As I look to push the envelope on real-time, multimedia (text, audio, images, video) journalism here at ZDNet, I'm also constantly trying to figure out what gear is the best gear to have if you're a journalist that has to do it all: Write stories (obviously, a well connected PC - I use a Thinkpad T42 and we publish on Wordpress); Take pictures (a good digital camera -- although we're still figuring out how to use them, Dan Farber and I both own Nikon D70s); Record audio (for me this means a dependable podcasting rig that works well for in-studio and on-location recordings where there might not be any power). Here's where things stand in terms of my thinking and what I'm currently using for ZDNet's IT Matters series of podcast interviews.
The Innkeeper goes toe-to-toe with the desktop version of the Telos' Telos One (the desktop model). They are "digital hybrids" that can "seize" a phone line and are often used in radio stations to put listeners on the air. The digital technology separates the caller's voice and the DJ's voice into two separate audio channels. Then, using the Innkeeper's built-in mixer, the DJ can remix the two channels into one output. But the key, in terms of final production quality, is that the DJ can separately control the level of each channel. This way, if the caller's voice is coming in a little weak, the DJ (or you if you're using the Innkeeper for a podcast interview) can crank it up a bit to put the callers voice more on par with the DJ's voice.
With the Innkeeper, you don't have to use its built-in mixer. On the back, it has two XLR (three-pronged) audio outputs. One for the full mix (if you are using its built-in mixer) and another that has just the caller's voice (made possible by way of the digital separation). The Innkeeper also has an XLR-based audio input that can be connected directly to a microphone, if you want. In that configuration, the Innkeeper takes your voice and feeds into the phone-line. Since its digital technology is feeding your voice into the phone line, it also knows what audio to filter out of the audio it's picking up from the phone line so all that's left is the caller's voice (typical phone lines send your audio back into your ear, which can ruin the audio if you're trying to record a phone call). Then, from the full-mix output on the back of the Innkeeper, you feed the full mix into a recorder.
In my case, instead of using the built-in mixer, I use my own: a Behringer MXB1002 (discussed here and pictured below) that I picked up on eBay for around $60. Compared to all other mixers out there, I really love this mixer for four reasons.
For starters, you can operate it on batteries, or plug it into a wall. Battery operation comes in handy if you have to setup somewhere where there are no plugs. There aren't many decent sub-$100 battery operated mixers that can satisfy the needs of a demanding podcaster both in the studio and out in the field. Shure has some high end ones (the FP series, seen at the bottom of this page) that professional movie makers, news crews and videographers use, but they're very expensive. Even on eBay where, in addition to the current FP33, older discontinued models like the FP31 and FP32 can be found. Second, it has five separate XLR based input channels (providing a lot of flexibility. The more inputs you have, the more granual control (gain, volume, EQ, etc.) you have over each audio stream you might add to your mix and ultimately the more control you have over the final quality of your production. An audio stream could be a microphone. Or, it could be a recording being played back from an MP3 player, a cassette recorder, or a computer that you want your interviewees to respond to.
Thirdly, the MXB has three separate audio buses (main mix, FX, and mono) to which each of the five inputs can be optionally fed (more on this in a second). Unfortunately, the mixer is discontinued and the only ones out there that I know of are used. Fourth, there's a bit more granularity to the LED lights that show the output level of the main mix than there is with the mixer I was using before: a Beheringer UB802. In terms of audio quality in your final production, you can't count on your ear to tell you if all audio is coming in at roughly the same audio level (which is what you want). You need something visual to prove it. Some mixers like the aforementioned Shure mixer have needles. Others have LED lights that hop up and down with the audio level. In either case, the more granular the viewability of the audio, the easier it is to make sure most of it is at or near 0 db or "line level."
In my configuration, I feed the output of my Shure SM58 cardioid dynamic microphone (at left) into the first channel of the mixer (see my must-read treatise on microphone selection) and, using the first channel's dials and knobs (on the mixer), I contribute it's audio to two of the mixer's three buses: the main mix (there's a slider for this) and the mono bus (uses a knob). The mixer can send any audio on the mono bus out to another device through a standard 1/4" mono phono jack. So, any audio on any of the five channels (including my voice coming through the microphone) that I want my interviewee to hear gets contributed to the mono bus by turning up the "mono knob" on those channels (each channel has its own knob).
