I would like to know wich camera system offer that nice color

[Jim Kasson]

I was thinking about this passage from Doug Kerr's article:

Gee, I hate to disagree with Mr Kerr, but I do.

[added later: Turns out I don't actually; Kerr constrained the solution space to cameras that use the sRGB primaries.]

What's wrong with this set of curves? (same picture as above)



As an aside, I have a little experience in trying to build a sorta-Luther-Ives camera.

When I was working at the IBM Almaden Research Laboratory in the early 90s as a color scientist, I consulted with Fred Mintzer and his group in Yorktown Heights who developed a scanning camera with the objective that the wavelength-by-wavelength product of the camera's RGB filters, the IR-blocking filter, and the CCD's spectral response would be close to a 3x3 matrix multiply away from human tristimulus response. The camera was used to digitize Andrew Wyeth's work, to capture artwork in the Vatican Library, and for some other projects where color fidelity was important.

Here is a a paper that has some system-level RGB responses. You'll note a high degree of spectral overlap of the red and green channels, just as there is overlap in the medium and long cone channels. You'll also note an absence of the short-wavelength bump in the red channel; this camera didn't do violet. Because the illumination was part of the camera system, the camera did not have to deal with illuminant metamerism the same way as a human.

Jim

[ErikKaffehr]

Hi Jim,

Thanks for making that clear. Another learning experience… I guess I owe Jerome an apology!

Best regards
Erik

Gee, I hate to disagree with Mr Kerr, but I do. What's wrong with this set of curves? (same picture as above)



As an aside, I have a little experience in trying to build a sorta-Luther-Ives camera.

When I was working at the IBM Almaden Research Laboratory in the early 90s as a color scientist, I consulted with Fred Mintzer and his group in Yorktown Heights who developed a scanning camera with the objective that the wavelength-by-wavelength product of the camera's RGB filters, the IR-blocking filter, and the CCD's spectral response would be close to a 3x3 matrix multiply away from human tristimulus response. The camera was used to digitize Andrew Wyeth's work, to capture artwork in the Vatican Library, and for some other projects where color fidelity was important.

Here is a a paper that has some system-level RGB responses. You'll note a high degree of spectral overlap of the red and green channels, just as there is overlap in the medium and long cone channels. You'll also note an absence of the short-wavelength bump in the red channel; this camera didn't do violet. Because the illumination was part of the camera system, the camera did not have to deal with illuminant metamerism the same way as a human.

Jim

[ErikKaffehr]

Nope, it doesn't make them better, it shows they are adequate.

Best regards
Erik

This straw man argument about contest winners that came out of nowhere is hilarious. What is the install base of canikon users vs MFDB? Of course you will see a disproportionately high number of canikon shooters participating in contests and subsequently winning them. Does that make them "better"?

[synn]

Nobody said they are inadequate. The choice of system for any particular assignment depends on a lot more factors than color.
Just pointing out that the awards argument is absolutely off point to the thread subject.

[Jim Kasson]

I was thinking about this passage from Doug Kerr's article:

Hold the phone, Erik. I just reread the Kerr passage you quoted. He added the additional constraint that the camera use the sRGB primaries.

• Had outputs would be the (linear) coordinates of the sRGB color space.

With that constraint, I do agree with him. You can't do it will all-positive spectral responses. But you never added that constraint, did you?

Jim

[Chris Livsey]

So, quite clearly some of the greatest photographers are happy with Canon and Nikon,

And some of them are paid to shoot with them 

[Jim Kasson]

Just pointing out that the awards argument is absolutely off point to the thread subject.

Yes. And I think the color accuracy (Luther-Ives) argument is a bit off-point, too, although as a reformed color scientist, it pains me to say so.

I don't think photographers desire accurate color in their images. I think they desire pleasing color. One of the reasons I think that is that many photographers cite the characteristics of certain films as examples of what they wish they could do with digital.

I once took a tutorial at an SPIE conference from Ed Giorgianni, the Kodak scientist, on color management in chemical (and a little digital) photography. Ed walked us through all the constraints that film and paper designers had to live with, and I came out amazed that the whole thing worked at all, much less that the results looked good.

I once measured the color accuracy of 20 different films using the Munsell chart as a target. They we all pretty awful by the standards of today's digital cameras. But there was clearly a method in the "errors". I put the word in quotes, because I think most all of the films were performing as designed. You could see that vivid color was important to the Velvia designers, to cite the most obvious example. Skin tones often were biased in the yellow/red direction (never in the green direction, obviously), and were slightly more chromatic than the dictates of accuracy would require.

