[FFmpeg-soc] [PATCH] jpeg2k : support higher sample precision

[Jai Menon]

On Wed, Jun 24, 2009 at 1:27 PM, Michael Niedermayer<michaelni at gmx.at> wrote:
> On Sun, Jun 21, 2009 at 04:35:20PM +0000, Jai Menon wrote:
>> Hi,
>>
>> Attached patch fixes decoding of files where component precision > 8.
>>
>> Possible test files are :
>>
>> codestreams_profile1/p1_04.j2k
>> testfiles_jp2/file6.jp2
>> testfiles_jp2/file7.jp2
>>
>> Posting here for comments.
>>
>> --
>> Regards,
>>
>> Jai
>
>>  j2kdec.c |   35 ++++++++++++++++++++++++++++++++---
>>  1 file changed, 32 insertions(+), 3 deletions(-)
>> 10381c8c4a6486b49b3013bc0683cb6d62f9161e  high_precision.patch
>> Index: j2kdec.c
>> ===================================================================
>> --- j2kdec.c  (revision 4434)
>> +++ j2kdec.c  (working copy)
>> @@ -54,7 +54,7 @@
>>      uint8_t cbps[4]; ///< bits per sample in particular components
>>      uint8_t sgnd[4]; ///< if a component is signed
>>      uint8_t properties[4];
>> -
>> +    int precision;
>>      int ncomponents;
>>      int tile_width, tile_height; ///< tile size
>>      int numXtiles, numYtiles;
>
>> @@ -225,6 +225,8 @@
>>      for (i = 0; i < s->ncomponents; i++){ // Ssiz_i XRsiz_i, YRsiz_i
>>          uint8_t x = bytestream_get_byte(&s->buf);
>>          s->cbps[i] = (x & 0x7f) + 1;
>> +        if (s->cbps[i] > s->precision)
>> +            s->precision = s->cbps[i];
>>          s->sgnd[i] = (x & 0x80) == 1;
>>          if (bytestream_get_byte(&s->buf) != 1)
>>              return -1;
>
> FFMAX

Changed locally.

> [...]
>> @@ -806,6 +815,26 @@
>>
>>          line += s->picture.linesize[0];
>>      }
>> +    } else {
>> +        for (; y < tile->comp[0].coord[1][1] - s->image_offset_y; y++) {
>> +            uint16_t *dst;
>> +            x = tile->comp[0].coord[0][0] - s->image_offset_x;
>> +            dst = line + x * s->ncomponents * 2;
>> +            for (; x < tile->comp[0].coord[0][1] - s->image_offset_x; x++) {
>> +                for (compno = 0; compno < s->ncomponents; compno++) {
>
>> +                    *src[compno] = av_rescale(*src[compno], (1 << 16) - 1,
>> +                                              (1 << s->cbps[compno]) - 1);
>
> av_rescale is too slow

So just (*src[compno]/((1 << s->cbps[compno]) - 1)) * ((1 << 16) - 1) ?

>
>> +                    *src[compno] += 1 << 15;
>> +                    if (*src[compno] < 0)
>> +                        *src[compno] = 0;
>> +                    else if (*src[compno] >= (1 << 16))
>> +                        *src[compno] = (1 << 16) - 1;
>> +                    *dst++ = *src[compno]++;
>
> av_clip()

Ah yes, fixed. I'll change it in the other case as well.

Thanks for the review.

-- 
Regards,

Jai