by Damián Santilli
I’ve always loved technology, and during my years at university studying translation back in the early 2000s, I quickly realized that I would always be able to link my passion for computers with translation. Additionally, I have always also been a movie buff, so that’s why I knew I had to become a subtitler. Luckily, I have been able to both experiment with and teach translation technology to students and colleagues and work as a professional subtitler since 2007.
When I started subtitling and dealing with my first clients, I was surprised to learn that they would not demand that I use any specific software. Nonetheless, I decided to dive into Subtitle Workshop and the like and spent several years subtitling with free software. Simultaneously, I was working as a technical translator using CAT Tools, and teaching them at university and other places, so it always baffled me that there wasn’t a subtitling-specific CAT tool. In recent years, as we moved towards translation environment tools (TEnTs), it became more and more apparent that
we needed to integrate translation memories (TM), termbases (TB), and all the other tools these systems offer, into subtitling environments.
In 2019, I decided to check the state of play of translation tools for my presentation at HispaTAV in Barcelona, and I was glad to see that the leading translation software companies offered some alternatives using TMs and TBs for subtitling. Today, I’m going to be revisiting the subject for Deep Focus.
We have the technology. Let’s use it!
As a subtitler and technical translator myself, I find it hard to believe that we spent so many years not using TMs and TBs in subtitling projects. Luckily, that’s changing in some areas, particularly for projects where a timed text file is provided. I’ve been able to translate some subtitling projects using Trados Studio, and I imagine that some LSPs are also using memoQ. But my main concern here is that, as a freelance subtitler, we don’t always deal with timed text files. When you’re working with direct clients, like an independent movie director or a production company, there’s seldom a need to create a timed text file in the original language, so you usually produce an SRT file in the target language. This is where we fall short, as no translation tool available offers a combination of speech recognition technology and TMs and TBs to create translation units when working directly from source audio and not from a timed text file. Nonetheless, I encourage anyone reading this article to use available tools when subtitling into your target language from a timed text file.
I will talk about possibilities for the future a bit more at the end of this article, but now I want to focus on what we can do with the tools we have available at the moment.
Using TMs and TBs for Your Subtitling Projects
Traditionally, you could always add an SRT filter to a CAT Tool to translate these files, but it really wasn’t such a good idea to do so, because you didn’t have control over simple things such as segmentation or time code editing, and you couldn’t even preview the video in the same software. All of that changed a couple of years ago with the introduction of Studio Subtitling in Trados Studio and the Video Preview tool in memoQ. So, let’s say you have a series of webinars you need to translate or a whole season of a TV series, and you’ll be needing to create a team of 3 or 4 subtitlers to make the deadline. If–and this is a big if– the client provides timed text SRTs, you should probably stop using Subtitle Edit, or even EZTitles and the like, and jump into Trados Studio or memoQ, and carry out the project using translation memories and termbases, which, by the way, you can easily share in real-time with your colleagues without having to buy a professional version of the software. Both Trados Studio Freelance version and memoQ translator pro will suffice.
Currently, Trados Studio offers, in my opinion, the most comprehensive utility for subtitling projects with its Studio Subtitling app. While memoQ only allows you to translate SubRip (SRT) files and custom-made XLSXs, it offers a module for creating custom filters for files like Advanced SubStation Alpha (ASS), SubStation Alpha (SSA), Spruce Subtitle File (STL), WebVTT (VTT), and other common subtitle file types. On the other hand, Trados Studio offers support for ASS, SRT, VTT, STL and YouTube (SBV). So, while you’re welcome to try memoQ if you’ve never used a translation environment tool for subtitling, I will swiftly guide you through Trados’s utility.
The first thing you need to do is visit the SDL AppStore website and login into your account, or if you have the latest version of Trados Studio, you will find the AppStore embedded right into the tool, and you can access it on the Welcome screen. Once you’re in the AppStore, you’ll need to download and install the Studio Subtitling app and the filetypes you’ll need. I suggest you install them all, just in case. After the app is on your Trados Studio, you just need to open a file for translation as you would normally, and then, in the Editor, you’ll have to click on View > Subtitling Preview, to activate the video preview window, and then View > Subtitling Data to see information like the start and end times of subtitles, characters, words, WPM, CPS, CPL, etc., which you can change, if you need to.
Then, you can create as many translation memories and termbases as you want, share them among your team, and take advantage of the same features you would use while translating any technical document. This can be extremely helpful, given the right working conditions, as you can reduce terminology problems, and depending on the subject you’re dealing with, you might even save a lot of time using Trados Studio’s upLIFT technology (TM matching based on fragments). This is also true, obviously, when working with memoQ, but remember, only with SRT files.
Okay, but I don’t work with timed text files. What can I do?
I’ve been passionate about the subject of using TMs and TBs for subtitling for a while now, mainly because I have plenty of direct clients, and as I mentioned before I don’t usually get timed text files. Recently, I have had projects where my client wanted me to translate their videos into English, but also needed a closed caption file in Spanish. In those cases, I was able to use a translation environment tool, but in most situations, I’ve been dealing with my files in Subtitle Edit and Ooona. At least they offer robust QA options and even machine translation.
So, I have to say that there’s not much you can do if you’re like me and want to integrate TMs and TBs into your workflows. And here, the question is: what would it take to use TMs and TBs for all of my projects? Well, the answer is kind of simple, and there’s no need to reinvent the subtitling wheel. We need software, whether it’s a translation tool like memoQ or a subtitling tool like Subtitle Edit, that adds a speech recognition feature that helps you create a timed text source file. I mean, it’s not that difficult, since YouTube, Facebook, and others are already creating timed closed caption files. You just need to have this in your subtitling software and then you’ll be able to create translation units, even though the speech recognition quality might not be the best.
So, Is the Future Almost Here?
Well, the problem is that we need three things: excellent subtitling software, a speech recognition feature, and a robust translation tool that can manage translation memories, termbases, and other common features. But those three things are usually sold by three different companies, and it might not be as easy as it seems to create one tool that combines Google’s speech recognition technology with Ooona’s professional tools for subtitling and captioning, and Trados’ or memoQ’s translation software suite.
Sure, some streaming companies are already using translation memories, termbases, and machine translation and Ooona may soon incorporate speech recognition technology, but there’s always going to be something missing, right? Maybe it’s a missing feature, or maybe you can only access those features if you work for a particular company.
Maybe the future is almost here, and developers will surprise us soon. I can only hope that, if this technology arrives in the next couple of years, it is accessible to freelance translators and not only to big companies.