Home / Opinions / Fixing Noisy Audio / Audio Extraction and Noise Profiling

Audio Extraction and Noise Profiling

The process will usually involve adding each clip to the timeline on its own, then heading to the output section of your software and exporting the file as audio only. We recommend a 48kHz, 16-bit WAV file to preserve quality as much as possible. If your editing software doesn’t allow you to export WAV audio on its own, an alternative is VirtualDub, or VirtualDub-MPEG2 VirtualDub-MPEG2 if your footage is in MPEG-2 format. But AVCHD isn’t compatible with VirtualDub, at least not without a complicated blend of extra plug-in software. So in this case you will be best advised to use a premium application.

For this tutorial, you will need to extract the audio from your video file.

Assuming you now have a series of WAV files, and have installed the latest version of Audacity, now is the time to fire up the software. Open the first file in Audacity.

As you can see, our example from an interview has a lot of background noise, primarily from the camera itself. There are no regions where the waveform is close to a single line. The areas where the interviewee wasn’t talking are still very busy, full of unwanted hiss.

The next task is to find an area of the file where only background noise is audible. Play through the file to locate an appropriate portion – just a second or two will do, so long as it’s representative of the noise found across the entire file. When you have chosen your portion, left click and drag across to highlight it.

comments powered by Disqus