Last update: July 17, 2006
Recording the TWiT podcasts is a work in progress. I'll update this page as we refine our techniques. Our goal: to record the voices of two to six simultaneous participants in disparate locations, maximizing audio quality while minimizing post production requirements.
This is how we've been doing it since TWiT episode 10:
I record almost all our shows from my small studio in Northern California. I'm using a mixer and a professional microphone, the rest of the participants are usually calling in via the Internet telephony program Skype
. We've tried Gizmo, GoogleTalk, iChat, TeamSpeak and Ventrilo, but Skype really offers the best sound with the lowest lag. Any lag between speakers really kills the flow of the conversation and requires excessive editing (see Episode 9, the raw
and the cooked
I'm shooting for a sound that's pretty close to having all of us in the same room. For five or fewer participants with sufficient upstream bandwidth on the host side (the Leoville Labs have a 5mbps down/768kbps up DSL line) Skype actually offers pretty darn good results - very little latency and surprisingly high vocal quality. Per Steve Gibson's suggestion I use a dedicated port for Skype (any port above 1024 will do), and open that port for UDP connections on my router. This eliminates the need for Skype Supernode support and seems to reduce latency.
All participants are required to have headset microphones. I recommend the Plantronics DSP-500
. We get such good results with these that I often overnight a pair to guests who don't already have one. At $50 you won't get better sound for the price. USB headsets seem to perform better than others.
For recording Security Now, Steve actually uses MAudio Fastrack
with a Heil PR-40 mic. When you combine his excellent audio with the low latency connection between us you get the best Skype can do - and it's pretty darn close to studio quality.
I also use the fantastic Heil Sound PR-40 microphone
for all my podcasts and the radio show. This pro mic requires good technique, but it offers excellent sound at a very reasonable price.
People often ask me how I record Skype and get my voice along with all the participants. You can do it in software with programs like Hotrecorder
for Windows or Audio Hijack
Mac, but I get better results using a mixer. I record my voice and the Skype output from the PC on separate tracks then mix them together in post production. If necessary I can process the separate tracks to improve the sound (although most of the time I don't need to).
Skype audio from a PC and my mic audio are mixed with a Mackie Onyx 1620
mixer with the Firewire I/O option. This board can handle up to eight mic inputs and can output up to 16 individual channels digitally to a computer over Firewire. I record the Firewire output on an AMD FX 64-based ShuttlePC with a 10,000 RPM Western Digital Raptor hard drive using Windows XP and Adobe Audition 2.0
. Audition is based on the venerable Cool Edit Pro and seems to have the best features for radio style multi-track recording and production. I find I can work very quickly in it.
I also record a stereo backup mix, just in case, on a Marantz PMD-670 compact flash recorder. That recorder has saved my bacon many times.
I do very little editing on the shows. Post-production mostly requires trimming the audio, mixing down the tracks to a single mono track, running that mix through Audition's compression plug-in to even out the levels (I compress it pretty hard to increase listenability), and mixing in the intros and music.
After mixing the show I save out an MP3 from Audition. For most of the shows I create a mono 64kbps MP3. For historic reasons, I offer TWiT in five flavors: a 64kbps MP3, a 32kbps MP3, a 16kbps MP3, a 64kbps Ogg Vorbis, and AAC. Audition can save into all the formats except AAC. I use iTunes to do the AAC conversion from a full-quality WAV master. I don't currently create an enhanced AAC version, but I'm looking into hiring someone to do so. The main deterrent is the time it takes - I generally record, produce, and post all our shows in just a few hours.
I add tags to the MP3s and Ogg files using Windows Explorer's property inspector. I use Windows Media Player to add album art to the MP3s. I use iTunes to add metainfo to the AAC file.
Once the files are prepped I upload them to AOL and Cachefly then update this site. I use an excellent Mac program called Feeder
to update the RSS feeds. Eventually we hope to update feeds directly from this site.
Our live setup for TWiT varies depending on the venue. Most of the time we use a Korg D3200 mixer
. The channels are recorded to the Korg's built-in hard drive then output to a Macintosh for mixing with Soundtrack Pro (Alex prefers it to Audition). We use the unobtrusive Countryman E6i
headset mic. It gives us good sound without being too obtrusive and since we videotape most live appearances, that's important. I find that headset mics also help keeps inexperienced broadcasters on mic.