Weekly log: 20120226 (or “The importance of accurate transcripts”)
No translator would dispute the importance of the quality of the source text, but if the source text is itself an audio transcript, then the source text itself would not really be the true source text per se but rather the results of a transcription (which is almost always at least somewhat edited). Worse, for languages where the spoken and written forms differ substantially (such as the case for non-Mandarin dialects of Chinese), transcription itself will then be in fact a thinly-disguised form of translation. In effect, for such languages, the translator will really be producing a translation from another translation, with no access to the true source (which is not even a piece of text but an audio or video file). Any transcription errors (which technically really are translation errors) will then percolate into the translation. And when it is time for a review, the reviewer will essentially be disconnected from the true source material. So as I was listening to the audio while reading the transcript I was reviewing, I was dumbfounded as to how the transcript did not in fact match the audio. Certainly, most of these discrepancies are not really that significant, but there are still some that will make you doubt whether the audio should have been transcribed just a little more faithfully (and how a more faithful rendition would have affected how the translator would have rewritten those words). How is the reviewer supposed to handle this situation? Last year I voiced out my desire to have “real” transcriptions — at least as an option — instead of “transcriptions” into standard written Chinese, a form of Mandarin. Translators really do not have a firm ground to base their translations on unless we can have an accurate grasp of what is actually spoken (and, preferably, how the words are spoken). Unfortunately, due to the fact that subtitles will necessarily end up being in standard written Chinese, my suggestion was never considered. It now looks like that a lack of accurate transcripts is really an even more serious problem for reviewers than translators.