In Copper issues 60 and 61, I interviewed Princeton University professor Edgar Choueiri about his groundbreaking developments in three-dimensional audio. Choueiri took us through the mechanisms underlying the perception of three-dimension audio, and the practical issues involved with realizing it in real-world audio systems. He described the limiting factors, and the approach he took to invent his patented breakthrough technology which is finally able to make true three-dimensional audio a reality. Along the way he made some quite startling claims regarding what his technology is able to achieve.
So it was that at the end of July I went down to Princeton to visit his laboratory and hear some of this for myself. Edgar first took me through his Plasma Space Propulsion laboratory, where he has a number of experiments running, each one involving a vacuum chamber you could hold a meeting in. The challenges of space propulsion are relatively simple to understand. Basically, you have to chuck mass out of the back of your spaceship, and the harder you chuck it out, the more thrust you get. With a rocket, it is a chemical reaction which determines how fast the mass is chucked out, and in the grand scheme of things it tends to be nowhere near fast enough. But if you first ionize your mass to convert it to a plasma, and then send the resultant charged particles through a powerful particle accelerator, you can effectively chuck mass out of the back of your spaceship with a lot more force. Right now this is the best known option for transporting significant payloads over interplanetary distances.
In one corner of the Plasma Propulsion Laboratory, behind a large white screen, sits a white Tesla Model S. Now, many of you may know that Elon Musk owns not only Tesla, but also Space-X, and that he recently launched a Tesla sports car into orbit. So you’d be forgiven for imagining that this must be some bizarre prototype of a plasma-powered space-going sedan. But you’d be wrong … this is a test bed for various exotic in-car audio systems. It’s just that there was no way to get a Tesla Model S into Choueiri’s third-floor audio lab! But, as things turned out, this was a foretaste of things to come.
So, next we went upstairs to the 3D Audio Lab, which actually occupies about three times as much space as the Plasma Propulsion Lab. The core of it is the main demo room, which has a small anechoic chamber at one end, and an audio set-up at the other end. My original plan was to have my HRTF (head-related transfer function) measured, but unfortunately the measurement setup had been temporarily dismantled, so that wasn’t an option. We focused most of the rest of the afternoon on a demo of the various capabilities of the BACCH software.
Choueiri’s BACCH software is a very complicated system. Complicated to understand, and complicated to use. The user interface is very dense, in a Pro Audio kind of way … but at least I wasn’t driving it. You can buy the BACCH software directly from Choueiri’s company Theoretica Applied Physics, and it comes in various forms, all of them, it has to be said, pretty expensive. At the top of the pile is the ‘BACCH-SP’, a complete solution in hardware, including a DAC said to be of a cost-no-object no-compromise design. I believe it sells for a whopping $54k. Then there is a scaled-down version, without the DAC, called ‘BACCH-SP dio’, at just under $20k. Finally there is a software-only solution called ‘BACCH4MAC’, in three versions priced from $2k -$5k. We auditioned the top-end ‘Audiophile Edition’ of BACCH4MAC (the three versions differ only in the bells and whistles added on, and not in the core 3D audio software). Needless to say, BACCH4MAC is only available on the Mac platform.
At the core of the BACCH software is the need to calibrate the combination of the individual user, the audio system, and the listening room. BACCH is a highly personalized listening system. So the first thing that had to be done was to perform a calibration of my ears in his system. For this we used a pair of carefully calibrated in-ear microphones which I placed in my ear as I sat in the listening chair. The calibration required minimal active participation on my part, other than sitting there as various test tones were played. All I had to do was move left and right as instructed, watched by the baleful blue eye of a webcam which was tracking my head movements. The tones themselves comprised three simple tone sweeps, each repeated twice. The whole process took about 20 seconds after which it didn’t have to be repeated.
Once the calibration was done we could get on with the business of listening. But before that, we got to play some games, since I had the in-ear microphones installed. You see, at that point in time, my head was basically a binaural stereo microphone, personally matched to my own HRTF. The BACCH system could therefore make ultra high-quality recordings of anything I chose to listen to. So we made a brief recording of Edgar walking around the room, talking. First he stood behind each of the speakers in turn, and between them in the middle. Then he wandered off to one side, completely to my left, where he opened a door and stepped out into the corridor where his voice echoed grandly. Then he stepped back in and came right up to my head and whispered in my ear. Finally he walked behind me and round to the other side where he whispered in my right ear. At the end of the recording he saved the result in a file on the BACC4MAC’s playlist.
Playing back this recording normally, without engaging the BACCH software, was not at all impressive. The sound occupied a diffuse volume between the speakers with only minimal motion detectable. However, with BACCH engaged the difference was night and day. I heard Edgar Choueiri pretty much exactly as though he was in the room. He was physically located with great precision, and as he walked off to the left … a long way to the left of the left speaker … his sonic image moved flawlessly too. Where he stepped out into the corridor I heard the exact acoustics of somebody speaking in an echoey corridor through an open door ten feet to my left. His sonic image then walked back into the room, came up to me, and whispered into my left ear. Finally, it walked behind me and whispered into my right ear.
