Ampliflare Audio 1.0: Offline Noise Removal, Now Open Source

Ampliflare Audio removes fans, hum, traffic and room noise from recorded speech. Podcasts, voiceovers and audiobooks come out sounding like they were made in a treated studio. A deep learning model trained on speech does the work, right on your computer. Nothing is uploaded, and it works offline.

The tools

Five tools, each built to run over a whole session of files at once:

• Remove noise cleans up speech in one pass. A Strength slider keeps more of the original room when you want it.
• Enhance evens out loudness to what each place expects: music streaming, podcasts, audiobooks or broadcast. It drops low rumble too.
• Cut keeps only the part you mark, and Split breaks one recording into pieces. Both show a picture of the sound, so the pauses are easy to find.
• Convert changes the format: mp3, wav, m4a, flac, ogg or opus.

Your originals are never changed.

Open source

The source code is public on GitHub under the GNU AGPL. Anyone can build it and use their build for anything, work included. The app comes with its own copy of ffmpeg, so there’s nothing else to install.

Every tool is free for personal use, up to 20 files a week. A license adds commercial use and removes the limit.

Ampliflare Audio runs on macOS (Apple silicon and Intel), Windows and Linux. For the full list of what changed in each version, see the changelog.