Here are some comments from radio professionals who make their living
recording interviews. You can find these on Transom.org, "an experiment in
channeling new work and voices to public radio".
Jeff Towne on iPod recording:
"It's still early days, but i don't have great hopes for the iPod as a
recording device. The Belkin mic has been criticized for poor sound quality,
think it records at a pretty low sample rate. The Griffin isn't out yet, but
it seems to have decent specs, and an external mic input, but it's still a
minijack, mono only, and there don't seem to be controls for input volume,
etc.... I rather doubt it will have a primo mic preamp or A/D converter."
Jay Allison on Belkin Voice Recorder for iPod:
"In the interests of science and Transom, I purchased one of these just to
make sure it sucks and it does. The incoming gain is fixed and set very high
so any close-miced interview distorts terribly. You'd need to place the unit
about 3 feet away to get non-distorted sound and then everything is just too
far off-mic, and the ticking sounds of the hard drive are quite legible. It
might be okay for dictation or evidence, but that's it. It actually records
drag and drop .wav files though, which seems promising for the future."
Jeff Towne on iPod recording:
"It's true that some of the specs look decent, as in recording 16-bit wav
files, and having a 20hz-20khz frequency response (frequency is not a valid
measurement of dynamic range, and 30dB is not a ratio, as Griffin states
them, it's always a little worrisome when the stats they proudly post don't
make sense!)
But here's the first problem, they also say:
Recorded File Specs (iPod Specified)
8kHz, 16 bit mono WAV files
A sampling frequency of 8 khz can't possibly give a true 20hz-20khz
frequency range. In fact it's hard to imagine any useable audio sampled at
8khz: there's a thing called the nyquist frequency, which is half the
sampling frequency, and is the frequency at which one needs to apply a
low-pass filter, to filter out "aliasing" effects inherent in digitizing
audio. So an 8khz sample rate would imply a 4khz filter roll-off, which
might create intelligible recordings, but not too great sounding....
The other variable is the Automatic Gain Control. Some recorders implement
this well, and can give nice sounding results, the AGC acting just like a
good compressor and limiter. But the stats given, if they are to be
believed, indicate a fairly long attack time and a REALLY long release time,
meaning that the AGC won't catch quick transients, and will stay low for a
long time after a loud sound. Ultimately what this will likely mean is that
we'll be able to hear the volume pumping. But it's unfair to guess at this,
we should wait and see.
I think it's a really cool concept, I'd love it if I could slap an accessory
on my iPod and get a nice recording, but sadly the mass-market is mostly
interested in just getting decent enough sound for voice memos, or
dictation, or recording a lecture. In those circumstances, high audio
fidelity is not a priority, as long as you can hear what's going on.
In the end, I'll be dropping $35 to experiment, and who knows, maybe we'll
get lucky and despite the specs, it could sound just fine for some
circumstances. We ran a really good piece here on transom that was done on a
microcassette recorder. So you never know... But I don't have high hopes for
"broadcast quality."
|