The Online Photographer

Check out our new site at www.theonlinephotographer.com!

Friday, May 25, 2007

James Clerk Maxwell's Big Mistake

by Ctein

Contrary to popular opinion, this was not the first true color photograph
ever made. It was experimental error.


There's a certain type of error in judgment that we humans are all subject to. When presented with information that contradicts our beliefs, we tend to be nitpicky and skeptical and aggressively compulsive about every detail. When handed information that confirms our beliefs, we're inclined to accept it without critical evaluation.

What does this have to do with photography? Well, James Clerk Maxwell is credited with having made the first true color photograph back in 1861. Most of us accept his results at face value, because they agree with how we now do color photography. There's one small problem with this. Maxwell's experiment was a failure; it had a major methodological blunder that rendered it meaningless. It didn't prove anything. Few of us caught that, and certainly Maxwell did not.

Here's the key problem, which should have bothered all of us the first time we read about this great experiment. In 1861, there were no panchromatic emulsions. The type of emulsion Maxwell used was sensitive only to blue light, with just the faintest hints of blue-green sensitivity. Nothing beyond that. No yellow, no red. While it's possible the Maxwell used some experimental emulsion that had red sensitivity, there's nothing about this in the detailed account of the experiments written by his assistant, Sutton (who did the actual labor). So, how in the world could they have successfully photographed a scene through red, yellow, green, and blue filters?

(And why yellow? Well, I think the likely answer to that is that Maxwell was also testing out the three-color versus four-color theories of human vision. Newton had proposed that there were four visual primaries: red, yellow, green, and blue. Later, others argued that there were three primaries: red, green, and blue. This was not a settled matter, so it is very probable that Maxwell saw this as an opportunity to test both hypotheses.*)

The unfortunate answer is that he didn't. The blue exposure was just fine—a matter of seconds. In order to make the green exposure, Sutton had to substantially dilute the copper chloride solution they used as a filter. Even then, the exposure ran to 12 minutes. A strong, deep-red filter of ferric thiocyanate resulted in an exposure of only eight minutes, though.

Maxwell should have caught this. Although he could not have known that his emulsions were 100% insensitive to red light (knowing that would take theory that wouldn't be devised for half a century) he would have known that the sensitivity to red light had to be very, very low.

So, why did the experiment work at all? Because ferric thiocyanate passes a considerable amount of ultraviolet light, and many red fabric dyes reflect in the near ultraviolet as well as in the red. By pure accident, they got a plausible-looking photograph. It was the right answer, but for the wrong reason; had they a chosen a different subject (say, a flower garden or a rainbow) it would have come out all wrong.

The greatest expert on light and electromagnetic radiation the world has ever known got tripped up by a simple experimental error because the results confirmed his expectations. Imagine how easy it is for us mere mortals to be caught in the same intellectual trap.

Posted by: CTEIN


*The answer, not established until the 1970s, is that both are correct! The human eye has three kinds of cells that respond to frequency bands primarily in the red, green, and blue. But the brain interprets the information from those cells as four distinct primary colors.

27 Comments:

Blogger Robin Dreyer said...

Great post! Both the historical and philosophical commentary are excellent.

10:36 AM  
Blogger David said...

actually Mike we have 3 colour sensors - Red green and blue cones - and the rods for B&W. (well 90% of us males have (10% are colour defective...but that's a different book!)

So we really have 4 channels, 3 colour and 1 luminance.

Best
David

11:31 AM  
Blogger MHV said...

All I have to say is... Felix culpa!

11:35 AM  
Blogger rob povey said...

Some of best shots have been completly accidental!!

Doing my bit got the "monkey" school of photography.

It's the image that counts not how you got there.

Regards,

Rob.

2:29 PM  
Blogger Leo said...

Recreated the experiment for a Scientific American cover back in 1962--We did the photo at Kodak, they screw up the processing of the film and the cover never ran. The article was published.
Had a lot of fun on the project anyway. They waited until press time to tell us they did not have the photo and had screwed up the processing. How I love this business, Leo

3:25 PM  
Blogger Morven said...

