M8 -- colour realism

[blakley]

Advertisement (gone after registration)

It's faintly embarrassing to even bring this up, since I presume everyone here already knows it, but black is of course not even a color, so saying "black is black" is sort of true and false at the same time. An object which reflects 650nm light (only) at an intensity 0.01% of the intensity of the surrounding objects will "be" red but "look" black. An object which reflects the same intensity of 550 nm light will "be" green but "look" black.

 

The main problem with the M8 is not just that it's recording the wrong color - it's that it's recording the wrong intensity (i.e. it's too sensitive in the reds, because of extended IR sensitivity) so things which are visually dark are pictorially lighter, and they acquire color.

 

So "black is black" is true for some objects - those which genuinely reflect NO wavelengths in the visual spectrum - but there are a lot of objects which are NOT black; they're another color but just really, really dark.

 

"Dark is dark" is true as long as we restrict ourselves to a specific spectrum. But objects which are "not black but dark" and "dark only in the visual spectrum but not in IR" look different to the M8 than they do to us.

[lct]

Theory and practice.

In the real world black is black when it looks so to anybody.

[blakley]

In the real world black is black when it looks so to anybody.

 

And what does that mean? Do you have some master list of "things that look black to anybody", or is your statement just a tautology? The whole point of the intial post was to say that things look different to different people in the real world. Everything looked black to Ray Charles.

[Peter Branch]

You may be interested in an amusing story.

 

Sometime in the ‘60s the Colour Group of the Kodak Research laboratories at Harrow in the UK were hot on the trail of “Perfect Colour Reproduction” and decided they needed a new “Test Chart”. All the usual issues arose. Fresh magnesium oxide for the white carefully prepared by burning magnesium in pure oxygen to prevent the very slight green colour associated with burning it in an atmosphere containing nitrogen etc. But the real problem was the black. It was decided that the only way was to build a “light trap” on the back of the test chart. This consisted of a series of holes and baffles all lined with the blackest velvet that money could buy. No light that entered the trap ever escaped; it was a “perfect black”.

 

The chart was photographed and the resulting images measured – but then the problem arose - where do you plot “Zero” on a logarithmic scale? There is no answer, you have in the conventions used to plot colour, to go to minus infinity on the “x axis”. So the data point for “Black” obtained only with significant time and effort was unusable! It was back to a black painted patch with a small but known reflectivity that could be plotted. Any such patch has some reflection and is usually coloured if viewed in the right conditions because the objective is to get the least amount of light reflected not to ensure that what is reflected has a neutral spectrum.

[lct]

OK i give up you win folks.

Black is not black and magenta is a kind of black not-so-black black.

[carstenw]

I have it on good authority that Ray Charles saw everything in magenta.

[Peter41951]

Advertisement (gone after registration)

OK i give up you win folks.

Black is not black and magenta is a kind of black not-so-black black.

 

......or green.

[erl]

I would throw into the ring (just for fun?) that if you can see it, it is neither black nor white. Black is only black when there is total absence of colour. Similarly, white is only white when there is total colour.

 

Now, I am going to duck off to work. You guys slow my production rate dreadfully. LOL.

 

Cheers,

Erl

[jaapv]

Yes i know this Grant there is no perfect black and no perfect white either.

But i don't refer to perfection here.

In the real life i see the difference between black and magenta and i expect that my cams show me this difference as well.

My Epson can do this, my Nikons as well.

I would like that my favourite brand is able to do it as well.

 

Actually, your Epson can not. It has the same magenta shift, only slightly less than the M8

[TimF]

Correction....Los Bravos.....NOT Boney M!

...and in the sixties too Andy! (or do you mean La Belle Epoque in the seventies, wrote a music bore)

[andym911]

sixties...my wife corrected me!

[lars_bergquist]

As far as dyes and inks are concerned I was led to believe that "black" is either very very dark brown or green or red. Is this true?

 

That varies. A black cat, or black 'panther' (leopard) is actually very dark brown. But somewhat like the petunia case, many 'black' dyes are really composite, so that the darnedest things may happen when you filter them.

 

This applies to visual colour filters only. The IR cutting filters that Leica wants us to experiment with cut off invisible radiation only. And because these filters really cannot be used on real wide angle lenses, I iterate my 'Carthago est delendam': Leica MUST improve the IR protection of the sensor (plus cure streaking and blobbing and erratic AWB) and spiffy up the firmware before I plonk down my hard-earned 42 000 SEK. That's final.

 

The old man who heard all the excuses half a century ago

[pklein]

For years, Kodak thought that red was orange. Here's a flash picture of my parents, the night they announced their engagement. It was taken in 1951 on contemporary Kodachrome, ASA 10 (that's *ten*), probably with a blue daylight to tungsten filter (81B?)

 

http://users.2alpha.com/~pklein/oldpics/EmMiltSofa51.jpg

 

And here's my aunt on the same film around the same time. In sunlight, the effect is less pronounced, but it's still there--bright red lipstick and a red tartan both have a distinctly orange tinge.

 

19Glady

 

I know these people personally and I can assure you that their lips were red, not orange, and that they did not magically turn red by the time I was born and grew old enough to notice.

 

Yes, the M8 sees infrared as purple or (pardon the expression) magenta. But this doesn't mean that color films are realistic. They all have a personality. If you have any doubt, try taking someone's portrait with Velvia.

 

This is why we have some films that render flesh tones as accurately as possible, and others that render a scene not like it was, but how you think you remember it was.

 

--Peter

[Peter Branch]

For years, Kodak thought that red was orange. Here's a flash picture of my parents, the night they announced their engagement. It was taken in 1951 on contemporary Kodachrome, ASA 10 (that's *ten*), probably with a blue daylight to tungsten filter .............(81B?)

 

There were actually many fewer degrees of freedom to "manage" the colour rendering back in those days and this is especially true when compared with the things that can be done routinely in the digital environment.

 

However I can assure you that great care was taken to ensure that IR, which is insidious, did not unduly affect the results. Where IR sensitivity was retained for particular reasons, e.g. Tech Pan, the information provided made this very clear and advised the use of - guess what - filters!

 

Given just three dyes to reproduce the full colour spectrum only certain colours, or more strictly regions of colour, can be reproduced with critical fidelity. Hence films specially made to get flesh tones as accurate as possible which inevitably have compromises elsewhere.