Srajan
Bach flower remedies. Like homeopathics not antibiotics, they’re deliberately subtle. Medics who call chiropractors fake doctors might call Bach flower remedies a New Age hoax. Meanwhile, BACCH power remedy promises to be potent and very real. It involves a nuclear physicist and very elaborate lines of code. At its most extreme, Dr. Edgar Choueiri’s crosstalk cancellation algorithm takes individualized measurements in the hot seat with calibrated in-ear microphones. That even tracks head movements. Like room correction, BACCH operates in the digital domain. It superimposes its response compensation on what our loudspeakers emit. According to speaker house Dutch & Dutch, their cardioid 8c monitor is the world’s first active DSP speaker to embed 3rd-party plugins. Today that’s the generic uBACCH algo. It embeds in each speaker’s Linux-based network card. This presently requires discrete Ethernet links to each 8c whose internal clock syncs to that of its mate via constant background pings over the LAN. The plugin currently requires Spotify or Roon. I’d hoped for UPnP through Audirvana but that remains TBA. Likewise for accessing the Choueiri plugin over AES/EBU. It’s planned but not here yet. Given that Spotify Premium tops out at 320kbps, that path is somewhat of a pearls-to-swine proposition. With the speakers’ upscale nearly €16K ask plus a €999 tariff for the optional BACCH code, the target audience isn’t low res. It’s SA/CD quality or better. That presently means Roon, period. I do neither that nor Spotify so re-upped a lapsed Premium subscription and signed up for a 10-day Roon trial which I’ve since cancelled. I much prefer Audirvana Studio.
What is BACCH? Simplified down–butchered?–to my grasp, it’s headphone crossfeed in reverse. Where crossfeed attempts to mimic speaker listening where each ear hears both channels, BACCH wants each ear to only hear its channel, not the other. It’s anti crossfeed so crosstalk cancellation. It simulates a virtual floor-to-ceiling wall that runs longitudinally from our front wall between the speakers before its free narrow end faces us. Now our right ear only hears the right speaker, the left only the left. Such binaural playback of standard music files over loudspeakers is said to reveal otherwise masked spatial data that are already encoded on our music but overlaid by crosstalk. Reality bites in that our algo isn’t a physical just virtual wall. To work best, it relies on very directional speakers for minimal room reflections; plus an optimized setup to diminish said reflections further still. Otherwise left-channel signal will reach our right ear via circuitous reflections and vice versa to undermine the extent of unmasking. Inherent in the concept too is lone-wolf appeal. BACCH only works in the hot seat and to a lesser extent, for a second listener sitting right behind the first. It’s decidedly no couch-wide affair for three or four listeners in a row but proper mancave stuff.
Astute readers already clicked. The Dutch & Dutch’s cardioid so heart-shaped dispersion makes it deliberately directional. The rear radiation of its 8-inch +100Hz/4th-order midrange exits through lateral slots behind the front baffle. Those slots cause broadband anti-phase cancellation towards the side walls. To that radiation pattern the tweeter’s 8-inch waveguide adds controlled directivity all the way up. When set up as intended close to the front wall, the rear-firing, sealed and paralleled 8” woofers cause nearly instantaneous reflections. That has them behave like virtual infinite baffle speakers aka in-walls. Their bass can only radiate forward into half space. Sound waves can’t wrap around a cabinet and emit rearwards into full space. The speakers’ internal DSP accounts for slightly late-arriving bass via a digital delay of the frontal drivers. We input front-wall distance, sidewall clearance and speaker-to-seat spread. DSP does the rest. Yet there’s more. Room EQ Wizard (REW)’s freeware is how Dutch & Dutch execute OEM room correction and voicing tweaks. A €499 remote log-in session puts one of their engineers in charge to program REW whilst we take 21 very specific measurements with the supplied microphone following our TeamViewer session prompts. Multiple tuning profiles like perhaps late-night, high SPL and normal listening can be accessed from the browser interface or app. The software controls of Spotify and Roon lock to the 8c’s virtualized browser/app volume control. Over the LAN that triggers 57-bit attenuation math inside each speaker just prior to D/A conversion. Sans BACCH, that’s also how it works for AES/EBU.
What does BACCH do? The published promise is of much enhanced spatial envelopment and wider clearer more specific soundstaging. Unlike binaural recordings intended specifically for headfi, BACCH does not rely on specialized albums, just normal stereo not mono productions. Like converting PCM to DSD on the fly to experience an entire local/virtual CD library as SACD, BACCH converts regular to enhanced stereo on the fly. So it’s also instantly defeatable by software slider. Hello, convenient comparisons for yummy couch-potato mash. Where’s the salt? We’ll get to that. To set our table, first some non-BACCH 8c commentary.
With 250-watt class D on the frontal array and another 500 switching bass watts, the 8c’s inherent gestalt to my ears was slightly overdamped so texturally dry. Combine superb driver control with dual-concentric type time coherence, very low distortion and zero passive xover confusion and you won’t be surprised that the 8c was a soundstaging provocateur of the first order. Set up in my typical free space as signed off by the designer, imaging was superbly layered and focused. With connective tissue or smear between images eradicated, the sound was impressively clean, clear, linear, extended and controlled. Without any room EQ above 180Hz and below it only notches (no boost) to address my room’s main 35Hz/70Hz modes which the 8c’s potent bass rode like giant breakers, I had full-power 25Hz reach and no hint of lean, bright or forward tonality as some still expect from ‘studio’ monitors.
