The Journey of Restoration

I gave this talk to The Photographic Historical Society of Rochester on May 19, 2005.

The Lost Beauty of Kodachrome

The Journey of Restoration Slideshow please the image above  to watch the slides as you read the paper.

Good evening. I know that for some of you, it was a tough choice to be here this evening with the last Star Wars film coming out today.  And while I can’t promise you a battle between good and evil with space ships and light sabers, I can promise you an interesting story that touches on such topics as the birth of modern dance, the battle of wills within the photo-giant Eastman Kodak, and even spend a few moments learning about a few ancient Hindu Gods along the way.

How many of you were at the Dryden last night to see “Becky Sharp”? What a wonderful restoration. For those who may not have seen the film, “Becky Sharp” is credited as being the first color feature film to have been originated using the 3-strip Technicolor process.

Tonight I’m going to talk about a color motion picture process that pre-dates 3 strip Technicolor by almost 20 years.

I work at KBTV in Kodak Office as an editor. When I’m not putting together training tapes or industrial videos, I’m hanging out with those lone survivors at Kodak that are trying to preserve the legacy of the company and to bring to light some of the company’s great achievements.

In 2003, I got a call from Research Librarian Ray Curtin. He asked me to help him find a home for some film he had discovered. This was a common occurrence at that time as the Business library was being closed down and dismantled and much of the materials in it were being sent to the U of R, George Eastman House, or discarded if a home could not be found. Ray knew of my historical interest in Kodak and when it came to early media, I would often try to save it myself.

When I got to Building 13, I had no idea what I would find:

There were early 16mm panchromatic film tests, footage of old executives including Stuber and Haste, we had a Kodachrome film of a Research Picnic from 1935, a time lapse shot of Jupiter rotating, even an early 16mm shot of George Eastman at his desk.

There was also an 800’ 35mm reel marked “The Flute of Krishna – Long Shots and Main Title Only”.

That name rang a bell.

As I recalled: “The Flute of Krishna” was a dance choreographed by Martha Graham in 1926 shortly after her departure from the Devenshire Dance Troup. Martha, famous as a pioneer of modern dance, for one brief year, taught at the Eastman School here in Rochester. A year later, she moved back to New York and started her own modern dance troupe that still exits today.

Back to the treasure:

Up to this point, everything I found was real easy to work with, 16mm safety film or 35mm safety reprints.

But “Krishna” was different.

It appeared to be a camera original, which meant it, was, of course, silver nitrate.

Standard procedure at Kodak for this type of hazardous material was to immerse the silver nitrate stock in a bucket of kerosene and remove it to a hazardous waste facility.

“The Flute of Krishna”, the earliest know record of Martha Graham’s choreography, was just hours away from such a fate.

Instead, some friends of mine at the Research Labs helped me ship the film to an outside lab where a fine-grain positive was struck. The new print was transferred to digital betacam and this is what it looked like:

Show Flute Telecine Transfer from DVD

A jumping-mess. But it was much more than that, “The Flute of Krishna” was an experimental film that was shot using an early 2 color process with a very familiar name: Kodachrome.

This was the camera original to a piece of early cinema history-as well as modern dance.


Research staff member John Capstaff shot “Krishna”,

during a dress rehearsal in April 1926 in Kilbourne Hall, using a color motion picture process he had invented.

This is the same guy who also perfected color filtration at Wratten & Wainright, and developed the successful amateur home movie system of 16mm reversal film, D76 developer, several motion picture cameras, and perfected the continuous developing system for 16mm.

However, color photography was probably Capstaff’s most elusive endeavor.

Originally, Kodachrome began as a still process in 1913 exposing two duplicate images through separate red and green filters onto panchromatic film. To duplicate this with motion pictures, Capstaff used a twin-lensed camera that recorded two images simultaneously onto the single strip of 35mm film. The jumping you saw was both the green and red record for each frame of motion.


What made Kodachrome a superior process was in two ways:

First it captured both records of green and red at the same time, thus allowing for perfect registration when it came time to print the film.

