I always knew they were inaccurate, but I never really had a problem with the blu Ray colors until I read this thread and saw how much better they could look. I love the improved contrast/white levels even more than the colors tho. White cores!
The worst blu Ray shots are the ones (such as the binary sunset scene for example) where they boost certain color channels, usually red and/or blue. It makes skin look like a glowing sunburn and R2 look neon. I look forward to seeing what you do with shots like that.
I think this still from what appears to be a Technicolor print probably captures what the Tarkin shots should look like imo (aside from too mucg green), and which ended up being the main reference source for my latest regrade:
That still is not from a Technicolor print, it is from a 70mm cell scan that was posted on this site: http://www.jedi1.net/.
Here’s the full image uncropped:
That site is interesting. Why do only the Jedi cells seem to have red fade? I though the 70mm prints of the first two were on fading stocks.
Are the other scans just heavily modified to remove the red fade, but they didn’t do so with Jedi? The color in some of the SW scans looks like it’s in good shape.
Those scans are definitely not from original exhibition prints; in the late 90s Lucasfilm distributed some limited edition film cells through the Star Wars Fan Club IIRC.
Here is a video about them: https://www.youtube.com/watch?v=bdfPhRXn_pU
On the back there’s some info “…hand cut from a master print taken from the original internegative…”. While I doubt the original 65mm internegatives for SW and TESB would have retained this much colour, the scans don’t appear modified. It seems that Lucasfilm does (or did who knows) have the originals on low-fade film stock. Oddly the cells for SW seem to resemble the Technicolor prints as they have similar colour shifts which leads me to suspect that a Tech print was involved at some point in the process.
That’s great! I can always use more references, and these are in pristine condition. If these 70 mm scans are accurate, and they appear to be, this is how the regraded Tarkin shots should look:
I think there’s too much green in that. It’s important to note that no matter what the colors of the walls or the lighting off camera, those lights behind Tarkin should be a fairly pure white, but I’m seeing a fair bit of green in them around the edges.
They needed a very slight correction for the white balance, but they’re pretty much the same, so I think they’re pretty accurate, and those walls are definitely not gray, but green.
Come to think of it, I think, I agree with darth lucas, that there’s too much green in the print, so I’ve decided to stay close to my original grading for this scene. There’s still some teal in the walls, but not as much, and the overall feel is a bit cooler than on the print, which I think is appropriate for the scene:
After reading the previous pages, I have to ask “the public” : What are you trying to achieve? Are you trying to get “the colours you prefer”, “the correct colours (whatever that means)”, “the most natural looking colours” or what?
The thing is everybody likes different colour palettes. There is no such thing as a colour palette which EVERYBODY would like so these discussions could go on forever with no result whatsoever.
Are you trying to achieve the look which George Lucas and Gilbert Taylor? Well good, but don’t forget that
Are you trying to make everything have the “real” colours? Then OK, production photos are PERFECT for this, but you are completely ignoring the original filmmakers “vision”.
That’s why I wanted to know what is the “goal” of this, because I have a feeling that Dr.Dre is trying to achieve something different than other people.
From my part I can say my goal is to approach as much as possible what was on the original interpositives. This is why the print scans, and film stills are to me the most reliable color reference, both in terms of color, contrast and brightness, although it should be noted that prints tend to have more contrast and saturation, which I try to take into account. I sometimes look at the JSC, because it was telecined at a time, when the source was still in good condition, and it shows in the colors, but I’m very much aware that capturing the colors with the telecine process is approximate at best. I would never trust the GOUT, because it was made at the time when the source material was in dire straits, which means they had to re-interpret the colors pretty much from scratch. The GOUT obviously has a revisionist 90s grading with it’s very pink skin tones. Of course, there’s always personal interpretation in trying to correct problems in print scans, film stills, etc, so my color grading will always be slightly different from somebody else’s color grading. I try to stay true to the source material, and hope that you guys will continue to call me out on shoddy regrades, such as pittrek and Harmy have done.
