Jump to content

What does Technicolor Really do? Dye Transfer, and then what?


Recommended Posts

I find the internet explanation of Technicolor 3 strip process inconclusive, and not very informative. 

Yes, scenes are filmed through a prism onto three strips of film at the same time. And then what? 

I'm assuming that is more than just stack all three and make a print, and Voila, you got this beautiful skin tone and print out of the box. 

So what does Technicolor - the company do after this. How do they mix/ grade the color to render the looks we know of.

And more importantly, why can't we duplicate the same look with later newer film stocks, let alone those far off digital emulations that just does not look the same, no matter how hard they try...

 

Does Technicolor have some kind of secret formula of mixing the chemicals to render those colors, like KFC's herbs for fried chicken? 

Link to comment
Share on other sites

At Technicolor they did quite a bit of on set QC before cameras even rolled with their color consultants, often to the point of fighting with the directors.

That said, the things you listed aren't trivial. A lot of the magic is in the camera and printing hardware. The film negative was just black and white, and the prints were the inks printed onto a blank. Just look at a documentary done on a three strip camera to see how the camera and dye process affected images in a unique way on their own, without on set tinkering.

If you're trying to replicate that look on digital cameras you won't get there with grading, LUTs or other post manipulation. Digital cameras just don't see a wide enough spectrum at the sensor level. You can play with the grade to try to make the image more colorful, but ultimately the colors you capture are tied to each other. Think of the image your sensor creates as a piece of fabric. You can stretch that fabric to a point, to make it bigger, but when you pull the edge hard enough the center moves too, or eventually tears.

With film stock you can get closer, but will still run into the same issue.

Background: I studied this for years and eventually developed (with my partner) hardware and a workflow to allow modern cameras to film in the three strip spectrum. We call it Optical Radiance. We have lots of examples at opticalradiance.com and on our Instagram.

Other people may chime in and point to the work done on "The Aviator" or David Mullen's wonderful work on "The Love Witch". I think both films are great, but neither looks like three strip. In fairness I don't think "The Love Witch" ever claimed they were trying to look like three strip, it was people reviewing it that kept using the phrase. And for "Aviator" Scorsese has said he was looking to evoke his own memory of what it looked like, he wasn't after being accurate.

Link to comment
Share on other sites

8 minutes ago, Andrew Schroeder said:

At Technicolor they did quite a bit of on set QC before cameras even rolled with their color consultants, often to the point of fighting with the directors.

That said, the things you listed aren't trivial. A lot of the magic is in the camera and printing hardware. The film negative was just black and white, and the prints were the inks printed onto a blank. Just look at a documentary done on a three strip camera to see how the camera and dye process affected images in a unique way on their own, without on set tinkering.

If you're trying to replicate that look on digital cameras you won't get there with grading, LUTs or other post manipulation. Digital cameras just don't see a wide enough spectrum at the sensor level. You can play with the grade to try to make the image more colorful, but ultimately the colors you capture are tied to each other. Think of the image your sensor creates as a piece of fabric. You can stretch that fabric to a point, to make it bigger, but when you pull the edge hard enough the center moves too, or eventually tears.

With film stock you can get closer, but will still run into the same issue.

Background: I studied this for years and eventually developed (with my partner) hardware and a workflow to allow modern cameras to film in the three strip spectrum. We call it Optical Radiance. We have lots of examples at opticalradiance.com and on our Instagram.

Other people may chime in and point to the work done on "The Aviator" or David Mullen's wonderful work on "The Love Witch". I think both films are great, but neither looks like three strip. In fairness I don't think "The Love Witch" ever claimed they were trying to look like three strip, it was people reviewing it that kept using the phrase. And for "Aviator" Scorsese has said he was looking to evoke his own memory of what it looked like, he wasn't after being accurate.

I'm still wondering about the process of what Techicolor does after the film is exposed. 

But you reply makes me questions a bit future that is not really revenant to this post in particular: What do you mean by colors are tied together? A digital camera sensor is all monochrome, it's the Bayer layers on top that filters the color of Blue, Red, and Green. Similarly, a color film stock operates the same way, with three emulsion layers of Yellow Cyan and magenta. 

So essentially, they are all  3-colored process stacking on top of each other. Are you saying, if I film the same scene with 3 monochrome digital cameras, with each prime-colored filter, and recombine the same scene on top of each other it will solve the problem? 

Link to comment
Share on other sites

In your research have you not yet discovered that the old Technicolor process no longer exists?   All those labs closed long ago.

The images coming out of a Technicolor 3-strip camera were black and white.  The colors were added in the lab by a complex dye-imbibation process, which is why they were so saturated and beautiful, not to mention fade-proof.

