21:27:45 RRSAgent has joined #css-color 21:27:45 logging to http://www.w3.org/2016/05/11-css-color-irc 21:27:55 RRSAgent: make logs public 21:28:00 dholbert has joined #css-color 21:28:03 zcorpan has joined #css-color 21:28:03 smfr has joined #css-color 21:28:23 trackbot has joined #css-color 21:29:08 ScribeNick: zcorpan 21:29:21 dbaron has joined #css-color 21:29:42 waiting for Godot 21:30:18 present+ 21:30:18 present+ 21:30:20 present+ cabanier 21:30:21 present+ Florian 21:30:21 present+ 21:30:21 present+ 21:30:26 present+ ChrisL_ 21:30:28 present+ 21:30:33 Present+ 21:30:34 present+ 21:30:47 present+ 21:30:51 https://drafts.csswg.org/css-color/#lab-colors 21:30:55 Topic: color 21:31:15 ChrisL_: we've had discussion in the last f2f for thing swe've wanted to do 21:31:26 ChrisL_: apple have fancy screens now nad want to take advantage of that 21:31:40 s/nad/and/ 21:31:42 ChrisL_: something something colorspaces 21:32:21 seth has joined #css-color 21:32:31 present+ 21:33:02 ChrisL_: XYZ is a colorspace that can describe all known colors 21:33:21 ChrisL_: XYZ fails to do white adaptation 21:33:51 ChrisL_: LAB sticks an axis to teh white point and if you take a perfect white surface, that's 100 21:34:26 ChrisL_: then two orthorgonal channels A and B that correspond to greenish and yellowish axises 21:35:07 ChrisL_: LAB is useful, you get it in physical measurement devices etc 21:35:21 Florian: in the luminosity direction, how white is 100%? 21:36:10 ChrisL_: there are different standards for standard daylight and the one sRGB uses 21:36:16 ... uses D65 21:36:26 (6500K) 21:36:33 ChrisL_: if you look in the spec, i've tried to keep the math to a minimum 21:36:43 ChrisL_: (before) LAB uses D50 21:36:50 esprehn has joined #css-color 21:37:16 ChrisL_: ... this is all just summarizing the spec 21:37:32 ChrisL_: issue about out of range numbers 21:37:52 ChrisL_: steps for how to convert ... 21:37:58 TabAtkins: what is a reasonable range for C? 21:38:52 ChrisL_: to go from LAB to ??? 21:38:54 ACTION ChrisL_ to add note explaining a reasonable range for C in lch() to the spec 21:38:54 Sorry, but no Tracker is associated with this channel. 21:39:16 TabAtkins: so conversion between a/b and c/h is just rectangular to polar 21:39:37 ChrisL_: this is using the color() function, see example 9 21:39:46 ChrisL_: it's like @font-face 21:40:01 ChrisL_: you give it a url 21:40:44 https://drafts.csswg.org/css-color/#color-function 21:40:55 TabAtkins: idents is the right thing 21:41:13 ChrisL_: ident and then as many numbers as needed 21:41:17 (inside of https://drafts.csswg.org/css-color/#icc-colors ) 21:41:24 Florian: why only 2? 21:41:42 ChrisL_: sRGB could be added but it doesn't give us anything new 21:41:53 ChrisL_: happy to add it if people want it 21:41:57 smfr: makes sense for consistency 21:42:12 ChrisL_: there's a section about grayscale 21:42:31 ChrisL_: consider replacing the grayscale section with a grayscale colorspace 21:42:53 ChrisL_: we had a briefer syntax 21:43:22 leaverou: what about thing like named color profiles ? 21:43:42 leaverou: spot colors 21:43:50 ChrisL_: does not allow overrange values 21:44:00 ChrisL_: we've handwaved about this for a long time 21:44:14 ChrisL_: there was a linear space that was easy 21:44:25 ChrisL_: there's something black that's not absolute black 21:44:46 ChrisL_: that lasted for 2 years and then they decided to use a wider gamut if you want a wider gamut 21:44:56 ChrisL_: we should have early clipping 21:45:10 dholbert: it's quiet about that 21:45:26 s/dholbert/someone else/ (wasn't paying attention, sorry) 21:45:40 (maybe that was rik?) 21:46:00 s/someone else/cabanier/ 21:46:05 smfr: do we want commas? 