logo Sign In

Info: Best OUT materials at Lucasfilm?

Author
Time
 (Edited)

This purpose of this thread is to collect some off topic posts from the First Impressions of the OOT … thread in General Star Wars Discussion, which more properly belong here in Preservation and Fan Edits.

Author
Time
Originally posted by: zombie84
Okay to end all the debate about what prints exist and what went on with the SE:

-Because the original 1976/1977 negatives were beginning to deteriorate, a fairly exhaustive restoration was needed to save them. The problem with Kodak Eastman 35mm negative stock is that it does not live long--the colours and tones are never as vibrant or true as what is actually photographed, and whats worse the stock fades away and deteriorates over time...much more rapidly than people first realised. In fact certain shots had faded so much that they had "gone pink"--i believe the technical terminalogy for this may be a yellow layer failure, although i am not certain. In order to get the shades correctly certain pieces had to be duplicated and then re-timed to eliminate the pink tint. The corrected pieces were then re-inserted into the O-neg reel, replacing the deteriorated originals. Other shots had gone so bad that they couldn't be re-timed and instead had to be replaced with frames from an Interpositive print. The footage was all cleaned and washed by hand--in fact many of the tatooine shots had sand built into the film!
-Once the final reel had been restored SFX shots were re-composited into the computer. To do this they scanned the original elements--the raw bluescreen model shots and the on-set plates, both of which were kept in Lucasfilm's extensive archive, at 2K resolution (the maximum technology allowed at the time). Compositing digitally eliminates matte lines and dupe grain from the optical composite process. Transition shots--wipes, dissolves--were also redone digitially. The crawl was redone digitally as well, as far as i know. After all these changes were done they were printed back onto 35mm negative film and re-spliced into the O-neg reel, replacing the original pieces
-then there were shots that were enhanced with CG. To do this the original negatives were scanned at 2K res and then ILM did their thing. The final CGI-enhanced shots were then printed back out onto 35mm negative film and the pieces re-spliced into the O-neg reel, replacing the original shots
-then there were the totally new shots. Some of these involved filming new elements (ie more extras, sandtroopers, new model shot of sandcrawler) in which case the film was probably scanned for some touch-ups but most were created digitally. The finished shots were then printed back onto 35mm negative film and inserted into the O-neg reel
-finally, since the O-neg shots are litterally the raw footage filmed from the camera, they need to be colour timed. In order to faithfully re-produce the original colours, George Lucas gave YCM Labs, the lab that did the restoration, a perfect technicolor print of the 1977 original. Unlike Eastman Kodak, Technicolor does not fade away over time, and not only that it has superior tone and colour information as well as much less grain. It also came to light that Lucas had technicolour seperation masters made for him in 1977 as well.
-Finally, the colour-timed O-neg reel was printed into a new Interpositive. Hence the 1997 SE is born.

You can see what happened here--little by little, the original film was consumed by new pieces. This is what Lucas meant when he said "its impossible to make a new OOT, the negative doesn't exist." Well, technically, it doesn't. You may also be wondering what happened to the original pieces that were replaced: well, me too. Undoubtedly, they were put back into storage, although the pieces that had deteriorated to garbage may have indeed been destroyed.

But we're not done here. Because an equally important process occured, probably in early 2004 for the S-SE (or whatever you want to call it).

So now we have the new O-neg configuration. But Lucas wants to change it again.
-So, the O-neg is scanned--but in HD resolution. And the O-neg of course is now the 1997 SE
-Because the O-neg is scanned and not the re-timed 1997 IP that means that all the colour-timing information is lost. So the film must be re-timed, now in a DI environment. Because Lucas is linking the films up to the prequels a very different look is decided upon--very constrasty and "modern", which particularly is different for ANH and ROTJ which were very softly lit originally. Apparently Lucas approved the final colur-timings personally
-from here the O-neg is altered in the digital environment once again, but unlike the 1997 version now the entire thing is digitized. The additions are made: gungans, ESB McDiarmid, Hayden, etc.
-then Lowry gets a hold of it. My understanding is that they were the last of the line but it seems to me that a more logical workflow would be to have ILM work from the cleaned-up Lowry files. But as far as i remember, ILM did their alterations before. Then the HD files are given over to Lowry. Lowry uses its clean-up algorithim to remove thousands of bits of dirt and grain through computer-controlled software. Some sharpening filters were also used to bring out detail in certain areas deemed soft.
-From here an digital HD master is finally produced of the final product. Thus is borne the 2004 S-SE. The films are now existing purely as data, and from this data the DVD down-conversions are made.

