Daniel Matte


Debunking misinformation about the Pixel 2 XL’s display

(For anyone following me on Medium, I appreciate the gesture, but my actual blog is here. I’ll write a follow-up on the display issues themselves soon.)

It is impossible to explain anything non-trivial on Twitter, so I basically give up at this point. Here is more than 140/280 characters, because consumers deserve to know the basic picture:

There is nothing inherently bad about OLED color accuracy, period. The Phil Schiller keynote slide on OLED displays was 100% nonsense. In Apple’s defense, they were really talking about the quality of their competitors’ displays in general. The slide made it look like these were inherent OLED issues, which was extremely misleading to consumers. Companies try to sell you a product, not teach and inform.

OLED displays have been better than LCDs for the past few years. OLED has improved massively over many years but still has tons of issues. Both these statements are true.

Samsung and Apple both always ship awesome displays. Google devices meanwhile have had both good and bad displays over the years. I would strongly recommend not reflexively being biased towards any brand or product, though, if you want to actually understand the technology.

AnandTech has covered OLED displays in extreme depth over the years, so if you want to learn about OLED and all displays my highest recommendation would be to read their articles (can search for various previous device reviews). You can also follow @nexusCFX on Twitter.

Android didn’t have color management until Oreo. Now it does. We’re good here. (There’s much more work still to be done for HDR support for Android P.)

Because they have color management (and other factors), both the new Pixels target the Display P3 color space (which is currently correct for “wide color”), not sRGB as many have stated. Google’s marketing even says this. How good the specific panel calibrations turn out to be is a separate question. Vlad Savov’s review unit for The Verge is clearly extremely green-shifted and looks awful. I won’t cover the work that needs to go into calibrating displays at the factory level here.

The 2 XL was always going to have a bad display. LG Display’s mobile OLED is many years behind Samsung Display’s and still suffers from a ton of issues. There’s not much Google can do here. And no, the answer actually isn’t “don’t source LGD OLED in the first place.” Google can’t work supply chain miracles.

The 5.0” Pixel 2 could have a well-calibrated panel. We have no data, so I have no idea!

There are a ton of potential issues with OLED in general, including off-axis color shifting, mura, moire patterns, uniformity more broadly, chromatic aberration, chromatic aliasing, and many more. These are all separate things, many of which clearly apply to the V30 and Pixel 2 XL’s displays.

No one has publicly tested either Pixel’s display. Brandon Chester tested the original Pixel’s display, which unfortunately targeted the worthless NTSC color space and should be understood entirely in its own light (it was bad): https://www.anandtech.com/show/10753/the-google-pixel-xl-review/3.

Google isn’t stupid or unaware, and Romain Guy (the head of Graphics on the Android Team) in particular is phenomenal. Hardware, however, has often been a big struggle for Google due to insufficient engineering resources. Or investment and caring about hardware, if you don’t want to sugarcoat it. (Don’t even get me started on insufficient RF testing.)

The Android Central text and video reviews were the most sensible ones I came across regarding display quality.

Vlad, Dieter, and the rest of the Verge are good people, and I’m definitely not criticizing them :) Those folks helped me get into tech back in the Engadget days.

You should never try to eyeball color accuracy or a calibration. Even professional display calibrators don’t trust their eyes: https://techspecs.blog/blog/2017/7/21/why-you-shouldnt-use-someone-elses-tv-calibration-settings.

To assess color accuracy, you need actual data from credible testing. Unfortunately the necessary equipment and software cost thousands of dollars, even at entry level pricing.

Panel variability is undoubtedly a part of the story. All silicon has variability.

The iPhone X’s panel also doesn’t look so incredible at a distance. The color accuracy I’m sure will be excellent, but there’s a ton more that matters. It needs to actually be tested.

I try to cover the little that I know at TechSpecs.blog (and on Twitter, sigh). I promise I try not to write anything that I’m not pretty sure about. I also wrote about the rough details around the Pixel Visual Core a couple months ago if you’re into learning about that sort of thing.

And if anyone reasonably thinks, “this is just some random guy, why should I trust any of this?” I wouldn’t blame you, but I have talked with all the major display vendors over the years and have been given some behind the scenes explanations. If nothing else just believe AnandTech, whose credibility should speak for itself.

AnandTech also isn’t dead, so stay tuned for some awesome stuff in the medium term :)

Topics of interest

More Related Stories