Ctein, could you point me to where you learned about Maxwell's mistake? The Wikipedia article on Maxwell does not mention this, and leaves the impression that this was indeed a true color photo.

Thanks!

4:39 PM  
Blogger Ctein said...

Dear David,


Under normal bright light, the fourth psychophysical primary is purely a function of the brain, derived from a differential analysis of the three cone responses.

But...

Under special circumstances, we have actually four retinal color channels. The rods will function as a color channel, but they do not normally do so because cone response suppresses rod response. Kind of an automatic gain control so that rod response doesn't overwhelm the visual system when the light gets bright.

Under very dim light, though, you can find a crossover region where photopic (cone) and scotopic (rod) vision are simultaneously active. The light of a full moon is sufficient for many people. Under that condition you see color created by FOUR visual sensors, not just three.

This hybrid color vision is very interestingly different from normal color vision. In a day or three, when the moon is full, I encourage readers to observe it for themselves. Take a set of color swatches you are familiar with out into the moonlight. It needs to be real color patches, not a printed reproduction of colors. Something like a Macbeth ColorChecker chart is perfect, but a box of Crayolas or Prismacolor pencils will work just as well.

Under moonlight, the colors will appear somewhat dim and desaturated, but they'll be clearly visible. What you'll find interesting is the relative qualities of colors-- how the reds compare to the purples and yellows, how the blue and greens compare to the cyans, etc., in both hue and brightness. You'll be rather surprised by what you see.


pax / Ctein

4:26 AM  
Blogger Ctein said...

Dear Morven,

OK, this is rather vexing. I just realized I don't know the author nor the title of the book I'm cribbing from. The book says that Ralph Evans at Kodak was responsible for attempting to replicate Maxwell's work and figuring out just what was really going on. But all I have is the pertinent three pages from the book that were emailed to me. The first page is headed with "Chapter 2: Early Concepts and Processes", but the image files of the pages don't have book title or author on them.

I could email you the image files, but I can't otherwise send you to the source material.

pax / Ctein

4:27 AM  
Blogger Leo said...

Morven and Ctein-This is probably the guy we were working with at Kodak to replacate Maxwells work--I don't believe he made a mistake, just didn't know the film wasn't green and red sensitive.
"The book says that Ralph Evans at Kodak was responsible for attempting to replicate Maxwell's work and figuring out just what was really going on". The article in the Scientific American Magazine (being some 40 years ago) was published between 1961 and 63.
You can probable make the same photos today using the. -- Fujifilm FinePix S3 Pro UVIR, 12.3 Megapixel (6.17 million S-pixels & 6.17 million R-pixels), SLR, Digital Camera (Camera Body) Specially Modified for Forensic/Scientific Applications--With the right filters you can do the UV part and get great if not correct to the human eye color photos.
Have fun, Leo

11:06 AM  
Blogger Leo said...

Morven--Found where it is --You will have to find the magazine.
"Evans R.M. (1961), Maxwell's Color Photography, Scientific Photography 205,117-128 (November, 1961).
Have Fun, Leo

11:24 AM  
Blogger Morven said...

Thank you, Leo and Ctein. I'll try and track that magazine down - I have access to a fairly good academic library, they may be able to get it for me. At least, so I hope.

3:14 AM  
Blogger Struan said...

The ribbon photograph is interesting, and especially so if you take the trouble to understand it in terms of modern science, but this posting's tone of shock and awe is sensational rather than informative.

Twenty years prior to Maxwell's lecture J. Herschel had published his experiments into the darkening of silver salts by what we call ultraviolet and infrared radiation. In that work he describes a variety of paper emulsions sensitive to all or some portions of the spectrum, including red and infrared. He also discusses the fact that red objects sometimes left impressions on photographic plates that were not red-sensitive and gives the exactly the explanation that many red tints also include ultraviolet rays. He also proposed the colour seperation photography that Maxwell later demonstrated in practice.

Phil. Trans. Vol. 130 pp1-59 (1840). It's worth a read.

