Steve-Noble
commented
3 years ago Thanks for your comments, James. I propose that we edit the SAUR document section discussing your paper to better describe how the study was actually conducted. We will also look at the option of including a discussion on the differences between captions for the deaf and hard of hearing, and translation captions. As long as it can be written in such a way to show a close connection to both accessibility and synchronization, it may add value. I will have to reread your paper with that view in mind.
Thank you for this piece. It seems like a good overview of the main factors in caption quality. A couple of items I'd like to provide feedback on.
In reference to my paper on caption rate, you state "The first phase of the study included video clips which were purposefully created for the study, where BBC reporters attempted to recreate broadcast quality news pieces on the same topic, but re-scripted each clip in the study so that it included more or less words which were spoken over the same 30 second period of time." This isn't quite correct. We didn't select pieces on the same topic. We searched through recent news reports for pieces that were approximately the right number of words. Within a couple of words for the desired WPM at 30s length. We carried out minor re-scripting on the pieces that needed it to get the exact number of words required, and verified the words were read exactly as written and at a constant pace across the 30s while recording. The only content factors taken into account when selecting news stories were that they had to be unlikely to be familiar stories in the local area where the study was conducted, and otherwise unlikely to generate a strong emotional response. The topics of all of the clips were, broadly speaking, random.
It might also be useful to have a small discussion on the differences between captions for the deaf and hard of hearing, and translation captions. Much of the research on the subject conflates the two. In particular, they make the assumption that people using captions have no other access to the speech. Many go as far as to present the captions without lips in view and without audio to restrict comprehension of the dialogue purely to the captions. As you have pointed out, the vast majority of users consume captions alongside other modes. Either with residual or even un-impaired hearing, or via lip reading. I believe that prior research that indicated high sensitivities to word rates would hold true for translation captions which, by the nature of translation, eliminates the ability to lip-read or use residual hearing. The novelty of our study was to include access to these other channels - eliminating influences other than word rate, and to provide a control in the form of hearing participants. I believed this made our study a more valid representation same-language caption use. But it has taken some time to convince some parts of the research community of this approach. And I do not feel this should detract from the findings of other studies when applied to translation captions.