[FFmpeg-devel] libavfilter/vf_scale: Ensure scaled video is divisible by n

This patch adds a new option to the scale filter which ensures that the
output resolution is divisible by the given integer when used together
with `force_original_aspect_ratio`. This works similar to using `-n` in
the `w` and `h` options.

This option respects the value set for `force_original_aspect_ratio`,
increasing or decreasing the resolution accordingly.

The use case for this is to set a fixed target resolution using `w` and
`h`, to use the `force_original_aspect_ratio` option to make sure that
the video always fits in the defined bounding box regardless of aspect
ratio, but to also make sure that the calculated output resolution is
divisible by n so in can be encoded with certain encoders/options if
that is required.

Signed-off-by: Lars Kiesow <lkiesow@uos.de>
---
 doc/filters.texi       | 12 ++++++++++++
 libavfilter/vf_scale.c | 13 ++++++++++++-
 2 files changed, 24 insertions(+), 1 deletion(-)

On Mon, Jul 08, 2019 at 05:43:40PM +0200, Lars Kiesow wrote:
> This patch adds a new option to the scale filter which ensures that the
> output resolution is divisible by the given integer when used together
> with `force_original_aspect_ratio`. This works similar to using `-n` in
> the `w` and `h` options.
> 
> This option respects the value set for `force_original_aspect_ratio`,
> increasing or decreasing the resolution accordingly.
> 
> The use case for this is to set a fixed target resolution using `w` and
> `h`, to use the `force_original_aspect_ratio` option to make sure that
> the video always fits in the defined bounding box regardless of aspect
> ratio, but to also make sure that the calculated output resolution is
> divisible by n so in can be encoded with certain encoders/options if
> that is required.
> 
> Signed-off-by: Lars Kiesow <lkiesow@uos.de>
> ---
>  doc/filters.texi       | 12 ++++++++++++
>  libavfilter/vf_scale.c | 13 ++++++++++++-
>  2 files changed, 24 insertions(+), 1 deletion(-)
> 
> diff --git a/doc/filters.texi b/doc/filters.texi
> index ee6a93ffbf..2de71d9820 100644
> --- a/doc/filters.texi
> +++ b/doc/filters.texi
> @@ -15231,6 +15231,18 @@ Please note that this is a different thing than specifying -1 for @option{w}
>  or @option{h}, you still need to specify the output resolution for this option
>  to work.
>  
> +@item force_divisible_by Ensures that the output resolution is divisible by the
> +given integer when used together with @option{force_original_aspect_ratio}. This
> +works similar to using -n in the @option{w} and @option{h} options.
> +
> +This option respects the value set for @option{force_original_aspect_ratio},
> +increasing or decreasing the resolution accordingly. This may slightly modify
> +the video's aspect ration.
> +
> +This can be handy, for example, if you want to have a video fit within a defined
> +resolution using the @option{force_original_aspect_ratio} option but have
> +encoder restrictions when it comes to width or height.
> +
>  @end table
>  
>  The values of the @option{w} and @option{h} options are expressions
> diff --git a/libavfilter/vf_scale.c b/libavfilter/vf_scale.c
> index 7aebf56ad8..3ce6cdd1d5 100644
> --- a/libavfilter/vf_scale.c
> +++ b/libavfilter/vf_scale.c
> @@ -86,6 +86,7 @@ typedef struct ScaleContext {
>      int in_v_chr_pos;
>  
>      int force_original_aspect_ratio;
> +    int force_divisible_by;
>  
>      int nb_slices;
>  
> @@ -237,7 +238,8 @@ static int config_props(AVFilterLink *outlink)
>          goto fail;
>  
>      /* Note that force_original_aspect_ratio may overwrite the previous set
> -     * dimensions so that it is not divisible by the set factors anymore. */
> +     * dimensions so that it is not divisible by the set factors anymore
> +     * unless force_divisible_by is defined as well */
>      if (scale->force_original_aspect_ratio) {
>          int tmp_w = av_rescale(h, inlink->w, inlink->h);
>          int tmp_h = av_rescale(w, inlink->h, inlink->w);
> @@ -245,9 +247,17 @@ static int config_props(AVFilterLink *outlink)
>          if (scale->force_original_aspect_ratio == 1) {
>               w = FFMIN(tmp_w, w);
>               h = FFMIN(tmp_h, h);
> +             if (scale->force_divisible_by > 1) {
> +                 w = w / scale->force_divisible_by * scale->force_divisible_by;
> +                 h = h / scale->force_divisible_by * scale->force_divisible_by;
> +             }
>          } else {
>               w = FFMAX(tmp_w, w);
>               h = FFMAX(tmp_h, h);

> +             if (scale->force_divisible_by > 1) {
> +                 w = ceil(w / (float)scale->force_divisible_by) * scale->force_divisible_by;
> +                 h = ceil(h / (float)scale->force_divisible_by) * scale->force_divisible_by;

you dont need float here and its better not to use float when not needed
so theres a chance less for platform bitexactness issues

Thanks

[...]

Hi Michael,

> > +             if (scale->force_divisible_by > 1) {
> > +                 w = ceil(w / (float)scale->force_divisible_by) *
> > scale->force_divisible_by;
> > +                 h = ceil(h / (float)scale->force_divisible_by) *
> > scale->force_divisible_by;  
> 
> you dont need float here and its better not to use float when not
> needed so theres a chance less for platform bitexactness issues

Updated the patch so it's only using integers as well when rounding up.
I hope this works for you.

Best regards,
Lars