TmpGuy;I've noticed pyTranscriber is very unstable with large files, which is unfortunate and annoying. You can either cut your pieces up, but the bad news is that you lose all your timing. I've found another solution is to edit the audio file and make sure you zero out sections without any spoken text at regular intervals. Part of the process pyTranscriber uses is to look for areas where it can break up the audio into separate files before transmitting them, and I believe doing this helps that process work more effectively, avoiding whatever triggers the crash. It's still a bit of a hit or miss process, but if you can give pyTranscriber a bit of extra help like that, it seems to work more reliably.
That being said, unfortunately, I've found the end result is so horrendously bad, I've all but given up on machine transcription until the technology significantly improves. For now I'm sticking to translation of Chinese subs, since there appears to be quite a few of those.
Thanks for sharing these tips. I greatly appreciate them.
I've also tried using Aegisub (my favorite sub creation software) and creating one dialog line for every sentence or tight paragraph.
Then I run the line back through with Google Translate open in my browser, while hitting the Microphone icon.
I used to use Google Translate with my cell phone and computer speakers, but that was only successful maybe 1/3 of the time.
I started clicking on the microphone icon in browser based GT a few days ago, and so far it seems to be much better, while still not perfect.
I'd say my new method is successful maybe 2/3 of the time, at most.
It's still imperfect, but every little bit of time saved makes it easier for me to create and finish new projects.