Exif ColorSpace tag value 2 is wrong?
Posted on 2006-08-14 15:36:06-07 by boscarol
At this page tag ColorSpace is listed with values 1 = sRGB, 2 = Adobe RGB, 65535 = Uncalibrated.
I think that value 2 (Adobe RGB) is wrong.
The specification for EXIF 2.21 (JEITA CP3451-1, section 4.6.5-B) states that "if a color space other than sRGB is used, Uncalibrated (=FFFF.H) is set.
In fact Exif 2.21 discriminates between sRGB and AdobeRGB not by means of ColorSpace tag, but of InterOperability Index tag.
Am I wrong?
Posted on 2006-08-22 13:04:28-07 by exiftool in response to 2784
You are correct about the EXIF specification. However, ExifTool also reads/writes non-standard information. I will have to do a bit of work to determine where I found the definition for the value of 2 -- unfortunately I didn't reference it in the code like I have been doing more recently. Let me look into this.
Posted on 2006-08-22 19:19:57-07 by geve in response to 2818
This is not only the case for this tag, but for all tags. Should we document which values are valid for which camera's. And with valid a mean proven. This is a tremendous job.
But in the other hand an example:
Tag: $0009 ImageSize for Pentax, we though we next was valid (comparison IDF and makernote values) 8 = 2560x1920 // S50, S55, S5i 22 = 2304x1728 // S4, S40, S45, S4i, S5n But making pictures with a Optio S6 showed: 8 = 2304x1728 //S6 verified 22 = 2592x1944 //S6 verified
Posted on 2006-08-23 11:54:29-07 by exiftool in response to 2821
I have not been able to find my source for the ColorSpace value of 2, but it is most certainly not standard EXIF. Also, none of my digital camera samples use this value, so it is very uncommon if used at all.
The ExifTool EXIF tag table documentation contains information from a wide variety of sources, including private and proprietary tags which are not part of the EXIF specification. Documenting these a bit more carefully is certainly a good idea, but would take some work.
Documenting all of the maker notes tags for every camera model as geve suggests takes this to a whole new level, and would require a vast set of test images (not only one image per camera as I am currently collecting, but one image per camera per possible setting or combination of settings!). Very useful, but essentially impossible to validate all possible values.
It is unfortunate that Pentax (and other companies) use the same tag value to mean different things for different camera models. This breaks forward compatibility of decoding software and makes the job much more difficult to provide a reliable meta information reader/writer.
(This article is based on a thread on the CPAN::Forum.)