August 2013 - ffmpeg-devel - lists.ffmpeg.org

[PATCH 0/5] Scalar product, take 2
by Christophe Gisquet 06 Jun '14

06 Jun '14

Following the previous comments, I have opted for padding of the buffers in order to use the DSP functions. The troublesome cases for wma lossless are still untested. Patches 3 and 4 still try to address them, 3 being more secure than 4 (which is the shape of the best solution in my opinion). TAK decoder was modified in the same fashion. Christophe Gisquet (5): ra144: use scalarproduct_int16 wma lossless: reuse scalarproduct_and_madd_int16 wma lossless: switching back to C dsp function wma lossless: pad coeff buffer with 0 takdec: pad filter coeff buffer for DSP functions libavcodec/dsputil.c | 4 +-- libavcodec/dsputil.h | 3 +++ libavcodec/ra144.c | 14 ++++------ libavcodec/ra144.h | 6 ++++- libavcodec/ra144dec.c | 3 +++ libavcodec/ra144enc.c | 6 +++-- libavcodec/takdec.c | 33 +++++++----------------- libavcodec/wmalosslessdec.c | 63 ++++++++++++++++++++++++--------------------- 8 files changed, 65 insertions(+), 67 deletions(-) -- 1.8.0.msysgit.0

3 26

[PATCH] tests: add overlay rgb and yuv444 tests
by Stefano Sabatini 04 Jun '14

04 Jun '14

Also rename overlay test to overlay_yuv420 accordingly. --- tests/fate/avfilter.mak | 4 +++- tests/lavfi-regression.sh | 4 +++- tests/ref/lavfi/overlay | 1 - tests/ref/lavfi/overlay_rgb | 1 + tests/ref/lavfi/overlay_yuv420 | 1 + tests/ref/lavfi/overlay_yuv444 | 1 + 6 files changed, 9 insertions(+), 3 deletions(-) delete mode 100644 tests/ref/lavfi/overlay create mode 100644 tests/ref/lavfi/overlay_rgb create mode 100644 tests/ref/lavfi/overlay_yuv420 create mode 100644 tests/ref/lavfi/overlay_yuv444 diff --git a/tests/fate/avfilter.mak b/tests/fate/avfilter.mak index 15c8595..8ac8d41 100644 --- a/tests/fate/avfilter.mak +++ b/tests/fate/avfilter.mak @@ -14,7 +14,9 @@ FATE_LAVFI = fate-lavfi-alphaextract_rgb \ fate-lavfi-il \ fate-lavfi-life \ fate-lavfi-null \ - fate-lavfi-overlay \ + fate-lavfi-overlay_rgb \ + fate-lavfi-overlay_yuv420 \ + fate-lavfi-overlay_yuv444 \ fate-lavfi-pad \ fate-lavfi-pixfmts_copy \ fate-lavfi-pixfmts_crop \ diff --git a/tests/lavfi-regression.sh b/tests/lavfi-regression.sh index d6aba4c..9335124 100755 --- a/tests/lavfi-regression.sh +++ b/tests/lavfi-regression.sh @@ -47,7 +47,9 @@ do_lavfi "fade" "fade=in:5:15,fade=out:30:15" do_lavfi "hue" "hue=s=sin(2*PI*t)+1" do_lavfi "idet" "idet" do_lavfi "null" "null" -do_lavfi "overlay" "split[m],scale=88:72,pad=96:80:4:4[o2];[m]fifo[o1],[o1][o2]overlay=240:16" +do_lavfi "overlay_rgb" "split[m],scale=88:72,pad=96:80:4:4[o2];[m]fifo[o1],[o1][o2]overlay=240:16:format=rgb" +do_lavfi "overlay_yuv420" "split[m],scale=88:72,pad=96:80:4:4[o2];[m]fifo[o1],[o1][o2]overlay=240:16:format=yuv420" +do_lavfi "overlay_yuv444" "split[m],scale=88:72,pad=96:80:4:4[o2];[m]fifo[o1],[o1][o2]overlay=240:16:format=yuv444" do_lavfi "pad" "pad=iw*1.5:ih*1.5:iw*0.3:ih*0.2" do_lavfi "pp" "pp=be/hb/vb/tn/l5/al" do_lavfi "pp2" "pp=be/fq:16/h1/v1/lb" diff --git a/tests/ref/lavfi/overlay b/tests/ref/lavfi/overlay deleted file mode 100644 index ca70370..0000000 --- a/tests/ref/lavfi/overlay +++ /dev/null @@ -1 +0,0 @@ -overlay 4e85d551db9e56e8faddc1fc70fd6a00 diff --git a/tests/ref/lavfi/overlay_rgb b/tests/ref/lavfi/overlay_rgb new file mode 100644 index 0000000..ce4e30b --- /dev/null +++ b/tests/ref/lavfi/overlay_rgb @@ -0,0 +1 @@ +overlay_rgb abea47d9e61d7eab40c4734294519524 diff --git a/tests/ref/lavfi/overlay_yuv420 b/tests/ref/lavfi/overlay_yuv420 new file mode 100644 index 0000000..3d12251 --- /dev/null +++ b/tests/ref/lavfi/overlay_yuv420 @@ -0,0 +1 @@ +overlay_yuv420 4e85d551db9e56e8faddc1fc70fd6a00 diff --git a/tests/ref/lavfi/overlay_yuv444 b/tests/ref/lavfi/overlay_yuv444 new file mode 100644 index 0000000..b83e605 --- /dev/null +++ b/tests/ref/lavfi/overlay_yuv444 @@ -0,0 +1 @@ +overlay_yuv444 16215d0b4054a44bbe7f6c46bd97d235 -- 1.7.9.5

