M Color

[Guest redge]

Advertisement (gone after registration)

Still waiting for RickLeica and the people who agree with him to post a single photograph to support his argument. He makes his point, if one can call it that, not only in relation to the RX1 but also explicitly, by extrapolation, to the A7r, which he apparently owned for a grand total of two days.

[Rick]

Sorry, but your analogy breaks down completely because (a) different sensors have different sensitivities, so it's not at all similar to using the same film in different cameras... and ( even raw captures are processed in-camera before they're put on file, which is why raw files from different cameras with the same sensor supplier still look different.

 

You miss the point soooo completely. Who cares about the details of Adan's analogy. The point is that you have to profile any camera. And watch for the effects of blue adding to red and getting magenta under tungsten light. Yellow has to be corrected if, you want to get close to correct color... or just leave everything alone and let the image have that nice warm yellow tint.

 

In the case of the M240 and LR profile, the red is over saturated and the primary must be increased in Hue if, you ever want correct color. That is just a start. Chris, tries to show what I guess he thinks are examples of good color and the faithful just push the thanks button. But, the artificial lighting he continually posts as good color are confusing for anyone trying to learn color correction.

 

Adan provided the best profile for my M9 and I used it for years and tweaked the color from there and then learned how to make my own profiles from there. I'd sure like it if he had a M240 so he could give some general profile information. I know he could help me get even closer.

 

Thanks for your post Adan.

[MirekE]

Yes, Leica and Adobe "provide" a single profile (either Embedded from Leica, or "Adobe Standard" from Adobe. But if you think you can trust the Leica or Adobe engineers, consider this:

 

They give you a single profile.

 

Here are just a handful of the possible spectra of light sources under which you might take pictures: http://www.lamptech.co.uk/Images/Ill...0SPD's.jpg

 

Now - does anyone in their right mind really think any single profile will work across all those spectra?

 

So - at BEST - all they are giving you is a generic starting point for "approximately correct" color under generic (not necessarily "daylight") lighting. If you want really good color (and this has been true with every single camera I've used for raw shooting - Nikon, Canon, Leica, Sony....) - you have to make your OWN profile(s).

 

These profiles are dual illuminant and should suffice for natural light sources with continuous spectrum. If you make your own, they aren't any different.

[Rick]

Still waiting for RickLeica and the people who agree with him to post a single photograph to support his argument. He makes not only in relation to the RX1 but also explicitly, by extrapolation, to the A7r, which he apparently owned for a grand total of two days.

 

You are simply wrong. I owned the A7R for almost 2 weeks. I clearly stated that I did not have enough experience with the A7R to extrapolate my conclusions to this camera. Please, don't put words in my mouth, and the wrong ones at that.

[Rick]

These profiles are dual illuminant and should suffice for natural light sources with continuous spectrum. If you make your own, they aren't any different.

 

That may be true, but it doesn't mean they are going to be right either.

[MirekE]

A good custom calibration profile for yellow light (tungsten and its replacements) will move the red-hue slider much further towards yellow than a "daylight" profile will.

 

e.g. for my M9 profile(s) in Camera Raw, the red hue calibration slider is set to:

 

+10 (more yellow-ish reds) in my daylight profile

+25 (even MORE yellow-ish reds) in my tungsten/yellow light profile

 

I don't think camera profiles should be used to compensate global tint caused by color temperature of the illuminant. This is what white balance controls are for.

 

In any case, I am still missing the point - if the subject is lit by tungsten light, every single pixel will be too warm, including the red ones. So the reds need to be corrected towards cold colors as much as whites. So you are starting with reds that are too orange and I don't understand why they should end up too magenta after white balancing.

[MirekE]

Advertisement (gone after registration)

That may be true, but it doesn't mean they are going to be right either.

 

I am just saying that these profiles cover more than suggested in the post I was responding to.

[hoppyman]

I really don't understand how anyone can judge any output (from any of the cameras) like this when there are so many variables from capture through developing to viewing. Each is entitled to their own preferences and views of course.

Here are a couple from me. 3 done with high end studio lights and last in natural light. I'll split across 2 posts.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

[hoppyman]

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

[Rick]

In any case, I am still missing the point - if the subject is lit by tungsten light, every single pixel will be too warm, including the red ones. So the reds need to be corrected towards cold colors as much as whites. So you are starting with reds that are too orange and I don't understand why they should end up too magenta after white balancing.

 

No, the red primary needs to be increased towards red-yellow. It removes red and allows you to add global blue in the WB Color Temp. Which gives less magenta. Which the M240 and the M9 are contaminated with.