I find it impossible to believe that Maxwell was not aware of one of the most important papers in the history of photography and electromagnetics. I find it even less likely that he was unaware of his own research: why would he be 'testing out three-colour versus four-colour theories of human vision' when he had already spent the previous decade proving that human colour was three-colour using considerably more sophisticated and subtle methods than colour seperations onto photographic plates? He published what we now think of as the CIE diagram of spectral coordinates in 1855, six years before this talk.

So far as I know, what Maxwell actually said during his presentation to the Royal Institution has not been preserved. However, if you want to support the use of words like 'blunder', 'mistake', 'error' and 'failure' you have to assume that one of the greatest scientists ever to have lived knew less about his own experiments than can be gleaned from browsing a third-hand history of photography.

I know who I would bet on.

PS: for what it's worth the observation that the balance of the spectral colours changes in twilight was published by the eponymous Purkinje in 1825. It is entirely possible that Maxwell knew about that too.

6:23 AM  
Blogger Ctein said...

Dear S,

Actually I am betting on Maxwell. It is indisputable that the experiment WAS a failure. It purported to demonstrate that capturing red, green and blue wavelengths photographically would yield a true color photograph. The experimental materials, to the best of record, were incapable of recording red wavelengths. That's it. Failed experiment, incapable of proving the hypothesis; it was fundamentally flawed.

As to WHY Maxwell created and then reported erroneous results... that's another matter, and it is one where I took the high, most assuredly not the sensationalistic, road.

Along with the work you mention, I would think that Maxwell would know of the ongoing efforts at spectrography. Spectroscopy was nearly half a century old; folks started trying to capture spectra with sensitized materials as soon as the latter became available. Simple adaptations of Newton's prism experiments were widespread, both for the sake of spectroscopy and for characterizing photosensitive materials.

Perhaps Maxwell had no interest in those experiments. Seems terribly unlikely to me, just as it seems terribly unlikely he'd not have known of Herschel's work. But y'know, anything's possible-- we're not inside M's head and the paper trail doesn't give us enough clues.

So, maybe, just maybe, Maxwell really didn't know all that stuff and made his simple error out of ignorance?

No, I don't really believe that any more than you. In my opinion that denigrates his skills and knowledge.

So, what are we left with? An experiment that could not have demonstrated that capturing red, green and blue images would create full color, because no red images could be captured (and the green's questionable), and a researcher who should have known that (and could easily have tested the sensitivity of his plates to check) but seems to have not.

I can explain that two ways. One is as I suggested in the original article, that Maxwell was caught up in the scientific moment and failed to catch the flaw that meant that his results, no matter how they corresponded to expectation, provided no scientific evidence for his model.

Physicists would correctly call that a mistake, an error, and an experimental failure. Personally, I'd add "blunder," because I think Maxwell was so good that this falls way below his own standards. But it's all innocent enough. It happens. Most of us get caught up in our hypotheses at one time or another and do a bum experiment or fail to calibrate an instrument correctly. It's an unconscious and unintentional slip of the mind.

And is is hardly sensationalistic to call it a mistake. It's the accurate word.

Still there's one more possible explanation. It's possible that Maxwell was fully mindful of all the ancillary research that you and I have brought up and didn't have it slip his mind at all. In which case he'd be running an experiment that he'd know couldn't possible work the way he claimed it did, and be presenting results to the Royals that superficially appeared to support his RGB model but that he knew did no such thing, scientifically.

Physicists have a word for that, too. It's "fraud."

It wouldn't be the first time. Newton falsified data to support his theory of gravity. Mendel did around genetic statistics. Etc.

If Maxwell has truly as great and encyclopedic a mind as we both believe, fraud cannot be dismissed as the explanation. But we do not know-- we lack the paper trail, the smoking gun, the purloined letter.

I will not accuse a scientist of fraud when a more benign explanation exists. That WOULD be sensationalistic. But I see only three explanations-- ignorance, unconscious slip, or willful deception. I chose the LEAST insulting of the three. If you can think of a fourth less insulting explanation I missed, I'll happily entertain it.

But, please, do not suggest I've tried to boost a molehill to a mountain. Not unless you can suggest an even smaller molehill.

And, remain mindful that I am writing a POPULAR column, not a ACADEMIC paper. I do not adopt a scholarly tone, and that's fully by intent , not unconscious slip [grin].

pax / Ctein

1:23 PM  
Blogger Ctein said...

