Re: [libav-devel] [PATCH] Add MagicYUV decoder

2016-06-17 Thread Diego Biurrun
On Fri, Jun 17, 2016 at 12:10:31PM -0400, Vittorio Giovara wrote: > On Fri, Jun 17, 2016 at 10:20 AM, Diego Biurrun wrote: > > On Mon, Jun 06, 2016 at 12:31:56PM -0400, Vittorio Giovara wrote: > >> --- /dev/null > >> +++ b/libavcodec/magicyuv.c > >> @@ -0,0 +1,469 @@ > >> + > >>

Re: [libav-devel] [PATCH] checkasm: Add tests for h264 idct

2016-06-17 Thread Martin Storsjö
On Fri, 17 Jun 2016, Janne Grunau wrote: On 2016-06-17 15:33:20 +0300, Martin Storsjö wrote: The idct tests are inspired by similar tests for vp9 by Ronald Bultje. --- Now using a fixed stride as a multiple of 16 (16 exactly), and testing all valid alignments within that. ---

Re: [libav-devel] [PATCH] ffv1: Error out on unsupported format

2016-06-17 Thread Vittorio Giovara
On Thu, Jun 16, 2016 at 10:44 AM, Luca Barbato wrote: > From: Jerome Martinez > > Transparency is supported only by YUV and within specific bit depths. > --- > > Eventually had a look, added a similar block for the other colorspace as well. > >

[libav-devel] [PATCH] Add MagicYUV decoder

2016-06-17 Thread Vittorio Giovara
From: Paul B Mahol Signed-off-by: Paul B Mahol Signed-off-by: Vittorio Giovara --- In this version - correct handling of raw slices - added size checks - any additional review comment Vittorio Changelog | 1 +

[libav-devel] [PATCH] Add TrueMotion 2.0 Real Time decoder

2016-06-17 Thread Vittorio Giovara
From: Paul B Mahol Signed-off-by: Paul B Mahol Signed-off-by: Vittorio Giovara --- Addressed Luca's and Diego's comments. Vittorio Changelog | 1 + doc/general.texi | 2 + libavcodec/Makefile

Re: [libav-devel] [PATCH 07/28] h264: move a per-field block from decode_slice_header() to field_start()

2016-06-17 Thread Janne Grunau
On 2016-06-09 10:29:53 +0200, Anton Khirnov wrote: > This is a more appropriate place for it. > --- > libavcodec/h264_slice.c | 16 +++- > 1 file changed, 7 insertions(+), 9 deletions(-) ok Janne ___ libav-devel mailing list

Re: [libav-devel] [PATCH] Add MagicYUV decoder

2016-06-17 Thread Vittorio Giovara
On Fri, Jun 17, 2016 at 8:32 AM, Anton Khirnov wrote: > Quoting Vittorio Giovara (2016-06-06 18:31:56) >> From: Paul B Mahol >> >> +s->slice_height = bytestream2_get_le32(); >> +if ((s->slice_height <= 0) || (s->slice_height > INT_MAX - >>

Re: [libav-devel] [PATCH] Add MagicYUV decoder

2016-06-17 Thread Vittorio Giovara
On Fri, Jun 17, 2016 at 10:20 AM, Diego Biurrun wrote: > On Mon, Jun 06, 2016 at 12:31:56PM -0400, Vittorio Giovara wrote: >> --- a/doc/general.texi >> +++ b/doc/general.texi >> @@ -672,6 +672,7 @@ following image formats are supported: >> @item lossless MJPEG @tab X

Re: [libav-devel] [PATCH 06/28] h264: only allow ending a field/starting a new one before finish_setup()

2016-06-17 Thread Janne Grunau
On 2016-06-09 10:29:52 +0200, Anton Khirnov wrote: > Doing this after ff_thread_finish_setup() is called is invalid and can > conflict with reads from the other thread. > --- > libavcodec/h264_slice.c | 34 +++--- > 1 file changed, 19 insertions(+), 15 deletions(-)

Re: [libav-devel] [PATCH 05/28] h264: store {curr, max}_pic_num in the per-slice context

2016-06-17 Thread Janne Grunau
On 2016-06-09 10:29:51 +0200, Anton Khirnov wrote: > While the value of those variables will be constant for the whole frame, > they are only used in two functions called from slice header decoding. > Moving them to the per-slice context allows us to make the H264Context > passed to

Re: [libav-devel] [PATCH 04/28] h264: decode the poc values from the slice header into the per-slice context

2016-06-17 Thread Janne Grunau
On 2016-06-09 10:29:50 +0200, Anton Khirnov wrote: > Copy them into the decoder-global context in field_start(). This avoids > modifying the decoder-global context during bitstream parsing. > --- > libavcodec/h264.h | 5 + > libavcodec/h264_slice.c | 41

Re: [libav-devel] [PATCH 03/28] h264: set mb_aff_frame in frame_start()

2016-06-17 Thread Janne Grunau
On 2016-06-09 10:29:49 +0200, Anton Khirnov wrote: > Avoid unnecessary modification of the decoder-global state in per-slice > code. > --- > libavcodec/h264_slice.c | 8 +++- > 1 file changed, 3 insertions(+), 5 deletions(-) ok Janne ___