I am going to point out the obvious here--the 2004 S-SE is only HD res. Apparently Lucas must think this is okay since thats what he filmed his prequels in but for anyone who loves the SE this is a major bummer.

I hope this clears up any confusion regarding the SE. This info comes courtesy of many articles and interviews from various sources. Go Mer I can't believe you are still asking about this--i know for sure i have explained this in the lengthy TFN thread at least ten times.

As for Lucas, yes, he did actually say that the original negative was destroyed--this is a twisted exageration based on the fact that the 1977 O-neg technically has been annhilated, but not in the sense that it has been chucked into a furnace, which is what he is trying to get people to think. It could easily be re-constructed, or, even cheaper and downright better IMO, simply use the Technicolour seperation masters to make a new DI--a perfect copy of a pefect IP. Nor would any of this be expensive, relatively speaking, and a DI of the seperation masters would probably run about $100,000 or so, which is absolutely peanuts for LFL.
Author
Time
Originally posted by: Mielr
This is an example of film that has 'gone pink':
http://i28.photobucket.com/albums/c229/queen1970/rebelcel2.jpg
Author
Time
Originally posted by: Scruffy
Zombie --

That largely agrees with and supplements my research, but I came to a conclusion that I'm not sure you did, or maybe was outside the scope of your post. In or around 1997, the O-neg was restored. Certain portions were removed for SE alteration, and we can assume that those film elements are lost. But every one was scanned at 2k so the SE alterations can be made. Hard drive space is cheap, so we can assume the digital versions of these portions are not lost. So, in 1997, we have 90% of the O-neg intact in the SE and the remaining 10% in the digital domain.

In or around 2004, the entire SE was scanned at HD resolution. At this point, the 90% of the SE that existed only on film enters the digital domain. Because those are the portions that were not substantially altered, the entire O-OT now exists in the digital domain at HD or greater resolution. Further work was done on the movies, but it isn't relevant to us, and I assume an untouched copy of the HD scan was kept for various reasons.

This gives Lucasfilm everything they need to produce a fine DVD/BD/HD DVD of the O-OT. They just have to find the relevant files and give some interns a few days with an editing suite to paste it together and color time it. (I don't mean to belittle film or video postproduction, but they don't exactly need ILM's A-team here.) The sound mix might pose some additional challenges, but with the laserdiscs and the DVDs there's enough sound material in the digital domain today to put together a quasi-authentic "good enough" stereo mix.

I think this would be easier than making a new DI from the Technicolor masters and less personally irksome to George. It might disappoint those of us who had been hoping for a full restoration, all three original sound mixes, etc. But for me, at least, it would be a buy.

There's a few potential problems with this scenario. The SE negative may have picked up some dirt between 1997 and 2004; Lucasfilm evidently skipped the film cleaning that time and trusted Lowry to fix it digitally. In this scenario, that won't happen. There's also the possibility that Lucasfilm didn't keep the initial stages of their digital work, that they just deleted them when they decided they were done. Such an act of wanton carelessness would boggle the mind; but Lucas does seem to be a packrat, and he learned how that pays off when he did the 1997 recomposites, so I think he kept them.

Thoughts, comments, flames?
Author
Time
Originally posted by: THX
zombie84, thanks for drawing together all your research into one post like that - an excellent resource. It certainly agrees with everything I know about the "restorations."

