on Dec 30, 2015 at 1:13:10 pm Last Edited By Rune Jacobsen on Dec 30, 2015 at 1:30:13 pm
I am editing a podcast episode in Audition. For one particular section of about 1h20m there are two people talking - I have two sound files recorded of this conversation, each with the audio of the respective talking head.
Audio file A is clean and crisp. Audio file B is also pretty good, except when participant A is talking, there is a slight echo of that audio in the B file. The echo is with a constant delay and can be removed by hand, but I would prefer not to have to do that for the entire section if possible. For the most part, A and B do not speak on top of each other, so in theory I would think I could subtract the A audio (respecting the delay of the echo) from B, and we should get a clean B signal.
However, I have no idea how to do this with Audition, if it is even possible.
Any ideas? I know, garbage in, garbage out, but this echo did not become apparent until after the recording, so I am kinda stuck with it.
Edit: I realize I can sidechain a compressor that mutes B while A is talking, and that will help a lot, but because there is a delay, won't the echo then be audible right when A stops talking, before B starts?
Have you tried the spectral pitch and frequency display to find the echo and analyze the range? Then by selecting if and muting or erasing , eliminate what you do not want?
Or by analysis you know the exact frequency to suppress. I would try to use a parametric EQ just to begin.
It sounds more natural than blanking out the space which you would then need to fill with room ambiance.
If you had to go in by hand just to get it done there are several ways to work. But the spectral pitch will show you right where it is and you could then mark the track, navigate to each instance and use the tool you found to work best for your material. Being able to manipulate the total spectrum with such fine precision is more "surgical" Some time spent getting used to working that way will be well invested.