Film-Tech Forum ARCHIVE: Uneven polyester dye fading

This is topic Uneven polyester dye fading in forum Film Handlers' Forum at Film-Tech Forum ARCHIVE.

To visit this topic, use this URL:
https://ft-forum.com/ft/cgi-bin/ubb/ultimatebb.cgi?ubb=get_topic;f=1;t=005398

Posted by John Hawkinson (Member # 1135) on 09-30-2003, 11:11 AM:

Continuation from Star Wars topic in the FITA (please don't clutter it with discussion).

I noted that Criterion print #A:0776 of Star Wars: A New Hope (1997 Special edition) on Kodak 2386 polyester had "periodic color shifts toward red, lasting anywhere from a few seconds to a few minutes. Perhaps the result of improper storage and infrequent rotation."

Brad Miller said, "John, the color shifting was on all prints. It was that way on the best of the remaining elements. That wasn't just the copy you screened."

That doesn't quite ring true to me, since I don't at all recall this color shifting problem in 1997 when the prints came out. It really seems like some parts of the print faded to red much more than others. Is my memory just crap?

On the other hand, John P. said, "Uneven dye fading is often indicative of improper storage. For example, film stored near a heat source will have a convolution repeat, where the side of the roll that was hotter fades faster. Or acidic by-products of hydrolysis from excess humidity settle on one side of the storage container, causing localized dye fading."

This occured to me, and I even commented on it. But it doesn't ring true for me. I would expect to see fading/brightening effects every few seconds, but not run OK for mintues at a time, and then shift. At least, if it was stored on 2k reels.

Are you suggesting that perhaps the outer half (rather than the left or right half) might see some kind of localized heat effect? That seems kind of farfetched to me. There was no localized fading in the sense that one side of the image was faded and the other was not. Furthermore, the transitions between fading and non-fading were sharp, sometimes across cuts; they were not slow gradient changes.

Anyone have anything to offer?

--jhawk

Posted by John Pytlak (Member # 331) on 09-30-2003, 11:35 AM:

I was just outlining examples I've seen of localized fading due to improper storage conditions. Usually it is a rather rapid "convolution repeat" related to the diameter of the roll, or a differential across the image if one side of the roll was subjected to more heat (e.g. can of film stored over a radiator) or acid vapors. A shift that occurs over several minutes of running time is more likely a printer balance problem, rather than fading. Or a faded element may have been used for part of the reel.

Posted by Mitchell Dvoskin (Member # 751) on 09-30-2003, 12:18 PM:

The lab quality control on the Star Wars Special Edition prints was all over the place. Some prints were almost red, others had color smears on one side, and others were perfect. It was the luck of the draw as to what print you got.

Yes, the original elements had faded and were color corrected. You can see that most in some scenes that no longer had true blacks, but on the good prints, the colors were not noticably off.

/Mitchell

Posted by Brad Miller (Member # 2) on 09-30-2003, 02:37 PM:

quote:
Brad Miller said, "John, the color shifting was on all prints. It was that way on the best of the remaining elements. That wasn't just the copy you screened."

That doesn't quite ring true to me, since I don't at all recall this color shifting problem in 1997 when the prints came out. It really seems like some parts of the print faded to red much more than others. Is my memory just crap?

I ran 2 prints of the SE edition and one of them was an EK print. Both looked about the same in regards to the color shifting. You were probably just too enthralled at seeing Star Wars again on the big screen that you didn't notice, sort of like how on a really good movie it is possible to lose track of which reel you are on or the "old CAP code". (We all know not noticing the new CRAP code is impossible. [Wink]

)

Posted by Manny Knowles (Member # 1171) on 09-30-2003, 03:20 PM:

I don't want to get too embroiled in this discussion but I recall a special presentation at USC by either Tom Hollman or Gary Rydstrom in which we were shown R-1 of "Star Wars" with the original sound and then with each element being added.

This was around the time that work was being done on the restoration and in the course of the lecture-presentation Hollman (or Rydstrom) briefly touched on the problems with negative fade.

Part of the problem was one of the negatives used for a lot of the location shots was brand new when SW was made and Kodak had to discontinue making that stock because it was plagued by extreme fade in a short time-span.

Next was the problem with the YCM separations. Evidently the Yellow separation wasn't quality-controlled and was overexposed.

However, you guys are talking about the SE and all of this was supposedly corrected. I seem to recall seeing the color shifting on opening night at Mann's Chinese in Hollywood -- admittedly, I was "into" restoration back then and so I was looking at it from that perspective.

I think they did a great job -- I saw what it would've looked like without the restoration -- but I wouldn't say that the restoration was flawless.

How much shift are you seeing? I'm asking because I know we tend to be more observant (see also: anal-retentive) than average moviegoers. Is this a mountain-molehill situation?

Posted by Stephen Furley (Member # 1277) on 09-30-2003, 03:28 PM:

Wasn't some of the effects work on that film done on CRI, which did have poor resistance to fading?

Posted by Manny Knowles (Member # 1171) on 09-30-2003, 03:33 PM:

Okay I went and looked it up and Kodak CRI #5249 was the culprit.

Posted by Stephen Furley (Member # 1277) on 09-30-2003, 03:36 PM:

That was quick, you must have read my last post the moment I sent it!

John, I'm sure you'll read this, when CRI was introduced, was it's problem with fading not known, or was it expected, but felt to be outweighed by the advantage of the generation saved?

I assume that in 35mm, all CRI printing must have been optical, to preserve the correct geometry of the negative. I know that CRI was used, contact printed, in 16mm to produce final prints with the standard geometry, base to lamp, for that gauge.

Posted by Manny Knowles (Member # 1171) on 09-30-2003, 03:39 PM:

Actually, I figured someone would make me earn that one so I started looking it up before you posted.

See what I mean? I was right about us being an anal-retentive lot. [Big Grin]

Posted by John Hawkinson (Member # 1135) on 09-30-2003, 04:49 PM:

Manny asks, "How much shift are you seeing?"

I don't have the print anymore, so I'd have to do it from memory, which is probably sufficient here...but what units would you like?

--jhawk

Posted by Steve Kraus (Member # 476) on 09-30-2003, 05:05 PM:

I thought all the CRI was removed; that that was a big part of the restoration effort.

Posted by Manny Knowles (Member # 1171) on 09-30-2003, 06:21 PM:

John -- Just do your best.

Posted by John Pytlak (Member # 331) on 10-01-2003, 04:12 PM:

The science of predicting future image stability based on "Arrhenius Testing" was in its infancy in the late 1960's, when CRI was developed:

http://aic.stanford.edu/jaic/articles/jaic31-02-005.html