My one nit was the earlier gestalt. Whilst I can’t be 100% sure, I’m reasonably certain that it was full bandwidth class D. I own a pair of OEM Pascal monos from Gold Note. They’re only allowed on sub 100Hz subwoofer duty; for exactly the same reasons. That dry overdamped control is ace for bass but to my ears less so as frequencies ascend. Naturally, my usual 250-watt 2.5MHz DC-coupled class AB monos wouldn’t fit inside a Dutch & Dutch. Running on ultra-low Ω, noise-free, cool and compact switching power is essential for an active monitor speaker concept. Saying so is no diss on class D. I loved the admittedly very costly Merrill and Alberto Guerra Designs GaN implementations. My kind of sound originally informed by single-driver widebanders + SET later transposed to hybrid dipoles + transistors + active cardioid sub simply wants a bit more textural elasticity and flow. It’s a rather narrow distinction and academic to most. To arrive at my ideal mix took me 20+ years of uninterrupted very involved experiments. For better than 95% of all listeners, the one-stop Dutch & Dutch with its premium transducers short-circuits all associated time and costs in a far simpler far more predictable solution; which is fully customizable in digital EQ. To get all of its endless boxes ticked plus my few extras… that’s a very tall order. Meanwhile, an 8c can grow with our evolving tastes whilst we studiously avoid the hamster wheel of buy ‘n’ sell. Just input a different voicing profile. That’s radically more strategic and effective than faffing about with cables, pucks and cones.
How about toggling BACCH on and off? Forget Spotify. Whilst it did something, I found it far too marginal. Once I had Roon for full to high resolution either local or off the cloud, that changed. The potency of said change still varied from album to album, though. So did my reactions. On some, BACCH created decidedly wetter more reflective venue acoustics. That played direct antidote to my full-range class D misgivings by injecting more connective tissue and textural pliancy. As always happens when there’s more inter-note stuff because a more active ambient field of recorded reflections lights up space, focus softened a bit, images grew bloomier and depth more pronounced. Yet the primary effect was an increase of recorded reverberance. On more complex and busier mixes, I could find the associated softer focus and decreased separation less advantageous. On yet other material, the effect included a slightly phase-y minorly swimmy aspect which I didn’t fancy. Occasionally I also netted slightly wider staging. I simply never heard any semblance of forward envelopment. No matter what, the virtual stage began well behind the speakers as it always does.
Then it seemingly removed the front wall which separates my listening room from Ivette’s studio as it always does. I didn’t get super-wide stereo which on the right music I already enjoy from proper but basic speaker setup. However, BACCH could produce extra depth plus enhanced audibility of space. Being emptiness and silence, space can’t be heard per se but through recorded reflections. More reflections equal more audible space. That’s just like the moon having no luminance of its own but borrowed sunlight. We see the most moon when our earth throws no shadow on it. In those terms, BACCH could be from a full to a sliver of a sickle moon. Rarely was it a new moon so utterly inaudible.
Overall I’d estimate my hit rate at 50:50. The algo wasn’t equally effective on all music. Where it was clearly effective, I didn’t always prefer it. Here it should be instructive to know that I don’t fancy surround sound or headfi crossfeed, period. It’s equally important to acknowledge that BACCH only unmasks what’s recorded; and what that is we won’t know until we compare. Are there cubits of gold hidden in them thar hills of our local/cloud libraries? Is it a mere pittance; or pyrite fool’s gold? We won’t know beforehand. It’s why a 10-day cooling-off period should be mandatory to let prospective buyers assess whether the €1K uBACCH license is for them. Being software code, it ought to be easy to embed a ‘self-destruct’ code should a trial period not formalize into a purchase. As to video plus BACCH, the 8c’s basic DSP incurs 2.3ms of latency, its fully optimized phase linearization an extra 32ms whilst the algo adds another 10ms.
No matter which side of this equation you come down on, one thing can’t be disputed. Even in its generic ‘u’ form, BACCH can definitely work. How much, on how much of our files and whether you’ll deem it worth the cost are different matters. What had me most at hello was the mere fact that it could be so noticeable. Logic demands that Dr. Choueiri has to be correct when he asserts that recorded 2-channel music already contains spatial cues we’ve never heard before. They must just be allowed to rise to the surface by removing existing obstacles. That classic speaker listening would be said obstacle was news to me. It won’t prevent me from continuing to enjoy it though. I belong to the genus that finds head-fi, conventional speaker fi and live music enjoyable each on its own most distinctive merit. Now I simply learned that on certain recordings—which ones other than those I tried I obviously can’t know—there’s more recorded ambiance to be retrieved. It’s a good thing that my hybrid open baffles already excel at extra spaciousness if admittedly also of my own room not purely the recorded acoustics. But I can live with that and be very happy. And before you ask, no, I’m not adding a uBACCH license for my music iMac into my existing speakers. I already enjoy superb depth layering, focus and soundstage width. There’s little room for greed to take root and want more. If on the other hand, I retired with a pair of 8c to sell off my mountain of separates and passive boxes and subs, I would spring for the plugin. Even at a 50% hit rate I’d consider it a worthwhile bonus. I’ve been BACCH’d; though not fully remedied from classic stereo 2.1.
Further information: Dutch & Dutch | Bacch | Full-length 6moons review