21:46:13 TabAtkins: no, because we want to allow alpha 21:46:22 TabAtkins: need to distinguish 21:46:35 ChrisL_: so no commas between numbers in examples 21:46:46 ACTION ChrisL_ remove commas between in the color() examples 21:46:46 Sorry, but no Tracker is associated with this channel. 21:47:24 ChrisL_: next section is P3 and Rec 2020 21:49:03 ChrisL_: these refs can be added to MQ4 21:49:23 ACTION ChrisL_ fix the currently-broken Rec.2020 reference in the spec 21:49:23 Sorry, but no Tracker is associated with this channel. 21:49:27 leaverou: if one is supplied no color profile, it will be ignored? 21:49:52 leaverou: @color-profile foo icc profile which is treated differently 21:50:04 hober has joined #css-color 21:50:06 s/leaverou/ChrisK_/ 21:50:13 leaverou: this might be confusing in some cases 21:50:56 ChrisL_: it will look wrong if you copy a color but not the profile 21:51:09 TabAtkins: Rec.2020 is not a valid ident 21:51:16 TabAtkins: Rec2020 works fine 21:51:26 ChrisL_: ty noted 21:51:36 TabAtkins: also ascii lowercase pls 21:51:41 ChrisL_: mkay 21:52:06 TabAtkins: match what MQ does 21:52:06 ACTION ChrisL_ fix Rec.2020 and Rec2020 to be rec2020, and DCI-P3 P3 to either (consistently) dci-p3 or p3 21:52:06 Sorry, but no Tracker is associated with this channel. 21:52:43 ChrisL_: colors in wide color space and teh device with a narrower space 21:52:49 ChrisL_: what do you do with the ones outside 21:53:02 ChrisL_: if it's an image with a nice gradient, things can look ugly 21:53:20 ChrisL_: you move everything proportionally so that it looks nice 21:53:26 ChrisL_: it's handwaving 21:53:41 ChrisL_: maintain the relationship between colors but they're all going to match 21:54:04 ChrisL_: it's fine for a photograph but not when you want to match up a color in an image with a css color 21:54:41 ChrisL_: saturated is useless 21:54:44 TabAtkins: can we drop it? 21:55:37 TabAtkins: can we only allow them to choose the ones we care about? 21:55:40 ChrisL_: yeah... think so 21:56:01 smfr: i don't know if we want to drop absolute color metric, might be useful for testing 21:56:27 ChrisL_: "this has to have an absolute luminance of this" 21:56:47 Florian: if you have several profiles with the same color space? 21:56:51 ChrisL_: last one wins 21:57:03 TabAtkins: they cascade as a unit 21:57:48 cabanier: this is basically unused. can we leave it as is? 21:58:02 ChrisL_: in svg2 you have to overwride the profile 21:58:16 ChrisL_: should we remove the rendering intent thing? 21:58:23 cabanier: just use the default 21:58:33 TabAtkins: need some way to choose the default 21:58:44 ChrisL_: perceptual is less good for css 21:59:00 smfr: why is rendering intent tied to color profile? 21:59:14 smfr: you want to specify rendering intent ... 21:59:29 ChrisL_: the way ICC profiles work, is you have a profile for your source data 21:59:38 ChrisL_: then you have a profile for your output device 21:59:42 ChrisL_: compose those two 21:59:55 ChrisL_: how are you going to handle out of gamut colors 22:00:10 ChrisL_: if you select perceptual ... 22:00:40 ChrisL_: it's less applicable to a wide gamut screen 22:01:04 ChrisL_: an image has an intent built in 22:01:16 ChrisL_: you can't override it 22:01:31 TabAtkins: if the profile has multiple intents you can specify the one you want 22:01:43 Florian: if the profile has multiple, does it specify hte default? 