So let’s just recap. With two speakers playing in front of me, I could clearly hear voices whispering directly into both my ears, and a voice speaking to me from behind my back. I heard the acoustics of a long corridor through an open door ten feet to my left. If that isn’t impressive, I don’t know what is.
At this point I wanted to talk about actual playback of music, but there isn’t enough space here, and in any case I will be borrowing a BACCH4MAC system to try in my reference system, where I will be focusing solely, and at length, on real-world music playback. But I have to say that the BACCH system did have a notable, and beneficial impact on all of the music that I did listen to, although it stands to reason that Edgar will have populated his playlist with various tracks that showcase the system’s capabilities to their best advantage. At home I will be playing whatever I choose from my own extensive library.
If you read my previous Copper interviews with Edgar, you’ll know I was quite taken by an almost outrageous claim he made about headphone listening. In effect, he said that a BACCH system can emulate the effect of listening to a pair of speakers, while using headphones. In fact, he claimed that a cheap pair of headphones can emulate the most expensive pair of speakers in the world, and you wouldn’t be able to tell the difference. Them thar’s fightin’ words, and, to mix metaphors, I wanted to taste the beef for myself. So I asked for a demo of the BACCH system over headphones.
To do this, we required another calibration run. This time, we used a pair of open-backed headphones, which makes the demo a lot simpler. Although he had a pair of Stax SR009s to hand (which we also use at BitPerfect), these require a more elaborate setup with their own amplifier, so we used a pair of Sennheiser HD 650s, a nice piece of kit that can be had for about $400. I put the in-ear microphones back in, and placed the headphones over the top of them. Then, with the web-cam watching my every move, we went through the test tones once more. Fifteen seconds later the calibration job was done.
Edgar chose a binaurally-recorded music track – I don’t recall what it was – and told me he would play it first through the headphones, then through speakers, and finally through the headphones with the BACCH software emulating the speakers. Played through the headphones, the binaural track sounded to me like all binaural tracks tend to do … a bunch of sounds swarming around inside a sphere about 2 feet in diameter around my head. Then he played the track though the speakers, and because the headphones were open-backed I could hear them quite clearly. I listened carefully so that I would imprint a good mental image to compare with when he switched to the BACCH-emulation of the speakers.
But there was no third track forthcoming. When I turned in frustration to ask Edgar to play the emulation that I was so keen to hear, he just grinned at me. “Richard, that was the emulation!” To say that my jaw dropped would be an absolute rank understatement.
It was quite an impressive afternoon, and quite a busy one, with a large part of it involving the installation of BACCH4MAC by Edgar on my reference Mac Mini which I had brought along for that purpose. I also picked up a pair of his ‘BACCH-BM-Pro’ calibrated in-ear microphones so I could perform my own accurate calibrations in front of my reference system.
Since we finished quite late, and my wife was expecting to meet me back at our hotel a good two hours earlier, Edgar walked me out and offered to drive me the short distance back to Princeton. We walked up to his car which was, naturally enough, a Tesla Model S. I had never been in one … but I have heard a lot about it! I asked him which version of the Tesla Model S it was, since they all look just the same, and it turns out he has the rare and almost legendary P100D. This version comes with what is officially called “Ludicrous Mode”, and is highly appropriate for a guy who runs a Plasma Propulsion Laboratory. Basically … if you can believe this … nothing this side of a Formula 1 car is able to accelerate off the line as fast as a Tesla Model S in Ludicrous Mode. The linear acceleration is said to be about 1.5G.
“Have you ever experienced Ludicrous Mode?” asked Edgar.
“No”, I replied, wearing a hopeful grin.
“Would you like to?” he asked.
Damn right I would like to! So he went through the process of engaging “Ludicrous Mode”, during which the Model S’s colossal touch-screen console goes through a Star-Trek style warp drive display of driving through a star field.
We are lined up at one end of an ordinary parking lot in Princeton University, at 7pm on a Tuesday evening. Edgar looks at me and asks
“Are you sure you want to do this?”. I nod. And off we go…
Oh my giddy aunt! Holy mother of God!! He really has fitted one of his plasma drives to this thing!!! The acceleration is beyond description. How is this possible? It is ferocious. It is relentless. My body is sucked violently into the seatback and my head clamped to the headrest. I sense the onset of tunnel vision. And it all unfolds in the unearthly silence of a crypt, with the plush comfort of a Bentley. It’s the parking lot of a university, for Chrissakes, and we’re doing Warp 9! My heart rate leaps to a level from which it won’t fully recover until the next day. I grip the sumptuous creamy leather of the Tesla’s seats with a tenacity that has surely left a permanent imprint.
Yes, Edgar, I do want to do this!