| Video Noise June 2001
Better Sound for Your Next Receiver/Processor
It looks like a threshold is about to be crossed. Motorola
DigitalDNA DSPs (digital signal processors) have appeared in the first home-theater
surround processor. DigitalDNA processors are truly next generation, making the SHARC and
others look like 8086 processors compared to a Pentium 4.
In previous columns I’ve warned about
receivers/processors that degrade the sound due to analog crossovers being inserted into
the signal path of the audio signal. The only way out is to perform the crossover
operations in the digital domain. However, DSPs used in most existing and past receivers
and processors did not have the horsepower to perform complex calculations, like
crossovers, in the digital domain.
With the arrival of processors employing Motorola
DigitalDNA technology (more info at digitaldna.com),
the horsepower and integration is finally available for every manufacturer to be able to
employ fully digital surround-sound management. This means that the Motorola processor can
control surround modes, delay times, the number of active speakers, the crossover applied
to each speaker (determined by speaker size), the volume level sent to every speaker, the
overall volume level, late night listening mode, tone controls, you name it. These new
processors can control each output-related function with precision up to 24 bits and with
no loss of resolution.
This takes many traditionally analog functions completely
out of the signal path, as everything can be done in the digital domain. Properly
implemented with high-quality power supplies and thoughtful circuit topology, this fully
digital control should make every receiver and processor that adopts the technology sound
better. You can do a whole heck of a lot to an audio signal in the digital domain and it
will still sound excellent. But altering the signal in the analog domain has always had
negative sonic consequences because every component (resistor, wire, capacitor, inductor,
etc.) has a "sound" and some kind of loss. That’s why many audiophile
products built to sound as good as possible have few extra features. That keeps extra
electronic components out of the analog signal path.
Of course there will be opportunities to screw up, even in
the digital domain. There will be nasty sounding "cathedral" processing modes or
a "stadium" mode where the vocals disappear into nothingness (nothing new
there). Going digital doesn’t prevent mistakes; it just makes the mistakes sound
better than they would have in the analog domain.
The advent of the ability to do all surround processing in
the digital domain is most exciting when it comes to bass management. These new processors
can be quite flexible, allowing you to select a crossover in increments of 5Hz for each
loudspeaker for example. So if your main speakers are -3dB at 35Hz, you pick 35Hz for the
main speaker crossover. Bass below 35Hz goes to the LFE subwoofer. You use as much of the
main speakers’ capability as you can and don’t lose any of the bass that should
be in those channels. If your center speaker is -3dB at 90Hz, pick 90Hz for the
center-speaker crossover and all of the center bass goes to the LFE subwoofer. The
majority of processors today either discard, under-power, or over-power bass in the LFE
channel either via mistake or trying to save money by keeping the bass-management circuits
as simple as possible. Digital bass management done correctly should produce a more
consistent and better overall surround-sound presentation for every home theater.
Besides picking the crossover point from a wide variety of
choices, the new generation of processors will also allow user-selectable slopes.
Technically you could send 35.00000Hz and above to the main speakers and 34.99999 and
lower to the LFE subwoofer. The only question is whether this brick-wall filtering would
sound better than a 1st, 2nd, 3rd, 4th, or higher slope crossover. In addition to the
slope being selectable, crossovers in the digital domain will not add time or phase error
to the signal because the analog components that introduce those changes to the audio
signal would not exist.
It’s likely that there will be competitors for
Motorola’s new processors, but that’s OK. There will be pressure to reduce costs
on these processors that will put them in the reach of sub-$1000 receivers soon. One of
the first products to use Motorola DigitalDNA DSPs is B&K’s new $2800 Reference
30 Processor (review to appear here in HT&S). This means some top-of-the-line
receivers (costing $3000 or more) are probably next. It’s only a matter of time
before these processors appear in sub-$2000 and sub-$1000 products.
The age of universal, fully digital surround processing at
low-to-midrange prices approaches -- I can hardly wait!
...Doug Blackburn
db@hometheatersound.com |