Dear Leo,

Thanks for tracking down all that info!

Do you have access to one of the Fuji UVIR cameras? It would be fun to see posted the results of a modern run of Maxwell's experiment, especially with other subjects.

pax / Ctein

1:25 PM  
Blogger Ctein said...

Dear S,

Hey, would you happen to know how to pronounce "Purkinje?" I'm always afraid to say "Purkinje shift" for fear of getting it really wrong. Czech is NOT a language I'm much familiar with-- for all I know it's pronounced "chumley."

Thanks!

pax / Ctein

2:00 PM  
Blogger François said...

According to Naomi Rosenblum in A World History of Photography, the first color print was made by Louis Ducos du Hauron in 1869. Diaphanie (Leaves) is a three-color carbon assembly print. Then he made a view of Angoulême in 1877, a Heliochrome assembly print. These are interresting pictures, but you have to see the Lumiere brothers color photos from the beginning of the XXth century. Their Autochromes are amazing, so beautiful. They did not only invent Cinematographe, they invented Autochrome !
François

3:49 PM  
Blogger Struan said...

Ctein,

I would pronounce 'Purkinje' with a soft 'ye at the end and a rolled spanish ñ as in mañyana. But I am no Slav, and the Czechs can get feisty about the beauty of their language compared to, say, Russian or Serbo-Croat.

I am waiting for our library to locate their November '61 copy of Sci. Am. so I can see what Ralph Evans wrote for myself. While waiting, I have been thumbing through the two-volume set of Maxwell's manuscript letters and paper drafts that CUP published a few years back. One thing that strikes me is that Maxwell simply doesn't mention photography at all. Not in his letters to people like Kelvin and Forbes, not in any of the three inaugural speaches when he took chairs at Aberdeen, London and Cambridge, and not even in the notes he wrote prior to the Royal Institution lecture where he showed the composite photo of the ribbon.

My feeling is that the fourth way you mention is that Maxwell wasn't presenting the tartan ribbon as the invention of colour photography at all. Instead, he was using the lantern slides to illustrate a point about human vision. Only later did the photographic world credit him with an 'invention' that he himself seems to have had no desire to claim. That would fit with a scientific mileau where Herschel's work was widely known and accepted.

One thing is clear: Maxwell himself certainly knew of Herschel's work - he refers to it several times in letters and there is even a comment in the manuscript draft of the R.I. lecture, although that latter could equally well be a reference to W. Herschel, not J.. It is surprising that more historians of photography have not picked up on J. Herschel's comment. It is made in the same paper where he announces the use of thiosulphate as a fixer, and the paper is certainly widely cited. That said, it would not be the first time that fundamental work was widely cited and never read.

Incidentally, the reference I gave should be "Transactions of the Royal Society", not "Phil Trans" which is a brain fart of mine caused by looking too hard for an online source for Phil. Mag. Mea culpa.

I'll post again when I have read Evans' own words.

Struan Gray

1:38 PM  
Blogger Leo said...

Ctein, Sorry thats all I need is one more digital camera. There isn't any reason why it shouldn't work-you wouldn't really need the 3 filters with the Fuji camera. When we, Bill Vandivert an I re did the experiment, we tested the exposures with polaroid film(one exposure was during lunch)--when we did the 3 separation negatives, we left them at kodak scientific division to develope them. They processed them to the wrong gamma so the whole thing got screw up..Here's another Kodak top secret-- most of the camera store C print samples hanging on the walls in camera stores back in the late 50's and 60's were really dye transfers. There should be enough UV sensitivity in regular film to get results I think thats all we used, probable Super XX.
That was Scientific American not (Scientific Photography).

5:57 PM  
Blogger Ctein said...

Dear S,

Thank you for the suggested pronunciation of Purkinje! I would remind people that the full moon is only a day or so off; an excellent time to observe this weird rod-cone color vision for oneself.

The article excerpting I worked from quotes some of Maxwell's Royal Institution paper. It goes as follows:

"This theory of color [RGB color vision] may be illustrated by a supposed case taken from the art of photography. Let it be required to ascertain the colors of a landscape by means of impressions taken on a preparation equally sensitive to rays of every color.