Other photochemical processes are single-strip and rely on color layers in the emulsion which appear during development, and although we get good results, nothing equals the old Technicolor process.

 

Link to comment
Share on other sites

7 minutes ago, dan kessler said:

In your research have you not yet discovered that the old Technicolor process no longer exists?   All those labs closed long ago.

The images coming out of a Technicolor 3-strip camera were black and white.  The colors were added in the lab by a complex dye-imbibation process, which is why they were so saturated and beautiful, not to mention fade-proof.

Other photochemical processes are single-strip and rely on color layers in the emulsion which appear during development, and although we get good results, nothing equals the old Technicolor process.

 

yes, I'm trying to dig out where this technique is documented and saved... but no luck, all the info I can find stops at "a complex dye-imbibation process" without knowing what exactly is it!! 

I don't mean to split hairs, but I think it helps me to understand color better. right now I simply don't understand why film has the film look, digital has the digital look, and no matter how hard I try, color grading just does not match it. 

Link to comment
Share on other sites

  • Premium Member

Technicolor Movies : The History of Dye Transfer Printing

Using extensive research and interviews with many of the surviving Technicolor technicians, the history of dye printing and the events leading to its demise are fully covered. (The Beijing Film Laboratory is the only facility currently using the process.) Included are diagrams of how the process worked and an extensive listing of U.S. feature films printed with it.

 

https://www.thriftbooks.com/w/technicolor-movies-the-history-of-dye-transfer-printing_richard-w-haines/1580854/item/42471286/?gclid=CjwKCAjwx6WDBhBQEiwA_dP8rdgKyRnGKVTCHdjrtc-KYbVMASGcw1oxWF49qbLvnMaoBjVdUZ0HaxoCsTkQAvD_BwE#idiq=42471286&edition=58111917

 

DAWN OF TECHNICOLOR: 1915 1935 By James Layton & David Pierce - Hardcover *NEW*

https://www.ebay.com/itm/184739959673?chn=ps&mkevt=1&mkcid=28

Edited by Frank Wylie
  • Like 2
Link to comment
Share on other sites

3 hours ago, Wendy Sanders McDonlad said:

I'm still wondering about the process of what Techicolor does after the film is exposed. 

Well, after its exposed the negative was developed and a positive printing matrix strip was made for each strip. The matrix strips were kind of like stampers coated with their individual dyes. They printed each color on top of each other on a blank. For earlier 3 strip films they also added a fourth half tone black and white image over that, made from the green channel, to add contrast.

3 hours ago, Wendy Sanders McDonlad said:

But you reply makes me questions a bit future that is not really revenant to this post in particular 

You asked why cameras/film can't do this anymore, so I believe its a relevant answer.

3 hours ago, Wendy Sanders McDonlad said:

What do you mean by colors are tied together? A digital camera sensor is all monochrome, it's the Bayer layers on top that filters the color of Blue, Red, and Green. Similarly, a color film stock operates the same way, with three emulsion layers of Yellow Cyan and magenta. 

Unless a sensor specifically says its monochrome (meaning it has no CFA) it is a color sensor. The sensor stack, including the CFA, should be considered a part of the sensor. In bayer sensors all 3 color channels are interpreted by comparing photosites against adjoining photosites. Green and Blue values are needed to determine what the reds look like, and vice versa for the other channels for the image. So they're essentially tied to each other. Even if you work with RAW files you are working with them after theyve been demosaiced. You can use different programs to get different demosaic algorithms, but the colors will still be tied to each other.

Film isn't quite as restricted, but the colors are still tied together by the layers that create them

3 hours ago, Wendy Sanders McDonlad said:

So essentially, they are all  3-colored process stacking on top of each other. Are you saying, if I film the same scene with 3 monochrome digital cameras, with each prime-colored filter, and recombine the same scene on top of each other it will solve the problem? 

You could do that. But you would need some kind of beamsplitter like the three strip camera's prism, so that all 3 sensors received the same image and also lenses that would work properly with that set up.

3 hours ago, Wendy Sanders McDonlad said:

And besides, the 3-Strip process, is really a 2-strip... since the blue and red are stacked on top of each other.. only the green strip is exposed on it's own. 

The three strip process was really three strip. It had two different light pathways after the prism, but the red and blue strips were sufficiently separated and recorded on different pieces of film.

Link to comment
Share on other sites

3 hours ago, Andrew Schroeder said:

 

You could do that. But you would need some kind of beamsplitter like the three strip camera's prism, so that all 3 sensors received the same image and also lenses that would work properly with that set up.

 

I wonder if anyone has attempted with this experiment. 

