doing my own captioning for a long video. I’ll also wait to see if youtube does it, and I did find a nice autocaptioner via otter .ai -very generous – 600 mins autocaptioning per month for free – but 30 mins at a time and this was an hour – didn’t realize. Still, one to watch. Trying out https://github.com/m1guelpf/auto-subtitle – using the ubuntu i have on my windows 10… it uses open ai’s “whisper” and ffmpeg – – i’ll be surprised if it works tbh, but so far, it gladly read the mp4, extracted the audio and it’s now 14% into captioning a very wordy video. I have the text for the video; the audio is output from a text to speech – but it’s not lined up, and despite trying in vain for an hour to create a scrolling text that lined up well enough ,I simply couldn’t — I got close; but then it would go out of sync. Mathematically it was scrolling at an average rate that might have worked but i just couldn’t get the math right and I tried… might not be possible as ffmpeg’s text display rate thing is odd – doesn’t actually scroll per se… and with different lengths of time per each speaking line of text, different syllables ,etc ,getting a mathematicall y perfect “rate of speed” almost wouldn’t matter anyway as it’s unlikely to keep the desired text within the window much of the time, even if it ended up lining up at the end – and I wasn’t going to render an hour video with scrolling text just to screw up a few times. Anyway, I’m 20% in now hopefully this works

doing my own captioning for a long video. I’ll also wait to see if youtube does it, and I did find a nice autocaptioner via otter .ai -very generous – 600 mins autocaptioning per month for free – but 30 mins at a time and this was an hour – didn’t realize. Still, one to watch.

Trying out https://github.com/m1guelpf/auto-subtitle – using the ubuntu i have on my windows 10… it uses open ai’s “whisper” and ffmpeg – – i’ll be surprised if it works tbh, but so far, it gladly read the mp4, extracted the audio and it’s now 14% into captioning a very wordy video. I have the text for the video; the audio is output from a text to speech – but it’s not lined up, and despite trying in vain for an hour to create a scrolling text that lined up well enough ,I simply couldn’t — I got close; but then it would go out of sync. Mathematically it was scrolling at an average rate that might have worked but i just couldn’t get the math right and I tried… might not be possible as ffmpeg’s text display rate thing is odd – doesn’t actually scroll per se… and with different lengths of time per each speaking line of text, different syllables ,etc ,getting a mathematicall y perfect “rate of speed” almost wouldn’t matter anyway as it’s unlikely to keep the desired text within the window much of the time, even if it ended up lining up at the end – and I wasn’t going to render an hour video with scrolling text just to screw up a few times. Anyway, I’m 20% in now hopefully this works

Leave a comment

Your email address will not be published. Required fields are marked *


eight + = 14

Leave a Reply