
OpusRangeCoder *rc)
 if (i <= f->coded_bands - 1) {
 int curr_balance = f->remaining / FFMIN(3, f->coded_bands-i);
 b = av_clip_uintp2(FFMIN(f->remaining2 + 1, f->pulses[i] + 
curr_balance), 14);
-} else
-b = 0;
+}
 
 if (ff_celt_freq_bands[i] - ff_celt_freq_range[i] >= 
ff_celt_freq_bands[f->start_band] &&
 (update_lowband || lowband_offset == 0))
 lowband_offset = i;
 
 /* Get a conservative estimate of the collapse_mask's for the bands 
we're
-going to be folding from. */
+   going to be folding from. */
 if (lowband_offset != 0 && (f->spread != CELT_SPREAD_AGGRESSIVE ||
 f->blocks > 1 || f->tf_change[i] < 0)) {
 int foldstart, foldend;
@@ -744,10 +743,7 @@ static void celt_decode_bands(CeltFrame *f, OpusRangeCoder 
*rc)
 cm[0] |= f->block[0].collapse_masks[j];
 cm[1] |= f->block[f->channels - 1].collapse_masks[j];
 }
-} else
-/* Otherwise, we'll be using the LCG to fold, so all blocks will 
(almost
-always) be non-zero.*/
-cm[0] = cm[1] = (1 << f->blocks) - 1;
+}
 
 if (f->dual_stereo && i == f->intensity_stereo) {
 /* Switch off dual stereo to do intensity */

___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog

[FFmpeg-cvslog] opusenc: do not signal digital silence

2017-04-07 Thread Rostislav Pehlivanov

ffmpeg | branch: master | Rostislav Pehlivanov  | Sat Apr  
8 00:48:28 2017 +0100| [c39fba703bf35da0b4b894b2b33978c205507341] | committer: 
Rostislav Pehlivanov

opusenc: do not signal digital silence

Apparently its only use is to enable comfort noise/error recovery.

Signed-off-by: Rostislav Pehlivanov 

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=c39fba703bf35da0b4b894b2b33978c205507341
---

 libavcodec/opusenc.c | 12 +++-
 1 file changed, 3 insertions(+), 9 deletions(-)

diff --git a/libavcodec/opusenc.c b/libavcodec/opusenc.c
index d80ae7a..4e0781b 100644
--- a/libavcodec/opusenc.c
+++ b/libavcodec/opusenc.c
@@ -249,9 +249,9 @@ static void celt_frame_mdct(OpusEncContext *s, CeltFrame *f)
 }
 
 /* Fills the bands and normalizes them */
-static int celt_frame_map_norm_bands(OpusEncContext *s, CeltFrame *f)
+static void celt_frame_map_norm_bands(OpusEncContext *s, CeltFrame *f)
 {
-int i, j, ch, noise = 0;
+int i, j, ch;
 
 for (ch = 0; ch < f->channels; ch++) {
 CeltBlock *block = &f->block[ch];
@@ -274,10 +274,8 @@ static int celt_frame_map_norm_bands(OpusEncContext *s, 
CeltFrame *f)
 
 /* CELT_ENERGY_SILENCE is what the decoder uses and its not 
-infinity */
 block->energy[i] = FFMAX(block->energy[i], CELT_ENERGY_SILENCE);
-noise |= block->energy[i] > CELT_ENERGY_SILENCE;
 }
 }
-return !noise;
 }
 
 static void celt_enc_tf(OpusRangeCoder *rc, CeltFrame *f)
@@ -811,11 +809,7 @@ static void celt_encode_frame(OpusEncContext *s, 
OpusRangeCoder *rc, CeltFrame *
 /* Not implemented */
 }
 celt_frame_mdct(s, f);
-f->silence = celt_frame_map_norm_bands(s, f);
-if (f->silence) {
-f->framebits = 1;
-return;
-}
+celt_frame_map_norm_bands(s, f);
 
 ff_opus_rc_enc_log(rc, f->silence, 15);
 

___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog

[FFmpeg-cvslog] mdct15: use existing complex multiplication macros

2017-04-07 Thread Rostislav Pehlivanov

ffmpeg | branch: master | Rostislav Pehlivanov  | Sat Apr  
1 19:41:52 2017 +0100| [db788de8ecdee271b59b04c9cbb97320ec28c9ea] | committer: 
Rostislav Pehlivanov

mdct15: use existing complex multiplication macros

Signed-off-by: Rostislav Pehlivanov 

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=db788de8ecdee271b59b04c9cbb97320ec28c9ea
---

 libavcodec/mdct15.c | 36 +++-
 1 file changed, 15 insertions(+), 21 deletions(-)

diff --git a/libavcodec/mdct15.c b/libavcodec/mdct15.c
index e038209..8c42ece 100644
--- a/libavcodec/mdct15.c
+++ b/libavcodec/mdct15.c
@@ -33,17 +33,12 @@
 #include "libavutil/attributes.h"
 #include "libavutil/common.h"
 
