dpx samples (possibly broken) look broken with FFmpeg, work fine with ImageMagick

[Carl Eugen Hoyos]

​http://www.simplesystems.org/users/bfriesen/dpx/ contains three dpx samples, one (16 bit) looks ugly with FFmpeg, ImageMagick and GraphicsMagick, the other two (8 bit) only look broken with FFmpeg.

$ ffmpeg -i DLAD_8b_1c_big.dpx
ffmpeg version N-63624-g4ba8560 Copyright (c) 2000-2014 the FFmpeg developers
  built on May 31 2014 12:24:49 with gcc 4.7 (SUSE Linux)
  configuration: --enable-gpl
  libavutil      52. 87.100 / 52. 87.100
  libavcodec     55. 65.100 / 55. 65.100
  libavformat    55. 42.100 / 55. 42.100
  libavdevice    55. 13.101 / 55. 13.101
  libavfilter     4.  5.100 /  4.  5.100
  libswscale      2.  6.100 /  2.  6.100
  libswresample   0. 19.100 /  0. 19.100
  libpostproc    52.  3.100 / 52.  3.100
Input #0, image2, from 'DLAD_8b_1c_big.dpx':
  Duration: 00:00:00.04, start: 0.000000, bitrate: N/A
    Stream #0:0: Video: dpx, gray, 255x185 [SAR 1:1 DAR 51:37], 25 tbr, 25 tbn, 25 tbc
At least one output file must be specified

$ ffmpeg -i DLAD_8b_3c_big.dpx
ffmpeg version N-63624-g4ba8560 Copyright (c) 2000-2014 the FFmpeg developers
  built on May 31 2014 12:24:49 with gcc 4.7 (SUSE Linux)
  configuration: --enable-gpl
  libavutil      52. 87.100 / 52. 87.100
  libavcodec     55. 65.100 / 55. 65.100
  libavformat    55. 42.100 / 55. 42.100
  libavdevice    55. 13.101 / 55. 13.101
  libavfilter     4.  5.100 /  4.  5.100
  libswscale      2.  6.100 /  2.  6.100
  libswresample   0. 19.100 /  0. 19.100
  libpostproc    52.  3.100 / 52.  3.100
Input #0, image2, from 'DLAD_8b_3c_big.dpx':
  Duration: 00:00:00.04, start: 0.000000, bitrate: N/A
    Stream #0:0: Video: dpx, rgb24, 255x185 [SAR 1:1 DAR 51:37], 25 tbr, 25 tbn, 25 tbc
At least one output file must be specified

[Christophe]

DLAD_8b_3c_big.dpx​ is supposed to be decoding with R and B swapped, as evidenced by 2 other software, GraphicsMagik and XnView.

There are various interesting metadata left to parse, like colorspace and orientation but that's irrelevant to this bug.

Fixed in 4ba45c189ccafdedcdb703e40441fbd24fe62989 then.

[Carl Eugen Hoyos]

DLAD_16b_1c_big.dpx was also fixed, thank you!

[Balling]

DLAD_8b_3c_big.dpx​ is supposed to be decoding with R and B swapped

Why? Nothing signals that?