Scruffy, I have been meaning to post something similar to what you said about existing hi-res digital files of the OUT being available. Unfortunately, however, I do not agree entirely with your post (I wish I did!).

Agreed: the '04 original, untimed HD scans are likely to still exist, and represent a useable resource for OUT digital restoration. However, the SE changes represent considerably more than 10% of that version, when you take into account all the digital recompositing and remaking of all the wipes and dissolves. Anyway, let's say there's something more than 50% and less than 90% of the OUT available from this source (depending on your POV on recompositing - I'd be happy to accept that aspect of the SE to get these scans).

Now, to fill in the gaps, you suggest using material digitized in '97 for the making of the SE. However, as zombie pointed out, all those recomps were made from the original elements and there would be no reason to scan & restore the material that was being jettisoned (the cut negative of the optical composites). It's possible that some of this stuff was scanned for reference for the digital compositors, but in any case it wouldn't have received the restoration that was done for the stuff that was going to survive into the SE.

So, you're still looking for a hi-res scan of quality elements of anything not in the '97 SE. Would the parts of the original cut negative that were removed have been kept? Again, probably. Do they represent quality elements for new scans? Well, if zombie's description of the '97 restoration is accurate (which I think it is), then probably not - you'd have to go the IP for some of that stuff. Again, those parts of the same IP that were (not) used in the '97 restoration are probably in optimal storage and could be scanned and used, but I don't think the entire OUT is in hi resolution on LFL hard drives right now.
Author
Time
Originally posted by: zombie84
Regarding what needs to be scanned for a complete OOT O-neg DI--yeah, its not that difficult. There are a number of ways a high quality OOT scan could be accomplished:

1) Raw scan of a reference print. Lucasfilm has 'em, and although a bit rough because of their age, they are maintained in relatively good condition. Just a normal 35mm release print. A bit of tweaking would bring out the colours and contrast and the most basic off-the-shelf filters would easily eliminate much of the scratches and dirt. From this we could even do our own frame-by-frame restoration if we do desired.

2) Raw scan of an IP. Even better than a release print, Lucasfilm has at least one IP, the best culprit probably being the 1985 IP made for the home video releases, though by now this would be getting a bit banged up. Again, some tweaking and modern software would spruce this thing up in a few hours. I know the SE had a few segments taken from an IP, so at least one of them--and undoubtedly the one in best condition--would have frames and probably whole shots missing from it. This would require some filling in.

3) Raw scan of Lucas' Technicolour print. IMO this would represent a better source than an IP--it would be less grainey, practically free of any scratches and marks since it hasn't been used much (if at all) and have all the colour information intact. A quick scan of this and we would have the best version of the OOT that anyone has ever seen, frankly.

4) A 4K DI of the Technicolour seperation masters. These I am not sure if they are Lucas' personal collection or in the LFL Archives--perhaps they are the same thing. Scanning each of the seperation masters and then making a Digital Intermediate of them would IMO represent the best possible version of the OOT, practically at IP resolution since they are struck directly from the IP (or IN?), with pretty much perfect colour fidelity and practically no dupe grain. They would undoubtedly be free of any kind of scratches since as far as i know they have never even been used.

