Lecture Audio Cleanup: a Local Adobe Podcast Enhance Clone
I clean up Digisemestr lecture recordings on my own Mac instead of Adobe Podcast Enhance. An hour of audio in about a minute, no limits.
About the project
I used Adobe Podcast Enhance to clean up lecture audio. It works great, but it has limits and every file has to be uploaded and downloaded by hand. For a whole semester of lectures that gets old fast.
So I ran a blind test: several open source models against Adobe on the same samples, without knowing which one I was listening to. DeepFilterNet3 came out best. And when I added an EQ learned directly from the difference against Adobe, it jumped from 2 to 4 points out of 5. Adobe got 5.
Today it runs as a tool in Claude Code. It cleans an hour of recording in about a minute and the output is exactly as long as the original, so it goes straight back under the video.
What you get
- No limits and no manual uploading to Adobe.
- An hour of lecture cleaned in about a minute.
- Same length as the original, fits under the video.
- Free, everything runs on my Mac.
How it works (under the hood)
- Noise removal with open source DeepFilterNet3.
- An EQ curve learned from the difference against Adobe.
- Loudness normalized to -16 LUFS with a limiter.
- Launched by a skill in Claude Code.
Built with
Want to build like this too?
I teach building apps without programming in the AI First course.
More from Tool
Shoutbase: Testimonials Manager
An internal tool for managing testimonials: a queue of who to ask, a collection of what we've gathered, and an overview of where we use them.
Toucan: A Shared Second Brain for the Team
A shared cloud knowledge base that the team and its AI agents draw from, a "second brain" connected to my local context1.
Running a Project's Marketing Through Claude Code
A marketing team as a set of Claude Code skills: a director, department managers, and specialists that run a project alongside me.