mverta

Speeding to Mos Eisley

There's something more to it; I've never seen another shot quite like it, because the ratio between the starfield and the explosion is utterly different on those laserdiscs than on any film source I've seen.  Even if you crushed down the luminance on the explosion - which still wouldn't yield anything close to what we see on the LDs - the starfield would be utterly wiped out, and we know that these LD's don't have spot exposure areas.  So, ultimately, we know that on the negative, all that is in there, but it never made it to a theatrical print, and only the LDs have it - even the 2006 OT DVD doesn't have it, though it does have a yellow coloring.  It's truly unusual on many fronts.

Yes my question wasn't whether it was on every print or not - it is - but whether it's a tear on the negative or original elements, that is damage vs. limitation.  If it's damage, then I don't care if it's on the negative or not, it's outta here.  But as I said in the video, I don't suspect it's damage, rather limitation, which is why I have the version with it in.

In fact, here is the version with it in, and with the actual colors seen on all prints:

TIE_on_LUKE_5_GRADED

Good Shooting, Wedge...

Spaced Ranger said:

What software do you use for this work -- commercially available?

  I'm using a combination of proprietary tools (registration, temporal data harvesting, tracking, etc.) and commercial software (Cinnafilm's Dark Energy Pro, PFClean, Resolve, Nuke, even After Effects).  I'm working in DCI-P3 colorspace, primarily on a Eizo CG277 monitor, and then about once a week previewing on a Dolby PRM-4200 and Christie 4K projector.

_Mike

X-Wing Flyby 1 & VFX Breakdown

For sure the vanishing lines don't match, and the painting is not "shot" with the same lens as the live plate.  That's not really arguable; the issue is does this kill the composite.  If I'm the compositor on this plate, even with this mismatch you can go a lot farther to selling it with proper internal balance.  The audience's attention and length of shot pretty much save it.  But absolutely, yes, you need only extend the lines of the courses in the predominant wall and see they're not aligned with the live plate.  I've just personally had WAY worse mismatches to deal with in composites before, and it's amazing what you can get away with in certain contexts!

The problem with the matte painting is not the geometry; symmetry in design is not a given.  The problem is compositing 101, whereby the internal luminence curve and incorrect ambient bounces aren't aligned with the live action plate.  If you had the elements separately and could recomposite them, you could sell this shot without changing the painting.  That said, it is important to note that part of the loose style of these paintings requires the obfiscation of its nature by film grain; it is the reduction of the grain which really accentuates the textural differences between the glass and the live-action.  When this shot is finaled, with the grain returned and balanced for Rec.709/sRGB, you'll see it really works much better.

ww12345 said:

So regarding the camera work on that dive, in your opinion does that make for a more realistic shot than say the one used in the new teaser trailer for VII? I agree - I think it's kind of like CGI in general, that if it's not used to do something that "could" happen that your brain sort of discounts it as fake without you even knowing it...

When I'm doing VFX work, I'm careful about the aggregate amount of non-real in a frame, which is more important than any one single element.  For example, the entire idea of a spaceship flying and shooting lasers is "non-real" essentially, so I consider it important to have a lot of things in the camera motion and lighting and physics which are based in reality to help compensate for that.  Too much and the entire thing becomes unreal, and cool as it may look, our brains absolutely disconnect from the experience.  This is part of why our films today are insanely cool looking, and totally forgettable; it's why they need a new Avengers movie every year or whatever, whereas we're still trading in on Star Wars almost 40 years later.  Your average frame of a sci-fi/action movie today is almost entirely artificial, from the CG environments and unbounded cameras to the impossible physics, and having every element color graded separately so they no longer interact with each other.  The result is almost painterly and surreal, and usually fantastic looking, but also so utterly non-real that we're unavoidably detached from it.  The human brain is the greatest bullshit detector in the universe when it comes to visual comparative analysis. We can't always tell what's wrong with an image, but after a lifetime of experiental, comparative data to weigh it against, we can feel it.  And when that feeling is disconnected, you're not "in" the film, you're watching the film, and when you're not "in" the film, it doesn't stay with you.  

X-Wings Dive!

Battalion 23?

Yeah, internally it's not a great composite... honestly it's amazing this process works at all given how variable the filmed part of it can be during the production of the composite.

No, I think he's right that it's the second, recycled shot.  

Ah yes, you may be right.  And generally speaking, I haven't recruited copied elements, no. It's mostly not necessary because of the law of diminishing returns. That is, after a vertical stack, it's usually as beefed up as you can see.  

Hey, guys - just a curious point about the X-Wings Roll shot - it's one of the shots on the Reliance reel and as I was doing my restoration I noticed two things about their version:

1) It cuts short exactly where the Special Edition's version does.

2) They painted out the Y-Wings.  Not cropped out; painted out completely and/or didn't recomposite them if they managed to go all the way back to the elements.

All of that is either conspicuously deliberate or good evidence it's not only Special Edition, but a Special-er Edition.

_Mike

And...

X-Wings Roll!

Throne Room! 

Hmm...  

"Harrison Ellenshaw One of MY favorites! Mike Verta, again you have done an amazing job. Great credit goes to Ralph McQuarrie for his beautiful production illustration that was the inspiration for this shot."

See, I get nervous about going against the artist's feelings, and he's been quick to give direction on shots in the past.  Still, let's give it another run, shall we?

Forum software fail.

If you guys would like to play with a full-res, uncorrected still, here it is:

https://www.dropbox.com/s/gay45s862low64g/Massassi_Still.tif?dl=0

You guys are at least half right; I'd left a node on that should have been off.

But it's probably not the black levels, it's that the garbage and grain and shake hide everything.  As soon as you can really see the flat tone of the painting, it looks like a painting. To me the big rock on the right side is a far worse offender, but again...  Apparently, George didn't like this painting either; part of why the foreground stuff was added.  From a compositing standpoint, the painting's luminance is mismatched from the live plate - it would need more contrast in just the painting elements, the foliage needs a change in tone, etc.  It's not a great comp, no question about it.  Harrison is reviewing now.

Does this forum not autoscale images or...?

Oh, no I don't spot correct those.

You mean like ships over explosions?

Also:  Massassi Temple

Quick semantics point: I can't call Legacy a preservation.  Only someone given the original negative and officially charged with preserving the film could say that.  I probably shouldn't call it a restoration either, for that matter; it is not an official project tasked with restoring the film.  Maybe reconstruction is the right word, who knows.

Either way, let's be clear that I don't clip out a single pixel of data, ever.  Everything that was in the original scan is get-able after the fact; it's just a question of where they sit in the values.  Now, regarding the garbage mattes, if they were visible 100% of the time on 100% of prints, like the artifacts under the landspeeder, I'd say they should stay.  But that's not the case.  And more importantly: when I set the pure black of space to 0, the ship mattes fall below visible black in P3 naturally, anyway.  I don't have to adjust them out.  So I'd say that ultimately, I'm leaving them in, but because my black is pure black, you just won't see them.

??  ... pretty sure they'd heard the word "robot" in 1977, but whatever...

Either way, garbage mattes will not survive my color picker.

Well there goes my theory...

I've seen that, too, but to that point, when I screened a few prints on a vintage projector, with vintage bulb and screen, which apparently had lower reflectance by default, they weren't visible.  It's been my assumption since then that improvements in screen reflectivity and/or increased luminance output was the culprit, but this doesn't explain your anecdote from "back in the day."  One thing is for sure which is that during the development of the shots, whatever they were screening the tests on was not exhibiting the artifact or they'd have compensated.

Where does that leave us? Certainly I have been setting my garbage-matte-black-point nice and low in the P3 curve...!