A Rant About Colour

November 25, 2013  •  4 Comments

Today, in a thread on the GetDPI forum, I posted the following rant. It relates to an image posted by a forum member who, having posted some images from her new Sony A7r, had been told by some people that they could see colour shading in the sky and by others that they either could not see it at all or that it was normal or natural.

"Since we're all having our 2 Cents, here are mine:

Cindy's file does indeed seem to have a tiny, tiny, tiny magenta issue. It is absolutely within 'normal' variance for almost any system.

By 'system' I mean 'soup to nuts' - from lens to sensor thru RAW pp to display thru to eye and brain.

As other people have noted, normal vignetting can create small colour shifts, as can the natural variance of colour in a sky, and the uneven distribution of IR and UV across a sky, especially when taken into account with the amount of filtering on lens (screw-on and coating) and on sensor.

Then there is the fact that even very good monitors can have uneven display colour and brightness. I, like Jono Slack, have rather slacked off the frequency of calibrating my monitors because even when freshly calibrated, my 30" Cinema Display, when viewed next to my Eizo 30", gives the game away: the Apple display is blue and the Eizo is red, relative to each other. Viewed alone, each looks very good.

Next is the issue of the normal variance of human colour vision. Naturally the medical profession has decided to categorise and pathologise this, preferring to see it in terms of Norms and Abnorms rather than variance, though countless studies have shown that 'colour blind' people can determine patterns that 'normal' people cannot, and can often distinguish colour variations that 'normal' people cannot. Variations in abilities are what sustain populations, let's leave it at that.

Next, the fact that we perceive colours culturally and with gender bias. We are all familiar with the complaint that asian manufacturers favour certain skin tone renditions that look odd to western eyes. We are also aware that females, (and experienced male photographers!) are more likely to be able to use accurate language to distinguish between magenta and pink, or purple, mauve and violet. Using accurate language places a requirement on the speaker to perceive more diligently: my colour vision has become a lot more accurate since I started trying to see accurately, rather than seeing like a British boy raised in the 60's when knowing the names of anything other than red green and blue would cause a large pink question mark to appear over your head...

In every step of the imaging chain, in every situation, there is room for variance from what is 'true', 'accurate' and 'normal'. Your file is very comfortably within the sum of those variances IMHO. And when something is clearly wrong it is clearly wrong - such as the colour casts in this white frame from an 18mm Super Elmar M on an A7R - it will be obvious:


Additionally you will not, ever, accurately bottom out (let alone be able to eradicate accurately in post) a colour cast with a shot of a blank grey sky or a white wall. The light falling on a wall will be almost impossible to get perfectly illuminated and, unless lit by very accurately temperature controlled flash, will be subject to the same variances as a sky. To do this properly (and even then not 100% reliably or accurately) you need a Lens Cast Calibration sheet as used by MF shooters. They cost a few dollars and save a lot of time and yet very few people bother to source them. If you use one, you will see that, for example, a file from a Sony RX-1 with in camera colour shading corrections turned on still has some colour shading. And that is a camera with an extremely closely tuned ecosystem of sensor, lens and bespoke processing.

Most work does not require perfect colour accuracy. Photographers who work in fields where it does, do not use zooms with adaptors on what are fundamentally consumer cameras. 

Inaccurate colour renditions can be part of a photographer's style. 

There are highly successful artists, in both photographic and other media, with highly personal colour vision - and yet their work can be enjoyed, like the work of a deaf Beethoven, by a wider audience. 

Rant over. The file is fine."


Comments

Tim Ashley Photography
I'm sure that's true Tim - and I also am certain that human colour vision is far more variable and far more culturally conditioned than is widely accepted.I own a print I bought at a photo gallery that has a really cyan sky and it bugs the hell out of me but no one else ever notices until I point it out and even then, some people can't see that it looks unnatural.

Where this becomes bothersome is when I make an edition of 25 prints and sell the first few, but keep the rest of the series unprinted ready for 'on demand' production. Worst case scenario, the 25th sells years after the 1st and by then I am on a new printer. Best case, I am on the same ink set and paper batch… I did have an HP printer for a while that made its own profiles for every new paper batch and ink set but I didn't find the results more consistent than just using my Canon iP6300 with the best profile I can find.

Then there's paper that I just can't get right so give up on: Kodak metallic seems determined to give me cyan skies!
Tim Parkin(non-registered)
I've heard repeatedly in the print industry and once in the screen industry that people want to avoid a green cast at all costs and because random variability will give green occasionally they tend to err on the side of magenta where possible. Also the standard for colour accuracy is deltaE (deltaE 2000 if you're a geek) where an error of 1 is visible by the naked eye. Most litho printers work to an error margin of 3. However even if you have a monitor that is calibrated to an error of under 1 and another monitor that is also calibrated to an error of 1 you will still possibly have an error of 2 between them - making the difference visible.
Anna T(non-registered)
I have been following the color debate over at GetDpi and found that the magenta colorshift was really minimal.

However as the owner of an iPad2 (non Retina display) and very recently of an iPad Air, well I was really deceived by the Retina display. There is a huge difference between both. The iPad 2 had perfect colors. My iPad Air has a strong yellow cast. Greens are never green, the pictures look muddy. The skies are turning to magenta (and that said from someone who prefer when skies aren't too cyan). A saturated blue can turn to purple. And reds are just not showing up.
To the point that I will try to get it replaced by Apple.

Showing pictures on an iPad is a great way to promote your pictures and you can always have it with you.
Edward(non-registered)
I'm having a similar rant about color gamut vs calibration issues on the new iPads. Non photographers complaining about minor color shift issues or only covering 67% of argh in the case of the mini. However I have shown using spyder gallery identical results with high color images, when using spyder gallery. Even if you're a pro to think a new ipad mini retina isn't good enough to show a portfolio. Well if clients are that critical break out a print portfolio.

Also light bleed on devices at 100% brightness in black rooms. Who's viewing a digital device like that?
No comments posted.
Loading...
Subscribe
RSS
Archive
January February March April May June July (24) August (11) September (8) October (4) November (5) December (4)
January February March (2) April May June July August September October November December
January February March April May June July August September October November December
January February March April May June July August September October November December
January February March April May June July August September October November December