I subscribe to a lot of podcasts — specifically, a lot of technical audio podcasts. And while the technical content is often quite good, the audio production is often
quite terrible.
For example, most recently I listened to an episode of the
Object Database Podcast discussing
Magma. It's a good discussion but it's almost
physically painful to listen to. Why?
Because the interviewer is on the right channel and the interviewee is on the left channel! But wait, there's more! It also sounds like it was done via a transoceanic satellite link, so there's a delayed echo of each speaker in the opposite audio channel! It's bad in the car, and actually headache-inducing with headphones.
Another couple of podcasts I listened to recently were just as bad, not because they so horribly abused audio balance, but because the interviewer and interviewees would smack their lips and cough and snort and whistle through their noses while the others were talking. It was a huge distraction from their message.
So how can you avoid these types of situations and make your podcast sound like
Rory Blyth: The Smartest Man in the World and
This American Life? It's actually pretty easy! It just requires a little bit of planning and set-up when you do an interview, followed up by some post-production work with an audio package like
GarageBand.
Always Record Multiple TracksIf I had to recommend a single cardinal rule for podcasters, this would be it. The quality of your podcast will ultimately reflect the quality of your source material, which means it'll be worth your while to invest in ensuring your source material is as high-quality as possible. Think of it like cutting wood: You can always cut a board down a little more, or sand it smoother, but you can't un-cut or un-sand it. It's better to start out with more bits and more tracks and then pare them down to the best part.
If you're doing an interview podcast, each participant — including the interviewer — needs to be recorded on a separate track, and their audio should be isolated from others' as much as possible. This will let you work with each participant's audio independently, and will make it much easier to produce a tight, properly-leveled, professional-sounding podcast.
If you're doing telephone or telephone-like (e.g. VOIP) interviews, you should still try to record interviewer and interviewee separately. If possible, get your interviewee to run some recording software on their end, and have them send you the local recording. It shouldn't be too hard to match the timing with your own side of the conversation and you'll have much, much better sound quality than what you heard during the interview.
Manage Your LevelsOnce you have your source material, you'll need to produce a podcast with it. You can't just slap it into an MP3 file and go if you want it to sound good. You'll have to apply some audio filtering to
level it. This will give your podcast an even perceived volume, and help you avoid blowing out your listeners' speakers either because your podcast's volume is too high, or because its volume is too low but the next track is at a normal volume.
Leveling is often accomplished via
audio level compression rather than by tweaking the volume on a per-track basis in your post-production software. If you're using background music or bumps between segments, your post-production software should also let you manipulate your levels to "duck" the music and/or bumps, so you can speak over them at a constant perceived volume.
Needless to say, don't abuse the fact that you're outputting to stereo. Don't try to play games with positional audio, unless you
really know what you're doing. It's all well and good to have the interviewer
just slightly to one side and the interviewee
just slightly to the other. It's quite another to have them on opposite sides of your head — especially with headphones! It's actually a bit stressful on the brain to try and integrate completely independent audio from both sides; after all, in nature, there are almost never completely isolated sounds that are only received by a single ear.
Remove Anything ExtraneousThe final thing you should do if you want to sound
really professional is be hardcore about editing out anything extraneous from your podcast. Most importantly, make sure that silence is really silence! Make sure that you, as an interviewer, are
never heard fidgeting while your guests are speaking. Get rid of coughs, throat-clearings, lip-smackings, breathing, and other distractions.
This is easy enough if you're all on separate tracks: While a guest is speaking, zero the volume on your track. And do likewise to guests' tracks while you're speaking. Be merciless in chopping out the "ums" and "ahs" and "uh huhs" — no matter whose they are — unless they actually add to the conversation. In short, do as much refactoring on the audio as you can without changing the meaning. This is one of the things that makes
This American Life sounds so polished, even moreso than the often-subtle background music or the timbre of Ira Glass's voice.
Once you have everything post-produced, you should be able to mix down to an MP3 or other audio file with an appropriate bit rate for your content and have a podcast that sounds great. Be like Ira, be like Rory, and you'll find it much easier to get the information that you're trying to convey across to your listeners.
![[info]](http://l-stat.livejournal.com/img/userinfo.gif){kind=small}