Windows app to trim the commentary at the end of a song

Question

Is there a windows app to trim commentary at the end of a song? Like this one https://www.youtube.com/watch?v=lVXziMFEqX0 The music ends at circa 3:42, then there's silence for at least 10 seconds, and commentary. I want to batch trim all my aac/opus files from such commentary (and the silence of course). Format support doesn't really matter, I'll convert them to mp3 anyways.

Many thanks in advance!

Answer 1

Rather than a single program that does this for you I have a combination that can do it for you. The combination is:

• Python - Programming Language Download and install from https://www.python.org/downloads/ selecting Python 3.6 or later and selecting "Add to Path" during the install.
• Numpy - Numeric Library Once you have installed Python you can install from the command prompt with pip install numpy
• MoviePy - Video Editing in Python Once you have installed Python you can install from the command prompt with pip install moviepy
• FFMPEG - Media Processing Tools Automatically downloaded and installed by MoviePy on first use.

Note that this process was inspired by this blog post which automatically clips the highlights of football matches based on the average audio levels.

You can then put together a python script something like the following:

import os  # Pythons default library for OS operations
import glob  # Pythons default library for file pattern matching
import numpy as np # for numerical operations
from moviepy.editor import AudioFileClip

def find_silences(audio):
    """ Find the silences in an audio clip """
    cut = lambda i: audio.subclip(i,i+1).to_soundarray(fps=22000)  # get one sec of audio as an array
    volume = lambda array: np.sqrt(((1.0*array)**2).mean())  # RMS of Array
    # Use the above to get a list of the volumes per second
    volumes = [volume(cut(i)) for i in range(0,int(sound.duration-1))] 
    # Get the times when the volume is zero
    zero_secs = [i for i, v in enumerate(volumes) if v == 0.0]

def process_file(fname):
    """ Process a file """
    audio = AudioFileClip(filename=files[0])
    silences = find_silences(audio)
    two_secs = [s for s, i in silences[:-1] if s+1 == silences[i+1]]
    if len(two_secs) == 0:
        print(fname, "no two second silence found!")
    else:  # We have at least one 2 sec silence
        clipped = audio.subclip(0, two_secs[0])  # Clip to that length
        clipped.to_audiofile('temp.mp3')  # Save to a temp name
        os.rename(fname, fname+".orig")  # keep the old just in case
        os.rename('temp.mp3', fname)   # rename temp to original name

if __name__ == "__main__":
    for FNAME in glob.glob("*.mp3"):  # All the MP3 files in the current directory
        process_file(FNAME)

Note that the above code is very rough & ready but should get you started.