Also, it recorded these images on a single strip of film, rather than on two strips, like the early Technicolor process, which was also emerging at the time.

It also made redheads extremely popular at the Research Labs.

“The Flute of Krishna “ was not Capstaff’s first attempt using Kodachrome as a color process for motion pictures. In fact, he had been working on this color process for over a decade starting with this first effort “Concerning $1000” shot in 1916.

SHOW CONCERNING $1,000 clip from DVD

After several tests had been conducted, it was decided that an attempt to make a short film with the Kodachrome process would reveal any weakness in the process. This edited excerpt from the film is also noteworthy in that it was a very early infomercial. It showed the process for taking, processing, and sharing pictures-in this case to win a photo contest and share the prize money with the inventor so he can finish his experiment.

Incidentally, if the scene in the garden looked a little familiar, it might be because those scenes were shot at Mr. Eastman’s gardens on East Avenue.

The momentum of Kodachrome was halted as World War I took over the energies at Kodak. It wasn’t until 1922 that Capstaff was allowed to resume work on his passion.

SHOW 1922 KRL film test from DVD

More testing and refinement resulted in more brilliant colors and truer skin tones.

Although not necessarily better acting (pouty girl).

It also got the attention of 20th Century Fox.

SHOW 1922 FOX film tests from DVD

A soundstage and a photo lab were set up on the studio lot for Capstaff and his crew to perfect this color process. But instead of redheaded office assistants, Capstaff’s camera could now capture some of the stars in the Fox galaxy.

Both of these films were preserved and restored by

Ember A. Lundgren , graduate of the Selznick School of Preservation.

Understanding this process, it became clear that to restore the film back to its color state, I needed to use a digital solution. Back at Kodak office, using an AVID video editing system, I digitized the entire eight minutes of that flickering mess.

Then isolated the two records into individual elements.

Show Kodachrome Demo from DVD

Here are the two color records separated and registered: green on the left and red on the right.  The lab that preserved the nitrate negative remarked that it was in exceptional condition.

The next step in this process was to introduce filters with complimentary colors to the original records: green is filtered red and vice versa.

Then they are combined and the result is a color motion image.

Not a true record of the full color spectrum, but a fairly decent representation, especially in the area of flesh tones and green landscape.

My first attempt to restore “Krishna” to color was based upon what information I had at the time-which was almost none. I did not have the opportunity to see the test films I just showed you and any information on the original Kodachrome process was almost non-existent.

While I knew red and green were used to record and later re-print the film, I did not know the exact hue of these filters.

All I could do was guess and rely on my past experience.

For me, that meant going back to my very early days as a photofinisher when I was in college. The rules were-get your whites as pure as possible, but never at the expense of your flesh tones.

With these broad rules in place, this was the result:

Show First color composite from DVD

Notice the main title card: the flowers look pretty good and the large center area is almost white. The set looks OK. There were no slates, no gray cards, and the first to appear was the main character, Krishna.

I could see the setting was somewhere in “Arabia”, so I guessed he was a very dark-skinned man.

Maybe black, maybe olive.

Then the dancers appeared and they looked ok in their outfits.

This looked like many early films I’d seen before: contrasty, grainy, in the traditional red/green wash.

But a pretty good job for a first attempt.

When I work on a mystery like this, I try to become immersed in the back-story of where it came from –and in the case of this film, why it was even made to begin with.

The more I thought about it, the more I realized something bothered me: It’s 1926. Capstaff had help to jumpstart the amateur film industry and still throughout all of that, he kept on working on this process.

This Kodachrome color process had its problems: in order to create the color, the actual base was double coated with emulsion-one side was the green side, the other was the red side, as opposed to Kodachrome’s rival Technicolor which exposed two strips, dyed them separately, on different, ultra thin bases (made by Eastman Kodak) and then cemented them together.

In double coating the base, Kodachrome was extremely susceptible to scratching. But its biggest detraction was the fact that it really didn’t reproduce color any better than any of the other 2 strip systems out there. It was Eastman’s most direct criticism about the process: it couldn’t reproduce the color blue.