-#include "avfft.h"
 #include "mdct15.h"
 
-// complex c = a * b
-#define CMUL3(cre, cim, are, aim, bre, bim)  \
-do { \
-cre = are * bre - aim * bim; \
-cim = are * bim + aim * bre; \
-} while (0)
+#define FFT_FLOAT 1
+#include "fft-internal.h"
 
-#define CMUL(c, a, b) CMUL3((c).re, (c).im, (a).re, (a).im, (b).re, (b).im)
+#define CMUL3(c, a, b) CMUL((c).re, (c).im, (a).re, (a).im, (b).re, (b).im)
 
 av_cold void ff_mdct15_uninit(MDCT15Context **ps)
 {
@@ -236,18 +231,18 @@ static void fft15(const FFTComplex exptab[22], FFTComplex 
*out, const FFTComplex
 for (k = 0; k < 5; k++) {
 FFTComplex t[2];
 
-CMUL(t[0], tmp2[k], exptab[k]);
-CMUL(t[1], tmp3[k], exptab[2 * k]);
+CMUL3(t[0], tmp2[k], exptab[k]);
+CMUL3(t[1], tmp3[k], exptab[2 * k]);
 out[stride*k].re = tmp1[k].re + t[0].re + t[1].re;
 out[stride*k].im = tmp1[k].im + t[0].im + t[1].im;
 
-CMUL(t[0], tmp2[k], exptab[k + 5]);
-CMUL(t[1], tmp3[k], exptab[2 * (k + 5)]);
+CMUL3(t[0], tmp2[k], exptab[k + 5]);
+CMUL3(t[1], tmp3[k], exptab[2 * (k + 5)]);
 out[stride*(k + 5)].re = tmp1[k].re + t[0].re + t[1].re;
 out[stride*(k + 5)].im = tmp1[k].im + t[0].im + t[1].im;
 
-CMUL(t[0], tmp2[k], exptab[k + 10]);
-CMUL(t[1], tmp3[k], exptab[2 * k + 5]);
+CMUL3(t[0], tmp2[k], exptab[k + 10]);
+CMUL3(t[1], tmp3[k], exptab[2 * k + 5]);
 out[stride*(k + 10)].re = tmp1[k].re + t[0].re + t[1].re;
 out[stride*(k + 10)].im = tmp1[k].im + t[0].im + t[1].im;
 }
@@ -272,7 +267,7 @@ static void mdct15(MDCT15Context *s, float *dst, const 
float *src, ptrdiff_t str
 re =  src[2*k-len4] - src[1*len3-1-2*k];
 im = -src[2*k+len4] - src[5*len4-1-2*k];
 }
-CMUL3(fft15in[j].re, fft15in[j].im, re, im, 
s->twiddle_exptab[k].re, -s->twiddle_exptab[k].im);
+CMUL(fft15in[j].re, fft15in[j].im, re, im, 
s->twiddle_exptab[k].re, -s->twiddle_exptab[k].im);
 }
 fft15(s->exptab, s->tmp + s->ptwo_fft.revtab[i], fft15in, l_ptwo);
 }
@@ -287,9 +282,8 @@ static void mdct15(MDCT15Context *s, float *dst, const 
float *src, ptrdiff_t str
 const int i0 = len8 + i, i1 = len8 - i - 1;
 const int s0 = s->pfa_postreindex[i0], s1 = s->pfa_postreindex[i1];
 
