Awesome, thanks! On Mon, Dec 3, 2018 at 2:21 PM Leif Andersen <l...@leifandersen.net> wrote:
> Glad you're interested in checking out Video. If you need a thing > that's not yet in the standard library, please poke me and I'll put it > in. Its usually pretty easy to add something at this point, and I tend > to do it as people need it. on the list you just sent, Video already > supports most of what you are looking for. I'll go down the quickly > state what forms I would use. > You can also find more about video at https://lang.video. Based > > > - Take an .mp4 video file and a .wav file, composite them into tracksp > > `multitrack`. > > > - Identify initial clap on both audio and video tracks, sync them > together > > Video currently can't automatically find the 'clap', but you can use > `playlist` to handle the syncing. > > > - Delete sound from video track > > `remove-audio` > > > - Merge video and audio > > `multitrack` > > > - Locate/delete inhale fuzz > > `highpass/lowpass-filter` > > > - Cut everything from beginning to 1s before the first spoken word > > `cut-producer` > > > - Locate and delete bad takes (a single clap sandwiched between "good > > to/from here" spikes (double click)) > > I'm not sure what you're asking for here. But probably `cut-producer` > and `playlist` can do it? > It's a trick for fast editing. Every once in a while when there haven't been any errors you pause and make a distinct sound; I like to double-click one of those "safety button pops up when bottle is opened" things, since it makes a very specific shape on the waveform -- a double spike where the spikes are extremely narrow. That sound means "everything is good up to here." When you make a mistake, you clap once, which makes a wide/fuzzy spike, then you re-do the part you just messed up. Editing out bad takes becomes a simple matter of finding a 'good' spike, then removing it and everything up to the next 'bad' spike. > > - Grab first five seconds of audio and use it as a profile for noise > > reduction across the entire audio > > Video currently offers highpass, lowpass, bandpass, compand, etc > filters, which seem to go a long way to automatically cleaning up > spoken word. It currently doesn't support taking a sample and doing > noise reduction based on that. I'd be happy to look into that if > needed. I presume you were thinking of using libsox for that? > > In the past I've used Audacity; I can continue using this and then use the cleaned-up audio with Video. > That would eliminate about 90% of the work and I could do the rest > > manually. > > Again, feel free to poke me if you wished something worked > differently, or want a feature that is missing. I also regularly check > Video's issues on github at: https://github.com/videolang/video/issues > > Good luck, and happy editing. > Thanks! > ~Leif > -- You received this message because you are subscribed to the Google Groups "Racket Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to racket-users+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
