Music in the Round #7
To recap my March 2004 column: Many early researchers and developers of stereo, particularly Blumlein and Klipsch, asserted that true stereo reproduction requires a minimum of three speakers in order to ensure a truly stereo—in a word, solid—representation of the central performers. In HT parlance, most of us have stereo systems with a "phantom" center channel. Although this has served us well, it only approximates what can be achieved. The creation of a center channel to fill and solidify the central image and permit the widening of the soundstage without bloating the instruments or creating "a hole in the middle" is something even traditionalists should appreciate. I'm old enough to recall being impressed by Paul Klipsch's three-speaker demos in the late 1950s, though my aural memory doesn't extend back that far!
Simply summing the left and right channels would: 1) create phase interactions with the left and right speakers; 2) reduce separation; and, as a result, 3) demand that the left and right speakers be placed farther apart. The result would be an unnaturally larger sound image—the proverbial 10'-wide violin. One can also synthesize surround channels, but this requires a paradigm shift in listening that many stereo traditionalists are not prepared to make. The ideal solution for a stereo system is a device that derives a center signal from a two-channel source, does not require tweaking or surround channels, and does not corrupt the directional cues.
Enter the Trinaural processor
One of the highlights (and award-winners) at the 2004 CES was Spread Spectrum Technologies' Trinaural processor ($1500). The demo, with three VMPS speakers and Ampzilla2000 power amps, was outstanding for its spatial and instrumental integrity. While all the devices bore the stamp of longtime audio innovator James Bongiorno, it was the Trinaural processor that was the system's most salient component. In this all-analog box, according to Bongiorno, "the stereo composite signals are algebraically revectorized into three front channels." Doing so, he says, eliminates "cross-coupled error signals," which, though necessary for two-channel reproduction, create spurious directional cues, especially when one is not sitting rigidly in the central sweet spot.
I found the Trinaural's instructions cursory, a bit confusing, and totally lacking in technical explanations. (They're available online, where they're still referred to as "preliminary.") I was also a bit put off by Bongiorno's bold warning that one should not compare Trinaural listening to two-channel stereo, and that one needs to re-learn how to listen. He warns that, on hearing the Trinaural for the first time, many will find the center channel too emphasized, too "hot." He then admonishes listeners not to turn down the center channel, and to listen for an extended period before judging the performance. To help ensure that this adaptation has enough time to occur, the Trinaural processor has no simple on/off switch.
But can it really take two weeks before one's ear-brain linkage has unlearned its old listening habits and adapted to the new? As a card-carrying neurobiologist, this bugged me. I know that the ear-brain has a remarkable ability to adapt, and that, with sufficient exposure, it can readily learn to accept as correct and normal almost any rearrangement of signal parameters. Besides, don't some of us listen to real music in the real world all the time? Nor are any connection diagrams provided, or adequate explanation of the fact that the Trinaural's Bypass function won't let you use the main input signals without Trinaural processing. I did eventually figure it all out, and Bongiorno does encourage users to call if they have problems.
Ready or not, here it comes...
I fed the main stereo L/R outputs of the Sony SCD-XA9000ES SACD player into the Spread Spectrum Technologies Trinaural processor, and used the latter's Bypass inputs for the Sony's L/R/C/subwoofer signals. Next, I reconfigured my system to replace the Paradigm CC-470 center-channel speaker with a Paradigm Reference Studio/20, while the L/R speakers were Paradigm Reference Studio/60s. I'd been meaning to make this change for a while, and Bongiorno's forceful urging to use a better center-channel speaker did the trick. I also listened for extended periods with three Paradigm Studio 60s across the front. Balancing levels with the Trinaural's built-in signal generator is easy, and does not require a sound-level meter.
In my very first listening session, I did feel that there was too much center level (I checked it with a meter), but I didn't find it too bright. The "centerness" was emphasized with orchestral recordings, but right from the get-go, solo voices with accompaniment were enticingly realistic in localization and solidity, with no changes in harmonic balance.
So I let it ride. For the next several months I did all of my non-multichannel listening, CD and radio, via the Trinaural Processor. My initial feeling of "centerness" faded quickly, despite the fact that I interspersed these listening sessions with multichannel sessions and weekly returns to my main two-channel system.
The Trinaural was consistently satisfying with a wide variety of two-channel music. I felt less constrained to sit in my "serious listening" position of the center of the couch. The performers were always firmly located in the middle of the front of the room, regardless of where I moved, although my perspective on the soundstage changed depending on my proximity to the left or right speaker. This, of course, is exactly what would have happened with real musicians in the room, assuming I had the audacity to walk around as they performed. Even more interesting was that there was less variation in sound and image in the vertical plane as well. We don't walk around while listening critically, but many of us fidget or slouch while seated; with the Trinaural, that was just fine (footnote 1).
Another improvement wrought by the Trinaural processor, probably as a corollary of its centered coherence, was a subjective improvement in the bass. Sure, there were now three sets of bass drivers, not two, but the improvement was not due to any change in bass levels that I could measure. Deep bass seemed more firm and extended. I suspect Jim Bongiorno would say that this is the result of the elimination of those nasty "cross-coupled error signals," as most of the bass seemed to come from the new center channel. The Trinaural processor also has a sub-80Hz subwoofer output and a selectable high-pass filter for the main speakers. But with or without using the sub, Trinaural-processed bass was significantly more satisfying than the unprocessed variety.
So why Bongiorno's dire warnings about needing to "relearn" how to listen? Switching between unprocessed sound and Trinaural processed sound (doable via the dual L/R outputs of the Sony XA9000ES or a pair of Y-connectors) affords the opportunity to determine that the processor neither introduces any audible distortion nor changes harmonic textures. However, switching back and forth was often disconcerting; I found myself listening for different things each time. The whole soundstage seemed wider without Trinaural, but so did instruments and voices.
Footnote 1: It also occurred to me that Trinaural processing might be just dandy for deriving side-channel signals from the front and rear signals of multichannel recordings. While the ear-brain may be able to synthesize a coherent phantom center channel from two channels, it is markedly more difficult to perceive the phantom sides required by most multichannel recordings. Imagine a pair of Trinaurals, each deriving an "algebraically revectorized" side channel, helping to create a seamless 360-degree sound space!—Kalman Rubinson