DrDre

I finally managed to extend the rotations method to be able to predict different frames. Most of the color corrections shown are based on a very similar reference. However, I believed that it should be possible to use the color correction model calibrated on one frame, to predict the color correction for frames in the same scene, as these have usually been altered in a similar way or have faded in a similar way. I tested my method on the bluray/digital download, and this is what I get.

First I calibrated the model on a 720p frame:

http://screenshotcomparison.com/comparison/140900

Then I predicted the correction for this 720p frame:

http://screenshotcomparison.com/comparison/140782

The results of the prediction are:

http://screenshotcomparison.com/comparison/141323

The prediction is a bit brighter, but it is very close.

Next I calibrated the model on another 720p frame:

http://screenshotcomparison.com/comparison/140902

Then I predicted the correction for this 1080p frame:

http://screenshotcomparison.com/comparison/141262

The results of the prediction are:

http://screenshotcomparison.com/comparison/141324

The prediction is a bit darker, but it is again very close.

So, to summarize, here are the color corrections, based on a color correction model calibrated on a different frame:

http://screenshotcomparison.com/comparison/141325

http://screenshotcomparison.com/comparison/141326

Or if anybody prefers any of the other color gradings from the GOUT, JSC, 97 SE, and 2004 DVD, based on Althor's compilation:

We aim to please:

In principle it is possible to retrieve the colors with a color matching algorithm, provided you have a good reference. In this case I matched the colors of the print to the frame posted previously from the Semi-Specialised Edition V2.2 (bottom):

Here's a set of color corrections at 1080p. Some are pretty awesome, if I say so myself ;-):

http://screenshotcomparison.com/comparison/141230

http://screenshotcomparison.com/comparison/141233

http://screenshotcomparison.com/comparison/141235

http://screenshotcomparison.com/comparison/141236

http://screenshotcomparison.com/comparison/141260

http://screenshotcomparison.com/comparison/141261

http://screenshotcomparison.com/comparison/141262

In a few days there will be another video sample for SRV14. I've found to enhance the detail, while keeping the grain reduction. 

Ps. Haven't had time to clean up the scripts for SRV12 & SRV13, but those will be posted in the next days also. 

Here's a video sample for SRV13:

https://drive.google.com/file/d/0B8_LYKyZDiajRlpJeHdldVh5ejA/view?usp=sharing

SRV13 is softer than SRV12, but does not have the pervasive static grain.

The strange orange blotches in the Alderaan astroids/tie fighter chase scene in Harmy's Despecialized Edition appear to be mostly the result of the color correction, and not a makeup issue or color timing issue of the bluray itself:

http://screenshotcomparison.com/comparison/140960

The video sample is done. I just have to add sound to it, and convert it to mkv, so it should be online this evening (European time).

The model I use automatically adjusts the saturation, tint, contrast, and brightness by estimating how the three color channels in multiple color spaces have shifted. If the original frame has a lot of contrast, it can become darker, and if it has a lot of highlights, it can become brighter. 

One thing I noticed is, how much depth is added to the frames. The bluray colors are really flat. 

It's also amazing how screwed up R2's blue panels look in the bluray. In that last frame Han's pants are also far more blue, than they were originally. Here's a prime example of how screwed up R2's colors are in the bluray. In this frame he simply looks awful in the bluray:

http://screenshotcomparison.com/comparison/140906

Three more great shots from Star Wars color corrected:

http://screenshotcomparison.com/comparison/140900

http://screenshotcomparison.com/comparison/140901

http://screenshotcomparison.com/comparison/140902

I haven't followed the thread, but in principle faded colors are not an issue, if the scan has enough color depth. You would need some color references, preferably as many as possible, for the best results. The reference can have a SD resolution. 

Here are three more corrections to the bluray. The difference for the first frame is huge:

http://screenshotcomparison.com/comparison/140778

http://screenshotcomparison.com/comparison/140781

http://screenshotcomparison.com/comparison/140782

Here are corrections for three more bluray frames, including the notorious Mike Verta frame:

http://screenshotcomparison.com/comparison/140675

http://screenshotcomparison.com/comparison/140677

http://screenshotcomparison.com/comparison/140679

There are still a few pink highlights in the last frame, but it's much closer to the original colors, and perhaps a few more interations will get rid of those highlights. 

Here are some improved color corrections:

http://screenshotcomparison.com/comparison/140656

http://screenshotcomparison.com/comparison/140657

http://screenshotcomparison.com/comparison/140658

http://screenshotcomparison.com/comparison/140659

http://screenshotcomparison.com/comparison/140661

http://screenshotcomparison.com/comparison/140662

http://screenshotcomparison.com/comparison/140663

Here are a few preliminary bluray color corrections. The crushed blacks of the bluray are an issue, but you get the idea:

http://screenshotcomparison.com/comparison/140626

http://screenshotcomparison.com/comparison/140627

http://screenshotcomparison.com/comparison/140629

http://screenshotcomparison.com/comparison/140630

http://screenshotcomparison.com/comparison/140632

http://screenshotcomparison.com/comparison/140633

http://screenshotcomparison.com/comparison/140635

Glad to here you're still here, Poita! Those scans look great! Good colors too! 

Yes, I'll clean up the script and post it along with the sample.

Another video sample will be ready in three days. I believe I've managed to get rid of the intrusive grain, but at a price of course. However, I believe it can be best judged in motion. I will post a sample for both SRV12 and SRV13 based on towne32's cleaned up GOUT, and then it's time for critique ;-).

Possibly, I will do some tests first to see how it works out.

Looks quite good indeed! I will be sharing the script soon.

I think the problem is, that the histogram matching method assumes that the individual color channels can be adapted independently. Therefore each value has a unique mapping. Although this usually is a good approximation, in principle there can be multiple mappings for each RGB color.

That's great. Actually, it's a many to many mapping, not a one to many mapping. You should save the rotations, and the mapping in each color space, and then apply each of the rotations and mappings to the new image, transform the individually mapped images back to your original color space, and average all the predicted images. 

Update: there's a bug in my code, so mapping multiple color spaces for different images/frames is not yet possible.

Great discussion, and very informative! Mike is such a professional.

The Legacy frame I used has a preliminary color grading. I was aware of that fact. I personally like the more saturated look with more contrast. It also matches the Senator print photos pretty well, but the color grading is by no means definite.   

The new samples posted by -1 look terrific, so the color correction on the old video sample is outdated anyway. It mostly serves as an example what can be done with a faded source, and a limited number of reference frames.