Grey Cards .... erk!!!

[Kirk Keyes]

Oh, by the way ... I've been posting using "subtractive" terms, Magenta, Cyan, Yellow and (should be 'K") Density.... because that is what I am working with at the moment. I just happened to think, though ... It isn't possible to have an "additive" grey. There is no, nor can there be, a "grey light", only different strengths of white light.

While we don't usually think of different levels of illumination as being different shades of grey, I think that if you talked to some people that are really experienced in lighting (i.e. theatrical or motion picture grips), they would tell you otherwise.

And when you are sitting in front of your computer, you are dealing with additive grey all the time. Your computer uses the additive RGB system for displaying colors, and any time you have equal amounts of RGB, you have shades of grey. RGB = 200,200,200 or 128,128,128 or 15,15,15 - they are all grey.

Earlier, I mentioned that there was a system that color sensitometrists use for calculating impurities in color printing systems, and I could not remember the name of it - it's called the "END" system - Equivalent Neutral Density. All color dyes in film (and paper) had a wavelength region of maximum density, but then they also have some absorption at all wavelengths, a secondary density, and this is where the problems come in. The END system allows one to make measurements of both of these regions, and to make corrections for the secondary density of the dyes. But it does require a reflection densitometer in order to use this system when printing. And for general printing, unless you have a grey target, it will probably not help. (But understanding the theory behind it may be good!)

Instead of a regular grey card, I would suggest that you get the one here: http://www.thedopshop.com/item.cfm?itemID=213 that Lee pointed you to. I think that having the RGBMYC ring-around patches at +5CC density would be a great assistance for what you are doing.

Kirk - www.keyesphoto.com

 

[Lee L]

I emailed the DoP Shop about the Fotowand Color Grey Card mentioned by Kirk in the last posting to ask about specifics. The lighter and darker patches of neutral grey are +/- 1/3 stop from 18% reflectance. So placing exposures at 1 stop intervals using the card will give you spots to read at 1/3 stop intervals through whatever total range in f-stops you decide to use.

Lee

 

[Ed Sukach]

I've been trying to buy a Kodak Grey Card Plus, Kodak Cat. #847-8174. It hasn't been easy. I've e-mailed our sponsor, Calumet, three times, as to availability/ orderability from Kodak, and have received NO answers.

Sometimes, I moonlight as a Legal Courier - last Monday I had to deliver a package to the Suffolk Superior Court in Cambridge, MA, a stone's throw away from Calumet in Cambridge.
So I visited - up front and personal. Disappointing. I talked to a counter clerk who, rather obviously, was not into film-based photography. He kept muttering about how their "stock" grey cards were effective for D****** work... how one had a panel very useful in achieving a "white balance". When I remarked that neither my Hasselblad 503Cx, nor my Dichroic Head on my enlarger had any provisions for "white balance" it resulted in a CLASSIC "deer-in-headlights" reaction.

Anyway - I started to post a method for dealing with scanner color balance ... and I sidetracked myself into this rant. See next message.

 

[Ed Sukach]

Eureka! EUREKA!!! Where is that Hecatomb....

Sitting here, typing away, and scanning for submission to APUG - I had an inspiration. As everyo .... MANY know, there has been a LOT of discussion about "color fidelity" and how to achieve it.
I was thinking about the fact that I do not yet have a CALIBRATED grey card, when the thought occurred to me... True, but I DO have the original grey card, and a photographed and printed image of it (see Technical Gallery). Not only that, but the original grey card would fit on my scanners glass. So, I scanned the grey card itself. Using the program that came with the H-P Scanner, I analyzed that image. That program will not give values in the usual sense - Red, Green, Blue - but will indicate values derived from Cartesian coordinates*, based on a color "wheel".

The Delta grey card I used analyzed out to: X +4; Y + 1.
The Printed image of that grey card: X +13; Y -3.

So, If I scan another photographic image (the color balance does not materially change from frame to frame, unless there is a change in the lighting - I know that from a lot of enlarging work with the ColorStar), and apply that "fudge factor"; the differences of those two analyses: X -9, Y +4, - I am going to be pretty darn close.