Re: [libav-devel] [PATCH 01/28] h264: pass a H2645NAL to slice header decoding

2016-06-17 Thread Janne Grunau
On 2016-06-09 10:29:47 +0200, Anton Khirnov wrote: > Replace the decoder-global nal_unit_type/nal_ref_idc variables with the > per-NAL ones. The decoder-global ones still cannot be removed because > they are used by hwaccels. > --- > libavcodec/h264.c | 2 +- > libavcodec/h264.h | 3

Re: [libav-devel] [PATCH 02/28] h264: move the block starting a new field out of slice_header_parse()

2016-06-17 Thread Janne Grunau
On 2016-06-09 10:29:48 +0200, Anton Khirnov wrote: > There is no bitstream parsing in that block and messing with > decoder-global state is not something that belongs into header parsing. > > Nothing else in this function depends on the value of current_slice, > except for two validity checks.

Re: [libav-devel] [PATCH] Add MagicYUV decoder

2016-06-17 Thread Diego Biurrun
On Mon, Jun 06, 2016 at 12:31:56PM -0400, Vittorio Giovara wrote: > --- a/doc/general.texi > +++ b/doc/general.texi > @@ -672,6 +672,7 @@ following image formats are supported: > @item lossless MJPEG @tab X @tab X > +@item MagicYUV Lossless Video @tab@tab X > @item Microsoft ATC

Re: [libav-devel] [PATCH 2/5] Add TrueMotion 2.0 Real Time decoder

2016-06-17 Thread Diego Biurrun
On Fri, Jun 03, 2016 at 03:51:29PM -0400, Vittorio Giovara wrote: > --- /dev/null > +++ b/libavcodec/truemotion2rt.c > @@ -0,0 +1,227 @@ > +/* > + * Duck TrueMotion 2.0 Real Time Decoder decoder > +#include > +#include > +#include > + > +#include "libavutil/imgutils.h" > +#include

Re: [libav-devel] [PATCH] checkasm: Add tests for h264 idct

2016-06-17 Thread Janne Grunau
On 2016-06-17 15:33:20 +0300, Martin Storsjö wrote: > The idct tests are inspired by similar tests for vp9 by > Ronald Bultje. > --- > Now using a fixed stride as a multiple of 16 (16 exactly), > and testing all valid alignments within that. > --- > tests/checkasm/Makefile | 1 + >

[libav-devel] [PATCH] checkasm: Add tests for h264 idct

2016-06-17 Thread Martin Storsjö
The idct tests are inspired by similar tests for vp9 by Ronald Bultje. --- Now using a fixed stride as a multiple of 16 (16 exactly), and testing all valid alignments within that. --- tests/checkasm/Makefile | 1 + tests/checkasm/checkasm.c | 3 + tests/checkasm/checkasm.h | 1 +

Re: [libav-devel] [PATCH] Add MagicYUV decoder

2016-06-17 Thread Anton Khirnov
Quoting Vittorio Giovara (2016-06-06 18:31:56) > From: Paul B Mahol > > Signed-off-by: Paul B Mahol > Signed-off-by: Vittorio Giovara > --- > Applied Diego's comments. > Vittorio > > Changelog | 1 + > configure

Re: [libav-devel] [PATCH 2/2] checkasm: Add tests for h264 idct

2016-06-17 Thread Martin Storsjö
On Fri, 17 Jun 2016, Janne Grunau wrote: On 2016-06-17 00:26:08 +0300, Martin Storsjö wrote: On Thu, 16 Jun 2016, Janne Grunau wrote: On 2016-06-15 23:26:16 +0300, Martin Storsjö wrote: The idct tests are inspired by similar tests for vp9 by Ronald Bultje. --- Rewrote the 8x8 dct to avoid

Re: [libav-devel] [PATCH 1/2] hwcontext_vaapi: Add driver quirks to the hwdevice

2016-06-17 Thread Mark Thompson
On 17/06/16 09:46, Anton Khirnov wrote: > Quoting Mark Thompson (2016-06-12 19:03:54) >> The driver being used is detected inside av_hwdevice_ctx_init(), and >> the quirks field then set from a table of known drivers. >> >> Also adds the Intel i965 driver quirk (it does not destroy parameter >>

Re: [libav-devel] [PATCH] lavfi: add a QSV scaling filter

2016-06-17 Thread Anton Khirnov
Quoting Diego Biurrun (2016-06-12 21:56:49) > On Sun, Jun 12, 2016 at 08:43:39PM +0200, Anton Khirnov wrote: > > --- > > Now aligning the framesctx (allocated) dimensions to 32. > > --- > > Changelog | 1 + > > configure | 1 + > > libavfilter/Makefile

Re: [libav-devel] [PATCH 1/2] hwcontext_vaapi: Add driver quirks to the hwdevice

2016-06-17 Thread Anton Khirnov
Quoting Mark Thompson (2016-06-12 19:03:54) > The driver being used is detected inside av_hwdevice_ctx_init(), and > the quirks field then set from a table of known drivers. > > Also adds the Intel i965 driver quirk (it does not destroy parameter > buffers used in a call to vaRenderPicture()) and