5) Scan taken from the O-neg. This presents quite a few problems.
The first method of doing this would be to re-construct the original negative. This would be done by finding the original pieces from the archives and re-inserting them into the reel, removing the SE pieces. This would require a bit of effort and care however, and I'm not sure how comfortable Lucas would be with "destroying", as he put it, the SE. In any case after the scan the SE could be re-assembled.
The second method would be to use the scans already done for 2004. This would mean that the original pieces from storage would be scanned in HD and then edited into the 2004 HD telecine of the 1997 SE O-neg. This would require a bit of effort to match frames and ensure that the shots transition correctly. It also would be debateable if the SFX shots that were re-comped for the SE should be re-scanned, in which case there is probably over a half hour of original footage to be scanned. Another problem would result from this--because we would be working from the initial 2004 scan of the 1997 SE O-neg, Lowry's clean-up would be lost. No big deal. The 1997 SE looked fantastic in terms of clarity. A bigger issue is this--because it is from the O-neg, the colour-timing again would be lost. I'm sure that the original printer light setting records don't exist, otherwise YCM would not need to rely so heavily on the Technicolour print to match colour, although this is debatable. In any event, the film is now scanned into a DI enviornment, so everything would have to be re-done by eye anyway. To me, this is a huge issue, as the proper image settings of the 1977 original are very distinct and almost never have been gotten right. I'm sure a colour-timer could get the film to look pretty close but there undoubtedly would be nit picky things that would be inaccurate that would just drive some of us mad, especially since this would be the last OOT-scan for many, many years (possibly until 2K resolution home theater is around). Using the Technicolour print that YCM worked off would be needed to ensure accuracy. By now however, this process is getting quite complicated, which normally would be par for the course in a classic film restoration but with Lucasfilm's attitude its unlikely that they would put so much effort into this.

I also don't think that LFL scanned all the original pieces for the 1997 SE--some of them they did, ie all the shots that had CGI enhancements, but a lot of it was simply the scanning of the on-set plates and the raw bluescreen model shots, so the actual original composite O-neg's were simply put away.
Author
Time
Originally posted by: THX
For whatever reason, Robert Harris rejected the dye-transfer print as a restoration source.

BTW, we're way off topic here - we should probably edit some of these last few posts into a new thread over in Preservation and Fan Edits.
Author
Time
Originally posted by: lord3vil
Scruffy, Zombie,

Thanks for a couple of really informative posts! While I've gathered from various reasonably credible sources over the years that Lucasfilm certainly has the means to do a full restoration of the OOT, its very interesting to get to know some more details!
Originally posted by: Scruffy
So, in 1997, we have 90% of the O-neg intact in the SE and the remaining 10% in the digital domain.
When I followed the coverage of the restoration work for the SE on various fansites on the net back in 1996 and 1997, I remember that I picked up from somewhere that the original films were eventually scanned digitally in their entirety, not just the selected shots that were to be updated with new CGI. What's been said here appears to contradict this though. Do we know within reasonable doubt that the movies were fully digitized only quite recently, as in after the turn of the century, or could there be indeed be older fully digital sources from the original SE work as well?
Author
Time
The idea is to have an easily readable resource for those interested in the potential for an official restoration and/or anamorphic DVD/HD release. If anyone knows any further information, please post it here. I'd appreciate it if people would refrain from the type of posting sometimes seen in General Star Wars Discussion, along the lines of:

yeah, Luca$$h could give us everything we want but he'd rather stick it to the fans = Luca$$h is teh suxxorz!!!!!!!111!!1!

That's just a personal request. Of course this does not apply to Stinky-Dinkins, who can post anything he wants, as long as it's funny.
Author
Time
Thanks! All the cool info in one place so I don't have to dig through topics.
Author
Time
Originally posted by: zombie84



4) A 4K DI of the Technicolour seperation masters. These I am not sure if they are Lucas' personal collection or in the LFL Archives--perhaps they are the same thing. Scanning each of the seperation masters and then making a Digital Intermediate of them would IMO represent the best possible version of the OOT, practically at IP resolution since they are struck directly from the IP (or IN?), with pretty much perfect colour fidelity and practically no dupe grain. They would undoubtedly be free of any kind of scratches since as far as i know they have never even been used.

That would depend on whether or not the separation masters were done correctly. Unfortunately, some separation masters were made, but not tested (like the ones made for Spartacus, for example) and when the time comes to do a restoration- it is discovered that the separations are unusable.

Also, if preservation/separation masters were made for ESB and Jedi, they would have been made from Eastmancolor/Kodak prints (since Technicolor prints of those films weren't circulated), so the quality of the separations would be affected by the quality of whichever Kodak prints were used (for instance- if the Kodak prints were starting to show any sign of fading, that would be reflected in their separations).

But, I would think that if the fading was just minimal, it could be corrected digitally.

