Adding the captions was an interesting and complicated process. YouTube does its own automatic captions, with often hilarious results:
Instead of "You cannot write English without us" the automatic text in the caption said "you kind of diet industry nationalist".
Another example was "by their first degree murder and" instead of "My name's Professor Grahamarian". Our website seems much more violent through YouTube instant captioning!
Capturing other languages and accents is even harder. Oscar Cicada's words "or, as it is known in Spanish, Tengo talento pero no mucho talento" became "a series now in Spanish being with a lentil I don't know much" and "that you must have talent to win" became "dot you must of Thailand doing".
Making changes to the instant captions can take several hours, as each set of words is divided like this:
00:06:16,970 --> 00:06:20,560
result there is also the contest
00:06:20,560 --> 00:06:24,430
indicate dot you must of Thailand doing
00:06:24,430 --> 00:06:29,870
a habitual action in the present
That means we couldn't just paste in our own transcript and allow YouTube to use our words. We either had to add timings to our transcript or fit our transcript to the timings. To make it even harder, the timings for the smaller video chapters didn't match the timings for the whole video, so each chapter had to be done separately.
For the film editors, adding captions was just as problematic, as each bit of text had to be entered manually and couldn't be copied and pasted from another file. That's why we didn't add built in captions to all our videos - it would have doubled the editing time! Built in captions are always there, too, so there is no flexibility to turn them off if you want to practise your English without any prompts.
We hope we've reached a happy compromise between built in and additional captions. If you have any suggestions, though, please add them in the comments box and if we make more videos we'll certainly bear them in mind!