logo Sign In

Color matching and prediction: color correction tool v1.3 released! — Page 26

Author
Time
 (Edited)

Thanks! Here’s a slight adjustment of the regrade, because the color matching algorithm (I use Technicolor print references, with some manual adjustments) was set with too much stabilization, resulting in dim blues:

Bluray:

Bluray regraded:

Author
Time

The colours are just fantastic. Seeing the comparisons makes you realize that the Blu-ray colour correction was done either by colour-blind people or by people who just didn’t care.

Author
Time
 (Edited)

The color correction tool v2.0 will be released soon, which includes a better color matching model. Here’s an example of matching the lobster frame to a Technicolor IB print reference:

Bluray:

Bluray matched to Technicolor IB print:

Author
Time
 (Edited)

DrDre said:

The color correction tool 2.0 will be released soon, which includes a better color matching model. Here’s an example of matching the lobster frame to a Technicolor print reference:

Bluray:

Bluray regraded:

I dunno, the new tool makes Han and Obi-Wan look kinda robotic.

Project Threepio (Star Wars OOT subtitles)

Author
Time

CatBus said:

DrDre said:

The color correction tool 2.0 will be released soon, which includes a better color matching model. Here’s an example of matching the lobster frame to a Technicolor print reference:

Bluray:

Bluray regraded:

I dunno, the new tool makes Han and Obi-Wan look kinda robotic.

I lolled

she/her
mwah

Author
Time

DrDre said:

The color correction tool 2.0 will be released soon, which includes a better color matching model. Here’s an example of matching the lobster frame to a Technicolor IB print reference:

Bluray:

Bluray matched to Technicolor IB print:

There’s definitely more red in the new one, but the artifacting appears to have been reduced.

Author
Time

Here are another set of frames of the bluray, that have been matched to Technicolor IB print references with the new color matching engine.

Bluray:

Bluray matched to Technicolor IB print:

Bluray:

Bluray matched to Technicolor IB print:

Author
Time

DrDre said:

Bluray:

Bluray matched to Technicolor IB print:

Is there that much crush in the prints?

Author
Time

Yes, prints tend to have more contrast, than the interpostive on which they are made. Blurays releases are usually based on a scan of the negative or an interpositive if they want to keep the original color timing. Prints get more contrasty with each generation, which can be seen on 16 mm prints, that are usually based on scanning a 35 mm print. The result is more black crush, and blowout.

Author
Time

DrDre said:

Yes, prints tend to have more contrast, than the interpostive on which they are made. Blurays releases are usually based on a scan of the negative or an interpositive if they want to keep the original color timing. Prints get more contrasty with each generation, which can be seen on 16 mm prints, that are usually based on scanning a 35 mm print. The result is more black crush, and blowout.

Okay. So this is going more for Tech print accuracy rather than day 1 intended colors?

Author
Time

Yup, these are just color matching tests. A proper color timing would probably have somewhat less contrast, and saturation, which is what I suspect Harmy and camroncamera are doing at this very moment. The soldier and Vader frames from a few posts back are my personal attempts at properly color timing the bluray, by color matching the bluray to the Tech with the software, and then manually adjusting them to the desired final result. I showed them to Mike Verta, to get his opinion on them, and he thought they looked good.

Author
Time
 (Edited)

Here’s a comparison between the bluray, the bluray matched to the Technicolor IB print, and the regrade based on the latter, with manual adjustments.

Bluray:

Bluray matched to Technicolor IB print:

Bluray regraded:

Bluray:

Bluray matched to Technicolor IB print:

Bluray regraded:

Author
Time

These shots are looking really good DrDre, it would be interesting to see how this looks when applied to the 35mm Silver Screen Edition.

Author
Time

Looks much better after your regrading. As poita has said, the green tint was likely due to poor quality control in the final days of the IB Tech production.

Author
Time

The “green tint” is due to misalignment of the color channels. It happens when the red and blue channels are perfectly matched up, but the green channel is slightly misaligned so it doesn’t blend properly with the other colors causing green fringe along the edge of things and in the shadows.

Author
Time

Darth Lucas said:

The “green tint” is due to misalignment of the color channels. It happens when the red and blue channels are perfectly matched up, but the green channel is slightly misaligned so it doesn’t blend properly with the other colors causing green fringe along the edge of things and in the shadows.

That accounts for fringing, but not for a green color cast across the entire shot.

You probably don’t recognize me because of the red arm.
Episode 9 Rewrite, The Starlight Project (Released!) and ANH Technicolor Project (Released!)

Author
Time

Well when the green channel is misaligned it would have the harsh green lines of fringing, but because of the misalignment, green would be showing through in gradients in places throughout the image where it is supposed to be blending with the other colors, which causes the green cast. You’ll notice you see it most in grey areas (death star walls, imperial uniforms) because that’s where you find the colors channels’ levels being mostly equal to each other, so when the green is misaligned or improperly treated (as I tend to suspect the Tech IB prints were in this case) it shows through alot more visibly. What people need to remember is this is a chemical dye process and little imperfections like channel misalignment and improper treatment of the elements affect the image in different ways than it would if you were working in something digital like photoshop or after effects.

Author
Time
 (Edited)

DrDre said:

Yup, these are just color matching tests. A proper color timing would probably have somewhat less contrast, and saturation, which is what I suspect Harmy and camroncamera are doing at this very moment. The soldier and Vader frames from a few posts back are my personal attempts at properly color timing the bluray, by color matching the bluray to the Tech with the software, and then manually adjusting them to the desired final result. I showed them to Mike Verta, to get his opinion on them, and he thought they looked good.

Was a little surprised to see my name mentioned, you guys have been doing all the R&D and heavy lifting while I sit back and read about your amazing work and lament not having the computational firepower to try anything myself.

Color tools have evolved a lot since my DaVinci/Spirit 2K telecine days and I would have something of a learning curve to climb to get back up to speed with all of you.

If your crop is water, what, exactly, would you dust your crops with?

Author
Time
 (Edited)

Here’s a couple of examples of regrading the Star Wars bluray, using the 35 mm print scans of the 1997 SE of team negative1 as a reference.

35 mm:

Bluray:

Bluray matched to 35 mm print:

35 mm:

Bluray:

Bluray matched to 35 mm print:

Author
Time
 (Edited)

Here’s a couple more examples of regrading the Star Wars bluray, using the 35 mm print scans of the 1997 SE of team negative1 as a reference.

35 mm:

Bluray:

Bluray matched to 35 mm print:

35 mm:

Bluray:

Bluray matched to 35 mm print:

Author
Time

Amazing DrDre as always. I love the potential of how this can be utilized in restoration projects.

Author
Time
 (Edited)

It’s interesting to note the color differences between the Tantive IV scenes on the bluray, the 1997 SE print, and the 1977 Technicolor IB print.

Bluray:

Bluray matched to 1997 SE print:

Bluray matched to 1977 Technicolor IB print:

Bluray:

Bluray matched to 1997 SE print:

Bluray matched to 1977 Technicolor IB print:

Bluray:

Bluray matched to 1997 SE print:

Bluray matched to 1977 Technicolor IB print: