Crafting CTR - A Look into the Making of a Cyber Talk Radio Episode, Part II
PART II: Mixing and Editing the Episode
Now, this is where the rubber really meets the road. If your raw audio quality is good enough, you can really make your audio shine if you know how to mix and edit well. As a podcast lover and audio engineer, it drives me insane listening to a podcast with bad audio, and even worse editing. There are several things you can do to make your recordings shine, and make your host and guest sound better than they did live in the interview.
The first thing I will do is find a section of the recording where I was rolling tape, but there is only silence. I am looking for the “room tone”. I will take a sample of this room tone and run it through a plugin called ReaFIR. What this process does, is it finds a “noise profile” of the sample you feed it, and then negatively subtracts it from the rest of your audio. This really helps get rid of annoyances like the sound of the air conditioner, for example. One caveat with this tool is that it can work “too well”, and it can really make your recording sound robotic. I recommend setting this to about 70 percent wet, set each noise profile individually for every channel, and adjusting from that baseline while listening back to the mixdown.
Once the noise profile is set, I will run the audio through compression. Now, there are tons of compressors out there, and like most other pro audio gear, you can spend a ton of money on hardware compressors. Luckily, Reaper comes packed with several that are more than adequate for the human voice. We aren’t recording John Bonham’s drums at the bottom of a pool here!
In terms of EQ, it’s really easy to go overboard. I usually do a low end cut and boost the high mids in a subtle way to bring out the natural bandwidth of the human voice. I keep EQing very simple for this type of application. And that’s really about it for the effects that go into an episode of CTR.
Now we get to editing. The first thing I will do, before I take a knife to the episode, is go through the entire timeline and automate volume to essentially automate volume to 0 on the audio tracks when people are not talking. This removes the natural reverb effect caused by the other microphone picking up the speaker’s voice, and also a lot of the ambient noise in the room. Once all of the volume automations are in place, it’s time to actually start the heavy lifting. I will listen to the interview from start to finish, making sure to pay attention to the marks I’ve made during the tracking process. At this point I will also utilize what I jokingly call the “eloquence enhancing process” or what Bret calls the “dummer” or the “de-ummer”. As I listen, I will remove a lot of the “uhhs” and “umms” that the host and guests will say, and also remove awkward pauses, stutters, and the like. This can be a very tricky process, as it is easy to make it sound like audio stitched together, but once you have enough practice you can make the edits sound shockingly natural. This is one of the most time consuming parts of the entire process, but also really makes your audio sound professional, and qualitatively heads and shoulders above many of the rest.
Once the edits are done and the interview segments are up to snuff, I will tack on the theme music to the beginning and all of the bumpers. Then I render out stereo .wav files of the entire recording to export into Adobe Audition, which I use for spectral noise reduction. In Part III, the final part of the blog series, I will be going over the spectral noise reduction process. To read the first part of the series, go here.