Bob Carver: Carving a Name For Himself Page 2
Carver: I've been attracted to a particular kind of soundstage that was described brilliantly by Harry Pearson years ago in The Absolute Sound. It was a soundstage that had front-to-back depth, extended in a large arc behind and around the speakers, and from time to time could actually produce sound images outside of the limits of the loudspeaker edges. It seemed, when I read that essay, that that was so absolutely correct. That's the way a soundstage should be. And at that time, the loudspeakers that I was familiar with were unable to develop a soundstage that even came close to doing that. I found that mostly I would hear a flat curtain of sound strung between two loudspeakers.
However, the loudspeakers where sound could also go out the back, the dipole speakers and the bipole speakers, made a soundstage that approximated the description that Harry Pearson had enunciated in his essay. I thought that was wonderful. The speakers that made the big soundstage were the Magnepans, the Bose 901s, the Dahlquist DQ10s; and if you think for a moment about the speakers that have lasted through the years relatively unchanged, the classic speakers all have had sound going out the back, bouncing off the rear wall to make a special use of room acoustics to help generate a large, three-dimensional soundstage. Those are the guys that have longevity. And I think it's because there's something particularly sumptuous, realistic, loveable, and believable about the soundstage that those speakers present. Naturally, I fell in love with the dipole speaker.
After reading Harry's essay, following his teachings, and trying to replicate that soundstage, I found that there were many components associated with that soundfield that had to work just right to make it believable deep into our ear-brain hearing system. One of the basic requirements was that the early arrivals be pristine and uncluttered. Which meant that early reflections, sidewall reflections, needed to be eliminated. Floor reflections, ceiling reflections, all of those things had to be eliminated to allow the first arrivals to arrive uncluttered so that the ear-brain can process both the amplitude and the timing cues associated with those wavefront arrivals.
The way to do that turned out to be by using a line-source driver. Like a ribbon. Because a ribbon does not radiate down, and it doesn't radiate up. And it doesn't radiate to the side very much. It radiates straight ahead and straight back. So there are no early-arrival reflections to muddle up the sound. Our ears hear the first arrivals, latch onto them, and build in our mind's eye a picture that has space and breadth and believability. Then the special use of room acoustics comes in about six milliseconds later. These are the longer-term reflections. If the loudspeakers are three feet from the back wall, by bouncing the sound off of the back wall, a multiplicity of arrivals occurs at our brain and our ears 6ms after the first arrivals and further enhances that lush, big soundstage that I wanted to achieve. That's why I chose a ribbon over anything else. And it works. It works. It can build an incredibly believable, beautiful, fun to listen to, spectacular soundstage with a planar ribbon loudspeaker.
Atkinson: But then you had to choose whether the line-source ribbon would be electromagnetically or electrostatically driven. Were you at all tempted toward electrostatics?
Carver: Well, when I first decided to do the loudspeaker, I actually built electrostatic ribbons and electromagnetic ribbons. There are two ways to make a diaphragm move: electrostatically or electromagnetically. The universe gives us those two choices. Maybe some day we'll have gravity-driven speakers. (Some theoretical work has to be done before we can harness the forces of gravity.) I built some electrostatic ribbons, and they worked fine. I also built the magnetic ribbons. They also worked fine. I ended up choosing the magnetic ribbon because at the time it seemed that it would be less problematical and less expensive to attain the equivalent results.
Atkinson: And did you discover along the way why magnetically driven ribbons are not very popular drive-units?
Carver: Yes I did! It's a lot harder than I thought to make a ribbon really work right. Ribbons don't look very good on paper. Even a perfect ribbon on paper has an erratic frequency response, due to the physics of the wavelaunch acoustics. A practical ribbon is even worse than that. It has resonances, buzzes, power-handling capacity problems, it has longevity problems. It's a very difficult design to get right. But amazingly enough, as rotten as a ribbon might look theoretically on paperand, by comparison, even an inexpensive Radio Shack tweeter looks magnificentwhen you put a ribbon in a living room and fire it off, some magic happens. There's a seamlessness, the voices are just pristine, just beautiful. That's the part that I worked on the hardest. Getting voices right.
When I started the design, the closest ribbon that would do it all was the Apogee ribbon. It's a wonderful ribbon, just a fabulous design. However, it has to be rolled off beginning at about 800Hz and its sensitivityexcuse me, its efficiencyis approximately one quarter to one eighth of the efficiency that I thought the marketplace would accept. That's proven to be true. When I first came out with the Amazing Loudspeaker, my ribbon was four times less efficient than it is now. And the marketplace wouldn't accept it. I mean, I sold Amazing Loudspeakers all right, but not nearly as many as I wanted to.