Yet Capstaff was able to produce images with such consistency, he was sent to 20th Century Fox in Holly wood to further test the process. Fox even talked about shooting a feature in Kodachrome.

What was Capstaff doing? Why did he come back to Rochester and shoot this dance? Did he realize the significance of Martha Graham’s choreography?

I don’t think so.

After viewing the print of “Krishna” at the George Eastman House, which I will show you in a moment, you come to realize a person who was fluent in the language of dance did not produce this film.

The editing is very jarring, even at times, unintentionally humorous.

This was clearly a film with another purpose.

It hadn’t been created to preserve something.

Rather, it had been created to prove something.

Although many scientists at Research kept copious notes concerning their experiments, none of Capstaff’s thoughts or intentions surrounding this film exists.

Maybe a different approach to the back-story might help us out.

I thought it might be helpful if I went back and did a little research on “The Flute of Krishna”-not the film, but the dance, or the story itself.

I have to admit, until now, the only Krishna I had ever heard of was Hare Krishna and had prejudicial visions of those guys in airports-or worse yet, the movie “Airplane”. I had no idea where Google was going to take me. As it turned out the trip was quite pleasant:


Krishna, whose name means: ‘One who is always in transports of joy’ is a Hindu God of passion in a physical human form.

Krishna’s flute is an extension of his beauty. Not only is Krishna able to create with his flute the most beautiful sound imaginable, but it also imparts the essence of Krishna’s intoxicating nature.

In the myth of Krishna, men lose their wives, often referred as “gopis”, to the intoxicating strains of Krishna’s flute.

And it is important to understand the gopis are not, loose women; they are virtuous, but are jarred loose from their self-control.

The abandon of the cowherds’ wives symbolizes something bigger: a universal power of music and imagination.

Martha Graham’s choreography appears to be loosely based on the myth of Krishna and the gopis, but it also includes references to another story: that of his childhood friend, Rhada.

It is believed that Radha symbolized the spring of Love that continuously comes out from Krishna and returns to Him.

In the myth where Krishna dances with the cowherds wives,

Each Gopi believes that Krishna is dancing with her, and only for her.

This caused feelings of jealousy with Rhada.

She ran into the woods, with Krishna following her. Radha saw a beautiful flower on the branch of a tree.

She wanted the flower but could not reach it.

Krishna told Radha to stand on His shoulder to get the flower.

At that moment Radha felt very proud of her position and felt that she was better than everyone — especially the gopis.

And with that, Krishna disappeared—leaving Rhada alone in the woods.

Krishna could only be happy when Rhada believed that everyone is the same in the eyes of God.

Upon realizing this, Radha began to cry because she missed Krishna so much.

In the end, he comes back to her and all is well.

So at this point, we had a much clearer understanding about the passion behind this film as far as the dancers were concerned. This was pretty powerful stuff –but I still didn’t get the feeling it was the reason for Capstaff to make this film.

Maybe I was getting TOO into this –maybe he just liked pretty women.

Then I found this on the Straight Dope Website. For those of you who have never heard of the Straight Dope, it is a nationally syndicated column that people write to a guy named Cecil who will answer any question posed to him about any subject.

In this case, the word “Krishna” appeared in a rebuttal to something Cecil wrote concerning the question:

Is there really a race of blue people?

Dear Cecil:

You published several possibilities about the supposed race of blue people.   One such possibility was that of the Hindu god Krishna, who is usually depicted as having blue skin. You dismissed this as being the product of “cheesy color reproduction in those airport handouts.” While I will not defend the quality of the handouts made by the Hare Krishnas, I will point out that Krishna is traditionally shown with blue skin in almost all the art of India.

This is true. Another interpretation was since a God is considered infinite, like the ocean or the sky, he would naturally adopt the color of the infinite-which is the color blue.

This last bit of information turned on the switch inside:

I believe Capstaff was purposely trying to photograph something with a lot of blue in it – like the skin of a Hindu God, to show George Eastman that he was too quick to dismiss Kodachrome as a viable color process.

