Well it only took three years, but I finally got to experience a firsthand demo of the Smyth "SVS" multi-channel virtual surround product. Having been in contact with them about the subject ever since I first read the interview with Steve Smyth in WideScreen Review (September 2004), I was this week invited to visit their relatively nearby facility for a personal demonstration of their just-about-ready-for-primetime product.
They're exhibiting at the Head-Fi show in Ft. Lauderdale this May 3-4 if anybody's interested in having their own demo, and are about to enter a small prototype production run (to see if there are any production problems). Thanks to my patience and long-time interest I've been promised "serial number 0001" of their long-awaited product!
So... after all of this waiting, was it worth it? All I can say is REMARKABLE! With initial marketing expected to target the pro audio customer, I myself can't wait to get one (even though I'm hardly a pro audio person).
First, the hardware.
As it turns out the current incarnation of the SVS box is much smaller than its original prototypes and photographs showed it, or at least it's much smaller than I imagined it would be. Obviously it's been physically redesigned from the earlier versions. It's now maybe 8" wide, about 1 1/2" tall, and perhaps 6" deep.
There's an information display panel (perhaps 3" wide) on the left-front of the SVS box, along with an array of associated small LED's to the right of the display panel. As with many multi-channel processor boxes five of the LED's are arranged around a "listening position" to represent which channels are in use or are being tested, calibrated, etc. The rest of the LED's are for other status, channel, or function indications.
On the back of the SVS box are eight balanced 1/4" TRS analog connectors
typical of pro audio (and which can also be used with unbalanced connections) for discrete analog channel input from an external discrete analog channel source (where the multi-channel decoding from a digital source is also assumed to take place).
Also on the back are eight balanced 1/4" TRS analog connectors for output. In other words, more than one of these boxes can be "daisy chained", or this box can be placed in the "5.1/6.1/7.1 analog path" on the way to an analog input multi-channel preamp/amp, etc.
And finally, also on the back is a standard L/R-stereo analog output pair of RCA connectors to go to the headphone amp.
Smyth's retail product is also going to include an "entry level" Stax headphone (looks to me like the SR-202) along with an "entry level" class-A headphone amp (non-descript, may be Stax but doesn't look like Stax SRM-252II, but perfectly adequate). Even if you don't need this gear you'll get it, because the also included system-critical "head tracker" (gyro-sensing gizmo that sits atop the headphone band) works well with this headphone's plastic band (which is over the suede band that actually sits on your head).
The head-tracker faces "forward" and emits an IR signal that is picked up by a small IR receiver that is placed somewhere directly in front of you, most likely directly above or below the TV screen, computer monitor, or center speaker you're facing. The IR receiver can either stand on its own small base or a spring-clip bracket swings down so that it can be attached to the top of a flat-screen TV or computer monitor.
The lower part of the face of the IR receiver includes eight small photocell sensors arranged in a circle. On top of of the face is a horizontal row of six (or eight) small green LED's that light up to represent the left/right orientation of your head (presumably you're wearing the headphones with the head-tracker on top). If you're looking straight ahead, the center two LED's will be lit. As you turn slightly from left to right the corresponding off-center LED's illuminate to indicate your off-center angle and the other LED's turn off.
The IR receiver has a thin cord attached (maybe 10' long?) which plugs into a small "HT" jack for it on the front of the SVS box. There's also a second/alternative headphone output stereo mini-jack on the front of the SVS box. I don't know if both front and rear headphone outputs can be driven simultaneously.
Also on the front of the SVS box is what looks like a USB-port, into which the head-tracker itself can be plugged when not in use, for recharging (as there's a rechargeable battery in the gizmo).
The front of the SVS box also provides a card-reader slot, for a standard SD flash card. This card is used to store calibration data which thus "personalizes" the performance of the equipment. The RAM in the box can store up to 32 different "personalizations", which if desired can be written to or read from the SD flash card for external transportability to another SVS box.
And finally, on the front of the SVS box is an input "mini-stereo" connector jack for a crucial component of of the SVS product: the high-sensitivity binaural stereo calibration microphone pair (one for each ear, joined onto a single stereo cable). This binaural calibration microphone pair will also be part of the delivered product.
Also included in the product is a handheld remote containing buttons to completely control the SVS box, including the ability to select specific/ALL channels for listening, instantly selecting one of four different "personalizations" for A/B/C/D comparison, etc.
Now for the experience.
