On 4/3/17, Ricardo Constantino <wiia...@gmail.com> wrote:
> On 3 April 2017 at 19:18, Paul B Mahol <one...@gmail.com> wrote:
>
>> On 4/3/17, Ricardo Constantino <wiia...@gmail.com> wrote:
>> > On 2 April 2017 at 22:13, Paul B Mahol <one...@gmail.com> wrote:
>> >
>> >> On 4/2/17, Paul B Mahol <one...@gmail.com> wrote:
>> >> > Signed-off-by: Paul B Mahol <one...@gmail.com>
>> >> > ---
>> >> > doc/filters.texi | 3 +++
>> >> > libavfilter/vf_pad.c | 14 ++++++++++++++
>> >> > 2 files changed, 17 insertions(+)
>> >> >
>> >> > + if (adjusted_aspect.num && adjusted_aspect.den) {
>> >> > + adjusted_aspect = av_mul_q(adjusted_aspect, av_make_q(s->w,
>> >> ^ changed to
>> >> sample aspect ratio locally.
>> >>
>>
>> Have you missed what I wrote above?
>>
>> > Shouldn't this av_mul_q be a
>> > av_div_q(adjusted_aspect, inlink->sample_aspect_ratio) instead?
>> >
>> > That way it won't try to expand 40/33 SAR 704x480 to 1034x480 but
>> > instead leave it as 704x480.
>>
>> Why div_q ?
>>
>
> with av_mul_q(adjusted_aspect, inlink->sample_aspect_ratio):
>
> $ ffmpeg -f lavfi -i "color=s=704x480,setsar=40/33" -vf
> pad=aspect=16/9,scale=iw*sar:ih -vframes 1 -f null - -v verbose 2>&1 | grep
> -E "(Parsed_pad|Parsed_scale)"
> [Parsed_scale_1 @ 000001e46c6c60a0] w:iw*sar h:ih flags:'bicubic' interl:0
> [Parsed_pad_0 @ 000001e46c6c5bc0] w:704 h:480 -> w:1034 h:480 x:0 y:0
> color:0x000000FF
> [Parsed_scale_1 @ 000001e46c6c60a0] w:1034 h:480 fmt:yuv420p sar:40/33 ->
> w:1253 h:480 fmt:yuv420p sar:3760/3759 flags:0x4
>
> with av_div_q(adjusted_aspect, inlink->sample_aspect_ratio):
>
> $ ffmpeg -f lavfi -i "color=s=704x480,setsar=40/33" -vf
> pad=aspect=16/9,scale=iw*sar:ih -vframes 1 -f null - -v verbose 2>&1 | grep
> -E "(Parsed_pad|Parsed_scale)"
> [Parsed_scale_1 @ 000001522e496720] w:iw*sar h:ih flags:'bicubic' interl:0
> [Parsed_pad_0 @ 000002b8a29c8260] w:704 h:480 -> w:704 h:480 x:0 y:0
> color:0x000000FF
> [Parsed_scale_1 @ 000001522e496720] w:704 h:480 fmt:yuv420p sar:40/33 ->
> w:853 h:480 fmt:yuv420p sar:2560/2559 flags:0x4
Yes, but that conflict with expand behaviour.
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
