Estimating the original colors of the original Star Wars trilogy

This thread is about reconstructing the colors of the original Star Wars trilogy, specifically the first Star Wars and The Empire Strikes Back. To this end I developed an algorithm to retrieve the colors of faded color prints without using a color reference, in an objective manner, and therefore unaffected by subjective opinions of what the films should look like. The theatrical colors of the original unaltered Star Wars trilogy are used as the central theme of this thread, as these are heavily debated, and in the case of The Empire Strikes back are virtually unknown, since most 35mm and 70mm prints of the era have faded. For Star Wars the only unfaded 35mm references are the surviving Technicolor prints, but their accuracy as a color reference for the average 35mm print of Star Wars that was projected in 1977 in theatres around the world is debatable.

As a demonstration of the algorithm’s capabilities, I corrected the uncorrected scan of a low fade print done by Team Negative1 for their Silver Screen Edition with the algorithm. I applied a global correction to the entire reel, without manual adjustments. Although a gamma correction is in order, and many shots can use some additional tweaking, I think the results give a pretty clear impression of the colors of the original 1977 Star Wars.

Here’s the video sample that showcase the reconstructed colors:

https://drive.google.com/file/d/0B8_LYKyZDiajZzByMFdrUjhNVHc/view?usp=sharing

This comparison shows the state of the print scan before and after the color reconstruction:

…and here are some additional screenshots:

================================================================================================

Original start of the thread:

For a while now I’ve been working on figuring out an algorithm, that will automatically color correct a print, and estimate it’s true colors. Although there are still some things to figure out, the outcome so far looks very promising, so here are a few preliminary results (top raw, bottom corrected):

Oh my god this is fantastic DrDre!!

CatBus said:

Does it deal well with filtered shots? i.e. the R2 canyon scene, which has an overall reddish cast over the whole thing, because it was filtered after it was shot. If so, how can you tell the difference between pink shift and red filter, or do you use neighboring frames somehow to see that this section is supposed to be redder than the others?

That´s something I´m still figuring out 😃.

Wow, that looks really great.

For the faded Eastman prints, can you apply those settings to the entire reel and get as good of results?

towne32 said:

Wow, that looks really great.

For the faded Eastman prints, can you apply those settings to the entire reel and get as good of results?

That’s another question that needs answering.

I think we now know the Death Star conference room walls have a very slight blue cast, while the chairs are gray, and we have the color of Tarkin’s clothes.

DrDre said:

towne32 said:

Wow, that looks really great.

For the faded Eastman prints, can you apply those settings to the entire reel and get as good of results?

That’s another question that needs answering.

Can you do similar tests on the (amazing vanishing) ESB Grindhouse release? Or is there not enough recoverable color information?

towne32 said:

DrDre said:

towne32 said:

Wow, that looks really great.

For the faded Eastman prints, can you apply those settings to the entire reel and get as good of results?

That’s another question that needs answering.

Can you do similar tests on the (amazing vanishing) ESB Grindhouse release? Or is there not enough recoverable color information?

I’m on the road, but I will try to run some tests, when I get the chance. The Grindhouse should be easy compared to the red shifted TESB examples I posted.

It also appears those Death Star elevator walls have no green in them, as Mike Verta said.

DrDre said:

I think we now know the Death Star conference room walls have a very slight blue cast, while the chairs are gray, and we have the color of Tarkin’shot clothes.

Not gonna say I told you so 😉

Just teasing haha. Great work as always DrDre.

It also appears Tarkin’s hair has more brown, than we thought…

DrDre said:

towne32 said:

DrDre said:

towne32 said:

Wow, that looks really great.

For the faded Eastman prints, can you apply those settings to the entire reel and get as good of results?

That’s another question that needs answering.

Can you do similar tests on the (amazing vanishing) ESB Grindhouse release? Or is there not enough recoverable color information?

I’m on the road, but I will try to run some tests, when I get the chance. The Grindhouse should be easy compared to the red shifted TESB examples I posted.

Color corrected ESB Grindhouse would be THE BOMB!!! Your technology is amazing!

This could really be a valuable asset.

That last shot of the Emperor looks too warm with possibly a yellow/orange cast to it. I’m not on my color managed display, but all the other images looked really nice, except that last one stood out to me.

camroncamera said:

DrDre, I don’t think I’ve read you mentioning it… have you heard of the Digital ROC color-restoration processing plug-in by Kodak subsidiary Applied Science Fiction?
http://www.asf.com/products/plugins/rocpro/pluginROCPRO/
I don’t know that the product has been updated for many years. My Nikon film scanner came bundled with a version, but it only works during the scanning of film. Worked fairly well. Reminds me of your work on restoring color to red-faded motion picture film scans.

I’ve never heard of it, but there are probably many similarities.

guiser said:

That last shot of the Emperor looks too warm with possibly a yellow/orange cast to it. I’m not on my color managed display, but all the other images looked really nice, except that last one stood out to me.

That shot has a delibirate blue cast, while the algorithm neutralized the colors. This is why it is probably better to correct multiple shots simultaneously.

Mavimao said:

Wow! How does this even work??

Short summary:

• Convert RGB space to CMYK space (color space for film)
• Align color channels by maximizing color channel correlation

Here’s a video example, using the badly faded 70 mm TESB sample posted by -1 a while ago.

Original:

https://drive.google.com/file/d/0B8_LYKyZDiajWFIwNU9UcFJqRG8/view?usp=sharing

Automated correction:

https://drive.google.com/file/d/0B8_LYKyZDiajOFg2aklhUlA4RG8/view?usp=sharing

Here’s a comparison between the results of the automated color correction algorithm for the ROTJ LPP, and the ROTJ Grindhouse, that features Harmy’s manual color correction. To make things clear, there’s no manual adjustmens in my corrections.

LPP:

Grindhouse:

DrDre:

LPP:

Grindhouse:

DrDre:

LPP:

Grindhouse:

DrDre:

F*** me Dre, stop being so awesome! I would worry about correcting all the frames individually, as I bet you would introduce flicker, but if you averaged the resulting LUT from several frames/shots and applied the same correction to a whole shot/reel, you really would have something there.

I would even argue that if you did it right, it would be just as valid of a color reference as a Technicolor print, or the JSC, which are both slightly wrong, but your result wouldn’t be any more wrong.

-G