[How to transcribe]
Transcribing means creating
same-language subtitles.
OK, but what do we need transcripts for?
Well, same-language subtitles
help in three ways.
They allow
Deaf and hard-of-hearing viewers
to access the talk.
They help to spread
the ideas in the talk online,
because once a video has a transcript,
it will show up in Google
when people search for related topics.
And finally, once a talk has a transcript,
it can be translated into other languages
and set free into the world.
So, OK, what's transcribing like?
Generally, while transcribing,
you'll type down what you hear
and set up the time
when the subtitle shows up
and when it disappears.
However, there are a few
other simple rules
to keep in mind while you're working
on your transcript.
Don't transcribe slips of the tongue
and obvious mistakes,
like when the speaker says
"we thinks" instead of "we think."
Make every subtitle
a clean little bit of text,
even if the speaker
changes their mind mid-sentence
or is struggling with grammar
when giving a talk in a foreign language.
Keep to the limits for line length,
subtitle length and reading speed.
And remember that,
just like when you're translating,
on rare occasions,
when adjusting the timing doesn't help,
you will need to compress
the text in the subtitle
to maintain the reading speed.
Try to synchronize the subtitle
with what is being said.
It's OK to have your subtitle run a little
into the time the next sentence is spoken,
especially if you need that
for good reading speed.
However, don't start your subtitle
more than about 100 milliseconds
before the next sentence is spoken.
Otherwise, you'll be giving the viewer
a weird sense of precognition
when they see the speaker's body language
doesn't match the subtitle.
And don't let the subtitle stay
on the screen longer than about 1 second
after the speaker says
the equivalent bit of speech.
Don't end the subtitle
with a bit of the next sentence.
As much as possible, make your subtitles
full sentences or clauses.
This will make them easier to follow,
but it will also make them
easier to translate,
since the grammar of the target language
may make it impossible
to divide the sentence
the way it's split up in the transcript.
Include sound information
for Deaf and hard-of-hearing viewers.
Using parentheses,
indicate where there's laughter,
music and applause,
identify off-screen speaker changes
and describe every other sound
essential to understanding the talk.
Don't have any subtitle display
shorter than 1 second and longer than 7.
For longer bits of music or applause,
just let the sound information display
for the first 3 seconds,
and then indicate when the sound ends.
If there is on-screen text
in the language of the talk,
like embedded subtitles
in a video played on the stage,
transcribe this text if it's possible
without overlapping other subtitles.
This will allow that on-screen text
to be translated into other languages.
And to indicate that it's on-screen text
and not what the speaker is saying,
use square brackets.
Do some research to get the right spelling
of the proper names used in the talk.
If you can't make out
what the speaker is saying,
ask in the "I transcribe TEDx talks" group
or in your language's group on Facebook.
And most of all,
remember that by transcribing,
you're doing the whole world
a great service,
because you allow the ideas in the talk
to reach a wider, global audience.
If you want to learn more
about transcribing,
check out our more detailed guide,
"How to tackle a Transcript."
And for now,
happy transcribing and translating!