Using an 1/4"-to-XLR adapter cable that I purchased through MarkerTek, I connect the mono bus' 1/4" output to the XLR input on the back of the Innkeeper. Any audio going over that cable will also go down the phone line to my interviewees. Then, from the back of the Inkeeper, I connect a female XLR-to-male XLR cable from the caller-only output (the interviewee) to channel two on the mixer. This brings the interviewee's audio back into the mixer. But, when I do this, I have to make sure that the mono knob on channel two is turned down to zero or the interviewee's voice will be sent into the mono bus, to the Inkeeper, and back down the phone line in a way that produces a slight echo that will not only ruin the final production, it will also cause the interviewee to stutter (very few humans can avoid doing this when their own voice is echoing back to them). While the mono knob is turned down, the slider that control's the contribution of channel two to the mixer's main mix is adjusted on a per interview basis to make sure that the interviewee is at or near line-level. This is done by asking the interviewee to talk for a short bit while I watch the aforementioned LED lights.
Note that channel one on the mixer, which is connected to my microphone, is also contributing to the main mix. With channel one, I first adjust the knob for gain so that the microphone is sensitive enough to pick up my voice, but not so sensitive that its picking up everything else (room noise, street noise, barking dogs, etc.). Then, I adjust channel one's slider to make sure my voice is being contributed to the main mix at a sufficient audio level. Now, with the main mix having my voice and the interviewee's voice and any other audio I might be contributing (through one of the mixer's remaining three channels) nicely adjusted to a consistent audio level, I use another 1/4" mono phone to XLR cable to feed the output from the left side of the mixer's main-mix-out jacks (a 1/4" jack) to one of the inputs on an Edirol R-4 four channel portable digital recorder (pictured below).
I spent a lot of time trying to decide on what recorder to use. Prior to having a dedicated recorder, I mostly used a computer to do the recording (using the open source-based Audacity software). But, after doing more than 100 podcasts, I learned the hard way that computers (both PCs and Macs) have this nasty habit of occasionally crashing while in the middle of a recording. This could be disasterous if I'm with someone and only have 15 minutes of their time and the comptuer decides it's going to quit on me 13 minutes into the interview. Several podcasts that I thought were going to be published never were. When it comes to guaranteed, always going to work no matter what, you're just better off with a dedicated digital recorder. Going back to the ability to visualize your recording levels, you're also better off if the one you pick shows you your recording levels. While I've played with a pocketable iRiver H320 (which includes a line-in jack and records directly to high quality MP3s), its lack of any visuals when recording were problematic. The other problem with the H320 is that, by itself, it's tough to use more than one microphone. While it's input is a 3.5mm mini-plug stereo input (which equates to two channels), going from that to two or more XLR-based (high quality) microphones in a way that the levels on each of the microphones can be individually adjusted and monitored is kludgy. Not impossible. But kludgy.
Because battery operation is key, the choices boiled down to three options all of which provide seperate recording level visuals for each channel. The 4-channel Edirol R-4, the two-channel Edirol's R-1, and M-Audio's two-channel Microtrack 24/96. At $1300 to $1500 depending on where you buy it, the R-4 may be too much for your budget (the R-1 and Microtrack are much cheaper). But, the other options involved sacrifices I didn't want to make. For example, whereas the R-1 and the Microtrack use 3.5mm mini plug jacks for its inputs, the R4 uses professional three-pronged XLR jacks. Four of them. One for each channel.
A lot of people I've spoken with think XLR is no big deal and that 3.5mm jacks are fine. But XLR jacks are built for durability. 3.5mm jacks are easily thrashed and over time, if you've ever seen one on the arm rest of an airplane seat, end up being destroyed. So, what do you do if the 3.5mm jack in your digital recorder wears out or gets destroyed? Send it in for an expensive repair (during which time you don't have a recorder)? Throw away the recorder and buy a new one? The other feature I relish in the R-4 are the four separate channels. This means that I can connect four microphones and individually control the level on each one of them without having to front-end the recorder with a separate mixer. That said, a seperate mixer does come in handy if I have to add even more microphones (eg: take my five channel mixer and pump it's main mix through one of the R-4's four channels and now I have a total of eight separately controllable channels -- all on battery operation). Lastly, the R-4 comes with a bevy of built-in editing features that allow you to do preliminarily clean up your audio before you export it (via a USB connection) to your PC for final post production and conversion to MP3 (the one major downside to the R-4 is that it doesn't record natively to MP3... only WAV).
Finally, for post-production (adding some music, introductions, etc.) and conversion to MP3, I use Audacity but am contemplating something a bit more industrial strength.