Here's a Web 2.0 moment: among the 300-plus reader comments on my (David's) blog about the Zoom H2 portable digital recorder were some intriguing audio links from filmmaker, podcaster, and musician Michael W. Dean. Exploring his site,, I discovered he was an accomplished author too, and asked if he'd be interested in writing for O'Reilly Digital Media.

Michael said he was speaking on a panel at the fourth annual Podcast and New Media Expo, and suggested he cover the show, which is designed to "educate individuals and companies about how to produce high-quality online audio and video content, grow a loyal audience, and market or monetize that content in creative ways." That sounded right up our alley, so I sent him off with H2 in hand to record insider tips and insights from the top minds in podcasting.

Today's episode of Digital Media Insider is the best bits from the many hours of audio Michael captured at the show. We'll be back next week with a behind-the-scenes photo gallery featuring more of Michael's Expo discoveries. Incidentally, all photos in this article were taken by Michael's wife, Debra Jean Dean. (DMI 10-04-2007: 36 minutes 46 seconds)

Interviewees (In Order of Appearance)

Tim Bourquin

Michael Dean interviews New Media Expo founder Tim Bourquin. Note the handle and windscreen on the Zoom H2 recorder, which is operating in 90° stereo mode.

Production Notes: Recording

As you'd expect from a podcast pro, Michael delivered a completely produced show. I then recorded a brief voiceover into QuickTime Pro with a Rode Podcaster mic, edited out my false starts and tongue clacks in BIAS Peak, and popped the voiceover and Michael's file into my Digital Media Insider template in Ableton Live. (The template contains intro and outro music, along with an Izotope Ozone preset I use to enhance my voice.) I exported the mix as a stereo AIFF file and converted it to MP3 format in iTunes.

Michael recorded the field interviews and ambient interludes with his Zoom H2 digital recorder, choosing 16-bit, 44.1kHz, stereo WAV format. He used only the front (90°) microphones, except on the Steve Ely interviews, which utilized both front and rear (120°) microphones (see diagram). Disabling the rear mics minimized the background noise endemic to a convention setting.

Whenever possible, Michael lowered noise further by conducting interviews in unused conference rooms, which he commandeered guerilla-style. That averted the need to use electronic noise reduction in post-production, while retaining an acceptable level of trade-show ambience.

He and his wife Debra Jean then recorded voiceovers in their home studio, which features sound-treated walls. Michael used a Rode NT1-A condenser mic; Debra Jean used an MXL M3-B "Silicon Valve" condenser. Both mics were equipped with shock mounts and pop filters.

The microphones plugged into an Alesis USB-8 mixer, which also provided the phantom power. The USB output of the mixer fed a Toshiba Satellite laptop computer with a 2.8GHz Core Duo processor, 1GB of RAM, and Windows XP.

Grammar Girl

Mignon "Grammar Girl" Fogarty splits a few infinitives between sessions.

Editing and Processing

Michael used Sony Sound Forge 8 to record and edit the voiceovers, as well as to edit the field recordings. The only special effects were a bit of Sound Forge reverb added at the very end of the train recording to delineate it from the next section.

He then processed the edited field recordings and voiceover files through CN Levelator, a free dynamics processor for Mac, Windows, and Linux. For the final assembly edit he used Sonic Foundry Vegas Pro 1.0, a vintage 1999 video editing program.

Michael output a rough cut to WAV format. He and Debra Jean took notes while listening, made final tweaks in Vegas, rendered the mix as a 16-bit, 44.1kHz stereo WAV file—and ran it through Levelator again. Finally, he boosted the overall volume 20% in Sound Forge, confirming integrity of the track with Sound Forge's Detect Clipping tool.

Tim Street

Tim Street of the French Maid TV video podcast.

Painless File Transfer

The final WAV file weighed in at 341MB, so I suggested Michael convert it to FLAC format to speed the transfer to me. (FLAC stands for Free Lossless Audio Codec; as the name implies, it reduced the file size by half without affecting the sound quality at all.) Michael used Audacity 1.3.3 (beta) to do the conversion.

He then compressed the FLAC file again with WinZip Pro 10. That didn't change the file size much, but zipping can help reduce download errors. Uploading the file took Michael about 70 minutes; I downloaded it in far less. Decompressing the FLAC file took me just 19 seconds in xACT (Mac).

I originally told Michael an MP3 would be fine, but because I planned to mix his audio into the Digital Media Insider template and render a new MP3, he strongly suggested transferring the file in a lossless format instead. "I feel MP3s made from MP3s sound bad," he explained.

The Nerdly Stats

Michael reports, "This half-hour podcast was culled from seven hours of raw audio recorded over four days' time. The post-production recording, editing, fluffing, folding, checking, and uploading took approximately 18 hours. The number of mouse clicks to complete this entire project: 15,825. (Source: my program MouseCount, available free for PC.) No geeks were harmed in the production of this podcast."

For more on the New Media Expo, stop by next week.

Return to