3 5

[PATCH] Fix nonsense non-mod16 AMV flipping code.
by Reimar Döffinger 02 Jun '14

02 Jun '14

It is obviously nonsense since it produces wrong results or even crashes (crashes should be encode-only though). Fixes trac issue #1092. Signed-off-by: Reimar Döffinger <Reimar.Doeffinger(a)gmx.de> --- libavcodec/mjpegdec.c | 3 +-- libavcodec/mjpegenc.c | 2 +- libavcodec/mpegvideo_enc.c | 5 ----- 3 files changed, 2 insertions(+), 8 deletions(-) diff --git a/libavcodec/mjpegdec.c b/libavcodec/mjpegdec.c index a841384..fe54b45 100644 --- a/libavcodec/mjpegdec.c +++ b/libavcodec/mjpegdec.c @@ -972,8 +972,7 @@ static int mjpeg_decode_scan(MJpegDecodeContext *s, int nb_components, int Ah, s->coefs_finished[c] |= 1; if (s->flipped) { // picture should be flipped upside-down for this codec - int offset = (linesize[c] * (s->v_scount[i] * - (8 * s->mb_height - ((s->height / s->v_max) & 7)) - 1)); + int offset = linesize[c] * (s->v_scount[c] * s->height / s->v_max - 1); data[c] += offset; reference_data[c] += offset; linesize[c] *= -1; diff --git a/libavcodec/mjpegenc.c b/libavcodec/mjpegenc.c index 9326d4d..4788c47 100644 --- a/libavcodec/mjpegenc.c +++ b/libavcodec/mjpegenc.c @@ -466,7 +466,7 @@ static int amv_encode_picture(AVCodecContext *avctx, AVPacket *pkt, //picture should be flipped upside-down for(i=0; i < 3; i++) { - pic.data[i] += (pic.linesize[i] * (s->mjpeg_vsample[i] * (8 * s->mb_height -((s->height/V_MAX)&7)) - 1 )); + pic.data[i] += pic.linesize[i] * (s->mjpeg_vsample[i] * s->height / V_MAX - 1); pic.linesize[i] *= -1; } return ff_MPV_encode_picture(avctx, pkt, &pic, got_packet); diff --git a/libavcodec/mpegvideo_enc.c b/libavcodec/mpegvideo_enc.c index 6d89cfe..bb4be28 100644 --- a/libavcodec/mpegvideo_enc.c +++ b/libavcodec/mpegvideo_enc.c @@ -1012,11 +1012,6 @@ static int load_input_picture(MpegEncContext *s, AVFrame *pic_arg) uint8_t *src = pic_arg->data[i]; uint8_t *dst = pic->data[i]; - if(s->codec_id == CODEC_ID_AMV){ - av_assert0(!(s->avctx->flags & CODEC_FLAG_EMU_EDGE)); - h= ((s->height+15)/16*16)>>v_shift; - } - if (!s->avctx->rc_buffer_size) dst += INPLACE_OFFSET; -- 1.7.10

3 13

[PATCH 0/3] sbrdsp: SSE2 functions
by Christophe Gisquet 15 May '14

15 May '14

Drop SSE-only functions as much as possible to avoid any clutter. pre_shuffle and apply_noise were therefore affected. Christophe Gisquet (3): x86: sbrdsp: implement SSE2 qmf_pre_shuffle x86: sbrdsp: implement SSE qmf_deint_neg x86: sbrdsp: implement SSE2 hf_apply_noise libavcodec/aacsbrdata.h | 6 +- libavcodec/x86/sbrdsp.asm | 150 +++++++++++++++++++++++++++++++++++++++++++ libavcodec/x86/sbrdsp_init.c | 20 ++++++ 3 files changed, 175 insertions(+), 1 deletion(-) -- 1.8.0.msysgit.0

2 14

[PATCH] rtpdec: make the NTP time values unsigned.
by flameeyes＠gmail.com 09 May '14

09 May '14

As per RFC3550, section 4, the NTP time is provided as 64-bit unsigned integer, so follow the same logic here. --- libavformat/rtpdec.h | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/libavformat/rtpdec.h b/libavformat/rtpdec.h index 86af2b9..87fd419 100644 --- a/libavformat/rtpdec.h +++ b/libavformat/rtpdec.h @@ -148,8 +148,8 @@ struct RTPDemuxContext { RTPStatistics statistics; ///< Statistics for this stream (used by RTCP receiver reports) /* rtcp sender statistics receive */ - int64_t last_rtcp_ntp_time; // TODO: move into statistics - int64_t first_rtcp_ntp_time; // TODO: move into statistics + uint64_t last_rtcp_ntp_time; // TODO: move into statistics + uint64_t first_rtcp_ntp_time; // TODO: move into statistics uint32_t last_rtcp_timestamp; // TODO: move into statistics /* rtcp sender statistics */ -- 1.7.1.1