Now, I am going to post images here, with a reasonable level of certainty, that they are correctly "color balanced."

Any comments?

* Does that "date" me? Anyone else remember "abscissa" and "ordinate"?

 

[mark]

Ed,
Calumet has always been pretty slow with my E-mails but their 1-800 # is free and the folks on that end do know what you are talking about. I also never trust the guy behind the counter. I want someone who can type a number into their computer and give me the real answer not a salesman who may or may not have a clue what I want.

I realize why you want color balance in your scans but as you said before it looks right to you in real life. Who cares what a scanner has to say since we know it detracts from the quality of the final image anyway. Most folks here know that and take it into consideration. If someone wants to piss and moan about your color balance from a scanned print on a monitor, let them. I personally have not noticed your posts having odd color.

 

[Kirk Keyes]

So, I scanned the grey card itself. Using the program that came with the H-P Scanner, I analyzed that image. That program will not give values in the usual sense - Red, Green, Blue - but will indicate values derived from Cartesian coordinates*, based on a color "wheel".

So it is giving values something like the L*a*b* colorspace values? Where the scanner is showing you the "a" and "b' values. Does it give you a "L" value - which is the shade of the color, lightness/darkness? That would be important to know as well.

The Delta grey card I used analyzed out to: X +4; Y + 1.
The Printed image of that grey card: X +13; Y -3.

So, If I scan another photographic image (the color balance does not materially change from frame to frame, unless there is a change in the lighting - I know that from a lot of enlarging work with the ColorStar), and apply that "fudge factor"; the differences of those two analyses: X -9, Y +4, - I am going to be pretty darn close.

Do you mean "So, if I print another photographic image and make color corrections to remove the difference of (X+9, Y-4) when the print is scanned?" That seems like a reasonable approach. Make another print and minimize those differences.

Now, I am going to post images here, with a reasonable level of certainty, that they are correctly "color balanced."

Possibly, but not necessarily. The scanner can tell you that there is not much difference in "a" and "b" value, but you still need to know the brightness/darkness. You should be able to get that close with your eye.

But then, even though you have measured the relative differences between the original grey card and the printed one, there are still issues with the absolute accuracy of your scanner's color balance/color calibration. So while the grey card and the scan may not measure different at one point, they still may not be accurately portrayed on the monitor. And even if they look good on your monitor, that is no guaranty that they will look anywhere close to good on someone else's monitor.

So the next question is, do you have a calibrated monitor and scanner? If so, then you have a good chance of them looking acceptably good on someone else's equipment, if not, then maybe not...

 

[Ed Sukach]

I really DO NOT CARE about the appearance of the print on my monitor. I DO care - vitally - about the appearance, color balance and all on the print that I frame and exhibit. That PRINT is my goal, not the digital image.

That said, the scans in fact do look *very* good - quite close to the originals.

I have *no* control over how they look on another's monitor - I am not responsible for their color accuracy, and I'm not going to knock myself out trying to satisfy *everyone*.

 

[Kirk Keyes]

I really DO NOT CARE about the appearance of the print on my monitor. [...] I have *no* control over how they look on another's monitor - I am not responsible for their color accuracy, and I'm not going to knock myself out trying to satisfy *everyone*.

Sorry Ed, since you wrote:

Now, I am going to post images here, with a reasonable level of certainty, that they are correctly "color balanced."

I figured you did care.

 

[Ed Sukach]

Sorry Ed, since you wrote:
I figured you did care.

Knee jerk reaction, written in great haste. MY bad. I went off to somewhere or other, still thinking about it.

I'll try to "correct" - a matter of "degree":

I care, peripherally - not PRIMARILY - about the color balance as presented to those watching the Critique Gallery. If I can eliminate - or at least reduce - some of the compulsive reactions to what is perceived as "bad" color, and redirect comments to the impressions of the work - other than color, I'll be more than happy.

I was FLOODED by previous "Too green* comments, and the severity of that discussion overflowed to my Netscape mail - something I don't need.

By scanning and comparison to the "original" grey card, I have SOME confidence that I wasn't too far off... so now, I'll let the chips all where they may.

