OTP Learning Series 05: Subtitle length and reading speed

This tutorial explains how to use the subtitle info box in Amara to work with subtitle length and reading speed. The links used in this video are:
http://translations.ted.org/wiki/How_to_break_lines
http://translations.ted.org/wiki/How_to_Compress_Subtitles

This video has been created for the volunteers working in the TED Open Translation Project. The TED Open Translation Project brings TEDTalks beyond the English-speaking world by offering subtitles, interactive transcripts and the ability for any talk to be translated by volunteers worldwide. Learn more at http://www.ted.com/pages/287

Let's talk about the new subtitle info box
in the Amara interface

and about how useful it can be
to transcribers and translators.

Any time you click a subtitle
to start editing it,

you will see this pop-up.

You may notice
that it starts with timing data,

but here, I'll focus on what you can do
with the other information in this box,

that is, the number
of characters in the subtitle

and the number of characters per second.

First, let's talk about...

42.

In languages that use the Latin alphabet,

if your subtitle
is more than 42 characters long,

you need to break it into two lines.

This makes the subtitle easier to follow

and prevents offline players
from breaking it in weird ways.

In the new info box,
you can see both the total subtitle length

and the number of characters
in each of the two lines.

When the total subtitle length in that box
goes over 42 characters,

it means that you need
to break the subtitle,

but also that you need
to think about how to break it.

After all, it's not like you want to have
one line of 42 characters

and then one line
with just one character in it.

So, break it in a way that the two lines
are as close in length as possible.

And let me give you an example.

This subtitle is 51 characters long.

Now, you could break it like this,

with one line of 14 characters
and another line of 36,

but it will be easier to read
if you make the lines more balanced,

like here, with one line of 27 characters
and another line of 23 below.

When breaking subtitles into lines,

also try to keep
"syntactic wholes" together.

What this means is that you don't want
to break up a phrase

that linguistically works as "one thing."

For example, in English,

you want to keep the article
together with the noun,

the adjective with the noun
that it modifies,

and a preposition
with the thing that it refers to.

You can find out more about line breaking
in a guide on our OTPedia

entitled "How to break lines."

Now, on to the other thing we can learn
from that subtitle info box.

The number of characters per second,

which is also known as...

The reading speed.

Because after all, it's not only about
how long the subtitle is,

but also about how long
it stays on the screen

and how much time
we give people to read it.

The ideal reading speed for languages
that use Latin script

is about 15 to 21 characters per second.

But where is this number coming from,
and why is it important?

Well, the reading speed tells you
how long your subtitle can actually be.

For example, if you have
a subtitle with 40 characters

which displays for two seconds,

you only require the viewer
to read at 20 characters per second.

This is easily manageable,

so 40 characters
would be fine in this case.

But if the same subtitle were to stay
on the screen for only one second,

the viewer would need to be able to read
at 40 characters per second,

which is impossible for most people
to keep up with,

and it would mean that the subtitle
needs to be shortened.

Fortunately, you don't need to calculate
all that stuff in your head,

because the new info box
in the Amara interface

will do the work for you.

Let's see an example of why maintaining
a convenient reading speed is important.

In this short clip from a TEDxNoviSad talk
by Dragana Marjanović,

the reading speed in all of the subtitles
is above 21 characters per second.

And what happened right after
the first year of our doing this project

was that I was awarded what was to be
my first very big architectural project,

which was to plaster up all of the surface
of the pretty four-story staircase

that was located in a nice building

located in the famous city
of Grenoble, France.

This, for me, was a four month long period
of work related practice

which was mostly in France
but also partly in Bulgaria

and after that was concluded,
I began to study again.

This time, it was the earth architecture
of Spain and Portugal.

When I was finally able to return home,
I immediately began my graduation project

which was about the earth architecture
of the Autonomous Province of Vojvodina.

So, anyone remembers
what happened in Bulgaria?

And where was the nice building?

For the previous clip,
I modified the English subtitles

created by Tatjana Jevdjic
and reviewed by Ivana Korom,

and made them longer, with reading speeds
much over the comfortable values.

Here is the same clip
with the original subtitles,

so, with perfect reading speeds
of no more than 21 characters per second.

After the first year,
we got the first big job:

to plaster the four-story staircase
in a building in Grenoble,

after which I had
a four-month work practice,

partly in France and partly in Bulgaria,

and then I was studying
earth architecture of Spain and Portugal.

I returned home and started
a graduation project

about the earth architecture of Vojvodina.

Much easier to follow, right?

The thing about the subtitle reading speed

is that in addition
to reading the subtitles,

the viewer needs to take in
other information,

like the speaker's body language
and intonation

and some on-screen content
like slides and pictures.

And it may be difficult to follow
the subtitles themselves

if they disappear very quickly.

Especially if the viewer
doesn't understand the original language

well enough to help them
figure out what's going on.

And thankfully, that new info box in Amara

helps you see when the reading speed
exceeds the 21 characters per second limit

and thus, it shows you where you can help
the viewer in following the talk.

To fix the reading speed,

some more advanced users may want
to adjust the timing of some subtitles,

but in almost every case
your main tool will be...

Compression.

Which means, trying to express
the same meaning in a shorter subtitle.

For example, an almost literal translation

like "Now, what I would like to give
you people here is yet another example,"

with 70 characters,

can be changed into...

"I'd like to give you another example,"

which compresses it down to 37 characters

and which may mean this will allow
the viewer to read it

before it's gone off the screen.

Trying to find a way
to compress the subtitle is fun,

and the new interface helps you

by showing you
when compressing is necessary,

but also when you can do without it.

You can learn more
about compressing subtitles

and find other compression
strategies and examples

in another guide on OTPedia,

called simply "How to compress subtitles."

I encourage you
to explore the new interface

and these new,
helpful tools that it offers.

And, for now,
happy transcribing and translating!