Only after a demonstration of how the equipment is to be used can you truly appreciate what this SVS box does. Remember, it is NOT a digital source decoder... as that is expected to occur external to the box in some piece of equipment containing a Dolby or DTS codec with discrete channel analog output. All that is fed to the SVS box is up to 8 discrete analog audio channels decoded from an originally multi-channel digital source (typically DD5.1 or DTS or any of the newer HD audio codecs). This discrete multi-channel analog input is what is then processed down by the magic software in the SVS box into 2-channel L/R stereo output intended for headphones.
Once the SVS box has been "calibrated/personalized" its output (from any source) through headphones is intended to "EXACTLY REPRODUCE the original listening environment" experience in which that exact same source would sound (down to loudness as well as tonal quality, along with 3D-sensation) when played through through true physical loudspeakers (and which were used to perform the calibration/personalization in the first place).
In fact, as amazing as it sounds I can attest to the fact that once calibrated/personalized you cannot distinguish the sound of the same physical loudspeakers which were part of the calibration process and the SVS-produced results through the Stax headphones. The same 3D-spatial front/rear/left/right cues that come from the physical speakers surrounding me in the demonstration room were precisely duplicated by the standard 2-driver L/R-stereo headphones... including when I turned my head!! Even loudness levels matched, as did tonal quality.
The "calibration/personalization" process starts off with the stereo binaural calibration microphones (each surrounded by a foam shell, just to get them to fit snugly) being placed in your ears. Then the SVS box goes through a two-part process.
First, a signal generator produces sound individually for each loudspeaker and the calibration microphones take readings as a full-frequency sweep is repeated eight times individually from each speaker in the room as you look straight ahead at the center speaker. The entire process is repeated a second time as you rotate your head to the left, looking at the left speaker, and then a third time as you rotate your head to the right looking at the right speaker. The result of this process is that the precise physical characteristics of your own personal ear canal structure are exactly quantified, as determined by the calibration microphones sitting inside your ears. Everyone's ear structure is slightly different, but the result of this "personalization" is that the SVS box now knows exactly how sound which enters your ears actually sounds... to you.
Second, the Stax headphones are now placed on your head and over the calibration microphones which are still in your ears. A different set of sound sweeps is now produced by the SVS box but this time played through the headphones. The calibration microphones now pick up that sound coming from the headphones and relay the quantification to the SVS box, which adjusts the output levels in realtime until the headphone levels exactly match what would have been the original loudspeaker levels. In other words, the headphones are "equalized" to the loudspeakers and other external auditory factors which made up this particular "listening environment", so that they are both (1) just as loud which is crucial, and (2) reproduce the 3D-characteristics of the environment.
In other words, the true purpose of the SVS box is to "duplicate exactly how you yourself hear sound, in a given listening environment"... meaning everything that went into that auditory experience such as speakers, sound reflections off of walls, floor and ceiling, the quality of the sound equipment and electronics, etc.
EVERYTHING that contributes to how you personally would hear and then describe what you heard in a given room or given listening environment, is precisely picked up by the binaural calibration microphones placed in your ears for the "personalization process" and absorbed by the SVS box and then processed into a "personalized profile" of that particular listening environment, based on how your particular ears heard things. This might just as well be described as a "quantified audiological analysis of your own personal ear canal" for that specific listening environment.
And again, I have to say it: IT WORKS! Once calibrated to my own personalization, I was able to easily take off the headphones (and tilt them down so that the IR emitter in the head-tracker tells the equipment to turn off the headphones and turn on the loudspeakers) and listen to the loudspeakers instead. Doing this A/B comparison repeatedly I swear you really cannot tell the difference. The Stax headphones, via the SVS processing, are duplicating exactly all of the tonal qualities and 3D-spatiality of the loudspeaker sound. THEY SOUND THE SAME!
Oh yes, forgot to mention that the SVS processor also has a built-in 16-band EQ controlled by the software, in order to duplicate the tonal characteristics of the loudspeaker and enviroment sound.
Ok, finally... how might this be used by an audio pro?
Well imagine a producer who listens to and mixes music in a studio, listening to recorded material through physical studio monitor loudspeakers in order to be "tweaked". Imagine now if all auditory characteristics of that precise studio listening environment AS IT IS ACTUALLY HEARD BY THAT PERSON could be "captured" digitally by an SVS box, and then put on an SD flash card and brought home where it could be placed into a second SVS box attached to headphones, where playing back that same recorded material through headphones would be AN EXACT AUDITORY DUPLICATE OF THE STUDIO LOUDSPEAKER EXPERIENCE! You wouldn't have to be in the studio in order to do the work, as the SVS box has exactly duplicated how the studio sounds... but through headphones rather than through loudspeakers.