Jim

[ErikKaffehr]

Hi Jim,

I was considering that constraint but I was not aware of the significance. As I said, it is a learning experience, thanks for widening my understanding of colour.

Best regards
Erik

Hold the phone, Erik. I just reread the Kerr passage you quoted. He added the additional constraint that the camera use the sRGB primaries.

• Had outputs would be the (linear) coordinates of the sRGB color space.

With that constraint, I do agree with him. You can't do it will all-positive spectral responses. But you never added that constraint, did you?

Jim

[eronald]

In some ways this discussion of three color cameras is disingenuous because (I don't understand it) and because there is no reason not to try and do multispectral acquisition, especially when it comes to artwork ...

But I do agree with the fact that photographers desire pleasing color; this is why camera manufacturers are proud of their Jpeg renderings ...

Edmund

Yes. And I think the color accuracy (Luther-Ives) argument is a bit off-point, too, although as a reformed color scientist, it pains me to say so.

I don't think photographers desire accurate color in their images. I think they desire pleasing color. One of the reasons I think that is that many photographers cite the characteristics of certain films as examples of what they wish they could do with digital.

I once took a tutorial at an SPIE conference from Ed Giorgianni, the Kodak scientist, on color management in chemical (and a little digital) photography. Ed walked us through all the constraints that film and paper designers had to live with, and I came out amazed that the whole thing worked at all, much less that the results looked good.

I once measured the color accuracy of 20 different films using the Munsell chart as a target. They we all pretty awful by the standards of today's digital cameras. But there was clearly a method in the "errors". I put the word in quotes, because I think most all of the films were performing as designed. You could see that vivid color was important to the Velvia designers, to cite the most obvious example. Skin tones often were biased in the yellow/red direction (never in the green direction, obviously), and were slightly more chromatic than the dictates of accuracy would require.

Jim

[ErikKaffehr]

Hi,

I have just posted this on GETDPI, they have a discussion on CCD vs. CMOS. I don't know how relevant this is to the discussion, but it may illustrate some differences between cameras and profiles.

The samples here were taken from ImagingResource. This is one of their test charts containing a popular reference image, I guess from Kodak, would be nice to have a reference image, but I have not found any. The images are shot with Penatx 645D and Pentax 645Z, and they are a couple of years apart.

As a side note, I think the ColorChecker card in the ImagingResource sample may be quite old.

There are three pairs of images included.

• Adobe Standard
• A DCP profile I generated for each using Adobe DNG Profile editor using a ColorChecker in a different crop of the same sample
• The embedded profile in the DNG file.

The images are here, I would suggest checking them rather than the screen dump below!

A screen dump is shown below:

[jerome_m]

Yes. And I think the color accuracy (Luther-Ives) argument is a bit off-point, too, although as a reformed color scientist, it pains me to say so.

I don't think photographers desire accurate color in their images. I think they desire pleasing color. One of the reasons I think that is that many photographers cite the characteristics of certain films as examples of what they wish they could do with digital.

This is actually the argument I wanted to bring to this discussion by citing Luther-Ives conditions. Participants in this thread are discussing the superiority (or not) of MF cameras by trying to determine which camera is the most "accurate". The idea of "accurate" is implicit in the discussion (see the part about using color checker charts, for example).

But this is not how camera works. All digital cameras have limited color "accuracy" and the amount and direction of deviation is carefully chosen by the manufacturer (mainly by selecting the filter array pigments). The deviation is chosen to give "pleasing" colours. The choices are different between general usage cameras (which, as a rule, must be able to work under lights of varied types) and for cameras mainly used in studios (which mainly use flash).

[ErikKaffehr]

Hi,

I may add some observations from a test I made about a year ago. What I did was to photograph one of Wolfgang Faust's IT8 targets under studio light, and used Argyll CMS and Patchtool to compare the converted raw files with measured reference values of the IT8 target. The IT8 target has 288 patches and covers a decent part of Adobe RBG volume (see attachement).

The article is here

What I have seen was the reproduction was remarkably good, just looking at Delta E values (CIEDE 2000, AFAIK).