"Let a plate of red glass be placed before the camera and impression taken. The positive of this will be transparent where the red light had been abundant in a landscape and opaque where it has been wanting. Let it now be put in a magic lantern along with the red glass and a red picture will be thrown on the screen. Let this operation be repeated with a green and a violet glass, and by means of three magic lanterns the three images be superimposed on the screen. The color at any point on the screen will then dependent on that and by properly adjusting the intensities of the lights etc., landscape, as far as visible color is concerned, will be thrown on the screen."

So, yes, he was trying to make a point about color vision, but he clearly meant to use true color photography to make that point. His description of what should be done in that regards is distinctly at odds with what his experiment actually did (or did not do, depending upon how you look have it ).


pax / Ctein
[[ Please excuse any word-salad. ViaVoice in training! ]]
=========================================
-- Ctein's Online Gallery http://www.ctein.com
-- Digital Restorations http://photo-repair.com
=========================================

1:03 AM  
Blogger Struan said...

OK. I have now read Evans' Scientific American article. Frustratingly, he does not give any references for his sources, but there is enough information there to see what is going on. He published what I assume is a more technical article in the Journal of Photographic Science (vol. 9, p243, 1961) but I cannot find that journal in any local academic libraries or online, and past experience has taught me that transnational interlibrary loan takes forever (I live in Sweden).

I think it is clear what Maxwell and Sutton did. The question is whether it is valid to label their work as a failed experiment. I don't think so: I think Ctein's language was overblown and unjustified, and even if he remains convinced that Maxwell's tartan ribbon photo was a failure I think it would have been possible to give a popular explanation without lapsing into populism. 'Blunder' is a term best reserved for more flagrant errors.

Physicists are more subtle than the laity generally suppose. I am one, and although I and my colleagues try to conform to the classic Bacon-Popper form of the scientific method when publishing (it's more persuasive, and has more status :-), when actually doing experiments we are more fluid and varied in our thinking than all the talk of falsifying hypotheses would suggest. In particular, we are as happy to investigate the bounds of validity of a theory as we are to poke about at its formal underpinnings; and a simplistic dichotomy between correct and false is often more hindrance than help. Physicists and engineers will happily carry on using a theory proved to be wrong for some applications if it remains workable for others. GPS decoders need to worry about relativistic time dilation; your dashboard clock does not.

This was even more true in Maxwell's time, when it was regarded as wondrous that the universe was at all amenable to analysis, let alone to the extent that the almost daily new discoveries showed it to be. Much of what we now regard as basic science was then still at the botanising stage of collecting data on phenomena and seeing if any patterns or correspondences showed up. It is pure condescension to criticise the people who struggled to invent our fundamental conceptual tools for not finding them as obvious as we do with our haughty privilege of two centuries of hindsight.

Until we know what Maxwell said we don't know what hypothesis he made about the tartan ribbon photograph, even supposing he made any. From my own reading it seems clear that he was not in the slightest bit interested in claiming to have invented colour photography, but instead was illustrating how monochromatic images from different spectral bands can combine to form a convincing colour image. The title of his talk at the Royal Institution was 'On the theory of three primary colours', not 'The invention of colour photography', and as I said in my last post he doesn't mention photography at all in his notes for this or his other lectures on the subject, or in any of the major articles published in connection with this work, or in those of the letters sent to other scientists during this period which are held in the Cambridge collection of his correspondence.

Sutton insists that Maxwell must have been 'unaware of the fact that wet collodion plates were not sensitive to green and red'. He also finds it 'surprising' that the coloured filters chosen by Maxwell and Sutton had such good monochromatic pass bands. Both points can be refuted if Maxwell knew of John Herschel's work on absorbance and filtration of the solar spectrum from the 1830s, and his later 1840 paper on the difference between the visual effect of the prismatic spectrum and it's 'chemical' effect on various silver salts, including in the UV and IR spectral regions.

