Richard Varga wrote:
1. The client has a video and wants translated subtitles - what's the best process of (transcription), translation, spotting?
I’ve set an ASR (automatic speech recognition) program for this purpose. Normally, I’ll just pipe the video file into it, and it’ll create the English transcript with appropriate segmentation according to the parameters I gave it beforehand. Then I’d do the translation manually, following the style guidelines either stipulated by the client, or the “standard” subtitling conventions for my target language when none was provided. Is the transcription 100% accurate? No. Does it need to be? No. The English transcript is only an auxiliary document to cross-reference the utterances I hear with the written texts, just as a way for me to minimize errors.
2. The client has a video and a transcribed text - what's the best process of translation and spotting?
If it was an English master template, I’d translate it as usual. But if it was a raw transcription, I’d charge them my spotting + translation rates. Raw transcripts aren’t going to help me all that much, since I still would have to create the subtitle boxes/events/segments/whatever-you-want-to-call-it myself.
When spotting, what works best for you? Immediately putting text into adequately "sized" events in the video? Or preparing the events by full sentences and then editing down/splitting?
[Edited at 2023-05-10 16:54 GMT]