4 3

[PATCH 1/4] av_d2q: Add a special case for |value| > MAX and |value| < 1/MAX
by Michael Niedermayer 12 Apr '14

12 Apr '14

Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> --- libavutil/rational.c | 2 ++ tests/ref/fate/parseutils | 2 +- 2 files changed, 3 insertions(+), 1 deletion(-) diff --git a/libavutil/rational.c b/libavutil/rational.c index 912abc3..62473b4 100644 --- a/libavutil/rational.c +++ b/libavutil/rational.c @@ -115,6 +115,8 @@ AVRational av_d2q(double d, int max) exponent = FFMAX( (int)(log(fabs(d) + 1e-20)/LOG2), 0); den = 1LL << (61 - exponent); av_reduce(&a.num, &a.den, llrint(d * den), den, max); + if ((!a.num || !a.den) && d && max>0 && max<INT_MAX) + av_reduce(&a.num, &a.den, llrint(d * den), den, INT_MAX); return a; } diff --git a/tests/ref/fate/parseutils b/tests/ref/fate/parseutils index 5fa3bf4..1482452 100644 --- a/tests/ref/fate/parseutils +++ b/tests/ref/fate/parseutils @@ -21,7 +21,7 @@ Testing av_parse_video_rate() '.23' -> 23/100 OK '-.23' -> -23/100 ERROR '-0.234' -> -117/500 ERROR -'-0.0000001' -> 0/1 ERROR +'-0.0000001' -> -1/10000000 ERROR ' 21332.2324 ' -> 917286/43 OK ' -21332.2324 ' -> -917286/43 ERROR -- 1.7.9.5

3 9

trac for dshow "todos"
by Roger Pack 07 Mar '14

07 Mar '14

I was contemplating creating a list of "known lacking features" and bugs and wishlist items for dshow. My first thought had been to create it as a wiki page. Would that be ok or should I file "everything" as trac items assigned to myself, even if they might never be implemented, any opinions here? Thank you. -roger-

2 2

[PATCH]Support qclp in mov before stsd version 1, v2
by Carl Eugen Hoyos 21 Feb '14

21 Feb '14

Hi! Attached patch fixes the following sample as reported by Compn on irc: http://web.archive.org/web/20071027002243/http://www.americaone.org/video/i… As the original patch, it does not break samples with variable frame size like h263.mov Contrary to the original patch, this patch does not break stsd version 1 samples with bitrate < FULL. Please comment, Carl Eugen

3 14

ffmpeg configure for ios - using asm options causes MPx (mp1, mp2, mp3) codecs sound problems
by Murat Sudan 30 Jan '14

30 Jan '14

I've made many tests to get good performance for video decoding with ffmpeg on low cpu idevices such as iPad 1 and iPhone 3Gs, and see that using asm option when configuring ffmpeg for cross compiling for ios is very very important for performance.. But when I configure with --enable-asm, the audio stream which using mpx(mp1,mp2,mp) codec sounds squeaky. If I configure with --disable-asm option, than mpx(mp1,mp2,mp) codec sounds well but then the performance is %50 percent of the previous one built with the --enable-asm option... I've also tested these with the ffmpeg versions that are 0.8, 0.9, 0.10, 0.11, 1.0,1.1,1.2 All of them produces same results... Can anyone have the same problem with me ? Also, any idea how can I fix this issue ? please advice me .. (I can not find any patch to fix this issue on the net or in ffmpeg forums.) -mg

3 3

[PATCH 1/2] lavf: set codec time base after finding streams.
by Nicolas George 27 Jan '14

27 Jan '14

For most formats, it is done before reading packets, but for formats without global headers, new streams are added afterwards. Ensure that a time base is set even if no decoder was found. Useful for parsers. Signed-off-by: Nicolas George <nicolas.george(a)normalesup.org> --- libavformat/utils.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/libavformat/utils.c b/libavformat/utils.c index 0f13304..903ada3 100644 --- a/libavformat/utils.c +++ b/libavformat/utils.c @@ -2947,6 +2947,14 @@ int avformat_find_stream_info(AVFormatContext *ic, AVDictionary **options) compute_chapters_end(ic); + for(i = 0; i < ic->nb_streams; i++) { + st = ic->streams[i]; + if ((st->codec->codec_type == AVMEDIA_TYPE_VIDEO || + st->codec->codec_type == AVMEDIA_TYPE_SUBTITLE) && + !st->codec->time_base.num) + st->codec->time_base = st->time_base; + } + find_stream_info_err: for (i=0; i < ic->nb_streams; i++) { if (ic->streams[i]->codec) -- 1.7.10.4

5 8