Maxwell explicitly refers to Herschel's absorbance work in the 1855 and 1860 papers describing his experiments into tricolour vision. The manuscript draft of his R.I. lecture also explicitly cites "Sir John Herschel". I do not have access to a copy of the actual lecture (published in the Proceedings of the Royal Institution, vol. 3 pp 370, and in Maxwell's collected papers, CUP 1890, and reissued by Dover), but I think the balance of probabilities is that he referred to them there too. It is not so much unlikely as incredible that he would not have read Herschel's 1840 paper, not least because it has direct bearing on Maxwell's electromagnetic theory of light waves, the first part of which was published two months prior to the R.I. lecture. Keeping up with the literature was a simpler task then, especially if the relevant literature was one of the most famous and cited papers of its time.

Herschel himself cites Fox-Talbot in his discussion about the potential for colour photography and the way red objects can register on red-blind emulsions. In a paper stuffed full with claims of invention and elaborate neologisms he doesn't claim any priority for himself in this particular matter. That Maxwell, a notoriously modest man, would avoid claiming to have invented the same thing seems perfectly reasonable. It seems equally reasonable that knowing of Herschel's suggestion he would adopt it to illustrate his lecture on human vision.

One jig-step it might be allowable to dance on Maxwell's grave is the fact that photographing a tartan ribbon proves nothing much about human vision because the colours in that sort of tartan are roughly the same as the primaries used, and worse, they are neatly separated in space. Maxwell could just as well have drawn the picture on a blackboard with coloured chalk, since in this case the optical combination of images achieves nothing beyond placing the coloured lines next to each other. A 'proper' experiment would an object with many secondary colours - particularly the colour we call 'magenta' which has no spectral equivalent - and show how they reappear when the red, green and blue images are combined.

To me, this actually reinforces the impression that Maxwell's use of the colour separations was as an illustration of a simple point about human vision, and not an attempt to claim anything in particular about photography. In that context it makes full sense that he would have used a system that worked, even if the details of why it worked were more complex than a naive observer might infer. This is not experimental fraud, but illustration. Einstein explained special relativity to general audiences by talking about steam trains moving at close to light speed: is he a fraud too?

Great scientists do make slips, but it is also common that they do not describe the detail of their reasoning because they find it obvious. I agree that we do not know for sure what Maxwell thought about tri-colour photography, and we do not know what he actually said that May day at the Royal Institution. I also am happy to accept that the recorded lecture as published, or some letter of Maxwell's in the archives could easily prove me wrong. However, and for me, there is ample evidence that Maxwell knew as much as Evans, Ctein and I do about the important facts of his own experiment, and that to describe it as a 'mistake' is simply arrogance.

This has gone on a bit. At the risk of being non-popular, I'll stop with some references to online sources. Those of us with library subscriptions to JSTOR can easily find the Philosophical Transactions of the Royal Society of London, but those outside the academy will need to consult microfilm or pay real money. The more obscure journals like those of the Edinburgh Royal Society and the Royal Institution are effectively buried unless you travel to one of the UK copyright libraries, or the insitutions themselves, although if you are lucky your local university may have them on microfilm. Phil. Mag., *the* nonafiliated journal of the C19th appears in databases as the snappily-worded 'London, Edinburgh and Dublin Philosophical Magazine and Journal of Science' but doesn't seem to have been digitised anywhere yet.

All the same, thanks to this Italian page Maxwell's 1855 and 1860 papers are available online (scroll to the bottom):

http://www.boscarol.com/pages/storia/20-maxwell.html

I can also recommend R.J. Mollon's excellent early history of colour science to get some idea of the historical and intellection background in which Maxwell developed his theory of colour vision:

http://vision.psychol.cam.ac.uk/jdmollon/papers/MollonColorScience.pdf


Struan Gray

4:55 AM  
Blogger Struan said...

Ctein, I wrote and posted that tome before your latest comment appeared, hence the odd disconjunction between 'OK' and referring to you in the third person.

I'm not sure we have enough here for a storm in a thimble, let alone a teacup. But it's always fun to speculate for low odds.

Struan

6:59 AM  
Blogger Ctein said...

Dear S,

Yes, I had figured out that our messages crossed in cyberspace.