-CMUL3(im1, re0, s->tmp[s1].re, s->tmp[s1].im, 
s->twiddle_exptab[i1].im, s->twiddle_exptab[i1].re);
-CMUL3(im0, re1, s->tmp[s0].re, s->tmp[s0].im, 
s->twiddle_exptab[i0].im, s->twiddle_exptab[i0].re);
-
+CMUL(im1, re0, s->tmp[s1].re, s->tmp[s1].im, s->twiddle_exptab[i1].im, 
s->twiddle_exptab[i1].re);
+CMUL(im0, re1, s->tmp[s0].re, s->tmp[s0].im, s->twiddle_exptab[i0].im, 
s->twiddle_exptab[i0].re);
 dst[2*i1*stride ] = re0;
 dst[2*i1*stride + stride] = im0;
 dst[2*i0*stride ] = re1;
@@ -310,7 +304,7 @@ static void imdct15_half(MDCT15Context *s, float *dst, 
const float *src,
 for (j = 0; j < 15; j++) {
 const int k = s->pfa_prereindex[i*15 + j];
 FFTComplex tmp = { *(in2 - 2*k*stride), *(in1 + 2*k*stride) };
-CMUL(fft15in[j], tmp, s->twiddle_exptab[k]);
+CMUL3(fft15in[j], tmp, s->twiddle_exptab[k]);
 }
 fft15(s->exptab, s->tmp + s->ptwo_fft.revtab[i], fft15in, l_ptwo);
 }
@@ -325,8 +319,8 @@ static void imdct15_half(MDCT15Context *s, float *dst, 
const float *src,
 const int i0 = len8 + i, i1 = len8 - i - 1;
 const int s0 = s->pfa_postreindex[i0], s1 = s->pfa_postreindex[i1];
 
-CMUL3(re0, im1, s->tmp[s1].im, s->tmp[s1].re,  
s->twiddle_exptab[i1].im, s->twiddle_exptab[i1].re);
-CMUL3(re1, im0, s->tmp[s0].im, s->tmp[s0].re,  
s->twiddle_exptab[i0].im, s->twiddle_exptab[i0].re);
+CMUL(re0, im1, s->tmp[s1].im, s->tmp[s1].re,  
s->twiddle_exptab[i1].im, s->twiddle_exptab[i1].re);
+CMUL(re1, im0, s->tmp[s0].im, s->tmp[s0].re,  
s->twiddle_exptab[i0].im, s->twiddle_exptab[i0].re);
 z[i1].re = scale * re0;
 z[i1].im = scale * im0;
 z[i0].re = scale * re1;

[FFmpeg-cvslog] opusenc: switch between intra/inter mode for coarse energy

2017-04-07 Thread Rostislav Pehlivanov

ffmpeg | branch: master | Rostislav Pehlivanov  | Sat Apr  
8 00:53:31 2017 +0100| [d0a3143193ff3284f3dfbcc173bb94c9b91077d0] | committer: 
Rostislav Pehlivanov

opusenc: switch between intra/inter mode for coarse energy

Saves around 5kbps.

Signed-off-by: Rostislav Pehlivanov 

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=d0a3143193ff3284f3dfbcc173bb94c9b91077d0
---

 libavcodec/opus_celt.h |  1 -
 libavcodec/opusenc.c   | 36 
 2 files changed, 28 insertions(+), 9 deletions(-)

diff --git a/libavcodec/opus_celt.h b/libavcodec/opus_celt.h
index 1ed57f3..f0d55d6 100644
--- a/libavcodec/opus_celt.h
+++ b/libavcodec/opus_celt.h
@@ -100,7 +100,6 @@ struct CeltFrame {
 int end_band;
 int coded_bands;
 int transient;
-int intra;
 int pfilter;
 int skip_band_floor;
 int tf_select;
diff --git a/libavcodec/opusenc.c b/libavcodec/opusenc.c
index 4e0781b..5f5700e 100644
--- a/libavcodec/opusenc.c
+++ b/libavcodec/opusenc.c
@@ -630,19 +630,20 @@ static void ff_celt_enc_bitalloc(OpusRangeCoder *rc, 
CeltFrame *f)
 }
 }
 
-static void celt_quant_coarse(OpusEncContext *s, OpusRangeCoder *rc, CeltFrame 
*f)
+static void exp_quant_coarse(OpusRangeCoder *rc, CeltFrame *f,
+ float last_energy[][CELT_MAX_BANDS], int intra)
 {
 int i, ch;
 float alpha, beta, prev[2] = { 0, 0 };
-const uint8_t *pmod = ff_celt_coarse_energy_dist[f->size][f->intra];
+const uint8_t *pmod = ff_celt_coarse_energy_dist[f->size][intra];
 
 /* Inter is really just differential coding */
 if (opus_rc_tell(rc) + 3 <= f->framebits)
-ff_opus_rc_enc_log(rc, f->intra, 3);
+ff_opus_rc_enc_log(rc, intra, 3);
 else
-f->intra = 0;
+intra = 0;
 
