First time I'm reading a LTT Labs article and I really liked the format. The left hand outline (which is like a polished version of Wikipedia's) was very nice. No cookie popups, only a simple Plausible tracker in the logs. RSS Feed. Nice and refreshing!
The double space after a period, with the added effort to avoid browser whitespace condensing, is an interesting style choice. Is it meant to mimic old academic publications?
While present in some of their previous articles sparingly, this is the first one to use it consistently.
Reading this, marveling at the beautiful place we seem to be with regard to display technology.
Maybe an expert can clarify. My sense is that we actually were seeing amazing specs on displays toward the end of the CRT era—and that there was a kind of LCD lull where we traded color fidelity, gamut and dynamic range for thinness. But we are now seeing display technology catching up to and perhaps exceeding the best of the CRT displays?
That’s not what happened. The phosphors chosen for CRTs were pretty much unchanged from the late 80s through the end of the CRT era. There were minor differences in formulation, but often that was at the expense of color accuracy for motion clarity.
The early LCDs were poor because of limitations of the backlight’s spectrum. Once LED backlights were introduced, this was no longer a problem.
Does anyone have link to a deeper technical dive on the limitations of CIE 1931 that Apple CMF tries to address? The CIE's JTC22 (D8/D1): Optimized Colour Matching Functions for Display Colour Consistency documents seem to be behind an access wall.
From what I gather, it's an attempt to address the problem of observer metamerism, but I'm curious to know how closely it's tethered to D65.
It’s independent of white point. The metamerism problem is fundamental to the methodology of the 2° XYZ colorspace because it doesn’t model the entire real response of the human eye. Instead it underestimates the response of the blue cone. This can lead to blue becoming purple as Y is reduced. This is because the experiment used lamps with colored filters (broadband light) and not lasers or LEDs (narrowband light).
It’s not just a matter of introducing a correction factor (although that exists, see Judd-Vos) because the solution is more complex than that. The practical take away is that color matching cannot just involve comparing a narrow point in isolation but also needs to have a baseline of controlled surround and specular reflections in addition to understanding the nature of the perceived color (reflective, emissive, narrowband, wideband, etc.).
Unfortunately CIE standards like the JTC22 (D8/D1) work here tend not to be released for free. Eventually the mathematical curves should be adopted by open source implementations. Hopefully this isn't encumbered by patents.
First time I'm reading a LTT Labs article and I really liked the format. The left hand outline (which is like a polished version of Wikipedia's) was very nice. No cookie popups, only a simple Plausible tracker in the logs. RSS Feed. Nice and refreshing!
The double space after a period, with the added effort to avoid browser whitespace condensing, is an interesting style choice. Is it meant to mimic old academic publications?
While present in some of their previous articles sparingly, this is the first one to use it consistently.
I didn’t know how much I wanted this in my reading. It’s scannable and somehow feels ‘just right’, is there a reason not to use it?
Reading this, marveling at the beautiful place we seem to be with regard to display technology.
Maybe an expert can clarify. My sense is that we actually were seeing amazing specs on displays toward the end of the CRT era—and that there was a kind of LCD lull where we traded color fidelity, gamut and dynamic range for thinness. But we are now seeing display technology catching up to and perhaps exceeding the best of the CRT displays?
That’s not what happened. The phosphors chosen for CRTs were pretty much unchanged from the late 80s through the end of the CRT era. There were minor differences in formulation, but often that was at the expense of color accuracy for motion clarity.
The early LCDs were poor because of limitations of the backlight’s spectrum. Once LED backlights were introduced, this was no longer a problem.
Does anyone have link to a deeper technical dive on the limitations of CIE 1931 that Apple CMF tries to address? The CIE's JTC22 (D8/D1): Optimized Colour Matching Functions for Display Colour Consistency documents seem to be behind an access wall.
From what I gather, it's an attempt to address the problem of observer metamerism, but I'm curious to know how closely it's tethered to D65.
It’s independent of white point. The metamerism problem is fundamental to the methodology of the 2° XYZ colorspace because it doesn’t model the entire real response of the human eye. Instead it underestimates the response of the blue cone. This can lead to blue becoming purple as Y is reduced. This is because the experiment used lamps with colored filters (broadband light) and not lasers or LEDs (narrowband light).
It’s not just a matter of introducing a correction factor (although that exists, see Judd-Vos) because the solution is more complex than that. The practical take away is that color matching cannot just involve comparing a narrow point in isolation but also needs to have a baseline of controlled surround and specular reflections in addition to understanding the nature of the perceived color (reflective, emissive, narrowband, wideband, etc.).
So they introduced their CMF set and referenced it a bunch, but Apple CMF 2026 is not published anywhere. Is this an open standard or not?
Unfortunately CIE standards like the JTC22 (D8/D1) work here tend not to be released for free. Eventually the mathematical curves should be adopted by open source implementations. Hopefully this isn't encumbered by patents.