 

My initial point of the RX1 was that it is not overly sensitive to red. So, it starts out not having this issue. Don't read into this last statement anyone... I still like the M color. It just needs more work than the RX1. That's all. The sooner you realize this the better your subjects are going to feel about their faces.

[Guest redge]

You are simply wrong. I owned the A7R for almost 2 weeks. I clearly stated that I did not have enough experience with the A7R to extrapolate my conclusions to this camera. Please, don't put words in my mouth, and the wrong ones at that.

 

Wow. Where's the post in which you said that you sold the camera to your brother after a couple of days? Suddenly not here. Presumably deleted.

 

In your original post you did say that you didn't have a lot of experience with the A7R, but you were plenty eager to prognosticate consequences for Leica.

 

Meanwhile, still waiting to see images to support what you are saying.

[jaapv]

Still waiting for RickLeica and the people who agree with him to post a single photograph to support his argument. He makes his point, if one can call it that, not only in relation to the RX1 but also explicitly, by extrapolation, to the A7r, which he apparently owned for a grand total of two days.

Quite irrelevant. Color is subjective. If Rick has tweaked his colour output to his satisfaction, as quite a number of respected photographers here apparently have, that should give pause for thought. None of them needs to prove themselves because their preferred colour output may not be your preferred colour output. Nevertheless, there are plenty of shots in this thread to make the point. As for the A7r I have seen about the same proportion of images that appeal to me and images that I consider duds on the internet as f I have from any other high and medium level camera of any brand.

For instance, Geof has posted four very good examples, two of which I would been extremely proud to have made -probably would not have had the ability to- and two of which I would have processed slightly differently -which would most likely not be better, just different. Somebody else might see this the other way around.

[Guest redge]

Quite irrelevant. Color is subjective. If Rick has tweaked his colour output to his satisfaction, as quite a number of respected photographers here apparently have, that should give pause for thought. None of them needs to prove themselves because their preferred colour output may not be your preferred colour output. Nevertheless, there are plenty of shots in this thread to make the point. As for the A7r I have seen about the same proportion of images that appeal to me and images that I consider duds on the internet as from any other camera. As I have from any other high and medium level camera of any brand.

 

I agree that it's irrelevant. That is my point, as it is with several other posts. Pity that he claims an objective standard, which mostly suggests lack of experience with analogue, and for that matter digital, printing.

[Rick]

Wow. Where's the post in which you said that you sold the camera to your brother after a couple of days? Suddenly not here. Presumably deleted.

 

In your original post you did say that you didn't have a lot of experience with the A7R, but you were plenty eager to prognosticate consequences for Leica.

 

Meanwhile, still waiting to see images to support what you are saying.

 

You are simply wrong again. I never gave the camera to my brother and a did not delete the post because, I said nothing of the kind. You have presumed wrong again. This is quite a deep hole you are digging here.

 

Would you like me to prognosticate how long you will have to wait before I post anything here that might help you understand color? I'll give you a clue, it has to do with snowballs.

[adan]

I don't think camera profiles should be used to compensate global tint caused by color temperature of the illuminant. This is what white balance controls are for.

 

Agreed. Profiles are not for setting global white balance. In fact , a profile CANNOT be used for white balance - it is simply the wrong tool. All the profile controls do is adjust the RELATIVE appearance of colors (reds can be adjusted to be more yellow WITHOUT adding yellow to greens).

 

However, a separate white balance profile is needed for CORRECTING non-global distortions to specific colors, caused by the act of white-balancing itself.

 

if the subject is lit by tungsten light, every single pixel will be too warm, including the red ones. So the reds need to be corrected towards cold colors as much as whites.

 

Not Correct.

 

Red and yellow objects do not reflect blue light (it is why they appear red or yellow). Use a light source that does not contain blue light (i.e. yellow light) - and such objects will still reflect essentially the same energies and wavelengths as if it were white light. The missing blue light does not affect their color, since they would not reflect it even if it was present. If it were present, they would absorb it (and appear red or yellow) and if it is not present, nothing happens (and they appear red or yellow).

 

So they will not be "too warm", for the most part. And where they are too warm, the extra warmth is much less - proportionally - than the extra warmth that appears in white, neutral or blue-through-cyan objects.

 

In other words, "reds need to be corrected towards cold colors as much as whites" is exactly what is NOT true. They need LESS correction (but that is not what happens, thus we end up with magenta reds).

 

It would be nice if there existed "intelligent white balance" that could independently and proportionally correct reds and blues and yellows and whites under non-daylight spectrums - but the computer would have to guess as to what the true light spectrum was, and what the real subject colors were. So instead we must use a two-step process - a global addition of whatever color is in short supply (usually blue) through the white-balance - and a well-built profile for that specific spectrum, to compensate for the mistakes the global correction introduces.

