The Thief and the Cobbler: A Practice in Editing

My desk while editing The Thief: Polished, 2022

A little over two years ago I started a project titled "The Thief and the Cobbler: Polished" which was meant to be my own reconstruction of the film The Thief and the Cobbler explicitly using finished animation and storyboards. Not only was this an extreme exercise in video editing, which at the time was something I wanted to do a career, but it was also something I felt had promise and had never been done before. The only person to really step up and take the film by the reigns and recut it himself was Garrett Gilchrist, creator of the Recobbled Cut. Ever since his attempts, Recobbled Cut Mark 4 has been considered the definitive version of the troubled feature. While the Recobbled Cut is an incredible feat of editing, preservation, and perseverance, I always had the feeling that another approach could be put towards Thief.

The Thief was never finished, at least not to Richard Williams' expectations in what he wanted. Though, I don't think he ever truly knew what he wanted. I think I know what he wanted spiritually, to create his magnum opus, his love letter to animation, but I don't think he ever truly figured out what was best for the movie itself. Due to his constant need for perfection and his lack of organization, the film was taken away from him and "completed" in a way even he knew was wrong. Because of this, no cut of the film will be accurate. There will always be something missing, a creative decision made by the editor instead of Williams, something not polished to how he wanted. Through the various cuts of the film that we were given in the 90s as it passed several hands and the little bits and scraps Gilchrist managed to scrounge up from trashcan finds and eBay listings, we have the pieces to a jigsaw puzzle that can be put together in any decent way and be considered "complete" because there is no picture on the box for what it's supposed to be.

So while watching Recobbled Cut 4 for the fifth or so time, I decided I wanted to try my own hand. I went through and downloaded all of Gilchrist's old "Thief Scrapbooks" giant zip files filled with newspaper clippings, concept art, and clips he had accumulated from Williams' work, as well as other uploads he had put up over the years. Without Gilchrist's multi-year obsession for this film we wouldn't have half of what's currently available. It was impossible to do it properly without his collection. With that in hand along with whatever else I could find online, I created my own little collection of Thief material. One could, theoretically, put it all together in all sorts of ways to create their own version of the film.

I created a Discord server specifically for the project and invited people that, at the time, I had been talking to at some extent about the project and were passionate about the film like I was. In retrospect, in all honesty, most of them weren't very helpful besides bounce a few ideas around. However, Notelu, someone who I have known for many years, was helpful in upscaling the low quality footage we had on hand and did it subtly. Key word is subtly. Later on I had decided to invite a particular person to the project to help out with upscale and restoration, though that later proved to be a mistake. Their methods were far more abrasive, with every upscaled shot removing key details. Though the straw that broke the camel's back was when they wanted to use Gilchrist's restorations instead of doing it all by hand. When I told them stealing other people's restorations is precisely not what we wanted to do, but after that they decided to leave. Perhaps it was for the best.

A comparison between the original and restoration of spire turn around shot.

From the beginning it was a challenge. Though, thankfully, Notelu had sent me a link to some software named Phoenix Finish. It was the only film restoration software I managed to find for free that I could actually understand. Using Phoenix, I managed to remove the dirt and scratches from the beginning crystal ball animations. In the above comparison, I also attempted in creating what I call a contrast map, though I'm sure it's a technique that already has a name and has already been used by many others. I took the original animation, converted it to greyscale, and leveled the contrast for deep shadows, placing it on top and blending it with the original. This meant I could restore a shot using only what I was provided, without blending it with other sources such as pencil tests for better visibility. This technique, in my opinion, leads the viewer to understand it as simply a damaged film rather than an obvious restoration. I also made changes to the beginning crystal ball shots themselves, many unfinished and because of that, removed. This meant I had to shorten the beginning dialogue and speed up the ball's rise. This worked to my benefit, no longer do you have to wait 5 minutes for the movie to start. Originally, that 5 minutes wouldn't have been a problem, because the credits were meant to appear above the ball as it came towards the viewer. However, I felt that no one is qualified to mimic Williams' golden typography he had intended, and decided not to include them. This was a sacrifice I made for better viewing, not for accuracy. Such as many changes I made. The Recobbled Cut already covered most bases for accuracy.