The images below compare the renditions of the IT8 patches (P45, with C1 and linear profile to the left Sony Alpha 99 SLT to the right)


Some interesting observations:

• The P45+/C1 "linear"  was more accurate than the "film curve"
• The delta E values used take also L into account. The gradation curves used affect the L-coordinate.
• LR 5 conversion on the SLT99 was more accurate than C1 conversion
• Greatest devition on the P45+ was on the deep blue fields, around DE=17. Note that the difference is not visible in the patchtool generate images. The reason is that the colours are pushed outside Adobe RGB. This may be an indication that tuning of those profiles is more visually made than measured.
• The outdoor daylight and studio flash profiles seem to be identical, Delta E =0
• The Portrait profile deviates very little from the other two profiles

So my take is really that both tested cameras were quite accurate. I would also add that it may make more sense to strive for an accurate rendition to an internal colour space and apply pleasantness in the output profile.

The result of this test was for me:

• For the Sony, I decided to stay with LR5 using Adobe Standard Profile, as it was almost as accurate as my own.
• On Phase One I found that LR5 with my own profile was quite satisfactory. So I decided on using that (C1 doesn't fit my workflow).
• But, I found both conversions remarkably accurate

There may be some areas that this test doesn't cover. Both skin and chlorophyl are said to have tricky spectral reflections, and I would suggest infrared filtering plays a role for skin.

Best regards
Erik

Yes. And I think the color accuracy (Luther-Ives) argument is a bit off-point, too, although as a reformed color scientist, it pains me to say so.

I don't think photographers desire accurate color in their images. I think they desire pleasing color. One of the reasons I think that is that many photographers cite the characteristics of certain films as examples of what they wish they could do with digital.

I once took a tutorial at an SPIE conference from Ed Giorgianni, the Kodak scientist, on color management in chemical (and a little digital) photography. Ed walked us through all the constraints that film and paper designers had to live with, and I came out amazed that the whole thing worked at all, much less that the results looked good.

I once measured the color accuracy of 20 different films using the Munsell chart as a target. They we all pretty awful by the standards of today's digital cameras. But there was clearly a method in the "errors". I put the word in quotes, because I think most all of the films were performing as designed. You could see that vivid color was important to the Velvia designers, to cite the most obvious example. Skin tones often were biased in the yellow/red direction (never in the green direction, obviously), and were slightly more chromatic than the dictates of accuracy would require.

Jim

[synn]

Erik, I remember you vicariously reminding everyone that "your threads are tech oriented and photography related discussions can happen in many other threads" every time someone asks about the relevance of such threads.

Let this thread be a reminder that unfortunately, you don't let photography oriented threads to be and will derail the discussion into more test charts and equations and other things working photographers have only a passing interest in.

The awards strawman is a new one though. I'll give you that!

[douglevy]

It's hard for me to quantify, but I finally had a chance to play with some 645Z files yesterday, shot by a good friend. I say it's hard to quantify because there are so many variables vs. how I shoot. He shot pentax/LR5 and I shoot Hasselblad/Credo 40/C1, but the Pentax files looked just like higher res Nikon files, the Hasselblad files look a TON better to me, crisper, sharper, better color. Again, totally subjective, but just my $.02.

-Doug

[Gigi]

….My take on the issue is that I would guess something like 60-85% percent of the leading photographers use Nikon or Canon.
If there is some collection of the worlds best images, probably something like 70% of them would be shot with a Canon and a lot of them would be with shot with the 16-35/2.8L.
….

Erik -

I'm sorry - but this is worthy of some comment….

First, there are so many "guesses, probably's and would be's" to make this statement almost indecipherable.
Then there is the intended logic behind this statement, suggesting (as best as can be determined) that a collection of "best images" may be determined by some statistical model, so there would be representation equal to market share by Nikon and Canon. This isn't right either. Weren't there some good shots taken in the 19th century?

Then Canon somehow gets 70% of those shots (not sure how that works). And some single zoom lens (I guess its the kit lens?) will have the lions' share of best images? Why? Aargh.

And finally "if there is some collection…"? Surely, there are many. Take a look at MOMA, Art Institute of Chicago, Magnum, or others that exist.  Each have a view point and reflect at least one (if not many) curatorial approaches. I would recommend studying them first…

Perhaps you were trying to say something else. As written, it projects a particular viewpoint that I hope you did not intend.

[ErikKaffehr]

Hi,

Something like this (clearly limited to news type photography):

Taken from here: http://hastalosmegapixeles.com/las-camaras-de-los-world-press-photo/

This link also gives some info, but again it is photojurnalism:
http://petapixel.com/2012/12/02/the-most-popular-cameras-and-settings-for-reuters-best-photos-of-the-year-2012/

Cameras:

Zoom lenses (note popularity of the 16-35):



Comments see below:


Best regards
Erik

Erik -

I'm sorry - but this is worthy of some comment….

