Pump Up Your Podcasting Voice

Surprisingly, one of the most popular segments at my recent presentation, The $5 Self-Publishing Workshop, was about podcasting. To demonstrate some of my production techniques, I played this excerpt from the Digital Media Insider podcast:

DMI 13 Excerpt (1MB MP3)

That prompted questions like:

• What microphone did you use?
• I have a Mac. Can I do podcasting with, say, GarageBand?
• How long should a podcast be?

My short answers:

1. Mic: “For that episode, I used a Rode Podcaster mic, but for Skype interviews, I've had good results with a Logitech USB headset. But I also process the recording afterwards with signal-enhancing software.” [In retrospect, I'd add that enthusiasm and mic technique are at least as important.]
2. Mac: “Yes. GarageBand on the Mac has many features designed for podcasting, including voice-enhancement effects. There's also a free voice enhancer program for Mac and Windows called Levelator.”
3. Length: “One of the things I enjoy about podcasting is that the length can be flexible. Just respect your audience; I edit tightly to maximize the good stuff.” [“The Art of Podcasting,” my cover story for Electronic Musician magazine, goes into much more detail on the aesthetics of structuring a show.]

Levelator (free) couldn't be simpler: You drop a WAV or AIFF on it, stand back, and out pops a bigger-sounding version.

Speaking of Levelator, I was curious how it compared with simply normalizing the signal (boosting the level until the highest peak is at the maximum value) and enhancing the signal with a more sophisticated tool like Izotope Ozone, which is what I usually use. Levelator has quite a following among podcasters, but it's rarely delivered exactly the result I've wanted.

Here are some tests I made on a National Public Radio clip. Of course, this clip has already been processed — and then crunched down to a low-res streaming format on the NPR site, where I grabbed it — but you can still hear the differences. All examples are 16-bit, 44.1kHz, mono WAV files. I embedded them with the Yahoo Media Player.

1. Original NPR Clip
2. Normalized to –0.5dB (i.e., just under the maximum)
3. Levelator-ized
4. Ozone-ized (Here I used EQ, multiband compression, loudness maximizer, and exciter.)
5. RMS Normalized to –10dB (an effect in BIAS Peak that boosts the signal while squashing the peaks, creating a super-loud version)

Which do you prefer, and why?