Hacker News new | past | comments | ask | show | jobs | submit login
The Search for Our Missing Colors (newyorker.com)
64 points by bananaoomarang on May 23, 2016 | hide | past | favorite | 21 comments



I wish people writing about new broad-gamut displays would at least include a brief mention of the downsides of extremely narrow-wavelength-band primary colors:

Such displays can create dramatic “observer metamerism” problems, where different people view the colors on screen significantly differently from each-other – even if they would see real-world objects relatively similarly – because slight differences in color vision are amplified by color sources with spiky spectra. Even worse, spiky spectra tend to cause a noticeable difference between foveal and peripheral color perception.

References: http://rit-mcsl.org/fairchild/PDFs/PRO30.pdf http://cias.rit.edu/media/uploads/faculty-f-projects/1304/do...

Just did a google search, and there’s also some discussion here http://www.avsforum.com/forum/139-display-calibration/190850...

Personally, I also feel like spiky-spectrum displays and light sources give me more eyestrain and seem less pleasant to look at. But I admit that such an effect might be entirely in my head.

I’m looking forward to the day when we have displays with 6+ primaries, or maybe even reflective displays (not backlit) with 6+ subtractive color primaries.


> give me more eyestrain and seem less pleasant to look at. But I admit that such an effect might be entirely in my head.

Even if it isn't entirely in your head, it's still entirely in your head. ;-)


In case anyone else was annoyed by the lack of technical details:

Old (Rec. 709) colour space: https://en.wikipedia.org/wiki/File:CIExy1931_Rec_709.svg

New (Rec. 2020) colour space: https://en.wikipedia.org/wiki/File:CIExy1931_Rec_2020.svg

Most of the "new colours" are greens and turquoises; reds can now be slightly brighter, but blue is almost unchanged.


Gah. It’s been 40 years everyone. Let’s agree to stop using the CIE 1931 (x, y) chromaticity diagram, which is horribly misleading, in favor of the CIE 1976 UCS (u', v') chromaticity diagram. [I would replace all the diagrams on Wikipedia myself, but it would take several days of effort, and probably a few weeks’ slog of talk page arguments.] The difference between these two color spaces is much less extreme than the xy chart makes it look.


I don't know enough about colour perception to argue the merits of one diagram vs. another, but am I incorrect in my assessment of what colours are entering the gamut (lots of green, a bit of red, not much blue at all)?


Sort of. The green area is not nearly as dramatically different as the xy diagram makes it seem, while the red and blue areas will appear more different in the u'v' diagram. (Of course, chromaticity diagrams in general aren’t a perfect tool.) The “Gah” is directed at Wikipedia, not at you.

You can see some u'v' diagrams alongside spectral power distribution charts of specific real-world displays in http://cias.rit.edu/media/uploads/faculty-f-projects/1304/do...


What if we wrote a scraper and autoconverter? Is automated editing of Wikipedia prohibited?


I'm pretty sure most of Wikipedia editing is automated (by volume), see, e.g. ClueBot. That said, if you don't get consensus to make such bulk changes, it will probably not go over well.


I look forward to one day seeing what these diagrams are actually supposed to look like...

I kind of suspect that once you're used to the wider gamut, other screens will look washed out. Like how low-res CRT screens look awful and smudgy today, even though I used to be able to spend hours in front of one.


And violet is still entirely missing!


You don't see it between red and blue, a little closer to blue?


You definitely can't "see" it on this chart, because your monitor doesn't display it. "The wavelength of the Rec. 2020 primary colors is 630 nm for the red primary color, 532 nm for the green primary color, and 467 nm for the blue primary color." Violet is a shorter wavelength, 380-420 nm or so.


That would be purple, I was referring to spectral violet, the color after blue and indigo when going down the spectrum.


* it's a tour of the exotic pigment samples collection at Harvard


Thank you for saving me a click.

As a designer I get my panties in a knot every time I read about color. At least when I catch misleading or ignorant headlines.


That comment only describes the video at the bottom, not the article.


As someone who is used to being unable to distinguish between similar shades of color, this strikes me as somewhat funny. So much ado about a few missing nuances.


but its my understanding that rec. 709 isn't the same as sRGB which is what most non-TV things are attempting to emulate/implement


iow, they are attempting, again, to emulate the capabilities long available in film. This is reason #9179 why film STILL is the best medium.


This is not about the recording side of things, this is about the display side of things. And if you display footage shot on film on a standard computer monitor, you won't magically have more color.

Also, modern high end digital cameras can record a lot more color information than film can...


As I said, for display, too, film beats them all.




Consider applying for YC's W25 batch! Applications are open till Nov 12.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: