3-strip Technicolor and Ultra-Resolution: from film to DVD

Discussion in 'DVD' started by Bill Burns, Jul 3, 2003.

  1. Bill Burns

    Bill Burns Supporting Actor

    Joined:
    May 13, 2003
    Messages:
    747
    Likes Received:
    0
    Trophy Points:
    0
    This is an aside to the Warner Legends thread, and to the release of the three films in question to DVD, and a very technical aside, but I'd like to iron something out. As this isn't about the discs themselves I thought it better to separate it into its own thread, where perhaps further discussion of Lowry's Ultra-Resolution process might be possible.

    Reading over the announcement of these films at The Digital Bits (http://www.thedigitalbits.com/#mytwocents) reminded me of a question I posed off-handedly in the Studio Feedback forum: at what point, in creating the best possible digital master for a 3- (or 2-) strip Technicolor motion picture, is the color introduced, and how is it then timed?

    "Ultra-Resolution," I believe, refers to a Lowry digital process of combining the three strips of early Technicolor. Three negatives (in 3-strip) are exposed to light through primary color filters (red, blue, and green). These create B&W (not color) film negatives, each with varying densities as created by the filters. When printed with colored dies (cyan, yellow, and magenta, the "compliments" to the primary colors through which the film negatives were exposed) and combined, this creates the final dye transfer print we might enjoy in a movie theatre.

    All well and good, and more can be found at The Widescreen Museum. But if Ultra-Resolution digitally scans the black-and-white negatives and combines them within the digital field, how and where is the positive color applied? Is it created digitally (positive color dyes for cyan, yellow, and magenta should certainly be within the "creative" capacity of computer technology), and if so, is the positive element used for the transfer also created digitally? In the printed form of dye transfer, I presume the dyes are imbibed (thus the term imbibition, or IB) to positive print elements, and these positive prints must then be combined and issued to theatres; transferring to multiple emulsion single film negative stock would presumably deprive the film of the dye transfer richness that marks the process, and such stock didn't exist when the 3-strip process was created, anyway, and thus combining the strips -- registering them -- in negative form would seem unnecessary, as they'd still have to be printed to positive for exhibition, and positive elements would be three combined strips in their own right (with dyes now applied).

    So my questions are these:

    1. In printed form, 3-strip is combined positively, with colored dyes applied positively, correct?

    2. In Lowry's digital process, used on the upcoming DVDs, how does this change? If they are scanning the strips as B&W negatives and combining them digitally, they must be adding the color digitally, correct, and creating their positive element digitally?

    3. Further, wasn't Technicolor itself alone allowed (by copyright or trademark law) to issue dye transfer prints? Wouldn't any studio wishing to create a new dye transfer print for exhibition be up a creek as a result? I think this is why Technicolor's labs had to re-open to make the dye transfer prints of Apocalypse Now Redux a reality, but those labs are closed again, yes? Could Technicolor license this technology to a studio lab if they so desired?

    4. If #3's answer is "yes," exactly where does the digital creation of a "computer dye" transfer print fit into all of this? And if Lowry's restoration work can only represent itself on DVD (unlike their B&W work on Sunset Blvd., etc.) or other forms of home video, then any film-restored B&W separation elements they generated could never be used for the creation of actual, theatrical dye transfer prints (unless Technicolor opened their labs again, or authorized the creation of such prints).

    I should clarify that 2-strip, in its final Technicolor subtractive form (again see The Widescreen Museum), exposed two sides of a double-emulsion negative. Unlike 3-strip, there is therefore only one negative in this form of 2-strip, a negative with two sides. It could then be printed as a positive with two dye-transfered sides and exhibited as such. No "registering" of different strips would be required. But digitally, you'd have to scan two separate B&W images and register them nevertheless, would you not, as scanning the image as a single "double negative" would prevent one from properly applying the two dyes (cyan and magenta) in positive form. Thus, I presume Lowry's system could be used, much as it is now for 3-strip, with 2-strip, but if that's not so, I'd love to know.

    I realize this is all very technical, but I'd like to get to the bottom of just what we're seeing on these DVDs (Singin' in the Rain, in its two-disc SE form, also utilized Ultra-Resolution), and just how this product relates to a physical dye transfer positive.

    Thanks to anyone with the experience in this field to comment, and if I'm overlooking any other questions in these matters others might like to ask, by all means chime in.
     
  2. Roger Rollins

    Roger Rollins Supporting Actor

    Joined:
    Jun 19, 2001
    Messages:
    931
    Likes Received:
    0
    Trophy Points:
    0
    I don't know the answer to most of your questions, but I do know that Lowery (sp?) has nothing to do with Warner's Ultra-Resolution process. The latter is a propriatary technology developed by WB. Lowery is an independent company that did cleanup some of WB's films but has no involvement in this process, nor in the 3 films just announced.
     
  3. Gordon McMurphy

    Gordon McMurphy Producer

    Joined:
    Aug 3, 2002
    Messages:
    3,530
    Likes Received:
    0
    Trophy Points:
    0
    I, too would like to know more about the process.

    Oh, Bill, my whispy water-colored memory recalled this site: "In 70mm!" today. [​IMG]

    Great, great site - can't believe I forgot about it! [​IMG]


    Gordy
     
  4. DaViD Boulet

    DaViD Boulet Lead Actor

    Joined:
    Feb 24, 1999
    Messages:
    8,800
    Likes Received:
    3
    Trophy Points:
    0
    Sounds exciting to me whatever it is. I too would love to know more about how it all works.

    Seems that if you digitally encoded each original color-channel B&W print you could then digitally convert to one of the 3 primary colors at any given intensity and digitally combine for a pefect "fit". I think this is what they are talking about. Theoretically, all of this can be done 100% in the digital domain with only the final result printed to film for restoration purposes.

    Looks like WB is also doing this work in high-resolution, which is great. I'd like for these fantastic efforts to make use of future HD formats and digital projection technology and not be limited to SD resolution.
     
  5. Robert Harris

    Robert Harris Archivist
    Reviewer

    Joined:
    Feb 8, 1999
    Messages:
    10,380
    Likes Received:
    7,645
    Trophy Points:
    9,110
    Real Name:
    Robert Harris
    Bill Burns wrote:

    "1. In printed form, 3-strip is combined positively, with colored dyes applied positively, correct?


    Correct, whether the printing matrices are derived from three-strip photography or via extraction from an Eastman color negative.


    2. In Lowry's digital process, used on the upcoming DVDs, how does this change? If they are scanning the strips as B&W negatives and combining them digitally, they must be adding the color digitally, correct, and creating their positive element digitally?


    I don't believe this has anything to do with LDI, but is an internally created WB propreitary computer program. Either positive or negative can be scanned and the appropriate color filtration added.


    3. Further, wasn't Technicolor itself alone allowed (by copyright or trademark law) to issue dye transfer prints?

    Only during the length of the patent. But the process is so expensive and complex that few would wish take it on.

    Wouldn't any studio wishing to create a new dye transfer print for exhibition be up a creek as a result?

    Totally accurate.

    I think this is why Technicolor's labs had to re-open to make the dye transfer prints of Apocalypse Now Redux a reality, but those labs are closed again, yes?

    Their opening had nothing to do with Apocalypse. Experimentation began in the early '90s with much input from Jeffrey Selznick. He and Technicolor engineer Richard Goldberg experimented with the Chinese varient of the process, printing a reel of Duel in the Sun and some other short bits. We were working with them and later with Technicolor, which created a beautiful reel of "Vertigo" in 1996 as a test.


    Could Technicolor license this technology to a studio lab if they so desired?

    Yes, but why?


    4. If #3's answer is "yes," exactly where does the digital creation of a "computer dye" transfer print fit into all of this?

    One could scan an Eastman Oneg and create the matrices from digital files if desired.

    And if Lowry's restoration work can only represent itself on DVD (unlike their B&W work on Sunset Blvd., etc.) or other forms of home video, then any film-restored B&W separation elements they generated could never be used for the creation of actual, theatrical dye transfer prints (unless Technicolor opened their labs again, or authorized the creation of such prints).


    You're mixing not only apples and oranges here, but throwing in a hand or two of bananas. Lowry has not yet done anything that can be construed as true "restoration work." Lowry does not create separation masters. These are produced by a handful of highly organized and technically talented labs and optical houses.

    I should clarify that 2-strip, in its final Technicolor subtractive form (again see The Widescreen Museum), exposed two sides of a double-emulsion negative.

    I believe this was two frames above and below, not two sides of an emulsion, although early Tech two-strip was glued together back to back, before the imbibitiion process was created.

    Unlike 3-strip, there is therefore only one negative in this form of 2-strip, a negative with two sides. It could then be printed as a positive with two dye-transfered sides and exhibited as such. No "registering" of different strips would be required.

    The example which you may wish to use is that of SE. Most Technicolor animation was exposed by sequential exposure, ie. three frames of the same information, each exposed through a different filter. Here you have no registration problems, as any shrinkage will be uniform. WB is currently doing yeoman work on their Merrie Melodies and Looney Tunes, working from original or protection SE material. The new preservation elements and the resultant transfers for DVD should be magnificent.


    I realize this is all very technical, but I'd like to get to the bottom of just what we're seeing on these DVDs (Singin' in the Rain, in its two-disc SE form, also utilized Ultra-Resolution), and just how this product relates to a physical dye transfer positive."

    What you're seeing is a beautiful representation of the film on DVD, but one which was never intended by the film's creators. The information was always imbedded in the black and white Onegs, but was never seen on screen due to a number of technical issues, which were designed into the process from origination photography through final release print.

    RAH
     
  6. DeeF

    DeeF Screenwriter

    Joined:
    Jun 19, 2002
    Messages:
    1,684
    Likes Received:
    12
    Trophy Points:
    610
    I am not erudite in the filmmaker's art, a la Mr. Harris, but I am someone who uses Photoshop every day. It seems to me your question is, at what point is color added to the process?

    If we start with 3 black and white positive lengths of film, representing the 3 filtered Technicolor safeties, and we scan each of these individually into the digital domain, they can be digitally corrected for shrinkage in the digital domain, and then each assigned a color, magenta, yellow, or cyan, and then composited together in the computer. Photoshop calls the color separations "channels" and one can look at separate channels or a color composite at any time.

    Then, presumably, the digital composite is printed back to film (if a new negative is to be made). It can be printed as positive or negative. Or, the corrected black and white channels can be printed back to film, still in black and white, and these will represent new safeties, having been corrected for shrinkage and fixed up digitally.

    I suppose it's possible to do all this work in the digital domain, print black and white "channels" back to film, and then combine them using the traditional dye-imbibition process. In the case of Singin' in the Rain, we got a new DVD as well as new prints from new negatives, so the digital process was combined at some point with traditional film.

    I don't know how the Ultra-Resolution process is separate from the Lowry process, but I think they both were involved in the recent version of Singin' in the Rain. Perhaps the Ultra-Resolution process involved the scanning of the three individual safeties, and then Lowry did its normal work of filtering out scratches, etc. I know Lowry has 300 G4 Macs in its office; I work on the same machines every day.
     
  7. DaViD Boulet

    DaViD Boulet Lead Actor

    Joined:
    Feb 24, 1999
    Messages:
    8,800
    Likes Received:
    3
    Trophy Points:
    0
     
  8. Bill Burns

    Bill Burns Supporting Actor

    Joined:
    May 13, 2003
    Messages:
    747
    Likes Received:
    0
    Trophy Points:
    0
    Great information, everyone, and thanks especially to Robert Harris and DeeF for their extensive info.

    Mr. Harris wrote:
     
  9. DeeF

    DeeF Screenwriter

    Joined:
    Jun 19, 2002
    Messages:
    1,684
    Likes Received:
    12
    Trophy Points:
    610
    I think you're mixing up your terminology, and making the process sound more complicated than it really is. I will try to elucidate what I understand the process is.

    1. Normal film transfer to DVD
    A contemporary film, when finished, is made into projectable prints. One of the prints might be used for transfer to the digital domain, in order to make DVDs. The print is run through a telecine machine, essentially scanning each frame, and then the frames are stored on a hard-drive. Depending on the resolution of each picture, the hard drive must be very large. Because a DVD can only hold so much information, the files must be compressed, altered in a way, and this is where the transferer's skill comes into question. For instance, a sky that is all one color can have a lower resolution, because there is no detail in that sky at all (that one can notice). Anyway, when the transfer is complete, the movie will be downloaded on disks and sold. It isn't remotely as high quality as 35mm film, but the monitors which will show it are also lower quality, so everybody gets a pretty good approximation of the original picture (presumably).

    2. Restoring a movie for DVD
    There are several ways one could go about this. The best way would be to restore the movie in the film domain, and then transfer the film per the normal procedure for DVD. Mr. Harris will know more about what this takes than will I. Usually, this is done piecemeal, meaning, the sections of the movie which are fine, are left intact, and the sections which need work will be worked on: perhaps pieces may be found from other prints, or perhaps pieces may be scanned, digitally, and worked on digitally, and then printed on film and cut back into the negative, or print. Eventually, a new print will be made, and then transferred to DVD, etc.

    3. Restoring a movie using Ultra-Resolution (as far as I understand)
    I believe that the original negative, the ONEG, for Singin' in the Rain was destroyed in a fire. It does not exist. There were plenty of prints available, but each of these has its own problems: scratches, color fading, perhaps projectionist's edits, etc. The original disk for Singin' was pretty good: but one could see scratches and no true blacks: Gene Kelly's tuxedo and hair often looked blue, revealing a fading of the yellow layer of the print (which may or may not have been a dye-transfer print).

    But it was discovered that black and white safeties, basically, 3 positive pieces of film representing each of the three colors, red, blue, and green, existed, which had been made from the ONEG. These 3 pieces were in good shape, but had normal problems of differentiated shrinkage. These 3 could have been printed into a Technicolor positive print using the dye-imbibition process -- but there might be registration problems due to the shrinkage.

    The solution was to scan each of the 3 safeties separately, at a high digital resolution, and then do all fixups in the computer. Basically, it was the first time an entire movie was scanned digitally for restorative processing.

    Once the filtering and fixing of the 3 safeties is complete, they are each assigned the appropriate positive color, yellow, magenta, and cyan, and then digitally brought together: composited. So, the entire movie, in color, exists on the computer, in a pretty high definition state, perfectly registered. Compression for DVDs will be a snap, because no telecine is required at this stage.

    In the case of Singin' in the Rain, after this whole thing was done, it was made into film prints. So, the digital file was printed back to film, either as a negative from which prints were made, or directly to positive prints. Either is possible.

    Remember, though, that as good as these prints are, they are probably lower resolution than the original negative and prints would have been. The digital domain limits the size of the files, for now. Many of these companies are working in 2K (Metropolis was restored at this level), but 4K and 6K are more feasibly like film, and 70mm quality might require something greater than 4K.

    Is any of this clearer?
     
  10. DaViD Boulet

    DaViD Boulet Lead Actor

    Joined:
    Feb 24, 1999
    Messages:
    8,800
    Likes Received:
    3
    Trophy Points:
    0
     
  11. Bill Burns

    Bill Burns Supporting Actor

    Joined:
    May 13, 2003
    Messages:
    747
    Likes Received:
    0
    Trophy Points:
    0
    Thanks, DeeF. Good info there. I had a handle on most of it, and I think there's more to #1 (I believe DVDs ideally, when available, resort to an interpositive element, rather than a release print in which contrast and grain gains become an issue, and the film must then be transferred down to a certain kind of tape before it's encoded for DVD? I've yet to determine exactly how an IP differs from a release print, though, so this may be wrong), but I certainly appreciate the detail. Thanks again.

    I wasn't aware Singin' in the Rain's original negative elements were lost (now that you mention it, it rings a bell -- is something said about it on the DVD?) -- it looks great, given the material with which they had to work. I know Snow White and the Seven Dwarfs made news back before Lowry, when Disney digitized the whole thing for restoration prior to the boxed set laserdisc (I think this was the first time it was done, back in the early 90's, for a feature film), but that's about as far as my digital restoration history goes in a film timeline. [​IMG] Anyway, keep the info comin', as I think there's more to investigate on the Ultra Resolution issue, but I have to run at the moment. I'll check back in later tonight or tomorrow. Thanks once again! The insight of all who've participated in the thread is much appreciated.
     
  12. PaulP

    PaulP Producer

    Joined:
    Oct 22, 2001
    Messages:
    3,291
    Likes Received:
    0
    Trophy Points:
    0
    So are Singin' in the Rain and the upcoming Warner Legends discs the only ones done in this process, or are there more?
     
  13. DeeF

    DeeF Screenwriter

    Joined:
    Jun 19, 2002
    Messages:
    1,684
    Likes Received:
    12
    Trophy Points:
    610
    One difference I can see between the Singin' in the Rain project, and something like Sunset Boulevard.

    Singin' in the Rain was produced from 3 black and white positive safety films representing the 3 colors. These would never have been used to generate prints, or anything else. So they had no wear and tear, or fading, or anything. Just shrinkage differences, since they are separate items that have been stored for 50 years.

    But Sunset Boulevard, and other black and white movies, had no safeties, or anything like that. The ONEGs were used for printing, many thousands of times. As generation upon generation was... generated, the neg and its copies start to lose definition -- basically, the most vulnerable areas in a negative are the dark areas of a picture, because they are the lightest areas on the negative. Over time, these areas lose detail, and become, well, just black all over.

    So, a movie like Sunset Boulevard needs to be restored, needs Robert Harris or someone of that calibre to research all the available source materials, and put together a neg/print, and then make a NEW neg, which I'll call NNEG, from which prints can be made. Part of this process assuredly is digital. Many scenes can be scanned into the computer and cleaned up, of scratches and dirt, and also contrast of grays and blacks, and to help with jitter. This is where Lowry comes in. I'm sure they were more involved in a project like Sunset Boulevard than they were with Singin' in the Rain, which really just needed a lot of computers and hard drive space.
     
  14. DeeF

    DeeF Screenwriter

    Joined:
    Jun 19, 2002
    Messages:
    1,684
    Likes Received:
    12
    Trophy Points:
    610
    In addition to the upcoming Warner Legends series, I believe that "Ultra-Resolution" process is being used for Meet Me In St. Louis and Easter Parade, two other Technicolor musicals. I don't know if this process would work for large format musicals, like Oklahoma!, for instance. You would need to work at a very high resolution, digitally, to match the current quality of 65mm or 70mm movies. But the old Technicolor movies seem ideal for this kind of work. Again, it isn't exactly "restoration" but a digitizing of the original 3 separate b&w films, and compositing them digitally rather than dyeing them and pressing them onto prints.

    By the way, if you look at the old Singin' in the Rain DVD, you can see a terrible tear or burn in the film, towards the end. This isn't on the new DVD, so it must be assumed that it happened to the neg at some point, which was made into a print, and that print was the source for the older DVD. No wonder they didn't feel they could use prints to refashion this movie digitally -- this "Ultra Resolution" process was their only hope.
     
  15. RobertR

    RobertR Lead Actor

    Joined:
    Dec 19, 1998
    Messages:
    9,955
    Likes Received:
    347
    Trophy Points:
    9,110
     
  16. Patrick McCart

    Patrick McCart Lead Actor

    Joined:
    May 16, 2001
    Messages:
    7,596
    Likes Received:
    257
    Trophy Points:
    9,110
    Location:
    Georgia (the state)
    Real Name:
    Patrick McCart
     
  17. DaViD Boulet

    DaViD Boulet Lead Actor

    Joined:
    Feb 24, 1999
    Messages:
    8,800
    Likes Received:
    3
    Trophy Points:
    0
    I watched the new Singing in the Rain on my friend's Sony 10HT at a viewing distance of about 1.4 screen-widths and never noticed any EE.

    ?

    dave [​IMG]
     
  18. Ken_McAlinden

    Ken_McAlinden Producer
    Reviewer

    Joined:
    Feb 20, 2001
    Messages:
    6,191
    Likes Received:
    80
    Trophy Points:
    9,110
    Location:
    Livonia, MI USA
    Real Name:
    Kenneth McAlinden
     
  19. RobertR

    RobertR Lead Actor

    Joined:
    Dec 19, 1998
    Messages:
    9,955
    Likes Received:
    347
    Trophy Points:
    9,110
     

Share This Page