First, there are so many "guesses, probably's and would be's" to make this statement almost indecipherable.
Then there is the intended logic behind this statement, suggesting (as best as can be determined) that a collection of "best images" may be determined by some statistical model, so there would be representation equal to market share by Nikon and Canon. This isn't right either. Weren't there some good shots taken in the 19th century? I should have stated that I mean award taking pictures today.

Then Canon somehow gets 70% of those shots (not sure how that works). And some single zoom lens (I guess its the kit lens?) will have the lions' share of best images? Why? Aargh. This was based on a list of award winning images I have seen a few years ago. An amazing number of pictures were shot with the 16-35/2.8L (that is not a kit lens, a recent price I have seen for the present version is 1699$US)

And finally "if there is some collection…"? Surely, there are many. Take a look at MOMA, Art Institute of Chicago, Magnum, or others that exist.  Each have a view point and reflect at least one (if not many) curatorial approaches. I would recommend studying them first…
I would say that in this thread we talk about modern equipment delivering a certain quality of colour

My point, is that harsh words are used on the colour rendition of Canon and Nikon, but these system still take a lot of the best images taken today. I can mention that I have no pony in that race. I never had a Canon, but I happened to have a Nikon FA some twenty years ago. Personally I am shooting Sony Alpha and Hasselblad V with a P45+.

I have not done a statistical analysis of award winning images, and equipment is not normally listed on first page, but I have the impression that Canon and Nikon dominates with professional photographers. But I must admit that I am much oriented in direction of landscape, wildlife and news photography.

[eronald]

It's hard for me to quantify, but I finally had a chance to play with some 645Z files yesterday, shot by a good friend. I say it's hard to quantify because there are so many variables vs. how I shoot. He shot pentax/LR5 and I shoot Hasselblad/Credo 40/C1, but the Pentax files looked just like higher res Nikon files, the Hasselblad files look a TON better to me, crisper, sharper, better color. Again, totally subjective, but just my $.02.

-Doug

I guess you have never compared a Canon processed with Lightroom to a Canon processed with Canon's own DPP ...

Edmund

[ErikKaffehr]

Hi Synn,

I am aware of that. That is the main reason I stopped the discussion on Luther-Ives, that was actually raised by some else. Another reason I stopped the discussion is that folks more knowledgeable than me say that I am wrong.

A small point to make I also contributed links to something like 40-50 raw images of real life subjects, that anyone can process with the raw processor of their choice. I wish more folks would do that, raw images are pretty rare.

I would also say, that I would not suggest that folks spend say 10-20 k$ on equipment based on web size images in sRGB colour space, again that is a reason to post raw images.

Again, looking at the spending aspect, it my be useful to buy a 60$ ColorChecker card and build a DCP profile and see if it suits you better. Using ColorChecker Passport may or may not be option. That is why write about DNG Profile Editor. DCP profiles don't work with Capture One, of course. But they work with Lightroom, AccuRaw, Iridient RawDeveloper and RawTherapee.

This can bee seen as a quite technical thread, BTW, colour rendition is all about technology.

Best regards
Erik

Erik, I remember you vicariously reminding everyone that "your threads are tech oriented and photography related discussions can happen in many other threads" every time someone asks about the relevance of such threads.

Let this thread be a reminder that unfortunately, you don't let photography oriented threads to be and will derail the discussion into more test charts and equations and other things working photographers have only a passing interest in.

The awards strawman is a new one though. I'll give you that!

[synn]

Of course. Anything and everything can be made into a technical thread if you want it to be, right? 

It was established within the first couple of pages itself that post production plays a big part in the output that the two references the OP provided. This had  nothing to do with Luther-Ives, CCD vs CMOS and certainly not world press award winners (Now really, how many PRESS photographers use anything other than DSLRs? Their choice of gear has more to do with lenses in specific focal ranges, high ISO shooting, portability, affordability and many other factors that have nothing to do with color rendition) and so on.

This thread was heading towards a rather productive direction and intentionally or unintentionally, you've derailed it into the usual rabbit hole. What do you think is the productive thing to do for the OP? Demo equipment he has not used yet and/ or improve his post production skills or read post upon post about theories that have zero effect in his daily work?

Please show some restraint with your technical overload. That is all we can ask for. This afterall is the Luminous LANDSCAPE, not the Luminous TEST CHART.

[Isaac]

That is all we can ask for.

We can just skip the stuff we don't wish to read ;-)