22:01:46 ChrisL_: yeah 22:01:55 cabanier: we set that as the default in the product 22:02:04 Florian: it's ok for photos and what we should use for the rest? 22:02:06 ChrisL_: yes 22:02:19 ChrisL_: named color profiles, instead of having a mapping from a number 22:02:24 ChrisL_: you have mapping from a name 22:02:35 ChrisL_: people can do their own ones 22:02:50 ChrisL_: this syntax does not cope with that but can be extended to do so 22:03:03 Florian: color() function, either a list of numbers or a string 22:03:15 Florian: the other extension is to allow you to name your colors 22:04:19 Florian: new descriptor, @color-profile foo { src:...; named-color: "myblue" 00255, ...; 22:04:29 Florian: maybe not necessary because of variables 22:04:45 leaverou: can you rename existing colors? 22:04:47 ChrisL_: yes 22:05:07 TabAtkins: don't object but it's strictly not necessary 22:05:37 ChrisL_: i like it but variables is enough 22:05:46 cabanier: print wiht specific ink... 22:05:52 Florian: can do that with this or with variables 22:06:27 Florian: the variable contains "foo" 00255 so it contains the righ tcolorspace 22:06:40 TabAtkins: are these profiles text files? 22:06:46 ChrisL_: no terrible binary files 22:07:00 ChrisL_: tool to convert to xml so you can hack and convert back 22:07:10 dbaron: some security vulnerabilities 22:07:37 esprehn: apple have fixed security bugs in icc 22:07:46 ChrisL_: should restrict to same-origin 22:08:07 TabAtkins: web policy, things should default to same-origin 22:08:23 ChrisL_: why are people using google fonts 22:08:28 TabAtkins: CORS, no problem 22:09:08 smfr: before you download you have to wait for ICC to download, like with fonts 22:09:19 smfr: need to describe what should happen before it is downloaded 22:09:35 smfr: would not impl except for ??? 22:09:38 esprehn: same for us 22:09:46 s/???/built-in color profiles 22:10:06 Florian: context-dependent profile 22:10:10 esprehn: won't impl 22:10:17 Florian: what do you fallback to? 22:10:26 esprehn: wouldn't paint, same as fonts 22:10:34 Florian: so everything's white? 22:10:36 esprehn: yeah 22:10:51 ChrisL_ (above): flash of uncalibrated color 22:11:01 esprehn: i'd like to see that you can specify a name so you don't need to download if you already have it 22:11:08 ChrisL_: fonts have local() for that 22:11:17 ChrisL_: fine to have that here also 22:11:42 ChrisL_: how do you name your profiles, typically don't have a unique name 22:12:01 ChrisL_: how do i define it? 22:12:15 TabAtkins: platforms can define what their names are and you can put it in the spec? 22:12:23 ChrisL_: so it's different per platform? -_- 22:12:42 TabAtkins: here's url, here's hash 22:12:51 esprehn: not a security issue, privacy issue 22:13:03 TabAtkins: fingerprinting is a lost cause 22:13:14 dbaron: you can binary test for any font 22:13:24 TabAtkins: fingerprinting is not fixable 22:13:51 leaverou: if i'm using super-RGB color space, makes sense to fallback to sRGB until it's loaded 22:14:00 TabAtkins: so fallback to a known name? 22:14:26 dbaron: is there a dictionary of downloadable profiles we can name? 22:14:43 TabAtkins: we can inline them in the spec, they're small 22:15:10 Florian: want to point to an image to use it's profile 22:15:44 ChrisL_: we can put an appendix of common profiles 22:16:00 ChrisL_: print profiles are not small 22:16:14 TabAtkins: in houdini we want these to be serializable 22:16:26 TabAtkins: represent something in the string OM 22:16:38 TabAtkins: here's the profile, it's 700 bytes of chars 22:17:34 iank_ has joined #css-color 22:17:48 TabAtkins: the color is considered invalid until the profile is downloaded 22:17:52 esprehn: not going to reparse 22:17:58 TabAtkins: hrm 22:18:07 TabAtkins: ok nm 22:18:33 ChrisL_: if you have a font