Juliet is the ultimate Audiovisual Localization Producer, she represents a collective of highly seasoned and skilled post-editing professionals and the most advanced localization platforms available today. She uses Artificial Intelligence to match the best talents with respective projects. Then, she provides the talents with top-notch Translation Memory, Glossaries, CAT Tools and all the good stuff AI has to offer so that the final product is delivered with quality.

Within our industry, when we say “audiovisual”, we are referring to TV shows, movies, video games, training/corporate videos, ads, social media videos, etc. Recently, the term also began to include audio-only (like podcasts and audiobooks) or visual-only (like graphics and presentations) assets. Localization is converting this content from one culture and language into another. It is more than just translating the words, we adapt the language so it can be perceived as closely as possible to the original intent by individuals from another country/culture, also known as locale.

Juliet helps implement AI into all the basic localization services, such as translation, subtitling and voice-dubbing. The subtitling service consists of analyzing the content, translating and adapting the information, and inserting the result on-screen as text, keeping the original audio in place. It sounds simple, but there are several rules that must be followed in order for the viewer to be able to read, understand and enjoy the content. Voice-dubbing consists of analyzing the content, translating and adapting it into a script-like format so that voice-actors (human or otherwise) can re-record an alternative audio track (with lip-sync or not) in another language. In the case of fictional content, not only does the talent have to vocalize the script they are given, but they must re-enact the scenes. For both subtitling and voice-dubbing, QA (quality assurance) usually ensues.

Our consulting fees vary based on size of business, industry, current Localization workflows and amount of original content to be localized. It usually ranges from 20 to 50% of the project savings in the quarter following the implementation of our plan. Current Localization average prices range from $0.10 to $0.25 (US)/word depending on language pair and volume. Video subtitling, between $20 and $30 (US)/minute. And voice-dubbing, around $50 to $100 (US)/minute. Our goal is to reduce these costs by, at least, 40%. Without compromising on quality, obviously.

Many people think of MT as Google Translate and go: "there's no way I can use that for my content! I've tried it and it's bad." And they are right. We are not talking about Google Translate here. We are going to choose a specific MT engine that works specifically to your needs, your terminology, your glossaries. We are going to feed it with your previously translated content so that it can learn how to do future translations with more and more accuracy as time goes by. That's what it's called Translation Memory and it belongs to your organization only, it's not public.

We start by learning about your business and Localization needs and mapping current processes. That usually involves external vendors and different internal departments/stakeholders. Once we understand what's going on, we recommend the best AI tools and systems to be applied in order to reduce costs by at least 40% without compromising current quality-and even improving it sometimes when human-translation or Localization best-practices are not being followed. Once the systems are up and running in your servers or the cloud, we onboard your team so it can start making the magic happen.

The Localization workflow usually involves the following components: CMS (Content Management System): where the original content to be localized is located. In Audiovisual Localization, it's usually a simple repository of assets (video, audio, artwork, text, metadata, etc). Generally we don't recommend changes in this one unless we feel it's too obsolete and/or it compromises the application of the new AI systems. TMS (Translation Management System): where the original assets are translated into target languages/locales. It keeps everything organized and all the stakeholders in the loop to make sure all the processes are done on time and on budget. It may or may not include MT (Machine Translation), TM (Translation Memory) and QA (Quality Assurance) components built-in. MT (Machine Translation): the engine that will be responsible for the actual translation of the content. There are over 20 high-quality MT engines available these days. The choice of one for your business will depend on many variables, such as budget, industry, workflow, compliance, etc. TM (Translation Memory): this is part of the MT engine and a very important one in it. It represents the "soul" of AI because it represents what it actually learned from your previously translated content and each interaction it has with humans so that future output is even more accurate than the one before. QA (Quality Assurance): this component usually ships with the TMS, but not necessarily. Depending on what kind of content you produce and what file formats you work with, we might pick one or another QA platform. Think about the QA system as a "grammar and spell check" for Localization, but with many other features. It can flag numerous errors beyond typos and poorly built sentences.