It may be of interest to note that I did investigate the "Calibrate Your Screen" route. To really effectively do that - avoiding subjectivity - I would be looking at a device to attach to my monitor, calibration software; eventually I would have to purchase that dreaded "PhotoShop", spend some time - All of which I would much rathe devote to my "wet film - printing" work.

 

[Ed Sukach]

The information about the Gretag-MacBeth Color Checker" and "... Color Checker DC" is contained in a .pdf file at [ Dead Link Removed ]. Once there, select "Information - then "Gray Card Selection - Download PDF."

Check his table there about "suitability".

 

[Kirk Keyes]

It may be of interest to note that I did investigate the "Calibrate Your Screen" route. To really effectively do that - avoiding subjectivity - I would be looking at a device to attach to my monitor, calibration software; eventually I would have to purchase that dreaded "PhotoShop", spend some time - All of which I would much rathe devote to my "wet film - printing" work.

Yes, you really do need a "color spider" to calibrate your monitor, it will come with software. But that doesn't mean that you will need to get "that dreaded" Pohotoshop program. The spider will measure your monitor's output and make a color profile that your computer will use when it runs - for all your programs, not just in Photoshop. Used ones are probably not all that expensive these days now that the spiders have had a couple of generations of evolution.

 

[Kirk Keyes]

[ Dead Link Removed ].

Yes, good article, I've read it before. Note the regular Macbeth Colorchecker is about as good as you can get in this comparison. Just make sure you store it in it's protective sleeve and it should not have any issues with UV stability for a long time.

 

[Ed Sukach]

Yes, you really do need a "color spider" to calibrate your monitor,....

Let me try this once more....

My goal is to produce a digital file that is an accurate ( ~ so squint a little) facsimile of the original print.

I scan and analyze a grey card in my PC.
I then photograph that grey card, develop the negative and print it. Upon scanning and analyzing that printed image, I get the *SAME* result as the scan of the grey card itself.

Whether or not that grey card is a "true" gray is of little importance. It has been reproduced accurately.

During all this, I have my monitor "set" to "black and white" only. No color whatever. The monitor itself will have *no* effect on the FILE and the color balance contained therein.

The monitor acts in the same manner as the viewfinder in the camera. It is useful to indicate a lot about what you are doing; but has no - to very little effect by it self on the exposure of the film, and the color balance.

 

[Kirk Keyes]

My goal is to produce a digital file that is an accurate ( ~ so squint a little) facsimile of the original print.

Unless you have calibrated/profiled your scanner, you probably don't have what most people would call "accurate" scans in the first place. I know, more money...

And since you don't have a calibrated monitor, and you don't even display any color on your monitor (what's the point of that, anyway), you can use your eyes as a check of the accuracy of the scan.

Whether or not that grey card is a "true" gray is of little importance. It has been reproduced accurately.

We appear to have different definitions of accurate reproduction... All you can really be certain of is that the scan of the grey card and the scan of the printed grey card have been digitized to the same RGB value. There is no certainty that they acrually represent "grey" - are the values all equal when you try this? Like 128,128,128 foro the value?

The monitor itself will have *no* effect on the FILE and the color balance contained therein.

The monitor does not affect the color balance of the file, but the scanner certainly has an effect, and there is no guarantee that it is accurate, or even close to accurate. My scanner produces rather magenta scans unless it is profiled through software using an IT8 standard. Then the colors are much closer to the original.

 

[laser]

Gray cards

Don't put more stricter requirements nor faith in gray cards then they are intended to provide. The cards are:

1. a reflective surface with specific diffuse reflection characteristics. Not specular reflections. Diffuse is generally 45 degrees from the source axis and the card should be tilted 15 degrees off the camera axis. See KODAK instructions for gray card use


2. They reflect red, green and blue light nearly equally. Humans can only see red, green and blue. By definition this is LIGHT. If the standard observer can't see it it isn't light. The cards reflect 400-700 nanometers uniformly. The uniform reflectance may extend beyond this range but that isn't a requirement. Manufacturer's ards may differ if they make different assumptions about the light source. Since it is a relative device this shouldn't have an adverse influence on your success when using them. The goal is to have the gray card the same in all your prints. Using multiple cards doesn't provide very much value. You are just trying to eliminate variance in the reproduction system by driving reproduction to a standard point. The reflectance of 18% was selected because this represents an approximate integration of "typical" scenes. See book by Ralph Evans. Title is something like Eye, Film Camera or Falk, Seeing the Light, or Bob Hunt's book on color reproduction.

