View Single Post
Old 02-25-2016, 09:09 AM   #11
Human being with feelings
CoreyScogin's Avatar
Join Date: Dec 2013
Posts: 9

Originally Posted by EpicSounds View Post
First I normalized each file to -23LUFS
Then I inserted your Automixer plugin on each track and found that a window of 300 works about right.

Maybe this setting would depend on each person's voice? Would you use 100ms for one person and 500 for another?

Result was a pretty good sounding podcast, or at least a much better starting point than I would have expected. Much cleaner sound than without and managed levels well when everyone is laughing.

Still need a limiter on the master and would probably still use some NR, comp, EQ and DeEssing on the individual tracks.
Great! I'm glad to have someone else use it and find it useful.

300ms may be a better default window. When testing I used a range of windows and didn't consciously settle on 500. That was just my initial guess.

I also tested this a bit more last night with a couple of podcasts. My plugin chain ended up very similar to what you suggest. A high pass filter, AutoMixer, DeEsser, Compressor, then a volume leveler (Vocal Rider) on each track and a limiter on the main bus.

When testing with a recent podcast, one issue that I found was in short sections of silence. One of the mic channels had significant ambient noise (recorded in a different location) so since neither were being used for a moment, the AutoMixer didn't keep that noisy channel attenuated and may even favor the noisy one (by design). In that case, a gate or downward expander would be needed to keep the ambient noise out.

I've been pleasantly surprised with how well this works especially in a multi-mic recording.
CoreyScogin is offline   Reply With Quote