How to transcribe TEDxTalks in 10 steps
Hey, TEDx organizers and translators, here’s everything you need to know about transcribing TEDx talks in Amara - in 10 easy steps! TEDx talks are being added into TED Open Translation Project a few at a time so we can make sure they’re properly translated and credited. This is how do it:
If you don’t see talks from your event(s), or talks you'd like to transcribe or translate please use this form to submit them! In the meantime, you might like to help your friends and fellow organizers from around the world by translating their videos into languages you speak
1. Create a profile on TED and register with Amara, our subtitling partner. When you create your Amara profile, set the languages you speak and can translate to/from. This will enable you to find tasks you wish to work on. This short tutorial will guide you through the process.
2. Consider joining the Facebook group of TEDx talks transcribers and Facebook group of TED translators, and/or a group for your specific language. You can find the list of groups here. Translators are very friendly and can help you with anything and answer all your questions.
3. Once your application is approved, find the talk you want to transcribe. It’s easy – go to the TED team on Amara, choose the "TEDxTalks Project" from the list on the left. You can search for a specific talk using the search box within the "Videos"/"Tasks" tab. Alternatively, set the filters to find specific types of tasks. For example, choose “Transcribe” tasks in "TEDxTalks" in "my languages" assigned to "no one". Once you find the task, click “Perform task”. This short tutorial will guide you through the process.
4. This document explains how to use the Amara transcription interface. To learn about the process of transcribing, watch this short tutorial, and then watch this tutorial on subtitle style. You will find an overview in points 5-10 below:
5. Talk title and description: every talk should be titled in the following format: Talk title - Speaker’s name at TEDx[EventName]. Some talks contain a disclaimer in the description, that should be left and translated ("This talk was given at a local TEDx event, produced independently of the TED Conferences."). The description field of the talk should contain 2-3 sentences describing the contents of the talk, and they should both be in the original language of the talk. If it's not too long, the speaker's bio can also be transcribed and translated. The text explaining what the TEDx program is should be left out.
6. Start transcribing from the speaker’s first sentence. Do not put in (Music) or (Applause) at the beginning of the talk.
7. The content of the transcript must be broken up into subtitle lines, and these lines must be synchronized with the video. The subtitle line cannot be too long, because the viewer must be given enough time to look at and comprehend the video. Line duration and length: A single-line subtitle of any length should not stay on the screen for more than 7 seconds. A subtitle cannot stay on the screen for less than approximately 1.12 seconds, even if it only contains a single word, because subtitles with a shorter duration will just be a flash that most viewers will miss. In most cases one subtitle consists of up to two lines at up to 42 characters per each, with a maximum 84 characters per subtitle. You will also see a "characters/second" value for each line - to enable most people to follow, this must be a maximum of 21 characters / second. Line breaks: One sentence delivered in 30 seconds in the talk will often need to be divided into several subtitles. Also, every subtitle longer than 42 characters must be broken into two lines. When breaking lines and deciding where to end a subtitle and begin another, do not split phrases and linguistic or grammatical wholes (e.g. don't end the subtitle or break a line after an article or in the middle of somebody's name). Importantly, the subtitle shouldn't contain both the end of one sentence and the beginning of another. Learn more about line-breaking here. If transcribing in English, read our English style guide.
8. Sound representation in a transcript is meant to enable deaf and hard-of-hearing viewers (as well as viewers watching the talk without the sound on) to understand all the non-spoken auditory information that is necessary to comprehend the talk to the same degree that a hearing audience potentially would. Sound information must enclosed in parentheses, with the first word starting with a capital letter (e.g. (Laughter)). The most common types of sound representation are (Laughter), (Applause) and (Music). Off-camera or non-obvious speaker changes need to be represented in the transcript (learn more here).
9. Editing/compressing the text, reading speed, synchronization You don't need to transcribe every word verbatim: do not include slips of the tongue, sentence changes ("I don't want to... What I want to" --> "What I want to") and obvious mistakes ("we thinks" --> "we think"). However, do include on-screen text (like subtitles in a video played on the stage) to allow it to be translated into other languages. Once you've transcribed the text as subtitles, proceed to the synchronization step. Hit the down arrow key to indicate where the subtitle should stop displaying, and then go back and make finer edits using the sliders on the timeline. You will see a reading speed value on every synchronized subtitle. The reading speed can't be above 21 characters/second. If necessary, you can compress the text (rephrase the line to make it shorter while preserving the meaning), to get a reading speed at or below 21 ch/s (learn more here.) However, you will usually be able to improve the reading speed by simply extending the duration of a subtitle (don't worry if it goes a little into the time the next sentence is spoken, since good reading speed is more important than perfect synchronization). Don't let the subtitle lag over a long pause if not necessary for a good reading speed.
10. Once you finish the transcription task, click Complete. Another volunteer will then review the transcript, correcting any mistakes. Do not attempt to review transcripts until you have transcribed at least 90 minutes of talks (to learn more about reviewing, watch this tutorial). After the transcript has been reviewed, a Language Coordinator will need to approve it, in a quick second review step. The transcript can then be approved and published, which makes it available for translation in 100+ languages!
Once you've started, consider reading the full transcribing guide.