[gdal-dev] Skipping PHOTOMETRIC=YCBCR when not reasonable

Rahkonen Jukka jukka.rahkonen at mmmtike.fi
Fri Nov 1 07:56:39 PDT 2013


Hi,

Yes, it is understandable because there are so many parameters which may be wrong. I still don't write any other scripts than Windows batch files and testing the return codes is not my business. I will continue by sorting the source files as well as I can and running the batch a second time without YCBCR for the images which were not converted with the first run. Not a big problem for me really.

-Jukka-

Even Rouault wrote:

> Jukka,
> 
> It is difficult to guess if the option specified by the user is something that he
> wants absolutely to be honoured (and in case it is not possible, an error is
> expected), or if it is just a hint that can be ignored if necessary. So the current
> behaviour seems OK to me.
> 
> Your script could just test the return code of gdal_translate and, in case of error,
> try again without -co PHOTOMETRYC=YCBCR. Or check before if the input
> dataset is 3 bands or not.
> 
> Even
> 
> > Hi,
> >
> > Gdal_translate throws
> > ERROR 6: PHOTOMETRIC=YCBCR requires a source raster with only 3 bands
> > (RGB) if user tries to apply YCBCR to non-RGB image. It means that
> > when I have a mixture of RGB and B/W images I must separate them
> > before making batch jobs.
> >
> > I think that nothing bad would happen if gdal_translate in such case
> > just throws a warning and continues by compressing black and white
> > image with the default photometric settings.
> >
> > -Jukka Rahkonen-
> >
> > _______________________________________________
> > gdal-dev mailing list
> > gdal-dev at lists.osgeo.org
> > http://lists.osgeo.org/mailman/listinfo/gdal-dev
> 
> --
> Geospatial professional services
> http://even.rouault.free.fr/services.html


More information about the gdal-dev mailing list