Spread Spectrum Technologies, Inc.
Trinaural Processor
Manufacturer: Spread Spectrum Technologies
Inc.,
716 N. G St. #2, Lompoc, CA 93436; 805/740-9902;
sstinc@earthlink.net;
www.ampzilla2000.com
Price: $1,500
Source: Manufacturer loan
Reviewer: David Rich
The name “James Bongiorno” may not be as familiar to
audiophiles as it was in the ‘70s and early ‘80s. He
has been out of circulation for a number years because of some
serious medical problems that brought him to death’s door
more than once. He now appears to be completely healthy and is
in charge of a company named Spread Spectrum Designs.
James is perhaps the most important innovator that
high end audio has seen in the analog transistor era. He has been
directly involved in almost every major breakthrough in power
amplifier design. In his field of specialization he is one of
the most creative electrical engineers on the planet. Bob Carver
is the only other name that comes to mind at the moment. Bongiorno
and Carver are both inflicted with some strange disease that mostly
limits their vision to things audio. Both would no doubt be very
rich if they had chosen to work in another area of the electronics
industry. Indeed, the term “spread spectrum” is the
enabling technology of 3G cell phones and wireless LANs. One can
only imagine what would have happened to Bongiorno had he taken
his company in the direction of wireless data systems. He could
have done just this, because his expertise spans to RF design
(he once designed a tuner called “Charlie” that contained
many innovations in the front-end design that found their way
into most high-end tuners later in the decade), but fortunately
for us audiophiles, he applied his genius to things audio.
Early in his career, James worked at Marantz. He
was contributing engineer with the Sid Smith as senior engineer
in the project for the Marantz Model 15 Power Amplifier which
was the first example of a topology that is virtually identical
to most modern power amps. Some high-end designs have moved past
this topology to a fully complementary one, which as it turns
out was developed by James in 1973 (Scientific Audio Engineering
31B). Slightly later a higher-power version called Ampzilla was
introduced by James as a kit design in Popular Electronics. Demand
for both the kit and assembled unit created such demands on James’s
time that he was soon running is own company, GAS (The Great American
Sound Co.), producing this 200-wpc unit and its baby brothers
with cute names such as Son of Ampzilla (first separate power
amp I ever owned -KWN). James also introduced preamps that incorporated
the new balanced topology. Other innovations included the introduction
of DC servo circuits (Theadra, 1975) and very low-noise phono
preamps that did not need a pre-preamplifier stage.
Disagreement with partners resulted in James’s
departure from GAS. Soon another Bongiorno company called Sumo
was in business. It was here that the third-generation amplifier
topology was hatched—the fully balanced topology. A 450-wpc
monster balanced/complementary amplifier came out of this effort
(Power, 1979). Other balanced amplifiers were produced by others,
but problems keeping the amplifiers stable soon led others to
abandon the topology despite the fact that it solves a whole series
of technical problems associated with amplifiers with the speakers’
output terminals referenced to ground. The reason Sumo was left
standing alone with this topology is it is extraordinarily difficult
to stabilize the DC and AC feedback networks.
Spread Spectrum Designs is now introducing a new
set of preamp and power amps. I have seen the amps’ schematics
and they represent a fourth generation in power amplifier design.
This design, like the third-generation Sumo design, is virtually
impossible to copy by the “let’s replace the yellow
cap with a bigger red one” crowd. Only an analog engineer
at the highest level could understand what is going on in this
amplifier, let alone create the concepts to realize the design.
Now, I am not going to get into an argument on whether
modern electronics (i.e. stuff copied from the Marantz 15 or Ampzilla)
has a sound that cannot be attributed to measurable nonlinear
operation or small frequency response errors at the speaker terminal.
All I will say is that if you believe it can sound different even
if they measure the same using traditional measurement methods,
then the fourth-generation power amp should sound better than
one that is developed from Bongiorno’s first- or second-generation
designs.
Luckily for me, this review is not about the new
power amp or the soon-to-be-released preamp. I will leave that
to others. Instead, in this review, I look at a whole new class
of components. Bongiorno calls his new design a Trinaural processor.
What it does is revectorize the composite stereo information into
a three-channel signal. The extra channel drives a center-channel
speaker.
OK, what is this “revectorize” stuff?
To understand what it is on the simplest level we first need to
recall from my little red book of Hi-Fi that two-channel stereo
creates a soundstage by varying the intensity of the left channel
relative to the right channel. Thus, the left always has right-channel
information and the right always has some left-channel info. In
theory, when both speakers play the same thing at the same level
(our little red book calls this “mono”) you have a phantom
center. You quickly find out this is only theoretical, however,
if you run mono pink noise through your system.