stack, how does it work when the primary is downloaded 22:18:41 esprehn: recompute, not reparse 22:19:04 TabAtkins: fallback in color() fallback is like font list 22:19:08 esprehn: that's implementable 22:19:18 TabAtkins: must provide a fallback 22:19:21 ChrisL_: fine with that 22:19:40 leaverou: want to be able to fallback to a different profile 22:19:44 TabAtkins: that's fine 22:20:44 TabAtkins: in the color profile rule you can have a list of fallback profile names 22:20:53 TabAtkins: also in the color() function you can have a fallback color 22:21:24 cabanier: have to always specify fallback? no optional, ok 22:21:36 Florian: can you use rgba() in the flalback? 22:21:43 TabAtkins: ya 22:21:58 ChrisL_: open issue, blending happens in sRGB 22:22:04 ChrisL_: which will give you clipping problems 22:22:09 ChrisL_: also doing maths wrong 22:22:19 ChrisL_: was going to add linear 22:22:31 ChrisL_: there's hardware support 22:22:47 ChrisL_: you can get the right result with LAB 22:23:03 ChrisL_: we need a thing that says how you composit things 22:23:43 smfr: compositing and blending has a similar problem 22:23:58 cabanier: the gamut is being worked on in the canvas spec 22:24:12 cabanier: need to specify, just 1 22:24:55 smfr: we can't change the default because that breaks web pages 22:26:20 ChrisL has joined #css-color 22:26:23 TabAtkins: subframes does it only work for top-most windows? 22:26:28 TabAtkins: iframe 22:26:40 Florian: if it doesn't say but the iframe does, does it propagate up? 22:26:45 TabAtkins: no. you can have two iframes 22:27:03 ChrisL: can't find linear blending in the spec 22:27:13 s/ChrisL/smfr/ 22:27:16 ChrisL: it's an open issue 22:27:41 Florian: when you've done your things in your colorspace, when you've done it, what do you export it into? 22:27:49 ChrisL: the output colorspace, whatever that is 22:28:02 ChrisL: LAB is convenient for this reason 22:28:16 Florian: if you use anything but LAB do you need to go through LAB first? 22:28:18 ChrisL: yeah 22:28:33 Florian: when that's specified, do the color math other than oclor blending also happen in that colorspace? 22:29:01 Florian: if we have LAB, can i switch to it and use it for ??? 22:29:08 ChrisL: yes but you get different result 22:29:12 Florian: ya that's teh point 22:29:31 cabanier: if you render in colorspace then all colors are in the same colorspace 22:29:55 ChrisL: i've seen examples of interpolation of ??? 22:30:01 ChrisL: can we pick a placeholder name? 22:30:11 TabAtkins: document-interpolation-color-space 22:30:15 smfr: working color space? 22:30:19 TabAtkins: yeah 22:30:44 cabanier: you map your colors so it's fine 22:30:49 ChrisL: 3x3 matrix 22:31:01 ChrisL: don't use 8 bits 22:31:10 ACTION ChrisL working color space 22:31:10 Sorry, but no Tracker is associated with this channel. 22:31:21 Florian: sRGB? 22:31:28 ChrisL: auto is sRGB 22:31:34 ChrisL: linear-sRGB 22:31:35 ChrisL: LAB 22:31:38 ChrisL: any others? 22:31:52 cabanier: the list of working spaces can't include linear? 22:31:54 ChrisL: why not? 22:32:01 cabanier: can't implement it yet 22:32:07 ChrisL: that's wrong. HW has it already 22:32:21 ChrisL: can do it in linear-sRGB with clipping 22:32:31 cabanier: need to change the blending spec 22:32:45 cabanier: so the output is bogus? 22:32:52 ChrisL: no, it's wrong now but people are used to it 22:33:32 ChrisL: yes you will get a different result, if you're not using linear space you're technically get the wrong result now 22:33:48 ChrisL: additive... 