This project is for tools I created to making my podcast easier.
When editing a podcast I start with two tracks, mine and my guest. I realized that I spent alot of time cutting up each track into chunks for editing. This script will check for silence and split into multiple mp3s. I recommend renaming the original file to something small and easy to read. I add these split files to the audio editor one by one so it is easier to remove filler language and put the different tracks together.
Usage:
Install Python 3.7
Install ffmpeg brew install ffmpeg
Create Python 3.7 virtualenv python3 -m venv audio_editor
Activate virtualenv source audio_editor/bin/activate
Install requirements pip install -r requirements.txt
Run Script: audio_silence_splitter.py audio_file.mp3
It takes a while. While you are waiting, you can listen to my podcast, @FromSourcePod
From the Source is an interview show about what tech jobs are really like, from the good, the bad, to the boring.
Apple Podcasts ~ Spotify ~ RSS ~ Web
Rate & review to support the show and fine scripts like these.
I also wanted a transcript of what was in those files. This would make it easier to edit, and then have a transcript I could post with the episodes. I tried a bunch of different speech to text tools, and landed on IBM which had both the best accuracy and ease of use. This uses the same splitting functionality as audio-silence-splitter because it is easier to transcribe short audio.
Usage:
Sign up for a free IBM Cloud Account.
Follow the steps for Speech to Text quickstart until you get an API Key.
Save as environment variable IBM_API_KEY (command depends on your environment)
Follow all the steps for the splitt, but run the transcript instead. This also takes awhile, so you can listen to a second episode of @FromSourcePod.
Run Script: audio_transcript.py audio_file.mp3