 

Imagine a red object of color values, let us say, 240 red, 100 green, 35 blue.

 

Shine yellow light on it. We'll use tungsten light, which covers the spectrum in proportions of about 100 red units to 65 green units to 13 blue units.

 

Under such light, that red object's color values will become 240 red, 65 green, 5 blue (native color x the amount of each color available from the light source). The tungsten light contains just as much red as white light does, and so that value does not change. It contains very little blue light, and so that value DOES change, a lot. Green falls in between.

 

(Keep in mind that yellow is a SUBTRACTIVE primary color - yellow light, or a yellow filter, CAN NOT ADD red, yellow or green to what is already there (or not there). It can only subtract blue, and if impure yellow, like tungsten light, a bit of green.)

 

Imagine next to the red object a gray object with values 180 red, 180 green , and 180 blue. Shine the same tungsten light on it, and it will register as 180 red, 108 green and 24 blue.

 

Click on that gray with the WB eyedropper, and it will add blue and some green to restore the neutral balance - 156 points of blue and 72 points of green

 

The gray is now restored to 180 red, 180 green, 180 blue.

 

The red object is "restored" to 240 red, 137 green, and 180 blue.

 

Oops, it has been shifted to magenta by the act of correcting the global white balance. Previously there was more green than blue (good caucasian skin proportions, as it happens) and now there is more blue than green - a magenta-red. It is also a bit desaturated - too much blue and green together wash out the red a bit.

 

This is exactly what can be demonstrated by photographing a Colorchecker under yellow light, and only fixing the white balance. The red will be skewed magenta due to the white balance addition of excess blue, and a bit pale or de-saturated.

 

Creating a profile for use with yellow light pictures generally involves shifting the red primary hue towards yellow to counteract that "extra" blue added by WB. As I noted in my previous post, my correction to reds under such a profile amounts to 250% of that used for daylight pictures. The units are dimensionless and don't directly correlate to the standard 8-bit 0-255 scale, but in general terms mean that the WB correction applied above, WITH MY PROFILE or a similar one, would reproduce the red object at something like 240 red, 120 green, and 60 blue. Additional resaturation of red would knock down both green and blue to arrive at 240 red, 100 green, 40 blue or something close to that. A better match to the original colors than simply doing a white balance with a daylight profile.

[anupmc]

You miss the point soooo completely. Who cares about the details of Adan's analogy. ...

 

I didn't miss any point at all. Adan cared enough to make that analogy, I simply pointed out the mistakes in it, period.

[lct]

Interesting thoughts and pics folks, thanks for sharing.

[Leonil]

Interesting images - a lot of the "incorrect" skin tones look fine to me. A lot of the colours are on par with other cameras getting accurate colour in the boxing images posted.

[adan]

I didn't miss any point at all. Adan cared enough to make that analogy, I simply pointed out the mistakes in it, period.

 

I'll accept the point that different digital cameras amount to different "films" - due to the variations in sensors or on-camera raw processing.

 

If you think that invalidates the whole analogy - so be it. I don't. Most cameras can produce very good color - IF the images are properly profiled and processed. The key word is "properly" - which in my experience means avoiding what the geeks automate for you, and doing it yourself.

 

My 6D needs different profiles than my M9 - but I get the same colors out of each, eventually.

 

To borrow from Ansel Adams - "The raw file is the score - the profile is the performance."

[Advertisement]

Hallo,

Look here: Leica M

[hoppyman]

Quite irrelevant. Color is subjective....snip

.... For instance, Geof has posted four very good examples, two of which I would been extremely proud to have made -probably would not have had the ability to- and two of which I would have processed slightly differently -which would most likely not be better, just different. Somebody else might see this the other way around.

 

Hi Jaap, and to expand on the samples that I posted, each woman has different skin tones, one of middle European ethnicity here then add tan/bronzer, one from Italian stock,-again add bronzer, one Caucasian and one Nordic. How on earth can anyone say this or that skin tone is 'correct'? For all of these I used an GretagMacbeth target and a WhiBal card (which give quite different WB samples although in theory they ought to be similar) Then add the light source, the sensor characteristics, Leica's interpretation of what is pleasing, where the specific image colour temperature and conditions fits with the dual illuminant profile process that Adobe uses, Adobe's interpretation of what pleasing plus the complex interaction of the hue twists in the DNG profiles, plus whatever done in developing then whatever the web browser interprets the versions uploaded and of course the individual screens and their output for however people are looking.

 

 

You can say fairly that you prefer this or that as produced by this specific conditions sure. That is all.