Link to comment
Share on other sites

  • Premium Member

The differences between three-strips Technicolor and digital imagery are

  • TC is a subtractive process with yellow, magenta, and cyan separations filtering white light. Digital pixels are red, green, and blue to mix after the additive principle.
  • The white light behind a TC print was from a discharge arc that brings radiation from ultraviolet through infrared, a broad band. On an LCD or in projection digital colour images appear white by the very narrow band RGB’s mixture. Our retinae are sensitive to red, green, and blue. That’s the reason for the strong appeal DI has. Still more exact and saturated imagery is feasible with LASER.
  • The TC three-strips camera produces three colour separations. The fact that the blue and the red separation are made on two films in contact with each other does not change the fact that the blue sensitive stock wore a red filter layer on top of it. Red and blue were thus recorded in separate planes. Compared to an electronic light sensor the separations are found several thousands of an inch apart along the optical axes.
  • A TC print could be made with dyes quite different from those in multilayer colour film with some consequences.
  • A TC print was generally projected by a Petzval design lens (four elements) or, to the better, by a double-Gauss system (mostly six elements). Color rendition, sharpness, resolving power, and contrast of the picture varied among theatres. But the very point of film projection is the light source.
  • Digital colour imagery comes into being through a grid or raster. That needs to be fine enough relative to the viewing distance in order to stay invisible. Contrary to photochemical film images that appear as random silver-wool granulation or colour-dyes in clouds the just-not-discerned rigid raster always leaves a glassy smooth impression, the so-called plastic look.

Film projection and video can’t be compared. Film recording and video recording also not. As long as you haven’t seen a TC print projected in a traditional cinema you won’t understand. Sorry, if this comes across harsh, but it is so.

Link to comment
Share on other sites

  • Premium Member

The bipack design in the 3-strip camera created some unique artifacts since the blue record on top had no anti-halation backing but was dyed red to act as a filter for the red record underneath, which tended to be the softest record as well. So you ended up with magenta halation around bright areas. 

The dye transfer print process is a whole other system that also had a look.

If one were going to digitally simulate the camera process, one thing that would help is to make the red channel softer and grainier before recombining them.

”The Aviator” took their RGB scans of the color negative film used and used filters (digital or optical?) to reduce the normal crosstalk between layers, apparently their study of 3-strip showed that the color records were a bit more separated, giving the primaries more saturation and purity but probably also causing some color shades to not reproduce as well (hence all the testing that old Technicolor movies did for costume and make-up). You need some color crosstalk to capture the in between shades. Which brings up the other point, that some of the look was due to the control exerted in front of the lens. People talk about Technicolor skin tones but that was as much a result of Max Factor’s work as it was the camera and print process.

  • Like 1
Link to comment
Share on other sites

  • Premium Member

Regarding the yellow, cyan, and magenta dyes used to make the prints, they might have been tweaked over time, and perhaps even for a special project, but the nature of printing means that you can't adjust the dyes for a scene, create a special shade of cyan for example, the entire reel would use the same dyes and in fact all the reels would, you can't change out the baths per reel and not really for a single feature being printed.  So that aspect was fairly standardized.

However, the process was prone to being too high in contrast so Technicolor flashed the b&w matrices I believe to some minor degree so I guess some control was possible there if you wanted the movie to be harsher and more saturated versus softer and more pastel (again, per reel, not per shot). And then there was the unique printing processed used half the prints for "Moby Dick" (1956), shot on Kodak color negative but released in desaturated Technicolor dye transfer prints.

Link to comment
Share on other sites

6 hours ago, David Mullen ASC said:

Regarding the yellow, cyan, and magenta dyes used to make the prints, they might have been tweaked over time, and perhaps even for a special project, but the nature of printing means that you can't adjust the dyes for a scene, create a special shade of cyan for example, the entire reel would use the same dyes and in fact all the reels would, you can't change out the baths per reel and not really for a single feature being printed.  So that aspect was fairly standardized.

However, the process was prone to being too high in contrast so Technicolor flashed the b&w matrices I believe to some minor degree so I guess some control was possible there if you wanted the movie to be harsher and more saturated versus softer and more pastel (again, per reel, not per shot). And then there was the unique printing processed used half the prints for "Moby Dick" (1956), shot on Kodak color negative but released in desaturated Technicolor dye transfer prints.

This is some insightful information. I wonder if there was any case they specifically made short-end reels just so they can tweak them scene by scene instead of the whole reel. But that would be splitting hairs on information got lost a long time ago. Thanks again David. 

Just like everything else, I find the whole coloring/filming process always started with a structure with more control and freedom, and gradually moved to more and more restrictive formats to color stocks, to digital... much like the political atmosphere, and the internet... something tells me it's not a coincidence.  

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...