Specification of writing a script
By: Vanu Srikanth, Andy Swan, Casey Williams, Patrick Pearson
Dubbing and subtitling are inherently creative processes. At Netflix, we try to make watching shows in every language as enjoyable as the original language, whether a member sees original or dubbed audio, closed captions, forced descriptions, subtitles or any combination of their choice. Capturing creative vision and subtlety in translation is important to achieve this goal. Creating a dub or subtitle is a complex, multi-step process that includes:
- Transcribe and timing conversations in the original language from a complete show to create a source transcription text
- Note the events of the dialogue with character information and other annotations
- Creating localization notes to guide further adaptation
- Translating a conversation into a spoken language
- Adapting translations to dubbing and subtitling specifications; Consider changes in reading speed and shot for subtitles, similar to actor lip movements in ex-dubs
Script files are the essence and driving force of the localization workflow. They carry conversations, timecodes and other information, as they travel from one tool to another to be adapted for transcription, translation and acting by voice artists. Dub scripts, audio descriptions, forced descriptions, closed captions, and subtitles all need to be written on complex tools that manage the time, position, and formatting of the text on the screen.
Currently, scripts are distributed on Netflix in a variety of ways – Microsoft Word, PDF, Microsoft Excel, Rich Text Files, to name a few. They carry important information such as dialogs, timecodes, annotations and other localization contexts. However, variations in these file formats and inconsistent ways of specifying such information between them have in the past attempted to facilitate localization workflows.
We have decided to clear this stumbling block by creating a new writing specification called Timed Text Authoring Generation (TTAL). This enables seamless exchange of script files between different authors and prompting tools in the localization pipeline. A TTAL file contains all relevant information such as script type, dialog, timecode, metadata, native language text, transcript text, language information, etc.
By defining vocabulary and annotations around time text, we try to simplify our approach to capturing, storing, and sharing materials across localization pipelines. The name TTL has been carefully crafted to convey its purpose and use:
- The name “timed text” means it carries the dialogue with the corresponding timecode
- It is used to write scripts in docking and subtitling “authors”
- The “clan” part of the name talks about how the script evolved from the time the show was made in one language when it was performed in another language by voice actors or subtitled in another language.
In short, TTAL is designed to simplify the task of script writing, so creative energy is spent in the dubbing and subtitling industry rather than managing adaptive and recorded script distribution.
We’ve been writing and exchanging TTAL scripts with our technology partners and English dubbing partners for the past few months, as well as running related workflows. We accept adapted scripts before recording and once the recording is complete. This workflow, illustrated below, has enabled our dubbing partners to deliver more accurate scripts at critical moments.
As an initial step, we worked closely with their dubbing technology providers using TSAL as the underlying format using JSON. We appreciate the efforts made by the developers of these products to run TTAL and provide our important feedback to improve it.
Third party tools that support script import and export to TTAL:
Adopting TTAL as the only way to exchange scripts with tools in the localization pipeline would be beneficial in multiple ways for all players in the ecosystem. This will continuously improve the capture of structured dub scripts which will enable us to better analyze and leverage the contents of the script, facilitate the flow of work and enable interoperability between tools in the localization pipeline. Ultimately, these will meet Netflix’s unwavering goal of meeting and maintaining a creative perspective throughout the localization process.
It’s just the beginning. We have laid a solid foundation for enabling inter operability by developing a specification for script writing. We have worked with some dubbing technology developers to include TTAL in their products, and have changed the specification according to the response of these early adopters. In addition, we have conducted workflows with our English dubbing partners.
This effort has proven that the timed text authoring fills an important gap and benefits the entire localization ecosystem of technology developers and content creators, from individual transcribers and script writers, dubbing and subtitling service providers. We firmly believe that enabling tools to exchange scripts seamlessly will eliminate operational hassles and provide extra time and effort for the subtitle and dubbing, translation and adaptation industry.
Finally, TTAL is a growing specification. As TTAL adoption continues, we hope to learn more and improve the specification. To make it even more mature we are committed to continued collaboration with our localization partners and tool developers. If you’re interested in incorporating TTAL into your developer tools, please contact us at TTAL_Format@netflix.com to learn more about this exciting new specification and how you can use TTAL in your workflow. Please watch this video to know how TTAL Export works on VoiceQue