The final title composite

I used a similar technique with the film title itself. Given it was renamed multiple times in the past, the best copy that can be found of the original animated title is on the 2009 DVD. I took that, greyscaled, leveled, stabilized, removed dot crawl, and placed on top of a still image of the logotype from a poster. It's probably the best rendition of the animated title card anywhere currently and probably the thing I'm most proud of out of the whole project.

The original workprint vs what was intended

Later on as I reviewed the workprint for the film I noticed that there were grease pencil markings at a few splices. Not every splice, just a handful. When mentioning this to Notelu she told me that they were left there by the editor, likely Williams himself, as a note for transitions. At the time I was very uneducated on 35mm film and how it functioned, and even now I find difficulty finding documentation on editor's marks on films. Even before I writing this I reviewed books on film editing from the time and none of them make note of specific marks that determine what kind of transition was intended. Nonetheless, it seems fairly obvious what they had in mind when reviewing the workprint.

Soon I'd notice more marks throughout the workprint. Lines crossing between each other between two scenes were dissolves, singular lines were fade ins/outs (or sometimes crossing lines- likely remnants from a past dissolve from an earlier workprint.) The scene with King Nod's dream of the soldier coming to warn him of One Eye made far more sense in how it was supposed to be constructed. The individual shots of the soldier and the growing eye were meant to dissolve into each other. Ironically, the lines that gave these instructions also obstructed the image, and I wasn't able to construct what was intended with what I had, given that the workprint was the only version of the film to include those shots.

As I was editing the film I would discover that many of the sequences that were never finished didn't really need to be included. While this, to many, may seem sacrilegious, I felt that it was worthy to try. Yum Yum never searches for Tack in the prison and only sees him again when he appears behind the curtain, that was the biggest change made. A few scenes were reordered as well. Another big change, and discovery, is when King Nod rejects Zig Zag's ransom to wed Yum Yum. This scene was never animated under Williams. When it was animated under Calvert, the lines were entirely rewritten. However, laying Zig Zag's animation on top of the audio from the workprint revealed that they had animated his reaction with that audio for reference. It is far from perfect and I prefer the original boards, however I included Calvert's animation anyway because of the goal of the project itself (not to use any unfinished animation) as well as how interesting the coincidence was in the first place.

Later on I discovered that Gilchrist's scans of the film elements that remain of the war machine sequence were very poor. Gilchrist himself agrees with this, and has been wanting to get it rescanned for quite some time. The problem is over-exposure, everything red is too red and because of that many details in the machine have been blown out and lost. Incredibly, I managed to find someone who had an older scan of these film elements that were done for a previous mark of the recobbled cut. While they were of lower quality, the white balance was far more accurate and easier to look at.

Unfortunately, by the time I made it to this point in the film, I was beginning to get tired of the project. Most of what I wanted to discover and learn I had done. There was little to keep me motivated to keep working, knowing that there will never be a "perfect" version of this movie to watch. I realized, most of the reason I was doing this in the first place was just to have a version of the film that I preferred to watch and enjoy without jarring changes in quality. I edited it like a producer instead of a historian. As my motivation dwindled, I decided to cancel the project. I had finished editing it, mostly, but I had hardly done any restoration, which was something I was really wanting to do in the beginning. (It didn't help that I got a new computer half way through that for some reason does not work with Phoenix Finish.) Because of this, the war machine scene I just put in from the workprint. I did not have the energy anymore to edit the entire thing together from the film scans.

A little less than two years later, for reasons I don't even remember now even though it was quite recent, I decided to release what I had done. I opened the old premiere file, tied up a few loose ends in editing and shipped it out on YouTube. I think I may be the only other person to "complete" a ground-up edit from The Thief and the Cobbler. Nearly every shot I placed by hand from various sources. I found a new respect for the Recobbled Cut, fan edits, and editors in general.

Every now and then I think about revisiting it, by finishing it how I intended. But I don't think it's time. So, for the time being, here is the "it's good enough" version. Complete without credits! Like how it was intended to end. (Much to nearly every animator's dismay...)