[FFmpeg-devel] [PATCH 1/2] acvodec/lipopenjpeg: Fix pixel value shift for 12-bit pixel formats

[Michael Bradshaw]

On Tue, Mar 3, 2015 at 1:19 AM, Vilius Grigaliūnas <
vilius.grigaliunas at gmail.com> wrote:

> This fixes pixel values not being properly shifted in
> libopenjpeg_copyto16 and libopenjpeg_copy_to_packed16 methods.
>
> Pixel formats like xyz12le need to be shifted by
> AVComponentDescriptor::shift
> to get the correct values.
> ---
>  libavcodec/libopenjpegdec.c |    6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/libavcodec/libopenjpegdec.c b/libavcodec/libopenjpegdec.c
> index 02b1ceb..1cd1b9b 100644
> --- a/libavcodec/libopenjpegdec.c
> +++ b/libavcodec/libopenjpegdec.c
> @@ -184,10 +184,11 @@ static inline void
> libopenjpeg_copy_to_packed8(AVFrame *picture, opj_image_t *im
>
>  static inline void libopenjpeg_copy_to_packed16(AVFrame *picture,
> opj_image_t *image) {
>      uint16_t *img_ptr;
> +    const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(picture->format);
>      int index, x, y, c;
>      int adjust[4];
>      for (x = 0; x < image->numcomps; x++)
> -        adjust[x] =
> FFMAX(FFMIN(av_pix_fmt_desc_get(picture->format)->comp[x].depth_minus1 + 1
> - image->comps[x].prec, 8), 0);
> +        adjust[x] = FFMAX(FFMIN(desc->comp[x].depth_minus1 + 1 -
> image->comps[x].prec, 8), 0) + desc->comp[x].shift;
>
>      for (y = 0; y < picture->height; y++) {
>          index   = y * picture->width;
> @@ -220,10 +221,11 @@ static inline void libopenjpeg_copyto8(AVFrame
> *picture, opj_image_t *image) {
>  static inline void libopenjpeg_copyto16(AVFrame *picture, opj_image_t
> *image) {
>      int *comp_data;
>      uint16_t *img_ptr;
> +    const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(picture->format);
>      int index, x, y;
>      int adjust[4];
>      for (x = 0; x < image->numcomps; x++)
> -        adjust[x] =
> FFMAX(FFMIN(av_pix_fmt_desc_get(picture->format)->comp[x].depth_minus1 + 1
> - image->comps[x].prec, 8), 0);
> +        adjust[x] = FFMAX(FFMIN(desc->comp[x].depth_minus1 + 1 -
> image->comps[x].prec, 8), 0) + desc->comp[x].shift;
>
>      for (index = 0; index < image->numcomps; index++) {
>          comp_data = image->comps[index].data;
>

Overall, looks good to me. Maybe the "+ desc->comp[x].shift;" part should
be moved inside the FFMIN macro? Seems okay as is, though.