Windows 10
ExifTool 12.29
Using this ARGS file:
-f
-a
-G0
-ver
-Format
-Software
-ImageWidth
-ImageHeight
-XResolution
-BitsPerSample
-ColorMode
-PhotometricInterpretation
-ProfileDescription
-ColorSpace
Here's the output:
12.29
[XMP] Format : image/jpeg
[EXIF] Software : Adobe Photoshop 22.4 (Windows)
[EXIF] Image Width : 1992
[File] Image Width : 1992
[EXIF] Image Height : 1512
[File] Image Height : 1512
[EXIF] X Resolution : 72
[EXIF] X Resolution : 72
[Photoshop] X Resolution : 72
[EXIF] Bits Per Sample : 8 8 8
[File] Bits Per Sample : 8
[] Color Mode : -
[EXIF] Photometric Interpretation : RGB
[ICC_Profile] Profile Description : Gray Gamma 2.2
[EXIF] Color Space : Uncalibrated
Using Adobe Bridge FileInfo to look at the raw metadata, I have verified that the photoshop:ColorMode TAG does exist and has a value of "1" (Grayscale). And the tiff:PhotometricInterpretation TAG has a value of "1" (BlackIsZero), not "2" (RGB) as reported by ExifTool. At least ExifTool reports the profile description properly.
I can provide a small JPEG file to test with if needed. Just didn't want to post it here since it is a file provided by one of my clients.
Thanks
The data shown in Bridge's "Raw Data" window is not tags that are embedded in the file. It's showing you what the XMP for that file will look like if you write out the XMP, either as a sidecar file or embed the XMP. The photoshop:ColorMode refers to XMP-photoshop:ColorMode, not Photoshop:ColorMode, and tiff:PhotometricInterpretation refers to XMP-tiff:PhotometricInterpretation, not EXIF:PhotometricInterpretation. Exiftool recognizes that it is a grey scale image as shown by File:BitsPerSample entry of 8, meaning one 8 bit channel. A color image would have 8 8 8, meaing three 8 bit channels.
The EXIF entries for PhotometricInterpretation and BitsPerSample are in error. I'm guessing the image was original in color and converted to grey scale. The converting program just didn't update the entries.
You can double check the file by running
exiftool -v3 file.jpg
And look through the output for the APP13 segment, which is where the Photoshop tags (https://exiftool.org/TagNames/Photoshop.html) are stored, and take a detailed look at the segment.
OK, thanks for the education. I must admit I didn't know the "FileInfo/Raw Data" panel in Bridge (or Photoshop) may not accurately reflect the current tags in the file. This is VERY important to know!
This highlights for me that there is a subtle distinction between using a tool like ExifTool to look at metadata vs other tools like Python/Pillow or Adobe Bridge. ExifTool reports what TAGS it finds (including ones with incorrect values), and does not really analyze/synthesize/interpret. Bridge gives me a higher-level (and more accurate) picture of the file's overall state. Python/Pillow I would put somewhere in the middle on this scale (it's not clear to me yet how it determines it's image "modes"). I started experimenting with ExifTool mainly because Python/Pillow cannot handle 16-bit TIFF files in Lab colorspace, and I needed a way (from Python) to retrieve a small set of metadata from image files to accurately characterize them.
Quote from: StarGeek on August 01, 2021, 08:17:47 PM
I'm guessing the image was original in color and converted to grey scale. The converting program just didn't update the entries.
Yes, you have guessed correctly. But the "converting program" was the latest version of Photoshop and I am dumbfounded that it did not properly sync/update the EXIF tags which are incorrect. I would expect that kind of omission from a lesser-known tool, but not Photoshop. I guess I'll pursue this with Adobe.
Thanks