Not only is the image not centered, but it is spread
between the channels, the size of the space varying with tonality
of the noise. This is the result of the room interacting with
the two speakers indifferent fashions and your brain’s ability
to create a virtual center. How many times have you marveled at
the beautiful sound stage of a recording only to find out you
had the mono button pushed in? You got fooled because of the phenomena
discussed above.
Let’s now plop a center channel as part of
a stereo system and put a mono signal through it (with some compensation
for the level increase in the center relative to left and right)
to get a REAL center. Now play the pink noise and it comes from
the center. Alas, we still have a problem—the left channel
still has right-channel signals present. Recall that we needed
this to make the soundstage with two channels—but with three
channels we now need to get rid of this crosstalk.
Consider panning a signal from left to right in
an ideal left-center-right (LCR) system. The signal starts in
the left channel and then the left reduces in level and the CENTER
increase in level as the image moves form left to center. The
right channel should be dead silent as we move from left to center,
but it’s not. Now as we move from right to center the left
channel should stay dead, but it’s not, and the center drops
in level as the right channel comes up in level. At full right
both the center and left produce no sound.
From the above it is clear that we have to remove
any signal from the right channel that is attempting to make a
virtual image LEFT of center. We need to do the same thing on
the left channel for any signal that is encoded to be placed RIGHT
of center. In high-tech terms, this is what we mean when we say
we revectorize the two channel signal into three channels. It
is possible to extend the process to five channels (no joke—the
latest 11.2 systems have 5 front speakers). But doing a two-channel
to five-channel revectorization is very complex and may only be
possible digitally. For now, we should be very happy that we have
a piece of electronics that revectorizes from two channels to
three channels. That is what the Trinaural processor does.
Please note that three channels is nothing new.
When the big companies did the first stereo recordings they recorded
and played back three channels. The end user, however, got only
two channels, with the mix engineer trying to figure out what
level to mix the center in. At the listener’s end some companies
proposed listening with three channels, but they only drove C
with L+R, which as I explained above, is no good because L still
contaminates R and R contaminates L.
If the Trinaural idea had been hatched in the ‘50s,
any system worthy of being called Hi Fi would have been required
to have three channels. Unfortunately, that never happened, and
TM (tape monitor rooms where the pros did the editing and EQ)
had the center channel removed and big 24-track tape recorders
replaced the 3channel recorders.
For the material originally recorded in three channels
there have been endless arguments about what level the center
channel should be mixed in at. One is left to wonder what magic
is on those tapes if we could hear the center encoded on a DVD-A
or SACD. I am told but have not actually seen in the flesh that
some re issues of material on Everest have actually been issued
in LCR with the surrounds silent. One hopes that Mercury and RCA
sessions that exist of three channels will also be released this
way, but the current state of the market makes any such release
unlikely. Luckily, we now have the Trinaural processor to recreate
the center on all our black vinyl and CDs, even when it did not
exist in the original mix.
In addition to the Trinaural unit itself, all
the schematics were made available to me. The build quality
and innovation in the design of the electronics is consistent
with the $1500 price. In fact, relative to what other small
American electronics companies charge, it may well be a bargain.
The innards are all made visible at http://www.ampzilla2000.com/trinaural.html
so have no need to go into gory detail here.
ICs are used throughout the unit. It would be impossible
to do this complex circuit with discrete components. I have always
favored discrete component design for systems aiming for ultimate
quality sound because of the flexibility such an approach offers.
Bongiorno has been busy designing electrical test protocols that
allow him to identify a few op amps that he claims (and I have
no way to verify) match the performance of discrete circuits.
The op amps he has identified are not the most
expensive units, which have the best measured performance by traditional
methods, but neither are they the cheap TL071 and NJM 2068 op
amps that makes up the vast majority of the stuff you purchase
at Best Buys and from your Ultra High end dealer. One has to wonder
how anybody could say that very expensive high end stuff using
these cheap op amps sounds better (as HP did in a recent issue
of TAS), but this is a concept I have beaten to death in The Audio
Critic and it has not had much traction with the true believers
The intellectual property involving the generation
of the center channel in the Trinaural processor is unique and
the presence of such IP alone could place the unit’s price
well above the cost of the construction. In this case no charge
for the IP appears to have been added to the price. The IP is
not protected (no utility patent on file) and could theoretically
be reverse-engineered by someone competent in the field without
taking the unit apart. It is important to understand that the
simple explanation I gave of process to revectorize the composite
stereo information to LCR is not complete—the actual IP in
the unit applies more complex signal processing to take into second-order
effects when the process is applied to real two-channel sources
and is used in normal-size listening rooms.
