In
search
of Ongaku: there are good reasons why 33 Bowls recordings sound
uniquely different from most others.
As a
technical backgound, yours truly worked in the telecom industry in the
'90s, you could call me an ex-insider. I was an alphageek straight out
of a movie set of “revenge of the nerds”, bunch of us in t-shirts and
jeans, on the team that designed the chipsets for the first candy bar
sized shirt pocket truly portable cell phones. We were just a bunch of
tecnho wizards making cool stuff. That design morphed into star trek
flip phones, then the ubiquitous touch screen “smart” phone took over,
which is another story, to be told later. But I was there at the
genesis. Then, jumped ship to another startup which set the standard
(called owning the socket) for chipsets that are the internet backbone,
gigabit mux/demux si-ge “silicon.” Pushed the envelope, owned the
sector, outside the box, and a host of corporate ciches. Then, another
startup that did the first bluetooth/wifi integrated solution, later
snapped up by one of the biggies.
Somebody had to make the hardware, the physical basis of the dot com and app eras.
We are several
generations into becoming accustomed to electronic artifice in our
music. However, we are still sensitive, including sub-consciously and
viscerally, to very small dynamic shifts,
glitches, or artifices in the parts per million range. Physical
resonance, energy storage, time smear
and consequences of such are reasonably well understood in
electro-mechanical devices, aka loudspeaker transducers and cabinets.
We have finite element analysis, vibration analysis, laser
inferometers, waterfall curves and all sorts of fun geek toys resulting
in all sorts of fun geek toys such as ring radiator and ribbon
tweeters, carbon fiber woofer cones, sorbothane damped monococque
speaker cabinets. However, a similar, more insidious phenomena exists
in what has been presumed to be a purely electronic domain: amplifiers,
including phantom powered microphones, pre-amplifiers, power
amplifiers, A/D and D/A buffers and current to voltage converters, and
the A/D and D/A converters themselves.
The germanium, and now silicon and Si/Ge solid state revolution, while
great for miniaturization, can
and
usually does pollute music with thermal tails. Along with the textbook
ideal of transistors being electrical devices as electronic switches
and amplifiers, they are also responsive to optical, acoustic, and most
importantly thermal input. Self generated thermal tails do not show up
in static analysis or measurements. However, in most designs, the
active devices wiggle around via self generated self heating following
a sloppy envelope of the power dissipated in response to signals
passing through. Settling time constants, while depending on the
package thermal characteristics, correlate typically with perceptually
significant
frequencies. Bipolar transistors have a threshold shift around 2
mV/degree C; field effect transistors around 6 mV/degree C. It does not
take too many milliwatts of variation to add a thermal tail
considerably above
the surprisingly small audible threshold of one part per million or
minus 120 dBV. This is exacerbated in typical long tailed differential
input feedback topologies, where the error signal is outside the
corrective feedback loop, and the thermal tails become a chaotic
perturbation of the input offset amplified by the closed loop gain,
rather than being reduced by loop gain. Worse, integrated circuits
couple thermal isofronts from the entire circuit "sloshing" around and
wiggling the parameters of sensitive devices, again at time constants
correlating with audible frequencies. Settling time specifications are
usually to 0.1%, sometimes to 0.01%, and rarely to 1ppm; even carefully
interleaved multiple parallel device layouts are going to exhibit a
settling time as the devices attempt to average out the "sloshing"
isothermal front.
These "sloshing" isothermal fronts also show up in mixed signal and
even strictly digital integrated circuit chips, as modulation of
reference clocks and clock distribution via thermal modulation of
threshold voltages. This is distinct from power supply impedance and
ground bounce issues. We are again, surprisingly sensitive to small
time domain errors, whether predominantly random or deterministic;
including 1/f phase noise, aka close in or low frequency jitter in
conversion clocks.
Back to the analog domain, the classic thermonic emission devices, aka
vacuum tubes and nuvistors,
while having a host of sometimes infuriating design issues, inherently
have less of these types of problems. The filament provides a large
thermal bath, the signal level can be much smaller relative to the
operating voltage, and most designs with either resistive or inductive
loads set the operating point towards the middle of the curve which is
closer to constant power mode. Also, there are no minority carrier
p-type devices, so electron density is generally higher. And, the
dielectric of a vacuum is the most linear available, reducing
inadvertent "splatter" modulation of miller effect capacitance.
The golden age of vacuum tubes did not happen until transistors were
invented or discovered; as if the era of the Sun setting corresponded
with the maturation of the technology and implementation of that
technology. Likewise, we may be entering the golden age of silicon as
photonics rapidly advances from laboratory curiosity to practical
applications; and non-biological room temperature quantum devices may
still be decades away, possibly superseding photonic devices in the
distant, and still hazy future.
But, for now, silicon solid state is here to stay. There are ways to
design for constant
power or stasis modes of the significant active devices in solid state
electronics; the results are stunning and startling. Clarity, nuances,
micro-dynamics, textures, the ebb and flow of music emerges from a
velvety black silence. Music breathes, it is felt, viscerally and
emotionally, it is involving. Similar in
profundity to some transcendent experiences, but with more brain cells
remaining intact.
This is so with both natural acoustic sources and intentionally
created artificial musical spaces; the old and the new can co-exist
harmoniously, enjoyed for what they are in a clearer understanding of
what is what.
Music enjoyment should be both effortless and involving.
As we have
evolved technologies faster than we can adapt, nuance in the beginning
of the recording chain becomes even more significant. Two things happen
in the brain/CNS of the listener when listening to overly compressed
music stripped of it's essence, textures or nuance: more "horsepower"
or processing in the brain is done to extrapolate the missing nuances. Although
minimalism can be a potent visual art-form, in sound and music it takes
more to perceive, and the extra brain/CNS processing leads to shorter
attention span, and the other thing that happens: less emotional
involvement, as fMRI scans show that compressed music lights up less of
the emotional centers of the brain.
So far, such transcendant audio is mostly in the realm of DIY geeks and a few very,
very expensive commercial designs. The dominant paradigm has not really
shifted to understanding the dynamic nature of nature, yet.
However,
all
is not lost.
The metaphor of a chain being as strong as it's weakest
link is not really apt. In music reproduction, the experience is that
the textures are only as nuanced as the most transparent link in the
recording chain, which happens to be at the source. As distribution
media and stereo systems evolve, enjoy. Update: MQA is more than just incremental evolution; it is a true paradigm shift.
There's magick in them there
bits.
Indeed, now that "Breathe" has been remastered in MQA, the magick has been made accessable to all.
And do protect your hearing. Consider it a long term investment.
Earplugs are essential for most movies and concerts, and other loud
venues.
