PERSONAL PROJECTS

SAMURAIS, GANGSTERS, AND DIGITAL PIRACY

Beginning in July of 2025, I began developing a fascination with a particular foreign film. Actually, to be more precise, a clip from a particular foreign film I happened to see while scrolling one day. Despite my best efforts, I could not find any information about the movie, the actors, or story in the description or the comments of the post. Eventually, with some reverse image searching, I did come across a name, Kaikyô, and found an accompanying imdb page with some information on the film. Released in 1982, Kaikyô is an adventure/drama based on the story of the harrowing construction of the first tunnel between Japan's northernmost island of Hokkaido from the island of Honshu. Along with the description was a poster, but no images or video from the film itself.

One of the fascinating aspects of the digital age, that many may not consider, is the immense amount of creative material that falls through the cracks. In a time where there is more information at our fingertips than one would have had access to in a lifetime, the idea that there are stories, communities, or immense creative undertakings that are lost is almost unthinkable. One may think, “if it was really worth seeing, someone would have saved it.” In reality, that is not always the case. In a time of growing corporate media consolidation, strict intellectual property litigation, and pay-walled streaming services, the barrier for media to be lost, at least to the average consumer, is not particularly high. So with that in mind, I was determined to get my hands on this film. Not just to satisfy my own curiosity, but to preserve a piece of media history and honor the work of fellow creatives.

However, it began to dawn on me that the films obscurity was not my only issue. Even if I were to get my hands on this film, that seemingly had zero digital footprint, being able to watch it was a different challenge entirely. In all likelihood, this film never had an english release, so if I were to get my hands on it, I would have to find a way to translate it. And, while I was correct in that assumption, the ability of translating the movie did pose as a similarly intriguing challenge. Preserving a film is one thing, but being able to introduce a work of art to an audience that was previously inaccessible is too tempting to pass on. What follows is my path to building and translating catalog of foreign films that seemingly never existed.

Before solving any of the other challenges, I would first need to get a copy of the film. So, how does one buy a film that for which you cannot search? To start, I began to look for eBay listing for movies from people who had acted in Kaikyô, which brought me to our leading man, Ken Takakura. As a young man Takakura was often seen portraying the street-wise tough-guy in many Yakuza films. However in 1982, the more mature actor found himself in the role of the stoic structural engineer whose personal ambition drives the story of Kaikyô. It was under Takakura’s name that I found a seven disc collection of films from throughout his career. In all likelihood a Chinese bootleg, this collection of Blu-rays not only included Kaikyô, but close to thirty others of varying obscurity. And while the work had increase drastically, so did the opportunity for new gems to be uncovered.

With the collection purchased, the next challenge arose, data encryption. While not relevant in the same ways in today’s subscription-based media ecosystem, physical media posed an issue to Hollywood distributors. Without encryption, “ilicit” copies of VHS tapes, DVDs, and Blu-rays would flood the streets, robbing Hollywood and its producers of their intellectual property. Encryption of this kind, and the arms race between distributor and consumer, is fascinating and deserving of a much deeper dive; however, for our purposes, this issues has largely been solved in the modern era. With the right hardware (I used this Blu-ray drive) and the right software, decrypting these discs is easy.

A movie is much more than a single file. On any disc there are files for video, audio, chapter selections, metadata, and most importantly for our purposes, subtitles. The containers that hold all of that information make for efficiently storage and transmission of this data by keeping it in one place. Now, the working theory for the whole project was that if I were somehow able to open that container and extract the subtitle file, translating it would be as easy as translating any text document. Unfortunately, what is stored inside a Blu-ray is not a simple text document, rather a series of images overlayed on the screen which contain the corresponding characters. For the purposes of the consumer, it makes no difference if the text we are reading is actual text or an image of text, but it does to the computer. In order for the computer to discern the text from these images an Optical Character Recognition software needs to be run. For this process I used a software called SubtitleEdit which is a shockingly robust program that allows you to visually edit subtitles within their video containers and also supports OCR analysis. This software also contained an rudimentary auto-translate feature linked to google translate through an API. This not only allowed me to export an accurate version of the “original” chinese subtitles (remember it’s a bootleg), but also a rudimentary english version as well.

With the two versions exported as text documents, the final translation can be made. By properly instructing a large language model to maintain the strict format, the original text can be fed in and the following translation can be cross-referenced with the rudimentary english version made earlier. Using a video playback software like VLC, this improved translation can be imported to the selected movie once converted to an SRT file.