LosslessCut works great if the keyframes of the video are not too far apart and if you already know where you want to cut. I posted something similar (a web app for automatic silence removal) a few months ago in Show HN: https://www.savvycut.com.
We also use keyframe based cutting for lossless cuts and a smart-cut mode that only re-encodes parts around keyframes (at least for h264 and vp8 codecs). I see that you also experiment with a similar smart-cut feature: https://github.com/mifi/lossless-cut/issues/126.
That seems to work quite well if you re-encode with the same parameters as the original but I guess there are a lot of edge cases.
And I wondered why we suddenly moved to the first page on google for some queries for our app that we posted on Show HN some time ago that competes with kapwing in a small niche (SavvyCut: https://news.ycombinator.com/item?id=28950702). It seems that someone made some room for us.
I guess if you rely on googles monopoly to run your business you should not be surprised if they can remove you at will.
For us, a lot traffic originates from competing search engines that show us on their first page even without creating any SEO spam.
Great idea to apply it to recorded meetings! That sounds like a good B2B use case besides content creation. I guess you can directly use it on the recorded video files from Teams, but I would agree a proper integration would be much better.
Thanks for the feedback aphit!
Great to hear that you found it useful outside content creation as such. You are right you can also apply it to something like presentations or screencasts. That's actually what I started with, explainer videos and talks.
I will have a look at the e-mail entry field as I wanted to customize the auth0 login screen anyway.
Hi HN! SavvyCut is a browser based video cutting tool that automatically cuts out non-speech parts of videos and creates jump cuts.
This can be quite a time saver for content creators that spend a lot of time manually trimming their videos in more complex video editors.
It preserves both your privacy and your bandwidth by running locally in the browser using WebAssembly without uploading anything to cloud servers.
It turns out that for many use cases (vloggers doing stuff while talking / gaming / outdoor noise) a simple volume threshold is not sufficient, so it uses an ML based approach to detect speech.
We also use keyframe based cutting for lossless cuts and a smart-cut mode that only re-encodes parts around keyframes (at least for h264 and vp8 codecs). I see that you also experiment with a similar smart-cut feature: https://github.com/mifi/lossless-cut/issues/126. That seems to work quite well if you re-encode with the same parameters as the original but I guess there are a lot of edge cases.