Or, imagine taking the SVS box and calibration microphones to multiple environments (stadium, auditorium, arena, office, etc.) and creating a personalization profile for YOUR ears for THAT listening environment, and recording it. Then you can play back any source material through headphones, choosing that particular SVS profile, and relive the experience of listening in THAT listening environment... while in your bedroom!
Or... even more potentially fabulous... imagine taking an SVS setup to a high-end audio dealer or elite setup, with $200K worth of audio equipment and the world's best sounding speakers. Now you sit down and calibrate a profile for THAT remarkable sound environment, and take it home with you... a digitally reproducible analog image of exactly how that listening experience, on that equipment and in that room, sounded to you!!! After that, you can play music through headphones in your bedroom and have the sonic quality as if you were listening on that $200K never-achievable dream system with the dream loudspeakers! Of course you'll have to talk that dealer into letting you "steal the sound of his $200K system", but if you can then you're home free. $200K "home theater in a box"... for $3K.
32 memories in the box, but really an infinite number of quantified listening environment profiles on the SD flash card medium.
Anyway, I repeat: this was a REMARKABLE experience. You honestly have to hear it for yourself to believe it.
I did request one other test, to convince myself that my own personalized auditory quantification was different from somebody else's quantification (which is why shipping the product with a "standard factory-preset" set of values is worthless and meaningless). There is no such thing as "a best set of values". It's YOUR ears and how YOUR ear canal shapes and influences sound quality on its way to YOUR brain, and it's a particular listening environment (one among an infinite array of possiblities) that is being calibrated and calculated. There is no such thing as "best". Meaningless concept, with this SVS capability which simply duplicates ANY listening environment as it sounds to YOU.
So I asked that one of the other office staff go through the same calibration/personalization, which I would than be able to compare in an A/B test (using the "profile" buttons on the remote).
And, just as described to me in advance, the different auditory results were MORE THAN SUBTLE. My own "profile" sounded clear and precise, and very realistic... to me anyway, and very accurate. In contrast, the other fellow's "profile" sounded artificial and over-3D processed, moved back in my head instead of right in front of me, and very unnatural. Amazing.
Of course I was obviously attempting to listen to how sound appears to the other person based on that other person's ear canal, which is naturally not how my own ear canal is structured... which is why the "personalization" of the SVS box to my own physical ear canal is exactly that, allowing it to precisely reproduce for me how a particular listening environment sounds to me and me alone.
I'm sure that the other person would have a similar reaction if listening through my SVS profile... even when the exact same listening room environment was used for the calibration. Kind of like trying to wear someone else's eyeglasses, or using their bowling ball, or wearing their shoes (already broken-in by their feet), etc.
Final note: price point is still planned at $3K, for everything I've described above. Production is forseen for Summer, retail availability in the Fall.
Possible future hardware upgrades: HDMI 1.3a connector, to receive decoded discrete multi-channel source as PCM via HDMI rather than through 8 analog connectors.
Final comments... I plan to use my current Stax SRM-T1S headphone amp and Omega-I headphones. I also mentioned to Steve that I planned to also use my DBX 14/10 EQ in-line (between the output of the SVS box and the input of the SRM-T1S), to provide additional tonal control as I currently do with my Dolby Headphone Pioneer DIR-SE1000C setup.
He was concerned about possible phase shift resulting from the EQ processing which would of course affect the 3D-effect of the SVS box, but if the EQ were "linear" and "netural" then it could certainly could be used without ill effect.
Also, it was pointed out that the SVS box output could be RECORDED (e.g. to tape or CD) for external playback (e.g. in your car), where presumably it would produce a similar 3D-effect. I used to do that years ago with my Carver Sonic Hologram processor in the tape loop of my stereo system, when I recorded cassette tapes with this added processing. Playing back these tapes in my car sounded 3D!
Now I don't know if playing back a recording in my car which was made out of the SVS box would "only sound right to my ears" and to nobody else, or good and 3D to everybody, well I don't know. And would it again seemingly duplicate that original listening environment (or whatever SVS profile I had selected when doing the recording)? Certainly this will be an interesting experiment.