- From: John Bowler <john.cunningham.bowler@gmail.com>
- Date: Fri, 28 Feb 2025 16:03:24 -0800
- To: "Seeger, Chris (NBCUniversal)" <Chris.Seeger@nbcuni.com>
- Cc: "public-png@w3.org" <public-png@w3.org>
- Message-ID: <CAP7U398v8bEpvijLQ4aHdrRSeDAtkK+2E=quzdgqNtc4aXzjtQ@mail.gmail.com>
On Fri, Feb 28, 2025 at 3:31 PM Seeger, Chris (NBCUniversal) < Chris.Seeger@nbcuni.com> wrote: > I think I mis-read and thought it was diffuse white metadata. I blame > myself :-) > Very old fashioned :-) 21496 is at DIS stage so maybe you had an earlier WD? My assumption about the ISO DIS stage is that it's all over apart from the shouting; IRC it goes to industry representatives at the DIS stage, or maybe just before. Annex-C contains general "binary metadata payload" definitions (C.2) There are a couple of interesting things about it that can be deduced with, I think, reasonable reliability from the ToC. First is that Annex A describes (informatively) the derivation of the gain map from _two_ input images; an HDR and the desired SDR (the "Alternate" image). So a cynical person like me would just describe this as a way of transmitting the SDR with high compression. Annex A is informative so it is also possible, in theory, to derive the gain map directly, possibly computationally, from the original HDR; computational tone-mapping or, in photographic terms, a computational approach to dodging and burning. The second is that the gain map is a set of scale factors for linear RGB (3.3 and 3.4) to convert the HDR to/from the SDR and, also, between the endpoints of the two colour spaces (from the "Note 1" in 3.4). This implies that the standard encodes some amount of chromatic correction; the white points of the HDR and SDR may be different. There is more discussion in Annex B of "colour conversion" but in general the requirements of chromatic adaptation cannot be handled completely just by scaling R/G/B. Well, IRC; simple scaling is the original Von Kries approach isn't it ChrisL? (I might very well have got that bit of arithmetic wrong!) John Bowler
Received on Saturday, 1 March 2025 00:03:39 UTC