Hacker News Reader

Is it possible for your app to take into account not just audio but also facial expressions? AKA do not cut out the parts where the speaker is silently making unusual faces or facial signals.

View on HN

Possibly! Not currently, for sure. It'll have (again, soon) a feature that lets you manually override Recut's choices though, so you could select a meaningful silent part and leave it in.

I think at some point any sort of automation is bound to get something wrong and it'll likely never be perfect, so my goal is to add enough manual control that you're never just stuck with whatever the app decided.

That's cool, hopefully it can be a feature request you take seriously, it may not even be that hard I think to throw in a facial recognition library or something. The issue with doing it manually is that I would have to basically scan the entire video manually and that sort of counteracts the purpose.

Might not be as great as you seem to think, you'd be unflagging silence that you'd want cut anyway. I imagine you'd get false positive like with a speaker frowning and be silent for a few second because he's thinking.

Point is you'll still have to review manually the result anyway for under- and/or over-cuts.

The places where it's silent and making faces should be easier to review than the places where it's just silent.

Reply on news.ycombinator.com