After seeing those 70 mm frames, I went back to the Owen/C-3PO frame, to try to get as close to the original color timing as possible. This also gives me an opportunity to share my color grading process. I don’t have an accurate reference for this shot, since the Technicolor scans have some issues for these scenes. However, I do have the uncorrected preview for reel 1, and consequently this frame for -1’s scan of the Spanish LPP, and another 70 mm reference frame of the same sequence:
70 mm frame:
LPP frame:
I use the 70 mm frame to correct the LPP frame:
Then I take the -1 frame I’m interested in:
I subsequently correct this frame with the color correction model I’ve optimized for the other frame:
Now I can match the bluray to this corrected frame:
Next, I correct the contrast, saturation, brightness, and white balance, resulting in the final color grading:
Given the process I’ve used, I believe it should be pretty accurate.
Can’t say I understand the process when it comes to the final corrections, but it looks pretty nice to me if a little high contrast.
Then again, I don’t think I’m quite as picky about this shot as most of you. 😃
Question : why add this LPP step instead of matching the bluray to the 70mm for the first frame ?
Question : why add this LPP step instead of matching the bluray to the 70mm for the first frame ?
The goal was to regrade the Owen/C-3PO frame. The assumption of the method is, that the degradation of the colors of these shots is very similar for the LPP independent of the shot. Therefore, the correction for one shot should still be pretty accurate for another shot, and can therefore be transfered to another shot. Since, the bluray was regraded on a shot by shot basis, the regrade for one shot cannot be transfered to another shot.
Can’t say I understand the process when it comes to the final corrections, but it looks pretty nice to me if a little high contrast.
Then again, I don’t think I’m quite as picky about this shot as most of you. 😃
I’ve reduced the contrast somewhat. 😃
… to get as close to the original color timing as possible. … correct the contrast, saturation, brightness, and white balance, resulting in the final color grading … it should be pretty accurate.
Thanks for your work-flow! That is always of major interest to me.
I did discover something about program processing over at the 2001: A Space Odyssey thread. Standard picture adjustments (as opposed to individual R-G-B adjustments with the same processing numbers) have a built-in color bias that actually distorts the color . .
The justification for this has something to do with how color is supposed to look on hardware. But notice the resulting “pink-ish” bias from the standard adjustment compared to the expected, clean coloring from individual R-G-B adjustments (it’s the same numbers – only RED is showing here on the drop-down menu, but GREEN and BLUE have the same settings). This might be skewing all our projects. (You may want to test to see if/how affects your processing.)
| … oh, and yes, I do love R-G-B! |  |
 |
What’s interesting is that the latest color grading for the Owen/C-3PO shot is very close to the previous color grading, that I arrived at using a still frame I found on the internet as a reference, using a similar process to correct the Technicolor IB print scans:
Previous color grading:
Latest color grading:
So, I’m confident the current grading is accurate. I think we can put this to bed 😉.
Spaced Ranger said:
… (You may want to test to see if/how affects your processing.)
… oh, and yes, I do love R-G-B!
Thanks for that bit of info! I luckily always use at least 20 color spaces, including RGB to do my processing.
Love this final!!!
Here’s the shot where Admiral Motti is being choked.
Bluray:
Bluray regraded:
- production photos are USELESS FOR THIS
I disagree. While production photos are rarely indicative of what the actual scene looked like in the film, and you should not be trying to MATCH production photos by any means, what they are great for is giving you a sense of the natural color of sets and costumes, which is very useful in trying to understand things like just how green Tarkin’s uniform should look under certain lighting. It’s not a source to be matched by any means, but they’re still useful as reference when all of the sources you are trying to match contradict each other in some way.
Just for comparison, here’s how Admiral Motti looks on a 70 mm reference:
versus the regrade
There’s some weird splotchiness going on in his uniform on that one.
There’s some weird splotchiness going on in his uniform on that one.
Yes, I see it too. I will correct it. Thanks!
