This is the current podcast setup in the Roadhouse studios—my little laundry room office.

As you can see, the podcast takes several items to produce. The first is the workhorse Linux desktop on the far left – you can just see the 21" monitor. This machine runs Fedora Core 3 and Audacity to capture the live audio from the board. There’s also a laptop in the lower right, ready to push music to the board. Finally, you can barely see the Sony Walkman. It provides music and transition effects for the podcast.

Of course, in any audio application, the mic is everything. In this case, I picked an MXL 990 cardioid condensor mic with shockmount. It’s pictured here with my homemade pop filter – a looped wire hanger covered with one of my wife’s black nylon stockings. This is a great mic with a frequency range from 30Hz to 20 MHz. It’s especially stout given the price – $59.99 from Amazon.com. In fact, it’s almost too sensitive. I’ve had difficulty quieting a room with two computers and have resorted to setting the mic on a folded t-shirt on the desk to further dampen any vibration. I really need to figure out how to soften or eliminate the sound of the fans from the computers – if you have suggestions, make them in the comments.

There’s also the mixer. I heard about this board on Michael Geohegan’s Reel Reviews. He called it  "the Podcaster’s Best Friend," and he’s right. This is the Behringer Eurorack UB802, an 8-in powered mixer with a great range of features. Its small profile makes it perfect for a desktop, and it’s great price of $49.99 on Amazon certainly fit the budget for the podcast. I looked locally first, but no one was carrying this board at any price. In fact, anything close to the features was well into the $200.00+ range. It’s a great personal choice for another reason: I much prefer knobs over sliders. I think knobs retain their quiet for much longer than sliders, which seem to start crackling fairly early. It’s not quite as easy to manipulate and cross-fade with knobs, but they’re really a better choice over the long haul.

On the software side, I’m using Audacity to capture the podcast live. Fire it up, set the levels and go. For the interviews, I’m using Skype
where possible, though I wasn’t any too happy with its performance in
the first ‘cast. Audacity will capture the live audio from Skype on a
line input from the mixer.

Once the ‘cast is done, I run it through the compression tool in
Audacity -  3:1 ratio, -12db and 0.1 attack time. That seems to smooth
out all the rough edges quite well, without completely flattening out
the sound.

Finally, I use lame to
encode from .wav to .mp3. I don’t need a pretty GUI interface for it,
just the command line. I use the following command to encode the final
.wav to a 64Kbps, 22.05 Khz .mp3 file, complete with id3 tags:

lame -b 64 --resample 22.05 --tt "The Roadhouse Podcast" --ta
"Tony Steidler-Dennison" --tg Blues --tc "The Roadhouse Podcast 001 -
February 19, 2005" original.wav roadhouse_001.mp3

That converts everything to a tidy filesize, again without degrading the sound too much. The first ‘cast was roughly 25Mb – about what I’d expect from the above encoding.

I’m sure I’ll get this process down a bit better in subsequent ‘casts, but this is the point from which it all started.