-if (f->intra) {
+if (intra) {
 alpha = 0.0f;
 beta  = 1.0f - 4915.0f/32768.0f;
 } else {
@@ -654,7 +655,7 @@ static void celt_quant_coarse(OpusEncContext *s, 
OpusRangeCoder *rc, CeltFrame *
 for (ch = 0; ch < f->channels; ch++) {
 CeltBlock *block = &f->block[ch];
 const int left = f->framebits - opus_rc_tell(rc);
-const float last = FFMAX(-9.0f, s->last_quantized_energy[ch][i]);
+const float last = FFMAX(-9.0f, last_energy[ch][i]);
 float diff = block->energy[i] - prev[ch] - last*alpha;
 int q_en = lrintf(diff);
 if (left >= 15) {
@@ -673,6 +674,26 @@ static void celt_quant_coarse(OpusEncContext *s, 
OpusRangeCoder *rc, CeltFrame *
 }
 }
 
+static void celt_quant_coarse(OpusRangeCoder *rc, CeltFrame *f,
+  float last_energy[][CELT_MAX_BANDS])
+{
+uint32_t inter, intra;
+OPUS_RC_CHECKPOINT_SPAWN(rc);
+
+exp_quant_coarse(rc, f, last_energy, 1);
+intra = OPUS_RC_CHECKPOINT_BITS(rc);
+
+OPUS_RC_CHECKPOINT_ROLLBACK(rc);
+
+exp_quant_coarse(rc, f, last_energy, 0);
+inter = OPUS_RC_CHECKPOINT_BITS(rc);
+
+if (inter > intra) { /* Unlikely */
+OPUS_RC_CHECKPOINT_ROLLBACK(rc);
+exp_quant_coarse(rc, f, last_energy, 1);
+}
+}
+
 static void celt_quant_fine(OpusRangeCoder *rc, CeltFrame *f)
 {
 int i, ch;
@@ -823,7 +844,7 @@ static void celt_encode_frame(OpusEncContext *s, 
OpusRangeCoder *rc, CeltFrame *
 if (f->size && opus_rc_tell(rc) + 3 <= f->framebits)
 ff_opus_rc_enc_log(rc, f->transient, 3);
 
-celt_quant_coarse (s, rc, f);
+celt_quant_coarse(rc, f, s->last_quantized_energy);
 celt_enc_tf  (rc, f);
 ff_celt_enc_bitalloc(rc, f);
 celt_quant_fine  (rc, f);
@@ -871,7 +892,6 @@ static void ff_opus_psy_celt_frame_setup(OpusEncContext *s, 
CeltFrame *f, int in
 f->silence = 0;
 f->pfilter = 0;
 f->transient = 0;
-f->intra = 1;
 f->tf_select = 0;
 f->anticollapse = 0;
 f->alloc_trim = 5;

___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog

[FFmpeg-cvslog] opus_rc: add macros for checkpoint rollback

2017-04-07 Thread Rostislav Pehlivanov

ffmpeg | branch: master | Rostislav Pehlivanov  | Sun Apr  
2 20:55:30 2017 +0100| [a66121d964558348ffcc2fb820853c9381a6a8d2] | committer: 
Rostislav Pehlivanov

opus_rc: add macros for checkpoint rollback

Signed-off-by: Rostislav Pehlivanov 

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=a66121d964558348ffcc2fb820853c9381a6a8d2
---

 libavcodec/opus_rc.h | 14 ++
 1 file changed, 14 insertions(+)

diff --git a/libavcodec/opus_rc.h b/libavcodec/opus_rc.h
index 821eead..627f832 100644
--- a/libavcodec/opus_rc.h
+++ b/libavcodec/opus_rc.h
@@ -109,4 +109,18 @@ void ff_opus_rc_dec_raw_init(OpusRangeCoder *rc, const 
uint8_t *rightend, ui
 void ff_opus_rc_enc_end(OpusRangeCoder *rc, uint8_t *dst, int size);
 void ff_opus_rc_enc_init(OpusRangeCoder *rc);
 
+#define OPUS_RC_CHECKPOINT_UPDATE(rc) \
+rc_rollback_bits = opus_rc_tell_frac(rc); \
+rc_rollback_ctx  = *rc
+
+#define OPUS_RC_CHECKPOINT_SPAWN(rc) \
+uint32_t rc_rollback_bits = opus_rc_tell_frac(rc); \
+OpusRangeCoder rc_rollback_ctx = *rc \
+
+#define OPUS_RC_CHECKPOINT_BITS(rc) \
+(opus_rc_tell_frac(rc) - rc_rollback_bits)
+
+#define OPUS_RC_CHECKPOINT_ROLLBACK(rc) \
+memcpy(rc, &rc_rollback_ctx, sizeof(OpusRangeCoder)); \
+
 #endif /* AVCODEC_OPUS_RC_H */

