Thinger

A Python script which isolates usage of a word in a given video using SpeechRecognition and extracts all uses of that word into a single video. Supposedly. Uses Google Speech Recognition as of the time of this'y here writing. Many engines are available via the Python dependency.

Python Dependencies:

To avoid dependency conflicts, install moviepy separately.

• pip install setuptools decorator==4.4.2 imageio==2.4.1 imageio-ffmpeg==0.4.5 SpeechRecognition pydub
• pip install moviepy==1.0.3

This script features three adjustable knobs (no warranty, all sales final):

Minimum Confidence Level

• 308: result_path, count = finder.process_video(target_word, output_path, min_confidence=0.65)
• Adjust the min_confidence argument to increase or decrease speech recognition sensitivity to the given word you're looking for in the video's audio.
• Note: It is in this main block where you can also configure the path to the video you want to analyze, and the word you want to isolate.

Timed buffer before and after the detected word

92: def __init__(self, video_path, buffer_before=0.00001, buffer_after=0.00001):

• Adjust the amount of time before and after the word isolation to include in the compiled video.