Categories
Questions

Question #14: What the F*ck is “White”?

Phew. It’s been quite a few questions to be able to finally, just about, almost, kinda, sorta discuss what is “white”. The key here is to attach “white” to terms that will not only make sense to you, the happy pixel pusher, but also be able to discuss it in ways that apply to your digital image work.

Why does that colour science stuff matter? Because that’s how we take our creative pixels and migrate them from one medium to another. For pixel pushers, that’s how we get from numbers to what we see.

That short film you’ve spend months or years on getting to YouTube without looking like it has an orange tint? Colour science around “white”. The advertising print run of ten thousand pieces not ending up speckled with blue dots where no ink white paper should be? Colour science around “white”. That nice graphic you designed for your YouTube influencer channel not skewing blue on the web? Colour science around “white”. Even the most simple case of taking an encoded and compressed nonlinear code value generic image file and displaying it correctly? Colour science around “white”.

Seems like an important question, so…

Question #14: What is “white”?


You might be wondering why we’ve waited until Question #14 to tackle the idea of “white”. The truth is, we couldn’t answer this in any useful way without plenty of foundation!

That is a lot of ground to be able to discuss just what “white” is. Heck, it’s not a stretch to suggest that the entire series so far relates to ideas around “white”.


As we learned in Question #13, we can discuss colour in relation to physical spectral mixtures. As we learned, the canonized and referential absolute mapping is a means of plotting visible light spectra to useful numbers. Given it covers all visible light spectra, no mixture of visible light exists beyond its limits. That means that “white” isn’t some mystical narwhal unicorn that lives in the clouds somewhere else. Anything we call “white” will be a mixture of visible spectra and, as a result, be on that map!

But wait… some colours look like colours, and other colours… well… don’t look like colours. They look as though they are without colour, or achromatic.

As you may have guessed, the idea of “whiteness” is indeed related to colours and chromaticities, which tells us that an idea of “whiteness” is a psychophysical bit of magic around perceptual chromatic adaptation.


If you have ever walked inside of a large department store after having walked around outside during the day, you may have noticed something interesting regarding “white”. From the outside of the store, your field of view was largely lit by the daylight sky. Typically, even with a cloud cover, the light is biased towards the bluer wavelengths of spectral mixtures. Inside, quite a few department stores use those long fluorescent tubes, which while they can cover a varying range of spectral output, they tend to be biased toward yellowy-orangey spectral mixtures.

If we start outside the store, our constantly shifting perceptual systems will have fully adapted to the blue weighted spectral mixture of light coming from the sky and clouds. This means that our perception of an achromatic object or light will be relative to that blueish spectral composition. Our perceptual systems will “dial” the actual spectral blue tint out such that it will create that sensation of not having any colour. So what happens to the rest of the scene?

When our perceptual systems adapt to a neutral achromatic axis it does something truly remarkable; it shifts, bends, warps, and adjusts every single spectral combination in our field of view. The entire rest of our colour sensation will be skewed to that adaptation! If we were to plot all of the colours in the scene to chromaticities, all chromaticities will appear to have shifted around the chromaticity diagram after our perceptual system has processed them!

As we look inside the store, the fluorescent tube lighting looks rather yellowy-orangey from the outside. But what happens when we walk inside…


As we transition to inside the store, plenty of incredible things are happening within our perceptual systems. At first, the store will appear that yellow-orangey, and the outside will, if we check over our shoulder, still have “white” objects and lights.

After a few moments shopping for that ten pack of caffeine boosters to get you through that last video edit, you’ve found yourself realizing the label on the product is in fact… sort of “white”! You twirl around not realizing that while you were on the caffeine hunt that the entire store appears perfectly normal! Except… outside…

As you look from your fully adapted state inside the store, the outside scene through the store’s windows now looks distinctly blue twinged! You’ve fully adapted to the store’s illumination, and the outside is perceived relative to it. All of the chromaticities will appear to have shifted and changed invisibly, behind the scenes!

And in all of that magic, no such thing as “white” exists, only changing states of perceived achromatic “whiteness”.


So what is white?

Answer #14: “White” is simply another mixture of visible spectral lights that our perceptual system has fully adapted to. As a colour, the idea of “white” has no grand, universal, absolute truth without considering the context of the observer. We can express the achromatic point on the CIE 1931 plot as a chromaticity.

But wait! Does anyone know why this happens!? What the hell is showing when I set my RGB to 1.0 1.0 1.0?!? The lighting at my desk has an impact on my work on my iPad Pro then!?? How do I change my print run’s white to match the paper, because the five thousand copies are due tomorrow?! What about changing the colour temperature in my photograph to match the others in the set?!?! HELP!!

More questions that will probably require more questions…

4 replies on “Question #14: What the F*ck is “White”?”

Hi! I want to thank you for this incredible work, it’s just amazing material. I have a question related to an image that I saw on the Chaos group website. If I understood the meaning of the article correctly, then there is no “color” for the human eye outside the chromaticity diagram, but the ACES triangle goes beyond it. It turns out that ACES somehow allows you to display even those values that the human eye is unable to see?
https://static.chaosgroup.com/images/assets/000/011/098/full_width_original/ACEScg_03.png?1601049151

Like

The standard observer model that yields that familiar horseshoe is based off of testing a few folks.

Literally nothing exists beyond the spectral locus, and relative to the standard observer model, those values are meaningless math mumbo jumbo.

It’s just a generally bad idea using virtual primaries due to this, and it puts greater stress on gamut mapping, as well as compositing.

No, ACES can’t do anything magical, and in fact it is a sloppy mess that doesn’t really work terribly well.

Liked by 1 person

Hi Troy, thanks for this fantastic guide! I have a fair amount of experience of working with colour over the years, but this has been a timely refresher and it turns out that I’d been labouring under a number of misconceptions which you have neatly blown away for me.
With reference to ACES, I’d love to know why you think it’s a sloppy mess? I don’t have any particular skin in this game other than the fact that it’s what everyone in my industry is using, so I need to understand it.

Like

Happy to hear that the nonsense pile here has provided some utility toward understanding.

I’d say that “What everyone in my industry is using” is grossly overstated. If anything ARRI K1S1 / ALF2 is the defacto industry standard, and I am personally aware of more than a number of folks actively trying to avoid ACES, including a good number of cinematographers and post house TDs.

It is simply nonsense. It has no bearing on vision, poorly executed in terms of picture formation, has no actual colour management in terms of tristimulus or appearance matching of pictures, and is fundamentally a dismal attempt at any form of a solution.

I am happy to explain further, but it is at once a short conversation, but a longer and nuanced discussion in terms of details.

1. The protocol does not work from a creative standpoint; limiting applying creative manipulations to open domain tristimulus makes some creative choices impossible.
2. It does not stand up to a shred of scrutiny regarding “management”; it is the epitome of device dependency.
3. The totality of the design and engineering is less than grade school student inept.
4. It is an absolutely shoddy example of picture formation from open domain tristimulus.

And the list goes on and on. Pure nonsense of the scale whereby literally every single individual who has shilled for it, reveals themselves to be of the most poorly informed variety, or facilitating a role as a larger corporate nonsense peddler.

Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s