SCD | Visualization | VEG
Visualization and Virtual Environment Facilities - Audio Info
The NCSA CAVE uses a 15" subwoofer, a Velodyne ULD-15S4, driven by the matching
ULD-15A4 servo-controlled amplifier. The servo systems are great. When
hooked up correctly, they use a "four-wire" hookup to the Sub: two to
carry current, and two to sense the voltage right at the Sub's terminals.
The sense lines carry no current, and so have no I-R or inductive voltage
drop. This puts the full force of the negative feedback of the amp to
work right at the Sub, and drives the damping factor through the roof.
It's effectively the same as using superconducting moster cables. In
practice, it gives you very tight, very controlled bass, even when it's
loud.
For the Atoms (and generally, too) we use Bryston model 2B-LP. But you
could also put Hafler amps in there and not go wrong. The Hafler amps are
MOSFET output, and the Brystons are Class A, both of which are very low
distortion.
Question:
> He suggested that we first get a system with a 4-channel amp (or 2 matched
> 2-channel amps) of about 100W per channel. This system would drive 4 small
> speakers mounted at the top of each corner of the CAVE- he thought the Atoms
> were good- he also likes the Electrovoice S-40. If this turns out to be
> good enough for our needs, we're all done! If we wish more bass, we could
> then get 4 passive crossovers, each of which would split the speaker signal
> into two parts- one going to the small speakers, and one going to a woofer,
> or unpowered sub-woofer, like the Paradigm SB-120, which would sit on the
> ground somewhere near each corner of the CAVE. Steve pointed out that this
> would give the bass notes a little orientation, which might be nice if
> generating noises of explosions and other more-or-less point sources.
Well, let's cover each point, one at a time. First of all, the human
perception of location is very weak below 100 Hz. This has been proven in
tests, but you can reason it out since the wavelength at 100 Hz is already
11 feet. In fact, those are the frequencies that you "feel" in your body
more than you "hear" with your ears.
You can try it yourself, but we've found that if the satellites respond to
below 100 Hz (most of this genre go down to about 80) so that the
subwoofer truly only carries sub-100Hz audio, then the overwhelming
majority of the localization, by a long shot, will be carried by the
satellites. This means that the placement of the subwoofer is not critical
to preserve the effect, and that one subwoofer, placed in the room
reasonably close to the CAVE, is the same as four placed one at each
corner. In fact, it may even be better, since the waves are coming from
one point instead of four, and so they won't produce an interference
pattern in the CAVE field.
Now about the crossovers: I prefer active crossovers over passive - they
don't ruin the damping factor on the Sub. Passives have very bad impedance
gyrations near the crossover point. Here's how we recommend doing it:
Use a high-quality mixer (we use a Mackie CR1604VLZ) to pre-process the
SGI (and other) audio into four main sends. These would go to the four
hi-crossovers and then to the four main satellite amps. Then use an Aux
bus to combine the 16 channels, or preferably the four main sends if you
can patch it right, into one mono mix. Send that to the lo-crossover and
then to the subwoofer amp.
If you're lucky enough to find a subwoofer that rolls off in a
complementary way to the satellites, and if you're relaxed enough not to
split hairs about perfect response flatness at the subwoofer-satellite
transition, then you can forgo the crossovers completely. If you must, you
can then use a touch of EQ (either in the mixer chain itself, or placed
between the mixer and the amps) to smooth out the transition band, but
also realize that you'll need an audio analyzer of some kind to perfectly
balance the system anyway. The balance among the satellites is much more
important than the balance between the satellites and the subwoofer. In
fact, it can be left as a user adjustment, since relative bass energy is
very much a matter of taste.