What kind of a person would do this: try to prove Eastman incorrect, or at the very least hasty in his conclusions.

Capstaff would.

He was the same guy, who in sold Eastman on the idea that it was possible to create a motion picture system for the amateur in 1916.

In a famous exchange between Eastman and Capstaff, Eastman makes the observation:

“Mr. Capstaff, you are asking me to invest money in a project which has met with commercial failure many times and yet, in view of all these failures, you still think I should invest in a new scheme for its popularization?”

“Yes, Mr. Eastman”, Capstaff replied, “because the Eastman Kodak organization can give the service necessary to make it completely successful”.

It was completely natural in Capstaff’s thinking not to prove the great businessman wrong, but rather to work on the problem until it could be solved – which in this case took about 10 years. By that time, the opportunity would simply present itself.

Martha Graham’s choreography was a perfect way to do this: there was color and movement. It was all inside, so he could have complete control over the process of filming.

And the fact that the dancers had fairly risqué outfits for the time didn’t hurt either.

What science backed up this experiment?

Again, there is very little hard data to support this except to try it out in the digital print process.

Theoretically, panchromatic film should be able to grab at least some blue data?

But how would recording and reproducing blue even be possible?

It stands to reason one cannot get a blue record from red and green filtration.

Or is it?

It is possible Captsaff didn’t use straight red and green filtration but instead tweaked the dyes to an orange / cyan mix. Essentially bringing blue into the filters and dyes themselves.

My first attempt to try this had these results:

Show Blue Skin for Krishna (Composite #2) from DVD

Notice the whites from before have now gone yellow, but other parts of the picture are starting to make more color sense.

Krishna is still not quite blue, but there is a lot more photographic detail in the over all image. The gopi’s outfits now have much more detail in them. The color was still wrong but the image was definitely improving.

The last step was to color correct for the whites-by adding a little more blue to the hue. To essentially color time the print.

And this was the result:

Show Eureka!

Then show enlargements of frames on DVD

Admittedly the color now is too good-too rich, however, the patterns of the main title look very accurate. Reds, greens, AND Blue.

Where there are whites-they are pure.

Skin tones are reasonable, detail in all of the cloth remains and, of course, Krishna is BLUE!

Even his lips are the right color.

I realize, given our 21st century arsenal of digital tricks, it is possible I was creating an image today that just was not possible in 1926. So to be fair, I tried to create rules of restoration grounded in a photochemical reality.

This was achieved by limiting image manipulation to only brightness, contrast and applying color changes to the hue controls of the color filters.

What I really needed was a second blind test-and I found it in the most unlikely of places. In a color process that was even older than Kodachrome.

Buried in some of old 16mm prints I had transferred was a Castle Film from the 60’s called “The 1890’s Live Again”.

Among the 35mm reduction print images was some footage that jittered and flickered very similar to the Kodachrome, but judging by its subject matter, this stuff was a lot older.

It turned out that this was actually a series of scenes shot on one of the earliest color processes known: Kinemacolor.

Show Kinemacolor demo from DVD

This is the footage as it was transferred. What causes the flicker here is Kinemacolor was shot at double projection speed, or 32 frames per second using a crude panchromatic film shot through a two-bladed shutter. Covering each opening of the shutter was a color filter in red, the other, green.

To create the illusion of color, a projector with the same type shutter as the camera, projected the images at 32 frames per second, adding the color back in to the record. As you can see, Kinemacolor rendered a very crude color image:

As I did before with the Kodachrome, I digitized this footage and separated the green pass and the red pass.

Then I used the same settings I developed for the first color composite of “Krishna”:

And then I color-timed the final composite:

Notice that things that are supposed to be blue are, in fact, BLUE! The water-the sky-and everything else looks correct;

Overlooking the overall cartoony appearance of the image, which I partially attribute to the degraded and over-printed original material.

It was about this point in my research when I learned that the George Eastman House was helping the National Film Preservation Foundation create a 3 DVD set devoted to illustrate the range and diversity of the silent era before talkies entered the scene.

