It's 1976. Philips introduce active home-audio speakers with motional feedback. Here's an advert from those heady days. Note the tag line simply years ahead. How many - years that is? How 'bout a whopping 39? Let's call it forty for a nice round number. 40 it is.

Now it is 2015. Kii Audio use HighEnd Munich to roll out their active Three speaker in an off-site mastering studio launch party. I already made noise about these as my top find of the show both for John Darko's site in my KIH #22 column; and in these pages with sighting bits here and there.


There's more even if details remain sketchy and conjecture. Look at the very basic block diagram at left. You'll surely spot the voltage and current loops between driver and what's called motion control. Infinity and today's Genesis would probably call it a servo. Velodyne subwoofers exploit an accelerometer on their cones to compare woofer behaviour to input signal for on-the-fly auto correction.


With Kii Three author/auteur Bruno Putzeys having famously worked for Dutch giant Philips who made him develop what became UcD only to fail to recognize its promise and future impact, thus leading the man to what became Hypex and subsequently Mola-Mola... it's convenient to make a genetic connection between their motional feedback from 1976 and Bruno's motion control of today. With technological progress in the intervening years particularly on the DSP side of things, one of course expects rather a new chapter on the basic idea, particularly since nCore already includes the driver in its feedback loop.


But isn't it tantalizing to think that a concept deemed perhaps too radical some forty years ago—or perhaps its implementation wasn't as yet sufficiently mature—would reappear in a very compact 6-driver active monitor today whose stated goal is a cardioid dispersion pattern whereby to minimize typical room interactions and focus more of the acoustic energy at the listener? The drawing at left shows the polar response of a cardioid microphone. The goal is plain. For this rather than an omni type, one wants maximized sensitivity towards the main sound source in front; and very diminished sensitivity to the back to avoid picking up background noise.


Replace the microphone with a speaker and the advantage of such a dispersion pattern is just as plain: most the energy goes toward the listener, far less behind the speaker and some less to the sides. Interactions with the front wall cause undesirable reflections particularly in the omni-radiated bass. This also loads up the room's front corners with high-pressure bass energy for boom, mud and HF suppression. If the cardioid 'microphone-in-reverse' appeal is crystal, the how isn't. How does one practically achieve such controlled dispersion?


In Munich and for this techno peasant, Bruno had hinted at time delay by way of phase shift as the means toward his DSP-driven heart-shaped sound propagation control. It would seem that a contemporary form of Philips' original motional feedback is part of this hi-tech picture as well. It would certainly go a ways toward explaining how he milks 20Hz reach from four small woofers without driving them into deep distortion. Thanks to contributors Marja & Henk for unearthing the above vintage advert. The plot thickens...

PS: Here's more from the man himself as harvested from the thread on Actives Hören. "I should explain that my primary reason for making a cardioid is an attractive speaker that works well anywhere and doesn’t require users to find the magic spot in their room that sounds good. It's user friendliness I’m after. The second reason is an idea that allows me to keep it going up to 700Hz and then, using ordinary methods, maintain comparable dispersion up to about 5kHz. So it’s not just a cardioid bass. Dispersion is constant over most of the band, much more so than on its predecessor. On top of that, the acoustic centre of the whole affair—except the tweeter of course—is a point about 5cm in front of the midrange. I was quite chuffed when I figured that one out. So that’s why we’ve got this fancy 'Active Wave Focusing' name instead of just calling it a cardioid bass box...


"I use current drive wherever it helps reduce distortion. The midrange for instance is entirely current driven. Some confusion is caused by the term "Entzerrung" which could mean reduction of nonlinear distortion but is more usually taken to mean equalisation (i.e. correction of linear distortion). A DSP can easily do EQ but removing nonlinear distortion is quite hard. It's possible but none (including Klippel's excellent work) is far enough along to be usable for a design like the Three. I need to have the LF extension on the web page corrected. The flat portion of the response is indeed ±0.1dB but the low corner is at -3dB as one would expect. There is one XLR that is switchable between AES and analogue. RJ45 connectors provide loop-through in case the speakers are used wired. In wireless mode there is only the 230V to plug in. Two rotary switches allow rudimentary EQ to match placement. The only classical crossover point in the filter is between mid and high at 2kHz. Below that there is quite a game going on with overlapping frequency bands to shape the wave front. This is why we say that the filter is not comparable to a classical crossover filter. The concept of crossover frequency is not defined in a design like this. All the DSP hardware and amps were designed new specifically for the Three. The processor is one of the bigger Sharcs.

"My listening room is 6 x 9 x 3m. It's difficult to say how big a room they'll handle since that really depends on how loud you want to listen at what distance. But I wouldn't be afraid to demo in a room twice the size of mine. Actually, I think it'll be more interesting to try really small rooms to hear the diminished impact of the room modes. I haven't had the chance of doing that. The next model up (the Two) is a floorstanding version with the same drivers but more woofers going all the way down. Higher models have not yet been decided on. I'd love to make a scaled-up version that you could actually use as a decent PA. The Four would be a scaled-down version, like 15x30x30cm and not more than half the price of the Three. This will be a serious challenge. Below 40Hz the Three is essentially an omni. At that point it follows the same physics as any speaker when it comes to SPL and cone movement. For the same SPL at 20Hz, it needs to move exactly as much air (in litres per second) as any other speaker including the Devialet Phantom. The same material replayed over the Three would flap the cones as much as it does the Phantoms' (divided by the ratio of cone area, I've got more cone area so somewhat lower X). What the Devialet sales lot apparently haven't worked out is that you barely hear 20Hz in isolation. If you play electronic music with a lot of sinusoidal bass, it won't sound loud and it will make the cones flap apparently aimlessly. If you play orchestral music, the concert bass drum can create an enormous sonic wallop with relatively little subsonic content. Of course you need to reproduce that subsonic content to make it sound realistic but it's not a lot of energy, really. At some point the wavelength becomes so long that when you place the speaker near a wall, the dipole component cancels and only the omni component remains. So pushing the cardio transition deeper is only meaningful if you are actually planning to use the speakers free standing. I think the main programme for the Two is controlling vertical directivity and extending LF SPL."