22:34:03 s/additive/additivity/ 22:34:03 ChrisL: measure the light of two lights 22:34:24 ChrisL: additivity means i can take the two numbers and get the right result 22:34:39 ChrisL: in linear space you get the right result with +, otherwise not 22:34:55 Florian: in LAB, linear matches perception, which is nice 22:34:58 ChrisL: exactly 22:35:21 smfr: MQ? 22:35:38 cabanier: should match canvas 22:35:54 cabanier: ... p3 2020, optimal (matches the screen) 22:36:08 cabanier: on iMac you would be compositing in p3 22:36:21 cabanier: the gamut most closely matches the output 22:36:31 ChrisL: ok so MQ 22:37:11 Florian: i think what's in the spec now which is TabAtkins's interp. of what we had in the mailing list, is fine with me 22:37:26 Florian: if people disagree say where and why 22:37:42 dbaron: this is a new type of MQ 22:37:54 dbaron: MQ where more than one of the values can be true 22:38:03 TabAtkins: yeah all the range based ones have that 22:38:14 dbaron: it's a set MQ rather than a discreet MQ 22:38:24 dbaron: might want to describe as a different type 22:38:27 TabAtkins: ok 22:38:39 ChrisL: are the ordered? 22:38:46 TabAtkins: only range features have an ordering 22:38:54 Florian: that you can do "less than" 22:39:37 dbaron: color gamut sRGB means ??? 22:39:49 s/???/it can do all of sRGB/ 22:40:14 TabAtkins: should we add a value for "narrow" for extra-low gamut (below sRGB)? 22:41:01 dbaron: concerned whether we CAN calculate it, if we have information to calculate it 22:41:34 Florian: should i load normal images, fancy images or super-fancy images 22:42:05 leaverou: percentages seems useful 22:42:11 smfr: no this is good enough 22:43:11 TabAtkins: boolean context are true for anything other than none or 0 22:43:16 TabAtkins: not (boolean) is true 22:44:01 smfr: you're asking about the OS, not the screen 22:44:34 TabAtkins: not talk about the output device, need to fix wording 22:44:55 Rossen: present to projector 22:45:08 Florian: two screens? 22:45:34 esprehn: every monitor has its own profile, if you move a window between two, we ??? 22:45:51 we repaint the window, which resolves the colors again 22:46:29 Florian: the incentive for UAs to lie is not high, it only makes you download heavier images than you need 22:46:37 TabAtkins: if there are multiple screens 22:46:47 Florian: why not the highest? 22:47:03 TabAtkins: in general you want the lowest 22:47:19 TabAtkins: want to see the same thing as what's on the projector 22:47:24 TabAtkins: below CSS 22:48:12 Florian: there are objections that this is not restricted to RGB spaces or the screen media type 22:48:33 Florian: if you have a whatever colorspace, if it's bigger then p3 then you match 22:48:38 Florian: people have disagreed with that 22:48:54 Florian: maybe add a cmyk 22:49:09 ChrisL: we have TVs which have ??? 22:49:25 Florian: if you have a sucky printer sRGB is what you get 22:49:41 TabAtkins: we can add specific named variants 22:49:59 Florian: for UAs that don't output in cmyk, this has 0 impl cost 22:50:14 TabAtkins: we dismiss the objections and go with the current design 22:50:17 Florian: happy with that 22:50:40 ChrisL: publish this? 22:50:45 TabAtkins: intent to publish 22:51:16 PROPOSED RESOLUTION: publish css-color and mq4 22:51:29 dbaron: there's impl interest for some things but not others 22:51:41 dbaron: might be worth trying to figure that out sooner rather than later 22:52:11 TabAtkins: figure out what we want to impl 22:52:35 ChrisL: color-mod() problems TabAtkins? 