Re: Fwd: VOTE ABSTAIN: eXIf 2017-0119
Glenn Randers-Pehrson <[email protected]>
| Newsgroups | gmane.comp.graphics.png.general |
|---|---|
| Message-ID | <CA+PdXcvicdzMD1o+1THjbem8MbHJsP_o5C6UiETB46VjLSNZRg@mail.gmail.com> |
On Tue, Feb 7, 2017 at 10:08 AM, Willem van Schaik <[email protected]> wrote: > > would it be enough to add a statement that "when information in the EXIF > chunk is in contradiction with information in any other chunk, it should > be ignored" > > Willem > No, for reasons explained by Pavel. It can be in contradiction with gamma information that has been removed. I do think the chunk proposal can be salvaged with a additional statement, but it has to be more than that sentence. I'm thinking that the specification should tell decoders that they must ignore any "editable" tags such as gamma (i.e., any tag that the Exif spec paragraph E.3 does not say "should be copied unchanged"). We have to allow encoders to write those tags, for simplicity, but we don't want decoders to use those potentially invalid tags to be used in reconstructing a new Exif profile. Fortunately many of those are also covered by other PNG chunks (resolution, modification date/time, gamma) and many others (dealing with the image compression, etc.) are of no use when the pixel format changes from JPEG or TIFF to PNG. There are a few that aren't, though, and I'm not sure how to handle those (subject location, orientation, processing instructions). Glenn ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, SlashDot.org! http://sdm.link/slashdot _______________________________________________ png-mng-misc mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/png-mng-misc