Even my own amplifiers had troubles driving them, to be honest. You know, if you'd wail on an original Amazing Loudspeaker with one of my amplifiers, which had enough power, sometimes it'd run out of thermal capacity and would go "Click!" The thermal switch would turn it off.
Atkinson: One difference between the Apogee treble ribbon and yours is that you tension the ribbon from all four sides, whereas the Apogee just has it lying floppily in the magnetic field. Does that give you sensitivity advantages?
Carver: Not sensitivity advantages. It does gives me a low-frequency performance advantage. To make the ribbon work right, I felt I wanted it to be perfectly seamless, perfectly crossoverless. I didn't want to put a crossover in the system. To do that required response down to 100Hz and to achieve 100Hz response requires a lot of air area. My ribbon design is therefore a large-area ribbon. As a matter of fact, there's as much area as a 12" driver.
The big problem has been that it's my first serious loudspeaker design. And it's been a lot more difficult than I thought. Once the drivers, the woofer system and the ribbon system, were working properly, just getting the tonal balance right was fraught with difficulties. That drives everybody crazy, not just me, because there is no right or wrong. It's not like a direct radiator system where you can measure it one meter on-axis and make it flat and that's the end of the discussion. Because of the acoustic radiation geometries associated with the ribbon wavelaunch, the tonal balance changes with distance. So you either have to have a way to compensate for that or you have to pick one and hope that your listener agrees with you.
Atkinson: Voicing a loudspeaker is not a trivial affair when it comes to a planar design.
Carver: It's easier with an electrostatic with a wide panel. It's much more difficult with a line source. Yet a line source is the more perfect transducer.
Atkinson: Once you had the concept of the ribbon going down to 100Hz, did you try other methods of matching it to a subwoofer before you went with your current one where you use high-Q bass drivers to compensate for the finite-baffle rolloff?
Carver: The first thing I tried was just a normal subwoofer. I couldn't integrate it.
Atkinson: Is that because of the mismatch between the omnidirectional radiation pattern of the subwoofer at crossover and the dipole one of the ribbon?
Carver: I honestly don't know. I just didn't care for it. I then tried dipole woofers; at the time I hadn't yet figured out how to make them have flat response without a lot of equalization. So I hooked up my panel to a normal low-Q woofer and EQed it till it was flat, and listened to it: it integrated seamlessly with the ribbon. I thought, "Am I imagining this?" So I hauled out a Quad, which of course is seamless in the crossover region at 100Hzit doesn't have a crossover. I measured the phase response and acoustic response of the Quad, and compared it with my speaker, paying particular attention in the crossover region: the two curves literally fell on top of each other. And I concluded from that experiment that the dipole woofer was a perfect match to the dipole ribbon, and was definitely the way to go.
Subsequently, I found that there were many advantages, the biggest one being that the response can be extended down to 20Hz and below without the severe efficiency penalty. The woofers are also cheap, because their magnets are tiny.
Atkinson: If you use purely electronic equalization to flatten the response, then you would have to use significantly expensive drive-units.
Carver: Expensive drive-units plus the electronics to equalize them.
Atkinson: From this weekend's experience, there would still seem to be some minor problems with the speaker which have been irritating you for some time.
Carver: I've been hacking away at the problems for two years. And they're down to residual problems right now.
Atkinson: I felt that with the low-frequency ribbon resonance suppressed, the sibilance emphasis seemed to become more audible. Is it fair to say that, as soon as you've solved one problem, that leaves another slightly more exposed?
Carver: I'll never be happy with my speaker design. There will always be problems to pursue and always problems to solve. That's why I love being an audio designer. I love it with a passion. If the problems went away, it wouldn't be fun.
In the case of the ribbon, the remaining problems involve controlling resonances. So what else is new? Controlling resonances is always a problem in loudspeaker design. A never-ending one. The time-domain performance of the ribbon emerges very naturally by its design, however, so I didn't have to work very hard for that.
Atkinson: As well as the smaller Silver Edition, I understand that one day you might introduce a twice-the-size version of the Platinum.
Carver: I'm working on the "Great Amazing" right now. Which will be a killer system. It'll be right up there in performance with the Silver Seven tube amplifier. I haven't prototyped it, but as we speak, the cabinets are being fabricated. When I get back home, I'll be assembling it and firing it off for the first time. I can hardly wait! And it'll undoubtedly teach me things during the development that I'll be able to apply to the current Amazings.