Re: CFD: eXIf 2017-05-31

On 2 June 2017 at 00:44, Cosmin Truta wrote:
> It is a fact of existing EXIF-processing software that thumbnails (as
> much as any other EXIF fields) are not required to be updated.

One more important thing to mention, in the bigger picture, is this:

We shouldn't keep agonizing over the least important EXIF fields. EXIF
is about photography, and the photography-related fields are the ones
that are important: camera, lens, exposure, flash, etc. This is what
remains unchanged, and what you need to care about, and trust.

So you can't trust the thumbnail? No big deal: rebuild it, or ignore it.

So you can't trust the gamma field? (Hint: you shouldn't.) No big
deal: ignore it. A gamma field, in my experience, comes with AdobeRGB
images, not with sRGB images (understandably so, because AdobeRGB has
a real gamma, while sRGB doesn't). And AdobeRGB images already have an
ICC profile in addition to EXIF, otherwise they wouldn't display
correctly.

So you found a gamma field in a sRGB image? I would be surprised, but
ok. No big deal: ignore it. You already know what sRGB is, and you
also know what is the approximation of gamma inside sRGB. Chances are,
you might have an actual sRGB ICC profile as well, but I found out
this not to be guaranteed (but yet again, not a big deal here,
either).

Speaking about colorspaces, I like it how Google handled it in WebP.
You either know what the ICC profile is (and you have it embedded in
the WebP image), or you don't, in which case you should assume it to
be sRGB. They did the right thing, in my opinion. Problem solved.

Sincerely,
Cosmin

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot