Over the past few weeks, my ZDNet colleague David Berlind has been in the podcasting kitchen whipping up a studio. The latest results are part of Berlind's Transparency Project. Berlind kindly credits me along with Dave Slusher of Evil Genius Chronicles and Phil Terrone of EndGadget with helping him get started, but in the spirit of full disclosure, I'd suggest my "contribution" was to pass the buck to Slusher, Adam Curry, and anyone else who'd been down this road before. Indeed, I've been lurking for months waiting for everybody else to work the kinks out. With Skype OS/X beta now supporting conference calls and GarageBand II shipping on Saturday, the time is nigh. Here's Berlind's Recipe, graciously assembled by David at my request because I'm too damned lazy to write it down:
A. XLR-based mic input #1 with gain control
B. XLR-based mic input #1 with gain control (used only when second mic required, eg: in "studio" interviews)
C. 1/4" left/right stereo input #1
D. 1/4" left/right stereo input #2
E. FX Send (mono)
F. AUX return (left/right stereo)
G. Stereo Headphones out
H. 1/4" left/right stereo control room out
I. 1/4" left/right stereo main mix out
2. Audio from #1A sent out, through mixer's FX port (1E) to the Mic-in port on a Griffin iMic. #1A's FX dial is turned up about 50% (controls level of #1A's audio that's routed through the FX port). FX dials on all other inputs is turned down so that the people on the skype connection hear my voice only. If a second person was in the studio and the person on the other end of the skype connection needed to hear them as well, I'd have to turn #2A's FX dial level up as well. The FX port is a mono port, with Radio Shack cables, I split the mono audio into two channels and then convert it into a stero mini plug that can plug into the Mic-in port on the Griffin iMic.
3. Griffin iMic connected via USB to USB port on Powerbook
4. Skype audio preferences set to listen to Mic-In line on Griffin iMic (therefore, it picks up my voice coming from the mixer).
5. Skype audio preference set to transmit audio (only picks up 'guests')
to Powerbook's built in headphone jacks.
6. Powerbook built-in headphone jack connected to regular input channel #1C on the mixer. Using cables from Radio Shack, I split the stereo mini jack coming out of the Powerbook into individual left and right 1/4" mono jacks (the only type of connection accepted by the mixer's non-microphone inputs).
-- so now, my voice is in the main mix, and the outbound audio from skype is in the main mix.
7. Main mix sent back out of mixer through #1I to the Powerbook's built-in Mic-in port (needed Radio Shack cables to convert from two 1/4" mono plugs to one stereo mini plug).
8. Current recording software is Audacity (open source, kinda flaky, could suddenly crash). Like skype, Audacity's preferences can be set to tune into a specific audio source and so I set audacity to listen to the Powerbook's built-in Mic in. Then I just record.
9. Adding Music or other MP3 audio to the mix
I've done this two ways and investigating 3rds.
1. One is to run iTunes on the Mac. Unfortunately, you cannot select where iTunes outputs to. It automatically goes into Mac's built-in audio which means that it's merged with the audio (what's coming out of Skype) that's going out of the Powerbook's built-in headphone jack in step 6. Technically, this works fine, but your guests won't be able to hear that audio which is a problem if say, I'm interviewing someone and I want them to hear a quote from someone else and comment on that. One work around is to turn up the FX dial on #1C. What that does is it resends anything coming out of Powerbook' headphone jack back into audio source that Skype is listening to. The problem with this is that if the guest speaks while this is happening, they get serious echo.
2. Cleaner simpler way: use another audio source and plug it into #1D on the mixer. Could be a PC, or an iPod or whatever. That's what I'm doing right now.
3. Find Audio recording software that does playback and record at the same time and that can pick and choose, per track, which of the system's audio channels to use. For example, when playing back a track, route audio from that track into a channel that Skype is listening to as well as to the main mix. When recording (at the same time as doing playback), set the record to where ever the main mix is. Most literature on recording software doesn't get down to this level of detail, so I have to ask around to find something like this.
Another thing I'm working with is this M-Audio Firewire 410 thing. I haven't plugged it in yet, but it's like a physical mixer with ins/outs, but the cool thing is that you can connect it to your mac via firewire and then all of those ins/outs become audio devices on the Mac. Then you get a software mixer with all the controls and slides. This could simplify things for me in that I may not need the Griffin iMic anymore. The iMic is cool, but I have a serious EFI problem when the Powerbook's fan comes (to cool off the processor) that results in a noticeable high pitched whine in the background of my recordings (and unfortunately the fan cranks up ALL the time with all the stuff I'm running on this baby). I'm hoping that coming in through Firewire eliminates this and provides me with some additional benefits bu haven't fully investigated it.
Finally, there are several "recipes" for podcast setups that can be found on the web. The two I like and that are almost the same are on engadget and dave slusher's site. They don't include Skype in the formula though. Just straight podcasting. They involve the use of two pieces of software -- LineIn and Soundflower -- to handle mixing and routing of audio into a main mix channel that lives inside the Mac, separate from the Mac's built in audio, and that turns up in the Mac's system preferences as another audio source (which is very cool). By the way, the Windows equivalent of Soundflower is called VAC (Virtual Audio Cables). The problem is that with all the software running... Soundflower, LineIn, Skype, Audacity.... things kept crashing. It was very sensitive to the order in which you loaded things and proved not to be nearly as reliable as just eliminating as much of the software as possible.
The mixer therefore serves as the replacement for all of Soundflower and LineIn. But that affects portability. If you want to go mobile with your podcast setup, I can't run the UB802 off of a battery. Nor will the M-Audio device work off a battery (it would have been very cool if I could draw power from the USB port for the M-Audio device). So, even if the M-Audio device ends up replacing the mixer, I still have a portability issue. To solve that, I'm also looking for a battery operated mixer that's AS capable as the UB802 (so if you know of one, please send me the details).
Update from David on M-Audio 410: There is an error in the recipe I sent you. My assumption was that this needed external power. It doesn't as long as you've got regular firewire connection (not the mini 4 pin kind). No idea how much gas it sucks if you're running off of battery.
The Soundflower/LineIn configuration did not appear to be as sensitive when Skype wasn't in the formula. So, to some extent, the external hardware (mixer, or M-Audio thing) is all about pulling a Skype connection into the recording (which for me is very important). So, for non-skype podcasts, you might be able to get away with nothing but the powerbook (ultimately very portable). I'm not sure how you'd bring in a good balanced, XLR-based mic like the Shure SM58 though. Perhaps a battery operated mixer is the way to go for really good audio. I also wonder if bluetooth might represent a good audio source on the powerbook (interesting thought).
Finally, even my mixer setup is somewhat sensitive to order. If a Skype call is active and Audacity is loaded after that connection is made, the audio from the call gets dropped. You can hear this in action in my podcast interview of Userland CEO Scott Young.