A New Study About Color Tries to Decode ‘The Brain’s Pantone’ | Hacker News Reader

A New Study About Color Tries to Decode ‘The Brain’s Pantone’

wired.com

37 pointsdangerman11 comments

The article describes an attempt to create a "one true colorspace" by using fMRI to look at people's brain activity as they view colors, which is an interesting idea.

One of the questions asked at the beginning of the article is “Why are there so many different color spaces?” he asks. “If this is really reflective of something fundamental about how we see and perceive, then shouldn’t there be one color space?”

There are multiple color spaces mostly because we generate, measure and communicate color spaces in different ways. Active displays (monitors) tend to use some kind of additive color space (e.g. RGB) where the default is "black" and we add various amounts of selective wavelengths to generate colors. Passive displays (printing) tend to use a subtractive space (e.g. CYMK) with a full spectrum (white) reflective surface (or backlight) and then selectively block certain frequencies. Paint colors have their own (proprietary) color spaces based what set of colorants and paint bases are available for mixing.

For transmission of images as data, we tend to use other color spaces (e.g. YCrCb) that actually do map to something like the average of human color sensitivity, as that provides less visual color distortion when quantized to fewer bits.

All these color spaces exist in parallel because they are optimized for different purposes.

And of course color perception varies from person to person. In addition to the various types and degrees of color blindness, a very small % of people appear to have super color sensitivity [1]. I also expect that an fMRI of someone looking at colors would give very different results depending on whether their languages support the blue-green distinction [2].

So I wish them luck, but I doubt there will ever be a "one true color space".

[1] http://www.digitaljournal.com/article/326976

[2] https://en.wikipedia.org/wiki/Blue%E2%80%93green_distinction...

There will never be 'one true color space', due to the variance in our eye makeup, but that does not mean we shouldn't try. This seems like a marginally better way of eliminating a 'middle man' to get to a truth. This procedure would have to be applied to many people to arrive at an average, which is the best these measures can do.

I read “eye makeup” very differently at first, and thought, sure, strange eye shadow could change the color of the light that ends up in ones eyes....

the article seems to skip over the entire field of colour science. there is a colour space for the human eye. it's called LMS. there is also a colour space, derived from it, called CIEXYZ. it's the colour space that all ICC profiles are specified in. they are RGB-like in nature and correspond to the three types of cone cells in the human eye. there are other colour spaces that derive from it, such as CIELab and CIELuv, which are luma-chroma versions of XYZ.

Indeed, I discovered from a picture of a heterochromatic cat that my blue-green perception is different in each of my eyes.

In other words, they're trying to answer the question of whether you and I are perceiving the same thing in our brains when we look at something that we both agree is "red"?

Sort of, but it depends on what you mean by "perceive". They are measuring responses in the visual cortex. What you feel about that is likely also similar to most people but I bet it's another layer in the brain doing that connection.

They are obviously not. Your representation of red is based on your experiences and mine is based on my experiences.

Experience but also biology. Blood is red and some people faint at the sight of blood, that can't be all attributed to experience.

Mabye it is a trait that we are predisposed to learn. But I doubt it is part of the default neural network as it is encoded by the genes.

Many of us are colour blind which confirms we're definitely not all perceiving the same thing...

Reply on news.ycombinator.com