Earlier quoted context omitted.
Nifty. I went looking for google translate style audio, and found the Advanced > Audio button, and hit a bug. Page load; for Chinese input, enter "中文 輸入"; Translate; Audio; surprised by English input and Chinese spaced having cat and ffmpeg command lines. Feature idea: permit repeatedly playing the audio for just a part of the phrase. Often when using ggl translate, I want to rehear just a syllable or word I'm unsure…
Sorry! Thank you for exploring the Advanced features, I put a lot of effort into those, but they need more explaining. I'm pleased that you actually looked at it though :). The audio feature doesn't actually play audio. It's supposed to generate input commands to put into ffmpeg. I would write more documentation, but my priority is to get the existing documentation translated to Chinese. I asked 4 friends over the pa…
http://opennmt.net/ links to a "Live system" with a Chinese option. Discussed on infoworld[1].
FWIW, its translation of "This is a test." was at least recognized by google translate.
> audio [...] using only client-side JavaScript
It looks like the audio starts out as small mp3 files, which are then concatenated? There are WebAudio-based libraries both for playing one file after another, and for concatenating them for download.[2] For the mp4 subtitles, hmm, well, even if nothing else, there's ffmpeg.js[3] (but fonts might be an issue?).
Thanks for your work. :)
[1] http://www.infoworld.com/article/3152050/artificial-intellig... [2] For example (merely the first one I saw): https://github.com/jackedgson/crunker [3] https://github.com/Kagami/ffmpeg.js