One of the first reactions to my new series of podcasts about the definition of quality in digital design was
very nice, but I can’t hear it because I’m deaf. And since accessibility is right there in my own definition of quality I decided to transcribe every episode. The first episode is published, and the transcript is done. But now after reading it I have a question about how detailed a transcript should be, especially for the people who really need it.
How detailed should a transcript be?
There a few kinds a transcripts. In literal transcripts you’ll read every uhm, every errr. Literally every sounds that’s on the recording. I assume this is not what anybody who’s interested in the content of the conversation wants to read.
Then there’s something like a wordly transcript. Uhms and errrs are removed, but sentences are published word by word. This can be annoying to read, since people tend to not finish their sentences, or mix a few sentences into one. It can be rather distracting, but at the same time it offers a good insight into a few details that might be lost in a summary.
The third type of transcript is a summary where spoken language is translated into sentences that are easier to read. This sounds like the type of transcript I’d want, but these transcripts are not for me. The problem with this kind of transcript is that a lot of important detail might be lost. The transcriber will have to interpret the things we say, which could be inaccurate.
My question to the accessibility community, and especially to those people who really need transcripts is, what kind of transcripts should I publish, and why. You can use the comment form below, send me an email at vasilis at vasilis.nl or talk to me on Twitter. Thanks!