Computer Monitors and Eyeglass Lenses: The Color Fringing Problem

If you use a computer and have a relatively strong eyeglass prescription — more than 4 diopters, positive or negative — you should read this!  Particularly if you have noticed annoying color fringing when you look at your monitor.

Some time ago I bought an LCD monitor with an RGB LED backlight.  I immediately noticed a strong color fringing effect: the colors did not always line up as I looked at them.  So, for example, a horizontal yellow line would split into red and green lines.

This effect was not unfamiliar to me; I had known for years that my eyeglass lenses caused Chromatic aberration.  Except at the optical center of a lens, the faces are not parallel, causing the lens to behave, to some extent, like an optical prism.  Different colors are refracted slightly differently.  The effect depends on what part of the lens one is looking through: at the optical center, there is no aberration; it's at a maximum (for that lens) near the edges.

And, I was familiar with the effect this had on reading text on a monitor.  The aberration causes the colors to separate a little.  The effect is particularly noticeable in light-on-dark modes, which I prefer when available.

But I had previously found the effect to be tolerable.  With my new RGB LED monitor, it was much worse, to the point where I found the monitor unusable.  I set out to understand what was going on and what, if anything, could be done about it.

The first question was, why was the effect so much worse with the new monitor?  The answer can be seen in these graphs:

Here we see the spectrum of the more common white LED backlights on the right, and that of the new RGB LED backlights on the left.  Also shown are the filter spectra.  What we really want are the products of each emission spectrum and the respective filter spectra; the product curves are not shown, but we can estimate what they would look like.  Look at the right graph first.  My eyeball says the peak of the product of the white LED spectrum and the green filter is around 560nm, while the peak of the product of the same emission spectrum and the red filter is around 610nm — about an 8.6% difference.

Now look at the left graph, for the RGB LED.  The green peak is clearly at about 520nm, and the red at 640nm.  That's about a 14% swing — some 60% larger!  Since the degree by which a prism separates colors is proportional to the ratio of their wavelengths, that means that red/green aberration is 60% worse with the RGB backlight.

The spectra of the even older cold-cathode fluorescent (CCFL) backlights varied depending on the phosphors used, but the net effect in most cases was that the green peak was also at a relatively long wavelength — relatively close to red — as with the white LEDs.  (I have a 10-year-old Philips CCFL monitor that I still use, and its green, next to that of the RGB LED monitor, is distinctly yellowish.)  Of course, the CCFLs were weaker in the blue than either LED backlight, and I believe their blue peaks were also at a longer wavelength (less deep blue).

A couple of digressions are in order at this point.  First, it is precisely the greater separation of the red and green peaks that gives the RGB LED backlights their desirable property: a wider color gamut, meaning the monitor can display saturated deep greens and reds that are not possible with a white LED monitor.  This can matter for purposes like professional photo editing (how much it really does matter is a question I'll leave to the pros).  Second, the location of the blue peak matters quite a bit less; I gather that the eye does not extract much detail information from the blue channel, so misalignment between the blue and green, or blue and red, is less subjectively distracting than red/green misalignment.  (Of course, the blue peaks in the white LED and RGB LED displays are at the same wavelength anyway.)

So, that answered the question as to why the aberration was so much worse with the RGB LED monitor.  This brought me to the next question: what can be done about it?

Again, the effect is caused by the interaction between the wavelength peaks of the monitor and one's eyeglass lenses.  People who don't wear glasses won't see it at all.  Two ways to eliminate it, then, would be to wear contact lenses or to have laser eye surgery.  I didn't want to do either of those, for reasons that need not concern us here.  Is there any other way?

It turns out that there is.  It requires delving into a different topic: eyeglass lens materials.

Different lens materials have different amounts of dispersion: the degree to which different colors are refracted differently.  Dispersion is quantified by the Abbe number: the higher the Abbe number, the lower the dispersion, and conversely.  Here are common lens materials with their refractive indices and Abbe numbers:

MATERIAL	INDEX	ABBE VALUE
Crown Glass	1.523	59
High Index Glass	1.60	42
High Index Glass	1.70	39
Plastic CR-39	1.49	58
Mid Index Plastic	1.54	47
Mid Index Plastic	1.56	36
High Index Plastic	1.60	36
High Index Plastic	1.66	32
Trivex	1.53	43
Polycarbonate	1.58	30

You can see that within each class of material — glass or plastic —the Abbe value is roughly negatively correlated with the refractive index.  This is bad news for those of us who need strong lenses; the lower the refractive index, the thicker the lens has to be for a given power.  So when we order lenses, opticians recommend the higher-index materials, which have the lower Abbe numbers, and thus higher dispersion: more chromatic aberration.   I don't know exactly which material my last pair of glasses had been made with, except that it was one of the high-index plastics, with an Abbe value quite possibly as low as 32.

Okay — so how can we use this information to obtain a pair of glasses that minimizes aberration?  Since the purpose of using high-index materials is to minimize lens thickness, and since lens thickness is positively related to the size of the lens (more precisely, to the maximum distance of any point on the lens from its optical center), it follows that an alternate approach to keeping the lens acceptably thin is to make it relatively small and roughly circular.  Such a lens could then be made with a high-Abbe-number material without being too thick.

I decided to give this a try.  I chose a round-style frame with a 41mm lens width.  My prescription calls for -5.75 diopters in both eyes.  I decided I wanted to try CR-39 since it has easily the highest Abbe number of any plastic.  Many online opticians will refuse to make such a strong lens in CR-39, but I managed to persuade one to do so by pointing out that I had chosen a small round frame.

The result was a huge improvement!  I can use the RGB LED monitor happily now.  (And even using my old CCFL monitor, or using my laptop which has a white LED backlight, is more comfortable.)

I strongly recommend this approach to anyone with the same problem.  Of course, you could also avoid RGB LED monitors, but that's getting harder to do as they become more popular; for instance, I think the latest iMacs use RGB LED backlights.  At the very least, anyone who wears glasses and is buying a monitor should be aware of the issue.