I shoot RW2 files. Is there any advantage to convert them to DNG or to keep them like they are now for raw therapee? (DNG is 30% smaller)
Converting is a bad idea as you may lose data, even without recognizing. Better keep the raw you produced in camera.
Better keep the raw you produced in camera.
@stephane-archer generally there is no advantage. Personally, I convert my ARW files to DNG precisely for the file size benefit.
Should you notice an issue with your camera’s DNG files compared to the source raw files, file a report: http://rawpedia.rawtherapee.com/How_to_write_useful_bug_reports
Although you always keep the raw file out of the camera if you want to keep the photo, there is sometimes an advantage to the DNG. Compare the temperature and tint values. They are more reasonable, and therefore easier to work with (for me), than the original raw from some camera models. I find it easier to adjust 5200/0.96 than 31000/0.6 !
I saw this phenomenon years ago in a few Sony .ARW files. It also happens with most Leica cameras (convert DNG to DNG with Adobe Raw Converter).
I’ve noticed some specific EXIF data being lost when converting my Nikon D600 raw files to .dng so I stopped doing that. I forgot which details exactly.
This sounds to me more like a bug in your raw converter (or unsupported raw format). There should be no difference in white balance values no matter which format if
- the raw converter reads the camera white balance, or
- the raw converter uses its own default values.
The only way I can imagine that these different values happen is if the raw converter uses different standard values for different formats, but this may then, again, be a bug.
“This sounds to me more like a bug in your raw converter” Perhaps I was not clear. Raw Therapee was used both on the native raw format and on a DNG out of Adobe Raw Converter. The former had wild temp and tint values, the latter had more reasonable ones – all in RT.
That’s what I assumed. If the white balance is different for the same image dependent on the raw format, this sounds like a bug to me.
RawTherapee’s temp and tint values depend on the
ColorMatrix. Could be this caused the discrepancy between the original raw file and the DNG, though without a proper report no one can tell.