Bug report: DNG RGBTables has the wrong tag id

Started by PixelDoctor, April 09, 2024, 01:00:30 PM

Previous topic - Next topic

PixelDoctor

The tag  names page (https://exiftool.org/TagNames/EXIF.html) lists DNG RGBTables as:

0xcd3b RGBTables

But according to the DNG spec (https://helpx.adobe.com/content/dam/help/en/photoshop/pdf/DNG_Spec_1_7_1_0.pdf), the actual tag id is 0xcd3f (b --> f).

DNG 1.7.1 also introduced a few other tags
0xcd49: JXLDistance
0xcd40: JXLEffort
0xcd41: JXLDecodeSpeed

Thanks!

Phil Harvey

Thanks.  I'll fix that and add JXLDistance, but JXLEffort and JXLDecodeSpeed have the same tag ID's as ProfileGainTableMap2 and JUMBF, so I can't add those.  Is this a typo in the 1.7.1 spec?

- Phil
...where DIR is the name of a directory/folder containing the images.  On Mac/Linux/PowerShell, use single quotes (') instead of double quotes (") around arguments containing a dollar sign ($).

Phil Harvey

I just got word back from Adobe.  These are typos in the DNG 1.7.1 specification that they were unaware of until I reported it to them.  (Interesting that nobody else has reported this because the spec. is about 7 months old now.)

The tags ID's for JXLEffort and JXLDecodeSpeed are in fact 0xcd4a and 0xcd4b respectively.

- Phil
...where DIR is the name of a directory/folder containing the images.  On Mac/Linux/PowerShell, use single quotes (') instead of double quotes (") around arguments containing a dollar sign ($).

Neal Krawetz

Quote from: Phil Harvey on April 11, 2024, 08:13:12 AMThese are typos in the DNG 1.7.1 specification that they were unaware of until I reported it to them.  (Interesting that nobody else has reported this because the spec. is about 7 months old now.)

Okay, my Adobe-bias is definitely going to show here. But it doesn't surprise me at all. Adobe is not known for high-quality solutions. E.g.:
  • Frequent critical vulnerabilities https://cve.mitre.org/cgi-bin/cvekey.cgi?keyword=adobe
  • XMP inconsistency due to library loading race conditions.
  • Inability to track XMP version to software release
  • Modifying metadata upon loading (never trust Adobe's built-in metadata viewer)
  • C2PA complexity and security issues
  • Lack of thorough end-to-end testing
  • PDF parsers must be built to support non-standard PDF structures, including inconsistencies from Adobe.
Based on the quality of things Adobe has released, I suspect they either lack internal controls for serious reviews, or only apply them halfheartedly. "How did that one get past quality control? Oh, it's Adobe." With DNG, nobody looks closely unless they absolutely need to.