Grow Your Market: Learn to Create Netflix-Quality Subtitles
| by Max Troyer
World-class subtitling: best practices for Netflix-quality subtitling, including creating open and closed captions and managing subtitling workflow.
Translation crowdsourcing is a growing trend in the localization industry.
Understandable, because it can be an efficient way to complete large translation projects in a short amount of time. It’s been successful for large companies such as Facebook and Google, as well as for nonprofit organizations such as Translators without Borders.
Crowdsourcing allows organizations to utilize the collective skills and knowledge of a community of users. But it can be complicated to manage a crowdsourced translation project. How to best recruit a sufficient volunteer community (quantity) and ensure excellence in end-translations (quality)?
Depending on your project, you may want to implement a vetting process in order to choose the translators who would be the best fit for your project. This process can be as simple as a survey to acquire demographic information, such as age, country of origin, or native language, or it can go as far as a translation test, if you wish, for only the most skilled translators.
Next, you will want to separate your content into small chunks so your translators do not feel burned out after each translation. In other words: microtasking. As users are likely not professional translators, having a short word or phrase to translate will ensure that they are able to handle the translations with ease.
Finally, when setting up your translation platform, make sure users have access to all of the necessary resources, such as any translation memories, termbases, or contexts for the translations at hand.
With thousands of users and translations constantly coming in, it would be both costly and impractical to bring in a professional to review every translation you receive. Instead, a system in which the community judges its own translations can really automate the process. How would this look? Once users submit a translation, it goes up for review. Your platform can be configured so that other users can vote or rate the translation. Finally, the highest-rated translations will stick around and the lowest-rated ones will be discarded.
Lastly, plan for trolls. In order to protect your translations from sabotage, you’ll want to set up some sort of filter system in order to prohibit certain entries. This can be done via a “blacklist” of words and phrases, such as profanity, in the target language.
Once you have highly rated translations by the community, the next step in ensuring high-quality crowdsourced translation is a final-review stage. This can be done either by professional translators ($) or by your top-rated users (if you’ve implemented a leaderboard, as mentioned previously in the section about gamification).
Although the above steps are labeled before, during, and after, each one can actually support the next in a continuous cycle. For example, let’s say there are a high number of top-rated translations by a specific user. In this case, your user would move up the leaderboard and become a high-ranking translator. The opinions of your high-ranking translators would likely be more valuable than those of lower-ranking translators. Therefore, implementing a weighted voting system would allow the votes of the high-ranking translators to be more determinant in deciding the best translations.
Similarly, the quality of the translations can ultimately shed light on the efficacy of the translation platform as a whole. Tweaks can periodically be made along the way in order to improve the overall process.
All in all, if implemented properly, these best practices can allow the translation crowdsourcing process to run smoothly and yield quality results.
Happy translating!
| by Max Troyer
World-class subtitling: best practices for Netflix-quality subtitling, including creating open and closed captions and managing subtitling workflow.
| by Sierra Abukins
Our localization experts partnered with the Office of Digital Learning and Inquiry to launch a new self-paced, short course—the first of many—on the subtitling industry standards pioneered by Netflix.
| by Jason Warburg
Middlebury Institute students organized a localization business case competition judged by industry professionals, testing their strategic thinking and creativity.