3. For exposure the reflectance of about 18% was selected because this represents an approximate integration of "typical" scenes. Not All scenes but represents typical scenes. For dark scenes increase the exposure and for light scenes decrease the exposure. For exposure meters include a K (constant) factor to provide some safety factor. For Zonies the card typically represents Zone V. (Caucasian flesh is Zone VI thus Fred Picker called his stuff Zone VI. Ansel's white Caddy had CA license ZONE X. Rumor has it that Minor White preferred his toast Zone
III.) See Mr. Adams' book The Negative.

For color neg film the manufacturers provide RGB densities that indicate proper exposure levels. Too low densities will have smokey shadows and too high density will have blown out highlights. With modern films there is quite a range that will produce good prints. In both case there also can be color mismatches. Again modern films have good neutral reproduction over a wide exposure range. See film manufacturer's film data sheet for pro films.


In summary gray cards are a reference, they aren't the Holy Grail. Any combination of RGB reflectance could work as a relative measure of the reproduction. The easiest combination is nearly equal reflectance of RGB. You may desire your "neutral" to be a little warm or cool along the illumiant axis. Our vision is pretty tolerant on this axis. We don't have much tolerance fro Green/Magenta variations.

Enjoy.....

 

[Ed Sukach]

Don't put more stricter requirements nor faith in gray cards then they are intended to provide. The cards are:

1. a reflective surface with specific diffuse reflection characteristics....

I'm not sure why this was resurrected ...

Yes, I know all that. My concern was that the SCAN agreed, reasonably, with the original print.

My strategy was to scan the original grey card; and to compare that scan with the image of that grey card as photographed and scanned. If there was close correlation, I could be fairly sure that I was "pretty good" close. If not, I could set up "correction values" to use in modifying the scanned image to get "pretty close".

My MAJOR concern lies in the "analog" color print, not the image on the scanner. At the same time, it would be ideal if any and all of my images appeared to be "correct" on every monitor everywhere ... but I will NOT bust my buns trying to make it so, realizing that I have *no* control over all the monitors out there.

Hope this clarifies things.

 

[Claire Senft]

I would choose one of the cards to be your standard. Having done that and having made a negative of it, I would make a print that is an exact match to it in color and density. This will be a major PITA but necessary for best control. You could also make this standard negative with the card and with the type(s) of skin tone you expect to photograph. Then of course you will wnat to use the chosen card as your standard in the future.

 

[benjiboy]

Colour vision

I agree! My wife, a graphic artist/illustrator/painter, can seemingly differentiate more colors than we mere mortals even think exist. I've long since learned to stop saying anything was grey. She'll say "No, it's _________!" :confused:

Cheers

David

Hi David, I once read that women in general can "see" more colour than men, that's probably why it always seems so important to them, with female artists I would think it is even more so.

 

[Roger Hicks]

Hi David, I once read that women in general can "see" more colour than men, that's probably why it always seems so important to them, with female artists I would think it is even more so.

Some people -- and only women -- apparently have an extra set of colour discriminators (I can never remember which are rods and which are cones) and therefore actually can see more colours. Then there are the ones who can see more because they care more, and the ones who just think they can see more; the last is a non-negligible group.

Cheers,

Roger

 

[r-s]

Re cones -- most people are trichromats, but some are monchromats and dichromats (varying degrees of colorblindness), and it is suspected that certain women are tetrachromats (due to gender-specific genetic factors).

There are some non-human animals with more than four cone types.

 

[FilmIs4Ever]

For women with these extra sets of color discriminators, what are they seeing? Is it past the ultraviolet or pass the infrared set of the spectrums, or something else? I need to hire a gal like this to balance colors for me ;-)