___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog

[FFmpeg-cvslog] ffmpeg: use av_stream_new_side_data() in init_output_stream

2017-04-07 Thread James Almer

ffmpeg | branch: master | James Almer  | Sat Apr  8 01:12:57 
2017 -0300| [b438a7868c8a672f103a6b50e9572f5d31f23ce3] | committer: James Almer

ffmpeg: use av_stream_new_side_data() in init_output_stream

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=b438a7868c8a672f103a6b50e9572f5d31f23ce3
---

 ffmpeg.c | 16 
 1 file changed, 4 insertions(+), 12 deletions(-)

diff --git a/ffmpeg.c b/ffmpeg.c
index 55467c0..e4b94b2 100644
--- a/ffmpeg.c
+++ b/ffmpeg.c
@@ -3458,22 +3458,14 @@ static int init_output_stream(OutputStream *ost, char 
*error, int error_len)
 if (ost->enc_ctx->nb_coded_side_data) {
 int i;
 
-ost->st->side_data = av_realloc_array(NULL, 
ost->enc_ctx->nb_coded_side_data,
-  sizeof(*ost->st->side_data));
-if (!ost->st->side_data)
-return AVERROR(ENOMEM);
-
 for (i = 0; i < ost->enc_ctx->nb_coded_side_data; i++) {
 const AVPacketSideData *sd_src = 
&ost->enc_ctx->coded_side_data[i];
-AVPacketSideData *sd_dst = &ost->st->side_data[i];
+uint8_t *dst_data;
 
-sd_dst->data = av_malloc(sd_src->size);
-if (!sd_dst->data)
+dst_data = av_stream_new_side_data(ost->st, sd_src->type, 
sd_src->size);
+if (!dst_data)
 return AVERROR(ENOMEM);
-memcpy(sd_dst->data, sd_src->data, sd_src->size);
-sd_dst->size = sd_src->size;
-sd_dst->type = sd_src->type;
-ost->st->nb_side_data++;
+memcpy(dst_data, sd_src->data, sd_src->size);
 }
 }
 

___
ffmpeg-cvslog mailing list
ffmpeg-cvslog@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog

[FFmpeg-cvslog] Merge commit '239d02eff3ffe9f7d40caa21dde50fb4a0e94c24'

[FFmpeg-cvslog] avisynth: Cast to the right type when loading avisynth library functions

[FFmpeg-cvslog] Merge commit '3ee5f25d37315b27f0e2d47aa69fc445545ee2e3'

[FFmpeg-cvslog] dxva2: Adjust printf length modifiers where appropriate

[FFmpeg-cvslog] mjpegdec: Check return values of functions that may fail

[FFmpeg-cvslog] Merge commit '212c6a1d70df011b6f2a2aa02f7677503287bd00'

[FFmpeg-cvslog] avcodec/h264: Check weight values to be within the specs limits.

[FFmpeg-cvslog] avcodec/tta: Don't try to read more than MIN_CACHE_BITS bits

[FFmpeg-cvslog] lavc/mjpegdec: allow failure while decoding APP

[FFmpeg-cvslog] h264: don't re-call ff_h264_direct_ref_list_init() w/ frame-mt.

[FFmpeg-cvslog] pthread_frame: make accesses to debug field be protected by owner lock.

[FFmpeg-cvslog] avcodec/utvideodec: add support for gradient prediction

[FFmpeg-cvslog] ffmpeg: use av_stream_new_side_data()

[FFmpeg-cvslog] lavf: use the new bitstream filter for extracting extradata

[FFmpeg-cvslog] avcodec/hevc_parse: ignore all non parameter set NAL units in extradata

[FFmpeg-cvslog] doc/libav-merge: remove line about extract_extradata_bsf usage

[FFmpeg-cvslog] vf_pad: center image on padded area if negative x/y

[FFmpeg-cvslog] opusenc: minor style changes

[FFmpeg-cvslog] opusenc: remove unused header entries and simplify normalization

[FFmpeg-cvslog] opus_celt: fix some style issues

[FFmpeg-cvslog] opusenc: do not signal digital silence

[FFmpeg-cvslog] mdct15: use existing complex multiplication macros

[FFmpeg-cvslog] opusenc: switch between intra/inter mode for coarse energy

[FFmpeg-cvslog] opus_rc: add macros for checkpoint rollback

[FFmpeg-cvslog] ffmpeg: use av_stream_new_side_data() in init_output_stream

25 matches

Site Navigation

Mail list logo

Footer information