22:52:41 TabAtkins: 1) syntax is confusing and crappy 22:52:45 https://drafts.csswg.org/css-color/#modifying-colors 22:52:55 TabAtkins: i agree partially 22:53:02 TabAtkins: it's overcomplicated 22:53:18 esprehn: aliasing for everything? can we not do that? 22:53:32 TabAtkins: 2) ??? is not super useful for real-world stuff 22:53:54 TabAtkins: modifying one color ... suggested some different operations that i want to put in here instead 22:54:05 TabAtkins: replace this with something like blending two colors together 22:54:19 leaverou: lightness or mix with white and black which works better 22:54:49 TabAtkins: color blending, color contrast, tint and shade 22:55:45 leaverou: i impl a polyfill, it occurred to me that there were lots of (((()))) like lisp 22:55:58 TabAtkins: the grammar is intended to reduce that 22:56:16 leaverou: suggest keywords 22:56:19 TabAtkins: agree 22:57:00 leaverou: crazy idea, what if we can use calc(color + semi-transparent white) then it blends with alpha blending? 22:57:06 TabAtkins: calc is for numerical things 22:57:26 TabAtkins: we have cross-fade as existence proof that we use a new function for specific things 22:57:48 TabAtkins: we need color blending for convenience and so you can ??? 22:57:57 Florian: device-cmyk 22:58:09 ChrisL: if you have calibrated device-cmyk 22:58:12 Florian: that's no problem 22:58:21 ChrisL: if you don't provide a fallback you get a bad color 22:58:26 Florian: how does it work? 22:58:39 ChrisL: it means here are some numbers, don't mess with them at all 22:58:57 Florian: if we composit and alpha-blend ... 22:59:13 ChrisL: if you use color() to do that, it looks better 22:59:35 ChrisL: inks are not transparent, can't overlay them 22:59:41 liam has joined #css-color 22:59:44 ChrisL: have a pattern of dots instead in a grid 22:59:51 ChrisL: overlaid in different angles 23:00:13 ChrisL: if you have a pale yellow and then few dots of black 23:00:41 ChrisL: that's what device-cmyk is for 23:00:49 ChrisL: don't want those black dots 23:01:31 Florian: why does output to PDF file have anything to do with the screen? 23:01:43 TabAtkins: don't have at a distance logic in MQ 23:01:52 Florian: if you're a headless device? 23:01:57 Florian: OS doesn't know what to do 23:02:11 Florian: what does the MQ do? 23:02:33 ChrisL: it should say yes to everything, can print to pdf in any profile 23:03:41 leaverou: can we get rid of device-cmyk somehow? 23:03:47 TabAtkins: it's implemented 23:04:39 Chris: black point compensation 23:04:45 ... white is easy 23:05:05 ... TVs can get much darker black 23:05:17 TabAtkins: can we not do something like white compensation? 23:05:23 Chris: no. the spec says why 23:05:34 ... same mutual axis 23:05:50 ... it's not pegged at 0 it's pegged at some dark 23:06:00 ... certain amount at the bottom for deeper blacks 23:06:06 ... do need to cope with that 23:06:13 ... wording about rendering intents 23:06:22 ... black point compensation matches or not 23:06:29 ... with or without black compensation 23:06:47 ... in PS it will ask if you want black compensation 23:06:53 TabAtkins: which do we want? 23:06:55 cabanier: turn it on 23:06:58 TabAtkins: ok. done? 23:07:18 cabanier: color conversion engine cna do it for you 23:07:23 cabanier: same thing happens with whites 23:07:30 cabanier: white stays white 23:07:51 RESOLVED: do black point compensation 23:08:04 ...when converting from profile to another 23:08:19 smfr: for print also? 23:08:23 Chris: yeah 23:09:06 Chris: doesn't affect p3, does affect 2020 23:09:24 TabAtkins: can i reformat everything? 23:09:28 Chris: sure 23:09:51 Florian: the only thing we haven't agreed on is the working color profile? 23:09:55 Chris: right 23:10:04 leaverou: device-cmyk 23:10:18 leaverou: stylesheet was used for printing to ebook 23:10:26 leaverou: using something similar to device-cmyk 23:10:38