The DVDs were to include short films that merited:


“The Flute of Krishna”.

Well, the work I had been doing on Krishna arrived a little late for inclusion in the DVD, which is too bad because the world thinks that the restored version of “The Flute of Krishna” should look, like this:

Show GEH restored from DVD

Note the colors of the title look reversed-and Krishna is not blue. He is green. In fact much of the image looks exactly like what Eastman was lamenting about. There is a lot of grain and scratches. Heavy contrast. This looks like what one would expect a film from 1926 to look like.

But we know better. We know that the equipment and film stock of that day was every bit as capable as the latest tools we have today to achieve exceptional photographic results.

The last item to be uncovered in my research that fully restores the original Kodachrome experience was with the identification of the composer that Martha Graham used: Cyril Scott.

All but unknown today he was considered the Father of British Modern Music in the early 20th century. He was also a devoted scholar of esotericism and Indian philosophy. Both of which are well represented in his compositions and arrangements.

After many hours of searching in the Silbey Library at the Eastman School of Music and the Martha Graham School in New York, the original music for this production appears to be lost.

But not all of Cyril Scott’s music has disappeared into obscurity.

One of his most famous pieces was called “Lotus Land, Opus 47, Number 1”.

In keeping with the spirit of the film: that dance needs music,

I used “Lotus Land”, performed by two different artists as a rough track, but it worked so well visually in the final composite,

I left it in.

And all of that effort has come to this:

Show Summary of process from DVD

You have all traveled with me on quite a journey and as a reward for sticking with me, I’d like to show you my final restoration of the film that lay undiscovered for over 70 years: “The Flute of Krishna”.

Show Director’s cut from DVD with full audio

Applause and roses. Then:

Before we leave Kodachrome I thought you might be interested in what happened to the process after “Flute of Krishna”.

Capstaff may have won the battle over reproducing the color blue, but he clearly did not win the war of developing a successful color motion picture system.

Eastman’s definition of a successful color system was one that could utilize a standard film camera with minimal or no change to the configuration of the camera itself. By 1928, there were thousands of motion picture cameras out there, many now in the hands of amateurs as well as professionals, and that is where Eastman felt the greatest impact of color could be made.

Kodachrome, like Technicolor, used a special camera to achieve its color reproduction. These expensive, one of a kind cameras made the marketing of the process very difficult, if not impossible.

Still optimistic, however, Fox Studios bought the technology from Kodak and renamed the process “Fox Natural Color”. The sale of the process allowed Capstaff to refine the process further, not to mention free up one of the most famous brand names in the history of photography.


Here are some of the last known scenes ever shot in Kodachrome, now the re-named Fox Natural Color process. Theses scenes seem to have been part of an attempt to shoot a feature using the process, but the production did not slate any of its shots and it is not clear what the plot of the film was.

Fox Studios would go on to test many looks for the process-without venturing into the areas of blue reproduction. Rather they would try to design their sets and wardrobe to capitalize on the color reproductive strengths of the process: reds and greens.

Sometimes with startling results (lumberjack).

Fox’s test were not limited to just color tests-


This is an early sound test from about 1929.

Note the clever slate:

As you see here, even Fox could see the wisdom of using the Far East as the basis for a story to be told using this color process.

By the 1930’s, many factors built up a case against the further development of this color process. “Becky Sharp” being one of them.

I did not realize that this project would expose me to so many seemingly unrelated parts of the human experience. Using 21st century tools, I reassembled a film that was shot to prove a theory, but in fact became the earliest know record of American Modern Dance.

But upon reflection, I believe it is that opportunity of discovery that propels each of us in the world of preservation and restoration.

For me, the story doesn’t end here. It is my intention to digitally restore this print, combine it with any additional surviving elements, and write the files back out to film.

Who knows what other treasures await discovery?

Thank you.

Tagged with: , , , , , , , , , , , , ,
Posted in Mike's Blog

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Cedarwood, Suite 380
6800 Pittsford-Palmyra Road
Fairport, NY 14450

Flickr Photos
%d bloggers like this: