Writing a Thesis on Digital Film Restoration (1 Viewer)

[Seppo]

Hi,




First, I want to apologise if this is in the wrong forum. If it is, feel free to remove it.

I study Information Processing Science at a local university, and this spring I’m planning to write a thesis on digital film restoration. My plan is to mostly introduce the basic steps of restoring films for Blu-ray nowadays, discuss the problem areas and perhaps use the fairly recent Coppola Restoration of “The Godfather” films as a use case. The “problem areas” that I’d like to investigate are basically the things how you can deviate from the original intent. This includes things like the overuse of Digital Noise Reduction, changing the color/look of the film etc. The length of the thesis will be about 25 pages.

Now I’m asking the help of all you fellow HTF members. I’m having a hard time finding proper “scientific” articles and sources on the problem areas I mentioned earlier. For example, has anyone ever held a conference or a seminar on the overuse of DNR? Maybe not. It sucks that because this is a thesis I can’t use regular articles found on websites.

After doing some research the sources I’ve found have titles such as “Historical Film Restoration and Video Coding”, “Perceptual Approach for Unsupervised Digital Color Restoration of Cinematographic Archives” and “Automatic Restoration Algorithms for 35 mm Film”. The only book on film restoration I’ve found is called “Restoration of Motion Picture Film” (2000). I can definitely get some basic information out of those sources, but those were all published between 1996 and 2000. Newer material would be appreciated.

There are some video documentaries I’m also planning to use. The main documentary would obviously be “Emulsional Rescue: Revealing the Godfather”. I think it’s a really good documentary on the subject matter. Other ones I was thinking of using are “Color Timing The French Connection”, “Prettier Than Ever: the Restoration of Oz” and perhaps “Miracle Cure: Restoring the Film for DVD” (found on The Who’s “The Kids Are Alright DVD). Any other good documentaries you might know?


So basically if you have any good sources or tips that you might think would be a good idea to use or “investigate” in my thesis, please let me know. Any help would be greatly appreciated.



UPDATE (December 1, 2011): See post #14 for a link to the finished "Digital Film Restoration and Remastering" master's thesis.

 

[BethHarrison]

Robert Harris posts in this forum. So maybe you may like to contact him so you could request an interview. Though he was in charge of the actual restoration of the films. I don't know if that included seeing it all the way through to Blu-ray authoring.

There was a good article in American Cinematographer on the restoration which quotes Mr Harris extensively:
http://www.theasc.com/magazine_dynamic/May2008/PostFocus/page1.php

The other thing I would do is get in contact with places like Lowry Digital, Deluxe Digital and Criterion and see if you can arrange interviews. Give them a list of questions and see what they say. They are the people on the front line actually using all the new digital tools to prepare films for Blu-ray.

 

[Robert Harris]

The Godfather(s) were taken all the way through to final approval of HD master.

To Seppo...

One may not need 25 pages unless the intent is to write a book on digital film restoration, and then you'll need a few more. The basics are extremely simple. World-wide search for inventory; safely bring elements to multiple locations in the same area for basic examination; detailed examination and comparison to approved continuity; selection of those necessary for scanning; selection of facility which can best handle the elements selected and their problems; finding workarounds for problems encountered with elements in scanning; examination of scanned footage; digital clean-up; conformation to original continuity; color and density; further digital clean-up and grain work in connecting unrelated elements; all with continuous screening via projection; final QC and projection of data at 2 or 4k; modification of color space and downrez for HD master.

The bottom line is to locate and use the finest reference print attainable, and not to modify the look or textures in any way. This means that any "automated" passes must be watched carefully for undesirable traits, digititus or additions of elements never in the film. One must also keep in mind the way in which film was created, ie. use all original paperwork, daily production reports, camera reports, etc, toward replicating and matching the look as printed, not necessary as harvested from the film elements used in the process.

Then one moves to audio, which has its own problems.

Feel free to contact me privately, if you need additional information.

RAH

Originally Posted by BethHarrison

Robert Harris posts in this forum. So maybe you may like to contact him so you could request an interview. Though he was in charge of the actual restoration of the films. I don't know if that included seeing it all the way through to Blu-ray authoring.

There was a good article in American Cinematographer on the restoration which quotes Mr Harris extensively:
http://www.theasc.com/magazine_dynamic/May2008/PostFocus/page1.php

The other thing I would do is get in contact with places like Lowry Digital, Deluxe Digital and Criterion and see if you can arrange interviews. Give them a list of questions and see what they say. They are the people on the front line actually using all the new digital tools to prepare films for Blu-ray.

 

[BethHarrison]

When a film is restored digitally, how do you get the right look for the film when it is printed back out to modern print stocks?

Are modern print stocks so different from the 1970s and 1960s stocks that the digital master needs manipulation so the prints look right?

For example, the first two Godfather films were printed using the dye transfer process, how do you replicate that digitally? Is it all by the way the digital master is made? Or are there different masters, one for video (i.e. for digital projection and to make the Blu-ray) and another that is altered so it looks right when scanned back out to film?

 

[Robert Harris]

 

[Adam_S]

The Pollyanna DVD has an excellent, 20 min feature on its unique restoration problems.

Metropolis is an interesting case study, as are many silent films (such as Greed, London After Midnight) that have been lost before or have been partially salvaged. Lost Horizon (the 1937 Capra film) is another film that has ran into significant restoration problems.

Tangentially, you could also add a section on the differing television and film standards around the world how issues like 3/2 pulldown and PAL speedup affect changes in films as they are telecined for video playback (all television being video at somepoint, afterall). One similar example, that's somewhat unique is Criterion's edition of Berlin Alexanderplatz, which was shot in 25fps for european television and has a pitch shift to compensate the different video playback speed in america because they weren't able to replicate the frame rate 100% accurately without major video problems for release here.

 

[Robert Harris]

 

[Seppo]

Quote:

Originally Posted by BethHarrison
There was a good article in American Cinematographer on the restoration which quotes Mr Harris extensively:
http://www.theasc.com/magazine_dynamic/May2008/PostFocus/page1.php

Thank you, Beth. I have actually read that article a few years ago and I wanted to use it for the thesis, but because it's a web article I thought I couldn't use it. Then I realized that I actually can use it because that article was released on a printed format. So thank you for linking it, I'll definitely use it as one of my sources now.

---------------------------------------------------------------

Originally Posted by BethHarrison
The other thing I would do is get in contact with places like Lowry Digital, Deluxe Digital and Criterion and see if you can arrange interviews. Give them a list of questions and see what they say. They are the people on the front line actually using all the new digital tools to prepare films for Blu-ray.

This is something I have been thinking about as well. The recommended length for this thesis is roughly 25 pages. It can be a few more or a few less. After this thesis I could start writing my master's thesis. The good news is that I can use this thesis and expand it to roughly 50 pages. The master's thesis is the one that requires empirical investigation. That's where interviews, for example, would be of great help. Thank you for mentioning those companies. I did a quick Google search and found contact information for all of them. I wouldn't hold my breath that they will reply to me, but I am going to give it a try at some point.

---------------------------------------------------------------

Originally Posted by Adam_S

The Pollyanna DVD has an excellent, 20 min feature on its unique restoration problems.

Thanks for the tip, Adam. I haven't seen the film nor the documentary. I'm now thinking about ordering the DVD from Amazon just for the documentary. Do you remember if it mentions digital tools and stuff like that? The emphasis in my thesis is, like the title suggests, digital restoration, so the more I can use the word "digital" the better.

---------------------------------------------------------------

Originally Posted by Adam_S

Metropolis is an interesting case study, as are many silent films (such as Greed, London After Midnight) that have been lost before or have been partially salvaged. Lost Horizon (the 1937 Capra film) is another film that has ran into significant restoration problems.

Tangentially, you could also add a section on the differing television and film standards around the world how issues like 3/2 pulldown and PAL speedup affect changes in films as they are telecined for video playback (all television being video at somepoint, afterall). One similar example, that's somewhat unique is Criterion's edition of Berlin Alexanderplatz, which was shot in 25fps for european television and has a pitch shift to compensate the different video playback speed in america because they weren't able to replicate the frame rate 100% accurately without major video problems for release here.

Indeed, those issues would be interesting to discuss in the thesis. However, the problem is finding good sources. Articles etc. found via Google aren't allowed, but search engines such as Google Scholar or IEEE can be used. It's a lot tougher to find good material on this subject matter with those search engines.

---------------------------------------------------------------

Originally Posted by Robert Harris

To Seppo...

This really made my week. Mr. Harris, I can't thank you enough for your informative, helpful posts in this thread. It really made me feel more inspired and excited about writing the thesis and for the past two days it has been the only thing I've been able to think about. If there is one person whose expertise I'd appreciate the most it would be yours. So thank you very much and if it's okay I will probably contact you with some questions related to my thesis this spring once I get the writing fully going.

---------------------------------------------------------------


Quote:

Originally Posted by Robert Harris

Then one moves to audio, which has its own problems.

Indeed. I'm not sure yet whether I will discuss audio restoration or not. It will depend more on how many pages I have once I'm done with all the picture related issues. Unfortunately, many of the restoration documentaries don't discuss sound restoration in detail. Off the top of my head, the ones that did provide some good information were "Obsessed with Vertigo: New Life for Hitchcock's Masterpiece" and "Prettier than Ever: The Restoration of Oz".

---------------------------------------------------------------


One thing that I'm struggling with is the use of pictures in the thesis. It's difficult to imagine doing a thesis on digital film restoration without any pictures whatsoever. Basically, what my instructor has said is that I can't copy a picture from a website and put it in my thesis even if the source is mentioned. It is a violation of copyright. However, if I can get permission from the owner of the copyright, then it is okay. At least that's the way how this kind of copyright issue works in Finland. Is it the same in the U.S.? Some of the pictures that I would like to use are from the American Cinematographer article that Beth linked above. I would like to use two pictures ("before" and "after") to help explain what it took to restore the movie and that particular scene. At the bottom under each picture it says "Frame grabs courtesy of Warner Bros. Motion Picture Imaging, 1972, 2008 Paramount Pictures. © 2007 American Cinematographer". Who should I contact to ask permission or is this all wishful thinking? The pictures would only be used in a positive way as I would try to show that the restoration of "The Godfather" films is how digital film restoration should be done.

---------------------------------------------------------------


Sorry for the long post, but this is hardly 1/4 of what I've had on my mind the past couple of days. I will probably post some updates, questions etc. once I start writing. Big thanks to all of you who've helped me so far.

 

[BethHarrison]

Here is an interesting discussion by Schawn Belston who does restoration work for 20th Century Fox:
http://www.thedigitalbits.com/articles/031709htfchat.html

The chat primarily concerns restoring The Robe (1953). If you ever need to demonstrate what modern digital restoration techniques can do, just get the original (2001?) DVD release of The Robe and compare it to the Blu-ray!

Belston would be another interesting person to get in contact with. Maybe one of the moderators here who arranged the chat would have his contact details? I'm am pretty sure he has (or at least had) a Facebook page, but I can't find it. It's possible this is his Twitter account, but don't hold me to it: http://twitter.com/SchawnB

He also mentions working with Audio Mechanics http://www.audiomechanics.com/ who were responsible for restoring the soundtrack, which included removing wow and flutter (distortion caused by the playback speeding up and slowing down). Apparently it was so processor intensive to restore the audio that only a few seconds could be done each day!

The April, 2009 issue of American Cinematographer has an article on restoring The Robe. Unfortunately the article isn't online, but you could buy the back issue (or find it in a good uni library):
http://wwww.theasc.com/ac_magazine/April2009/toc.php

Here's another short artlce:
http://www.studiodaily.com/filmandvideo/currentissue/10226.html

My recollection is that the film elements were scanned at 4K converted to 2K to restoration, before being scanned back out to film. 2K was considered adequate for the film because it was one of the earliest Eastmancolor films, and it was filmed with a late 1920s anamorphic lens, so the image just doesn't have much resolution. The big challenge of the restoration was the fact there were many missing sections of the negative that had been replaced with a huge variety of dupe elements. So the restoration team had to work very hard to give the image over all consistency.

The article also contains a choice quote by Belston saying to the effect that the digital tools are improving so fast that what took a year to do on The Robe restoration could now be done in a few months.

Regarding reproducing images, film historian Kristin Thompson and a team of lawyers analysed legal issues concerning the reproduction of frame enlargements and publicity stills from films in academic research papers. You can read it here:
http://www.cmstudies.org/index.php?option=com_content&task=view&id=60&Itemid=110

Thompson also wrote a blog post about changes since that report was written due to technological developments:
http://www.davidbordwell.net/blog/?p=2127

Basically using frame enlargements is legal based on the purpose of the reproduction. If you were to reproduce pictures of Humphrey Bogart from Casablanca to sell as posters, then that would obviously be illegal (both with Warner, who owns the copyrighted work 'Casablanca', and the Bogart estate, which owns Bogart's 'image'). But if you are doing it so you can write a thesis on digital film restoration, then that will be considered fair use because it is a scholarly work. At least that is how the law applies where I am!

 

[Seppo]

Thank you very much, Beth. I just ordered a digital version of the April, 2009 issue of American Cinematographer. I'll probably get to read it sometime this week; they'll email me when it's available for me.

I emailed American Cinematographer asking permission to use the pictures from the "Post Focus" article. As expected, they said they don't own the rights to the images; that belongs to Warner Bros. and Paramount Pictures. She said she'd forward my request to the editor who handled the story for a studio contact.

If I have close to 25 pages by the time I've gone through picture restoration, I'll probably then save sound restoration for the master's thesis.

Here's my planned structure for the (25-page) thesis:
1. Introduction
2. Starting a restoration project & searching for film elements
3. Scanning of film elements
4. Image cleaning
5. Color grading/correction
(6. Sound restoration)
7. After restoration/Finished product
8. Conclusions

 

[Seppo]

Well, the first thesis is pretty much done. A total of 29 pages, 7 700 words, four pictures, 25 sources (ranging from books to conferences to studies to magazine articles to documentaries). I used the digital restoration of The Godfather films as the main use case. This means that each chapter ends with a description how the main things discussed in the chapter worked out in that particular restoration.

First chapter is the introduction chapter and the last one provides conclusions to the overall thing. In the second chapter I talked about the bad condition a lot of popular, older films have been in, the search for film elements, going from photochemical restoration to digital and the DI process.

Third chapter deals with the scanning of film elements. I talk about the resolution of scanners, resolution of 35 mm film, workflow (i.e. 4K-2K-4K) and the differences between optical printing and digital scanning. Fourth chapter is about cleaning and fixing the image. This includes things like defects overview, detecting one-frame defects, removing them by copying information from other frames, the use of the “blur” tool, automated scratch and dirt removal, grain removal, the handling of production mistakes. Fifth chapter shortly deals with color grading. This chapter could’ve been longer and more comprehensive. It talks a bit about the ethical issues and gives an example how video-based source material can be fixed if there are color bleeds or something like that. The sixth chapter talks about what happens after the picture has been manipulated digitally. The cost of restoring digital data files, ethical questions of film restorations etc. Overall, it’s a bit of a mixed salad as you end up gathering information from various sources and try to create a coherent whole. The struggle with doing a scientific thesis is that everything has to be refered to – there can’t be any own opinions or things that you know are true but can’t prove them if there is no source. I don’t know at all if I succeeded or not – that’s up for the reader to decide if the whole thing makes any sense whatsoever. There are a lot of things that were left out; for example, I didn’t get to talk about sound restoration at all because I run out of pages. I’d like thank HTF for all the support and advice. It has been inspirational. For example, one very fortunate thing happened after purchasing the digital version of the April 2009 issue of American Cinematographer. I was supposed to get the link for the magazine in two weeks but that didn’t happen. I contacted customer service and they apologized for the error and gave me the link and the rights to read all the other previously released digital versions of AC! This meant that I got to read all the issues from April 2010 to early 2007 for free. I typed in “restoration” in the search engine and bang, several useful articles dealing with restoration provided some very useful information. So thanks for recommending the article, Beth. Now it’s time to start doing research for the master’s thesis. The plan is to write the whole thing this summer. Length will be somewhere between 50-100 pages. Doing interviews would probably be the best way to do it. So far I have three excellent sources: two are HTF members, and one works at Reliance Mediaworks (formerly Lowry Digital). My instructor said that there should preferably be maybe 7-10 people that I could interview. I sent emails to Warner Bros. and Criterion, but haven’t heard back (yet). I’ll probably try to contact Deluxe Digital. Shawn Belston would definitely be an interesting person to get in contact with.

Any tips who else I should try to contact?

 

[Flemming.K]

Great read this thread. If your thesis ever gets available for the public to read, please post a hint. I would love to read it and learn about the subject.

Good luck!

 

[Seppo]

Thought I'd post an update as it has been a year to date since my previous post. Time does fly.


