[gdal-dev] Bigger file size from 64-bit than 32-bit

[Iván Sánchez Ortega]

El Jueves, 16 de enero de 2014 13:52:37 Jukka Rahkonen escribió:
> Samples can be downloaded from
> 
> http://latuviitta.org/downloads/FLPIS_K003412.jp2 (18 MB, original JP2)
> http://latuviitta.org/downloads/gdal_translate_32bit.TIF (9.6 MB)
> http://latuviitta.org/downloads/gdal_translate_64bit.tif (57 MB)

> Builds I use are specially made by Tamas and I can ask for details if
> someone gives a hint about what to ask.

I've had a look at the files. With both gimp and imagemagick, I see the 64-bit 
image garbled, with obvious compression artifacts.

For example, when I run 'identify gdal_*', I get:

Image: gdal_translate_32bit.TIF
  Channel statistics:
    Gray:
      min: 0 (0)
      max: 239 (0.937255)
      mean: 91.8711 (0.360279)
      standard deviation: 23.4782 (0.0920712)
      kurtosis: 7.73665
      skewness: 1.39568

Image: gdal_translate_64bit.tif
  Channel statistics:
    Gray:
      min: 0 (0)
      max: 255 (1)
      mean: 117.769 (0.461839)
      standard deviation: 112.01 (0.439255)
      kurtosis: -1.78098
      skewness: 0.182984

Which means both images are greatly different.


I'd say your 64-bit build has some weird, nasty bug.


-- 
Iván Sánchez Ortega <ivan at sanchezortega.es>