Re: [FFmpeg-devel] [PATCH 5/5] avcodec/vc1: store zero MVs for all blocks in a MB

2018-05-28 Thread Carl Eugen Hoyos
2018-05-20 11:05 GMT+02:00, Jerome Borsboom : >>> Direct prediction for interlace frame B pictures references the mv in the >>> second block in an MB in the backward reference frame for the twomv case. >>> When the backward reference frame is an I frame, this value may be unset. >>> >>> Signed-off-

Re: [FFmpeg-devel] [PATCH 5/5] avcodec/vc1: store zero MVs for all blocks in a MB

2018-05-20 Thread Jerome Borsboom
>> Direct prediction for interlace frame B pictures references the mv in the >> second block in an MB in the backward reference frame for the twomv case. >> When the backward reference frame is an I frame, this value may be unset. >> >> Signed-off-by: Jerome Borsboom >> --- >> libavcodec/vc1_blo

Re: [FFmpeg-devel] [PATCH 5/5] avcodec/vc1: store zero MVs for all blocks in a MB

2018-05-20 Thread Carl Eugen Hoyos
2018-05-20 3:45 GMT+02:00, James Almer : > I see no emms in either ff_clear_blocks_mmx() or anywhere in > vc1_block.c, so either nobody tried this decoder on machines > without sse I do have access to an SSE-only system, I doubt FFmpeg was used or tested on MMX-only systems in a very, very long t

Re: [FFmpeg-devel] [PATCH 5/5] avcodec/vc1: store zero MVs for all blocks in a MB

2018-05-19 Thread James Almer
On 5/19/2018 9:13 PM, Michael Niedermayer wrote: > On Sat, May 19, 2018 at 08:23:59PM -0300, James Almer wrote: >> On 5/19/2018 6:40 PM, Michael Niedermayer wrote: >>> On Fri, May 18, 2018 at 05:06:36PM +0200, Jerome Borsboom wrote: Direct prediction for interlace frame B pictures references t

Re: [FFmpeg-devel] [PATCH 5/5] avcodec/vc1: store zero MVs for all blocks in a MB

2018-05-19 Thread Michael Niedermayer
On Sat, May 19, 2018 at 08:23:59PM -0300, James Almer wrote: > On 5/19/2018 6:40 PM, Michael Niedermayer wrote: > > On Fri, May 18, 2018 at 05:06:36PM +0200, Jerome Borsboom wrote: > >> Direct prediction for interlace frame B pictures references the mv in the > >> second block in an MB in the backw

Re: [FFmpeg-devel] [PATCH 5/5] avcodec/vc1: store zero MVs for all blocks in a MB

2018-05-19 Thread James Almer
On 5/19/2018 6:40 PM, Michael Niedermayer wrote: > On Fri, May 18, 2018 at 05:06:36PM +0200, Jerome Borsboom wrote: >> Direct prediction for interlace frame B pictures references the mv in the >> second block in an MB in the backward reference frame for the twomv case. >> When the backward referenc

Re: [FFmpeg-devel] [PATCH 5/5] avcodec/vc1: store zero MVs for all blocks in a MB

2018-05-19 Thread Michael Niedermayer
On Fri, May 18, 2018 at 05:06:36PM +0200, Jerome Borsboom wrote: > Direct prediction for interlace frame B pictures references the mv in the > second block in an MB in the backward reference frame for the twomv case. > When the backward reference frame is an I frame, this value may be unset. > > S

[FFmpeg-devel] [PATCH 5/5] avcodec/vc1: store zero MVs for all blocks in a MB

2018-05-18 Thread Jerome Borsboom
Direct prediction for interlace frame B pictures references the mv in the second block in an MB in the backward reference frame for the twomv case. When the backward reference frame is an I frame, this value may be unset. Signed-off-by: Jerome Borsboom --- libavcodec/vc1_block.c | 6 -- 1 fi