Last week I finished the first cut of my master's thesis titled Digital Film Restoration and Remastering. Roughly 62 pages, written in English, and yes, I'll upload once it has been graded by my university. Tomorrow morning we're gonna go over with my thesis instructors. I'll make all the necessary adjustments and then send the thesis to some of the interviewees who requested to validate it in order to make sure I haven't used any of their comments incorrectly. I'll do a presentation of the thesis at a seminar later this month or early next month. In any case, early(-ish) June is when I should finally be finished with the whole thing.


The main research question was: "How is digital technology used in the restoration and remastering of motion picture film and what kind of ethical debates might there be?" Qualitative research method was utilized in this thesis, and qualitative interview was used as the data gathering tool. In short, I basically interviewed 12 experts in the span of one year, the first one in April 2010 and the last one this April. With the exception of one, all of these were email interviews. The amount of questions could range from 3 to 15+. The questions ranged from technical to ethical to general. The quickest time between sending questions and receiving answers was under 30 minutes, and the longest time was nearly 6 months. There were also over half a dozen people who agreed to answer, but that was the last I heard from them.


I traveled to Helsinki last July. That's where the most substantial interview for this study took place. This sound restoration expert was on vacation and visited the city for half a day. We managed to arrange a meeting on a six-star cruise ship. I made an audio recording of the 105-minute interview. A really nice guy, gave a bunch of great information, and was very patient with my nervous ticks and crappy questions. The empirical research has been full of great moments, but by far this was the highlight. It was good that the only live interview was done with an audio restoration expert instead of picture restoration, as sound often gets less attention.


Basically the thesis is divided into two bigger chapters: "Literature Review" and "The Study". The other chapters are: "Abstract", "Introduction", "Research Method" and "Conclusions" (2 or 3 pages each). "Literature Review" basically consists of information gathered from 48 references. Pretty much all of "The Study" chapter is constructed from the responses the experts shared in the interviews. As this thesis attempts to be scientific, it's pretty much all about referencing what others have said. None of my views or opinions are expressed, which I think is a good thing.


When all is said and done, however, I don't think it's a good thesis. First of all, it contains zero new information. I think "Well, duh!" would be a common reaction among readers, particularly on this forum. However, it is designed in a way that you don't need any knowledge of the subject matter beforehand, and most of the terms and technical things are explained. Algorithms are not presented.


I'm just not a good writer, and it was an extra challenge to write it in English. Referencing what others have said without directly quoting has been particularly difficult, as I can't efficiently reference other peoples' texts in my own words. Instead, the changes made are very, very cosmetic (e.g. changing the order of words and using synonyms). This makes the end result look more like over 60 people wrote it instead of one. Also, if I couldn't get information on some issue or if an interview's response is too diplomatic, I can't go on and presume something (which I wouldn't do in any case), even if it seems logical and makes perfect sense. Again, one has to make do with the elements available.


Another thing of concern is that frankly I don't know if I've provided even a plausible look at the most typical challenges faced in film restoration. Each film has similarities and differences. There were cases when I got a "depends on the project" type of answer. I'd say the most efficient way to study this subject matter would be in the form of participant observation, where the researcher is actually involved in a project. He/she would make notes, document things and interview the people doing the work. The case study - actually, at least two different case studies would be recommended - would then be published.



Feel free to comment or ask if you have any queries.

 

[LiseLise]

Hi, I recently found a homepage, which publishes theses. They seem to sell books about almost every topic. Maybe this was is interesting for you? http://www.grin.com/book/2326630/digital-intermediates-for-film-and-video Bye.

 

[Seppo]

Hello all, If anyone wants to take a look, I've uploaded the final, graded version of my master's thesis here. EDIT: Link updated.

 

[DavidJ]

Thanks for sharing. I look forward to reading it.

 

[Seppo]

You're welcome. I hope it won't be a complete waste of time, and of course any type of feedback is much appreciated.

 

[benbess]

You're welcome. I hope it won't be a complete waste of time, and of course any type of feedback is much appreciated.

I couldn't get it to read. Any chance you'd just cut and paste some of your conclusions here? Just maybe 2-3 good pages? If so, thanks....

 

[Seppo]

Ben, I uploaded the thesis to another place. You can take a look at the PDF file directly here. Let me know if it doesn't work.