Author
Time
After a very short restauration session:

http://img160.imageshack.us/img160/5425/rebelcel2nxb6.jpg

If the image would have been better (no jpeg compression, higher resolution), the results would have been better. Han has a Pink shirt, because the RED layer lacks a lot of detail due to jpeg compression.

I just rebalanced the RED, GREEN and BLUE layer by hand. No further changes were made, i did not retouch the image. I know that film works in the CMYK domain, but the heavy JPEG compression worked against that.

@Mielr : if you have better scans or more pictures, BRING EM ON!
Author
Time
Here is a much worse faded print of Ben Hur:

http://img238.imageshack.us/img238/4368/benhuranamorphic70mmsmallue2.jpg

after carefully adjusting the layers:

http://img68.imageshack.us/img68/6103/benhuranamorphic70mmsmall2re0.jpg

I know, not perfect, since I am not a professional.

Faded image is from www.widescreenmuseum.com
Author
Time
Amazing what you can do with simple methods.

How does Eastman film stock work? Does it have separate "luminance" and "chroma" like video? In that case, how much does the black/"luminance" fade in comparison to the chroma?
If the black really is well preserved, I think for preservation reasons maybe it would be possible to combine the luminance from pink film stock with chroma from a less detailed source.
Author
Time
Originally posted by: Darth Lars
How does Eastman film stock work?

Pretty sure most film stock works the same way: three separate layers, one each for R, G, and B (or C, Y and M). So, no, there's no separate chroma and luma in film.

Author
Time
Originally posted by: Vigo
After a very short restauration session:

http://img160.imageshack.us/img160/5425/rebelcel2nxb6.jpg

If the image would have been better (no jpeg compression, higher resolution), the results would have been better. Han has a Pink shirt, because the RED layer lacks a lot of detail due to jpeg compression.

I just rebalanced the RED, GREEN and BLUE layer by hand. No further changes were made, i did not retouch the image. I know that film works in the CMYK domain, but the heavy JPEG compression worked against that.

@Mielr : if you have better scans or more pictures, BRING EM ON!

Vigo, you did a better job than I did:
http://i28.photobucket.com/albums/c229/queen1970/rebelceremony.jpg
I don't have any good software on my computer, though.
I have a few other scenes from Star Wars that I'll try to scan in later.

That Widescreen Museum is cool, isn't it?

Author
Time
Originally posted by: Karyudo
Originally posted by: Darth Lars
How does Eastman film stock work?

Pretty sure most film stock works the same way: three separate layers, one each for R, G, and B (or C, Y and M). So, no, there's no separate chroma and luma in film.


Chroma and Luma are technically digital video terms. 35mm Kodak negative stock has three layers of colours, RGB, which record the respective information of those spectrums.

Seperation masters are three black and white masters made from each colour value. One is a record of all the red value information, the other the green and the other the blue. The combined master of these three colour sepation records would theoretically give you a perfect replica of the original. George Lucas had these made for Star Wars, although I'm not sure about the two sequels.
Author
Time
Originally posted by: THX
For whatever reason, Robert Harris rejected the dye-transfer print as a restoration source.


Just found this on nightofdarkshadows.net (this is why you can't use a dye-tranfer print as a restoration source):

"While Technicolor prints do not telecine well (the density of the dyes block light and you end up with muddy blacks and poor contrasts), its helpful as it can be used as a reference in the film-to-tape transfer color timing session." Darren Gross

Author
Time
You would think that would be less of a problem with scanning. Still, there's more than one IP, possibly useful separation masters, and the majority of the movie already in hi-res digital data.
Author
Time
I think they used a dye-transfer print for the 'Apocalypse Now-Redux' DVD (I guess it was scanned rather than telecined?)

Author
Time
Yes, I think Harris was referring to a film restoration when he said that.
Author
Time
I was actually referring Darren Gross' comment, not Robert Harris'.

Author
Time
Sorry, I meant when Harris rejected the dye-transfer print as a restoration source.