What is Camera Color Science really?

roxics

Veteran
I've heard this term Color Science thrown around a lot these days. It's like the new Dynamic Range or Shallow DOF or whatever new technical term people learn and then becomes all the rage. But what is it really? We're not all engineers.

My assumption based on conversations I've read is that it's the particular tuning a camera manufacturer does to the way that sensor renders color. Yet it seems to be something they can carry over to various cameras, sometimes with sensors made by different manufactures. So it must be adaptable if every manufacturer has its own fairly uniform color science rather than every camera having its own. Yet it also seems to be something that happens early in the chain to the point where even raw files have that color science. Which means it must be a deeper thing than a LUT which seems to go on top of color science.

Speaking of LUTs, I'm under the impression that picture profiles cameras come with (neutral, vivid, etc.) are basically just built-in LUTs in everything but name. Am I wrong about this?

Why are there not multiple color science packages offered by the same manufacturer to change the way their cameras render color? Sort of like a digital filmstock that goes to the very bones of the camera to change the look? Why are we stuck with one color science per manufacturer? Maybe open it up to the public to sell/trade different color science packages the way they do LUTs. Why is that not a thing?
 

Fascinating article. Thank you.

So basically in a nutshell the color science is derived from the filter purity over the photosites (how close to primary they can get with the dyes used) in combination with the image processing... that involves a lot of math.

I guess my remaining question would be, for a raw file there is still some processing going on then? Because even raw files seem to have a particular color science to them.
 
BTW what camera was he talking about when he references the video of the two guys in the coffee shop? I remember seeing that video, I just don't remember what camera it was.
 
Fascinating article. Thank you.

So basically in a nutshell the color science is derived from the filter purity over the photosites (how close to primary they can get with the dyes used) in combination with the image processing... that involves a lot of math.

I guess my remaining question would be, for a raw file there is still some processing going on then? Because even raw files seem to have a particular color science to them.

There is more than just 'color' purity, it is which band in the color spectrum to acquire R, G, and B, and how 'good' the cutoffs are (for example blue filters not having a good cutoff for IR, leading IR 'poisoning' of the blue that is captured.

Then there is the processing of the 'raw' RGB captured to produce the resulting image and such processing as to compensate for 4:2:0/4:2:2 type of sensors, along with any non-linearities that the process may have...
 
Fascinating article. Thank you.

So basically in a nutshell the color science is derived from the filter purity over the photosites (how close to primary they can get with the dyes used) in combination with the image processing... that involves a lot of math.

I guess my remaining question would be, for a raw file there is still some processing going on then? Because even raw files seem to have a particular color science to them.
Raw varies too. Off other discussions that were had here on DVXU, it seems that the difference in filters and processing power is by "weight class" - i.e., the higher end cameras have mo' better ones.

What's harder to peg is how the technology has filtered down over the years or where it's actually been introduced on the lower end models. For example, when F65 was introduced about five years ago, it was the only oversampled 4K (17.6 MPX) sensor out there. Its colors are still considered superb (to wit, "Cafe Society"). Alexa's colors, also deemed fantastic, came off another oversampled sensor, this one in 2K/1080p.

However, since 2011-12, many consumer cameras have began to use this technology. Even a $1,000 A6300 is using 6K to deliver 4K and, with external recording, has fantastically vivid saturated colors. This makes for a weird disconnect where the low end cameras have the required oversampling but have a plethora of other problems, while Sony's and ARRI's current high end lineups are getting very, very, very long in the tooth. Which means that the next generation of products that may be unveiled at the IBC or whereabouts will probably make all the talk of color science moot.
 
Well, for me it is an over simplistic explanation of color science back in my camcorder dv tape days. Sony cams gave a cool bluish tint out of the box and Canon gave a warmer look more pleasing as far as skin tones are concerned. Today it is much more sophisticated.
 
Once an individual has experience using a variety of consumer, prosumer, and professional SLRs, mirrorless, P&S/fixed-lens (small and large sensor), and various cinema systems - inexpensive and expensive - from all of the different companies out there, one will understand color science much better even if it's in their own way.
 
Back
Top