Final Cut It Out

An automatic silence remover for Final Cut Pro X.

Origins

This project has its origins in jashmenn's final-cut-it-out.js script.

All credit goes to him and those who have left follow-up comments on his Gist.

Alternatives

Since publishing this project, FxFactory released SilenceScount, which runs in seconds within Final Cut Pro, works on multicam, compound, and regular clips, and has many more options for customization.

The downside? It's $99.

So buy the plugin if you'd like, but if you want a free way to remove silence and gaps, this project has your back!

Prerequisites

This script is meant to work with Final Cut Pro X (on a Mac), and requires the use of ffmpeg to detect silent portions of the clip.

You can install Final Cut Pro X from the app store, and ffmpeg using Homebrew: brew install ffmpeg.

You should clone or download this project to somewhere on your computer, and in the Terminal, change directories into this project's directory.

Usage

The following is how I use this script. You could have a simpler or more complex workflow, depending on your needs.

First, export a project timeline with the silent portions to an 'Audio only' file or a low-resolution video file that ffmpeg can process.

Second, run the following ffmpeg command to output silent portions into a text file:

ffmpeg -i [VIDEO.MP4] -af silencedetect=n=-35dB:d=800ms -f s16le -y /dev/null 2>&1 | tee silence.txt

Third, run this script to trim all the silent portions (it will take a little while if you have a long video):

./final-cut-it-out.js silence.txt

Finally, adjust the remaining gaps using the select/trim tool as desired.

Bash Alias

If you want to make it easier to run the whole process (once you're comfortable with the automation), you can set up a Bash or ZSH function (e.g. in .bashrc or .zshrc), like:

# Final Cut It Out - Silence Detection and Removal
function fcio() {
  if [ $# -lt 1 ]; then
    echo "You must specify an audio or video file for processing."
    return 0
  fi

  pushd /path/to/final-cut-it-out > /dev/null

  echo "Detecting silence using ffmpeg..."
  ffmpeg -i \$1 -af silencedetect=n=-35dB:d=800ms -f s16le -y /dev/null 2>&1 | tee silence.txt

  echo "Removing silent gaps in Final Cut Pro..."
  ./final-cut-it-out.js silence.txt

  echo "Finished removing silent gaps."
  rm -f silence.txt
  popd > /dev/null
}

Then after exporting an audio or video file representing your timeline, you can run:

fcio /path/to/file.m4a

Configuration

Before running this script, you should always have a backup Project/timeline in FCPX since this will make destructive edits to your video timeline! Cmd-Z should not be relied on in case of disaster—you've been warned!

I currently record at 23.98p, so some of the settings within moveToTimecode() have been changed accordingly. The original script was written with 60p footage in mind.

Someday it would be nice to modify the script to account for different frame rates... but for now, you'll need to modify the script if you shoot at 30p or 60p.

The other parameters you may wish to tweak, based on your own speaking style and the rhythm of your speech:

• In the ffmpeg command, silencedetect=n=-35dB:d=800ms: change the dB setting to a higher or lower threshold if you have softer or louder audio and are getting parts cut out that you do not want cut. Adjust the d value (how long a portion is silent before it is marked for deletion) as needed.
• In the script, adjust the startMargin and endMargin to your liking. I like a little bit of space before and after the audio, but some people like a bit more of a 'jump' cut style so should reduce these values a bit.