Re: [EXTERNAL] Re: [PNG] Cancel upcoming meeting?

Sorry, mis-send.  I blame Google.

On Fri, Feb 28, 2025 at 8:25 AM Seeger, Chris (NBCUniversal) <
Chris.Seeger@nbcuni.com> wrote:

> The ISO document only describes the Diffuse White Metadata Element
> (Section 4.6.5) but doesn’t create an exact standard for it’s
> representation in images or movies.
>

I thought this and the paywall discussion was about ISO-21496, here:

>ISO/DIS 21496-1(en)
>Digital Photography — Gain map metadata for image conversion — Part 1:
Dynamic Range Conversion
>https://www.iso.org/obp/ui/en/#iso:std:iso:21496:-1:dis:ed-1:v1:en

But that has no "section 4.6.5" and it contains "Annex C
<https://www.iso.org/obp/ui/en/#iso:std:iso:14708:-6:ed-2:v1:fr:sec:C>
(normative)
provid[ing] information about storing the gain map in several file formats."
"Several file formats" seems to mean "baseline JPEG" however I am inclined
to deduce from the remainder of the public (non-paywalled) part of the
document that the format uses EXIF and, elsewhere, I think use of the APP2
marker is mentioned.

I haven't reverse engineered the Adobe test images but, given the lack of
facilities in baseline JPEG, I would guess that PNG could simply do exactly
the same thing as it does with EXIF metadata; copy it over verbatim.  That
would provide complete support and just require a new chunk name
(unsafe-to-copy before PLTE).  Existing libraries that conform to PNGv3
would have no problem adding another chunk with the same underlying
encoding and, presumably, existing decoders that already handle eXIf would
have no problem with just a different black of EXIF tags.  The whole
paywall thing becomes moot (still annoying, but moot) as apps can just
copy'n'paste existing JPEG support code.

John Bowler

>

Received on Friday, 28 February 2025 23:22:40 UTC