In case you weren't aware of it, my training is also in physics (and I was one of the better experimentalists in my class {Caltech, ' 71}, he said immodestly). Whenever I read an experiment, I can't help but think like an experimental physicist. And so I am fully in agreement with you that we cannot and should not judge this work by 20th-century standards. I did my best to take into account the frame of mind and the instrumentalities of the time.

But I agree with you that it would take substantially more research to come up with an answer that satisfy everyone (rather obvious, considering how different our reactions are to essentially the same information). Probably there is a good PhD thesis in here for someone going for their degree in history of science-- trying to pull up enough primary sources to figure out just what was going on here.

Indeed, it is a tempest in a thimble. Were it not, this would not be half so enjoyable a discussion. If there were real stakes involved, we'd have to take this much more seriously.

Anyway, I think we've run this topic to death. I've enjoyed the exchanges greatly; I hope you have also.



pax / Ctein
[[ Please excuse any word-salad. ViaVoice in training! ]]
=========================================
-- Ctein's Online Gallery http://www.ctein.com
-- Digital Restorations http://photo-repair.com
=========================================

2:47 PM  
Blogger Ctein said...

Dear Leo,

Maybe I can convince Fuji to loan me one at some point. I have to figure out a better reason, though, than just adding a comment to this column, but it would be entertaining.

On a slightly different matter, do you or Struan know of anyone photographically simulating the Purkinje effect? It just occurred to me that it ought to be possible-- a scotopic-balance filter would be close enough to rod response for the point of demonstration. The only real issue would be figuring out what weightings to give the four images (RGB and rod) and what hue to assign to the rod. But Photoshop would make it easy to experiment with that in layers and channels. Also might make an amusing project, if it hasn't been done already.


pax / Ctein
[[ Please excuse any word-salad. ViaVoice in training! ]]
=========================================
-- Ctein's Online Gallery http://www.ctein.com
-- Digital Restorations http://photo-repair.com
=========================================

2:53 PM  
Blogger Struan said...

The CIE has defined a standard luminous efficiency curve for scoptopic vision. You could use that to construct a scoptopic colour space, although the idea of scoptopic colours is a nice oxymoron in itself.

I have played with the idea of using one of those four-lens passport cameras to take three-colour-plus-luminance seperations onto B+W film. You could do something similar with a scoptopic filter for the luminance channel, and thus make daylight look like twilight. Making twilight brighter so you can see the colour differences clearly is harder because you have to reverse the effects of reciprocity.

The manipulations are easier in digital, but for general use you would need an imaging spectroscope or a monochrome digital camera and a filter wheel with more than three colours. Capturing RGB and then converting to RGBL in some scoptopic colour space is a many-to-many-to-many transformation, and I can see the subtleties getting lost.

And the subtleties are worth persuing. I photograph a lot with film in twilight, and love the surprises the world and my camera throw at me. Moonlight and skylight have quite different colour temperatures to daylight, and combined with the effects of reciprocity the colour rendition in different atmospheric conditions can vary wildly. Secondary colours, like the cyans and pinks and oranges you get 60-100 degrees away from the sunset point, are particularly gorgeous. Among established photographers Richard Misrach is good at capturing these effects, although Cape Light is probably the Platonic exemplar.

This has been fun. Thank you for a pleasant exchange, and for stimulating me to take a fascinating trip into the archives.

2:10 AM  
Blogger Struan said...

I meant to say, I am taking a few days away from the internet, so if this horse requires further flogging I'll pick up my whip again after the weekend.

A final tidbit (I promise): One reason I am grateful for the prod your article gave me is the discovery of Evans' work in what he calls 'brilliance'. It explains why photographs of iridescence are always so disappointing.

Off to dream of numinosity.....

Struan

2:18 AM  
Blogger John said...

Struan, I know this is probably several years late! But I just came across this conversation, as I prepare for a lecture commenting on Maxwell's demo. If it is of interest, I have a copy of Evans' paper (J Phot Sci 9, p243) and could scan it.

11:29 AM  
Blogger Mike Johnston said...

John,
You're right, this is a very old post. I did check, and unfortunately I don't have an email address for Struan.

Thanks though.

Mike

2:58 PM  

Post a Comment

<< Home