The IP in the Trinaural processor could be implemented
in a DSP-based AV receiver for considerably less than $1,500 retail,
but questions would arise as to the sonic degradation that would
result from DSP processing, at least among the audiophile community
to which the unit is marketed. The Trinaural processor under review
here is an all-analog system intended to be driven from audiophile-grade
CD players and preamps in conjunction with audiophile-grade power
amps and speakers. Outputs are single-ended or true balanced (no
simple inverting opamp used to create a quasi-balanced output).
Inputs are single-ended. The Trinaural processor also has a built-in
subwoofer crossover centered at 80Hz. The low-pass filter (subwoofer)
is “on” all the time; the “subwoofer switch”
merely turns on the high-pass filters for the three front channels.
To use the Trinaural processor you need a something
that will act as a two-channel preamp to feed signals to the new
gizmo. That can be an integrated amplifier or a receiver if it
has line outputs after the volume control. If you do not have
that you can (and James is going to kill me for this suggestion)
pad down the speaker outputs to line level or use the headphone
jack output. What you do with the three wires coming out of the
processor is run them through any multi channel receiver with
5.1 analog input jacks for SACD players ($200 at Best Buys) You
will be using only 3 of the 5 channels since you will be connecting
only three speakers. In most cases when you put a receiver in
5.1 mode and leave it in that mode, you lose the front-end switching
functions except for the volume control, so think of this $200
as a 3-channel power amp with lots other funny controls up front
that you will not be using.
Now if you want to use the same system to listen
to 3-channel Trinaural signals and 5channel signals from DVD,
SACDs, DVD-A, etc., you will have some very interesting setup
issues. The most significant of these is you must place a center
channel speaker that front of your video monitor. You need to
move the video monitor to the left or the right of the center
speaker. The issue is not negotiable. Lucky for me I do not do
video and I find Trinaural processed CDs sound better than most
multichannel SACDs and DVD-As.
The simplest path to a home theater/ Trinaural system
is to have the AV receiver acting as the power amplifier in the
Trinaural system. In home theater mode you use the AV receiver
directly and bypass the preamplifier (or whatever component is
being used for only it preamp function) that is switching the
analog signals ) You use the A/V receiver directly by connecting
the AV receiver to the DVD player using a digital link. If you
have SACD in your system, you will need to make sure the receiver
you have supports two sets of 5.1 inputs. Things can get complex
quickly when you want to use video switching and other AV advanced
features. (You will need a six-year-old to help you to figure
this out, because this becomes way too complex for a 40something
Ph.D. to understand…)
Although the Trinaural processor is designed to
be used with the best available high-end components, I did the
sensible thing and used the unit in a system with a total cost
(exclusive of the processor) of less than $2,000. A Denon CD-R/
RW was the source for both analog and digital signals. Analog
signals were routed to a Yamaha integrated amplifier used as a
preamplifier only. The Yamaha was used to level-match the Trinaural
processor output level to the direct digital signal paths used
in the evaluation. The output of the preamp entered the Trinaural
processor. The output of the Trinaural processor was routed to
the 5.1 direct input of the AV receiver. The digital output of
the Denon CD player was also routed to the AV receiver using an
SPDIF digital interconnect. Three AR 302 speakers where attached
to the AV receiver. The speakers were deployed in an arch with
L and R angled in. This is consistent with recommendations in
the instruction manual for the Trinaural processor as well as
for instructions for proper SACD/DVD-A playback.
Note that you will find your L and R speakers angled
in more than you would do for a two channel set up if you follow
the instructions. Do not worry if you are changing your speaker
deployment. I played with other speaker deployments for hours
on end, including my favorite two channel placement for L and
R, and in the end I found that those who have designed the LCR
systems know where the speaker are supposed to go.
In my evaluation no subwoofer was used nor did I
think one was needed if you are using fullrange speakers to produce
music rather than the sound of a car crash. Remember, you have
three woofers playing now in three different places in the room.
You can get to higher SPLs without stressing the speakers, and
standing wave patterns may be smoothed out as result of deploying
three speakers in three different places in the room relative
to room boundaries.
The AV receiver was set for Dolby Pro Logic enhancement
of the incoming digital stream. Rear-channel processing was disabled.
The receiver allows remote switching between the direct inputs
being sent from the Trinaural processor and the DSP based Dolby
Pro Logic signals. This made comparison between the two methods
of generating the center channel much easier. (Note that this
is a sighted, not a blind test. I could not come up with a way
to deploy my double-blind ABX comparator into this test environment
although proper evaluation comparison of acoustical signals requires
that a double-blind test must be run.)
The Trinaural signal was level-matched to the Dolby
Pro Logic digital signal using pink noise. LR to center speaker
levels were matched per instructions of the Trinaural processor
using that unit’s internal noise generator and center-channel
level control. No level or distance correction was applied at
the AV receiver because all speakers were identical and all speakers
were placed at the same distance to the listener both on the horizontal
and vertical planes. A few notes about Trinaural processor setup
before I get to the sound. First, for some unknown reason the
switch on the Trinaural processor to optimize center channel level,
using a Radio Shack sound level meter, does not have a lock position.
The job of adjusting center level requires two people—one
to hold the meter at the listener’s chair and one to push
on the darn switch and toggle back and forth. At least the second
person can adjust the center level under the command of the person
reading the sound level meters speeding up the iterative process
of getting the center level correct.
Dramatic improvement in sound quality has been
observed by this reviewer in the past when moving from two-channel
to three-channel Dolby Pro Logic sound reproduction. This improvement
is only achieved for three IDENTICAL speakers deployed at the
same vertical planes (i.e., no center-channel speakers on top
of the TV set). Distances to the listening chair must be matched.
L/R speakers must be angled towards the listener. LCR deployment
requires the LC and LR to be spaced identically apart. C must
be dead center to the listener. If correct speaker deployment
is not followed it is impossible to image the speakers.
When things are set up properly, however, depth
of sound is dramatically increased. The size of the soundstage
is increased. Tonality of instruments becomes significantly closer
to what is heard in a live performance. Given the wider soundstage
(width and depth) and the decoupling of the sound from the speakers,
inner parts of a large symphonic work that were not heard on two-channel
reproduction are now made clear. Parts with complex unison orchestration
are heard with the multiple instruments more easily identifiable.
Dynamic range is enhanced. Clarity of vocal reproduction from
solo to massed chorus is improved, with words more clearly understood.
Harsh string sound and sibilance on vocal sections is reduced.
So, if Dolby Pro Logic is that good, is the Trinaural
processor really better? Well, yesssssssssss.
Comparison of Dolby pro logic with the Trinaural
processor showed the Trinaural processor to sound much cleaner
than the Dolby process. Deployment of instruments in the soundstage
appears to more varied. The Dolby process uses steering logic
to ensure that center images appear only in the center (LR cut
off). This is required for movie dialog, but the steering logic
affects musical reproduction negatively. On the other hand, the
Trinaural processor presents a mono signal in all three channels,
spreading the sound across the entire sound stage. This leakage
of mono signals into L and R when using the Trinaural processor
is an artifact of some of the second-order correction mechanisms
applied in the Trinaural to mitigate effects of real rooms and
real recordings. The L and R leakage effect is not noticeable
on stereo recordings, where the center is very well defined using
the Trinaural processor.
To restate, the best sound I got from Trinaural
processor occurred when setup instructions for the Trinaural process
are followed as closely as possible. In addition, high-quality
matched speakers (preferably full-range three-way designs) are
a far preferable way to go in comparison to three mini monitors
and a sub. As I said above, three big speakers launching low frequency
sound does wonders for standing wave patterns in a room. Also
note that I found in my tests with the speakers I had available
to me that using a different speaker in the center position than
what you have in LR, even if it correctly placed in the horizontal
and vertical plane, does not cut it.
OK—evaluation over. I can conclude the Trinaural
processor represents the current state of the art in three-channel
reproduction, and three-channel reproduction represents the first
significant advance in sound reproduction since the introduction
of the CD player.
I now put in a safe harbor statement before I move
on to giving additional sonic impressions. The conclusion that
Trinaural is the next best thing to the invention of the CD player
may not be valid with non-optimal placement of the LCR speakers,
such as the use of timbre-matched center-channel speakers on top
of a TV set with the two other speakers aligned on the sides of
the TV. For this reason, the Trinaural processor should purchased
with a money-back guarantee to ensure it works in your system
if you are not following ALL the rules.
I also need to disclose that I did not compare the
Trinaural processor with the newer Dolby Pro Logic II music processor
and DTS Neo 6 music processor (as well as some other less well
known processors) since they were not available for this review.
Please note that the IP used in the Trinaural processor is the
result of years of work in an attempt to improve music reproduction
by adding a properly placed center channel. No consideration to
uses of the Trinaural processor in an AV environment was made
in the development of the unit. This contrasts strongly with the
design goals used at Dolby and DTS. I did not try the Trinaural
processor with movies.
Nor did I attempt to evaluate the improvement achieved
in the Dolby/DTS systems when two or three rear speakers are employed.
The Trinaural processor has no rearchannel synthesis capability.
The rationale for not evaluating the effect of rear channel synthesis
in this review was that the development of the rear channel is
a synthesis process with frequency tailoring (perhaps dynamic)
and added signal delays (perhaps multiple) used as part of the
process. In addition, extraction of rear-channel ambiance from
a stereo mix is a guess made by the processor’s use of a
matrixing of the two channel incoming signal with an attempt to
develop a matrix that produces signals that appear uncorrelated
to the events occurring in phase in both of the stereo channels
(the hall ambiance). Dynaquad was the simplest form of this: the
rear speakers were connected across the two front channel plus
inputs. Out-of-phase signals went to the rear.
This is in contrast to converting LR to LCR. The
procedure for doing this involves no synthesis or guessing techniques.
Nor is any linear or non-linear signal conditioning used. Three-channel
systems only improve the wave launch to the listener by creating
a real (not virtual) center and by eliminating crosstalk between
the L and R speakers.
When you go to a live music concert, all the sound
appears to come from up front. Any ambience from the hall is not
localized or apparent as you listen to the performance. The same
cannot be said about most two-channel rear deployments. That is
one reason the AV folks keep pushing the number of rear channels
in AV systems up and up and up. I found that as long as you are
close enough to a proper LCR speaker deployment the instruments
sound as though they are in a concert hall. Move too far away
from the speakers in a large room and it sounds like you are listening
in the Academy of Music in Philadelphia. In such a listening environment,
rear-channel ambiance may be required.
Your editor now wants me to give you some lovingly
sculptured impressions of what the Trinaural processor does to
the best-sounding CDs I have in the house. I have spent 3700 words
getting to this place so you would think I need another 3700 now
to what the editor asks, but I do not. The reason for this is
that Trinaural will take almost all the music you have in your
house and make it better. Sometimes a little better and sometimes
a lot better. The improvement happens all the time. The more you
listen, the less you think “hi fi.”
Occasionally, magic moments happen. Today it happened
with the Chandos recording of Ottorino Respighi’s Concerto
Gregoriano. I had not heard this CD in a few years. What a wonderful
piece this is! Think The Lark Ascending but with a greater range
of emotions. For 33 minutes the violinist took me to a better
place. It is type of intense experience you often get live, not
through electronics. It is not the Trinaural processor that makes
the magic. It has to be the right piece, in the right performance,
at the right time, and you cannot always press replay and get
it to happen all over again. Now the adventure goes on and in
the next day or week you will find another disk you have not heard
in years takes you to the magic place. Badly produced recordings
unfortunately only sound a little less bad in Trinaural mode,
so the free lunch goes only so far.
In the end what Trinaural does is lets you get back
to being a music lover. The love of music is why I got into the
hobby and I suspect the same for most of you. The things we have
fought for all the years that drove us crazy with hi fi systems
are gone reduced or eliminated depending on the recording. With
Trinaural the speakers do not localize. Front-to-back depth is
a very real thing. Woodwinds take on a size and character that
one hears live but never in two channels. Brass glow with full
tone but not the harsh edge. Violin sections really sound like
sections, as you can hear the individual players, not some edgy
blob that oozes from the center of speakers. With the Trinaural
processor you can stop that part of your life that involved tweaking
components, speaker placement, etc., and get on with what you
really want to do – listen to music.
Soon you forget about all the cares and worries
of an audiophile. Some CDs sound good and some sound better than
good and a few still sound bad but you do not really care. Your
whole CD collection is set before you. Pick a CD, any CD. Forget
what you remember about the sound of the disc you picked. It is
not going to sound like that now. In most cases it is going to
sound good enough so you will listen to the music, not the sound.
Get the TV set out of the room. Put three full sized
speakers in the room in an arc around your listening chair. Go
to Circuit City and get the required electronics and then send
a check to Spread Spectrum Designs. Go forth and reconnect with
the magic that you have created in your collection of recorded
music. Be the music lover you were always meant to be.
