09:12 pq: emersion, whoo! \o/
10:07 zubzub: oh a discussion about X12 on HN
10:09 zubzub: thme comments 🥲
10:09 emersion: yeah, do not recommend
10:10 emersion: "but simon, what were you expecting"
10:12 zubzub: I read some of it knowing full well what to expect 💩
10:13 zubzub: but there were clearly no new insights to be gained after a minute of reading or so
10:14 davidre: Same mistake as reading phoronix comments :D
12:18 wlb: weston Issue #728 opened by Pekka Paalanen (pq) GL-renderer forgets to destroy GL context https://gitlab.freedesktop.org/wayland/weston/-/issues/728 [GL renderer]
12:27 wlb: wayland Issue #361 opened by Pekka Paalanen (pq) wl_display_disconnect has misleading documentation https://gitlab.freedesktop.org/wayland/wayland/-/issues/361 [Documentation], [IPC library]
12:59 pq: oh hey, wlb works again for weston too. :-)
13:48 pq: swick[m], I'm attacking the rendering intent next for the protocol. Are you continuing with the HDR static metadata?
13:48 swick[m]: sure, seems easy enough
13:49 pq: will need image description events for the metadata too
13:50 swick[m]: yeah, noticed that as well
13:52 pq: I think I've finally understood why I read from somewhere that one should never fill a chromaticity diagram with colors.
13:59 vsyrjala: one should just leave it blank or what?
13:59 pq: yes
14:01 pq: especially around the center, any specific x,y point does not have a specific color appearance. However, if you fill the diagram with colors, you are implicitly giving each point a specific appeareance.
14:01 vsyrjala: i suppose. but it does provide the "this way towards reddish" sort of hint
14:02 emersion: maybe only use colors for the primaries?
14:02 vsyrjala: but yeah, remembering that the colors aren't real is hard
14:02 pq: Which x,y point looks the most "white"? That is dictated by the viewer and their environment, it's not any one x,y point.
14:02 swick[m]: that also changes the appearance of the primaries
14:03 emersion: right, but maybe it's more obvious that it's just a hint that way
14:03 pq: emersion, yes, denoting primaries with colored point is fine, in that it doesn't give a wrong idea.
14:04 emersion: what's the difference here between the white point and primaries?
14:04 emersion: the white point is more of a lie than the primaries?
14:05 JEEB: H.273/CICP primaries points contain a white point as well, so most likely that's why that exists there, too
14:05 pq: In that very specific context, primaries are your R, G and B lamps, and white point is the chromaticity they emit when each one has equal power.
14:05 emersion: hm, and so there's no single white point on the graph?
14:06 pq: correct
14:06 emersion: it's more of a line in the 3D space?
14:06 pq: a monitor has a single white point, and you find it my setting all color channel to equal values.
14:06 pq: or simply to max values
14:06 JEEB: yup
14:07 pq: another monitor may have a different white point
14:08 pq: if you have only primaries, your definition of a color space is not complete without the white point, because the white point puts the primaries into the same "scale".
14:09 pq: Each primary has its own "scale of strength", and when you start mixing them to produce other, ahem, colors, you need to know their relative strength.
14:10 pq: That is all colorimetry, physically measurable quantities, like CIE 1931 x,y.
14:10 pq: Appearance, or what does a certain x,y *look* like, is another matter.
14:11 pq: That depends on what you, the viewer, have adapted to.
14:13 pq: If I was living in a room lit only with pale green lights, and you came in from outdoors, I might say "this paper looks white" while you say "how can you live here when everything looks so green?" :-)
14:13 pq: It's an exaggeration, but it's real.
14:15 pq: Ever thought why seeing a TV through a window while you are outside looks so blue, yet when you watch TV inside it looks normal?
14:15 wlb: wayland Merge request !298 opened by Simon Ser (emersion) client: fix wl_display_disconnect() documentation https://gitlab.freedesktop.org/wayland/wayland/-/merge_requests/298 [IPC library]
14:17 pq: The TV is emitting the same x,y chromaticity regardless of where you are, but you see it differently.
14:18 kennylevinsen: if you swap pale green for orange it becomes a completely regular scenario with excessively warm white lamps...
14:18 pq: yup
14:19 kennylevinsen: although less likely to catch people's attention due to lying on the planckian locus which we are accustumed to for white balance
14:19 pq: I must say, after spending an hour or so in a pale green lit room, walking out to see the nature's Autumn colors in daylight was mindblowing.
14:20 kennylevinsen: Hah, an interesting color perception experiment :)
14:20 kennylevinsen: And a peculiar choice in interior lighting
14:21 pq: and a few seconds later, the impression was gone :'-|
14:21 pq: well, it was a university color lab
14:22 pq: I was wondering why the professor threw us out with such speed and haste, but now I know. Adaptation happens fast.
18:25 wlb: wayland Merge request !299 opened by Simon Ser (emersion) Draft: server: make wl_event_loop optional https://gitlab.freedesktop.org/wayland/wayland/-/merge_requests/299 [IPC library]
19:06 swick[m]: pq: I don't think I can add any of the luminance metadata before we have not defined the reference display and reference viewing environment in the protocol
19:08 swick[m]: for PQ this is all tied closely together but everywhere else it already breaks apart
19:09 swick[m]: even for PQ you have to assume that the mastering display viewing environment is the same as the reference viewing environment
22:00 wlb: weston Merge request !1181 opened by Robert Mader (rmader) clients/simple-egl: Implement fractional-scale protocol support https://gitlab.freedesktop.org/wayland/weston/-/merge_requests/1181