Country of Origin
Reviewer: Srajan Ebaen
Financial interests: click here
Grimm Audio SB1 with Aurai Lieutenant speakers.
If you sublet part of your rental, you no longer foot the whole bill. Your sub tenant covers some of it. Replace bill with bandwidth. Now the same holds true for the labor division between sub and main speaker. In my ongoing adventures of the sort, I've come to a few conclusions which seem worthwhile sharing. In no particular sequence, here goes:
• timing is key. As modern subwoofers add more sophisticated room correction and complex EQ functions, they often incur lengthier latency from heavier DSP. Every millisecond of delay equals ~33cm of physical offset. For example, 8ms of digital latency with Arendal subs mean that if they set up right between our speakers, their sound will reach us late as though they sat nearly 3 meters farther away. Late bass is slow bass so we want to insure that latency is low enough to be offset by moving our sub closer to the seat by the necessary distance. Some subwoofers offer high-pass outputs for the main speakers and associate with them selectable digital time delay. That can put the sub on equal time regardless of latency or placement. It just involves A/D then D/A conversion of the speaker signal since time delay is done digitally.
Our Dynaudio S18 with its 2.5ms latency sits 83cm closer to the seat.
• isolation is audible. The physical strokes of big woofers can act like a minor or not so minor jackhammer on our floor. It depends on how well the subwoofer addresses resonance attenuation. Here force-cancelling driver arrays have an innate advantage. Still, effective isolation footers can make a difference. The lower and louder we play, the more mud and resonance follow when a subwoofer's physical action couples to the floor and migrates structurally. This is very much compounded by suspended flooring of upper storeys which don't have the benefit of being damped by foundation concrete. An easy test is to lose our shoes. We sit barefoot in our listening chair then play thumping bass. If our feet sense even faint pulses—they'll invariably limp behind the music beat—we've got cross coupling. Purpose-engineered isolation footers or platforms can probably still diminish it. The same applies to our electronics and main speakers. They too benefit from being cut off from floor-transmitted vibes which become exasperated by powerful subwoofers.
Relative to firing in our out, I preferred this orientation of very close proximity to brick'n'mortar walls. Sidefiring also duplicated the transverse woofers of the monitors to be preferable to front firing.
• orientation matters. Upfiring, front-firing, downfiring. Subs come in all flavors. Depending on how our speakers orient their own woofers, having that of our sub match can make a small but meaningful difference. That's because our hearing is very sensitive to transients as the first-wave attack. This registers differently if it hits us directly from a front-firing woofer or first reflects off the floor or a sidewall. Ideally we want no textural difference between our bass and the remainder of the bandwidth. If our LF are more striated and ripped than the midband, it's just as noticeable if perhaps less objectionable than if they're more bloomy and redolent. Just because our sub is front firing doesn't mean we must set it up that way. Likewise for a twin-woofer affair like KEF's KC62 whose drivers aim sideways. We can easily turn it around so one woofer looks at us, the other at the front wall. Here distance can matter again. If we run two sidefiring subwoofers for example, should they aim in or out? One orientation will have a quicker first reflection than the other. If we listen carefully, we might hear a difference to have a preference.
A rare RiPol subwoofer, this one from ModalAkustik in Germany.
• double crossing is virtuous. Subwoofers either tack on with a low-pass filter or dovetail more precisely by adding a high-pass filter to the mains. Single crossing hopes that the speaker's natural in-room roll-off is a perfect mirror image to the sub's roll-in; and that this handover is phase consistent. Double crossing insures both. In addition, removing low bass from speakers diminishes their excursions so minimizes associated distortion; and keeps their voice coils cooler to improve dynamic range. Double crossing with a low and high pass improves our speaker performance. Augmentation mode can't. This begs the question. How do we generate the necessary speaker high pass? Some subs include one but commit us to routing our purist signal through them. If that's unacceptable, we'll need either an external analog crossover à la JL Audio, spl, SublimeAcoustic or Wilson; an embedded digital alternative à la Bel Canto, Linn or TotalDAC; or an external digital crossover à la Behringer. That again could be sub par like routing speaker signal through inferior subwoofer electronics.
• where to cross over. Most subwoofers filter with 4th-order Linkwitz-Riley slopes. Those define by their -6dB point. The standard speaker spec lists at -3dB. To calculate our -6dB point, we multiply by 0.7 for ported, by 0.6 for sealed speakers. If our ported speaker is -3dB at 40Hz, it'll be approximately -6dB at 28Hz. If our sub's lowest filter setting is 50Hz, there's far too much overlap unless we also filter our mains with a matching high pass. If we double cross, we enjoy far more freedom on exactly where to cross. In my still limited experience and all other things being equal, a 80Hz transition seems to benefit rather more from using two subwoofers than an equivalent 40Hz filter.
Another RiPol sub, this with 2 x 15" woofers from sound|kaos of Switzerland.
• why to sub. If properly integrated, a subwoofer makes low-volume listening far more satisfying. It adds blackness to the color palette for more saturated colors and enhances the sense of recorded space quite profoundly. If the sub's playback quality over our chosen bandwidth is superior to that of our speakers—it ought to be or why bother—a sub improves our midband with less overlay of mud and bloom/boom. In the simplest terms, more bass clarity equals more vocal clarity. Of course a sub's volume is adjustable independent of the mains. It adapts to our room, taste and mood. A passive speaker's bass is fixed. We hope for the best but have no control over it. And to not belabor the obvious, in most cases a subwoofer will also add raw bass extension. On certain tracks with 'abnormal' bass, we'll likely hear things we've not heard before. But little music breaches 40Hz in significant ways. Even modern compacts tend to reach there without much attenuation. So this last reason on why we want to sub really does come in last at least in my book.
• why not to sub. In properly mixed live gigs, I've never heard bass pressurize the venue. Why should I want my playback bass to pressurize my room? With their acoustic claustrophobia, sealed headphones are bad enough with their unnatural ear pressure. I'd not pay good money to introduce that effect into my speaker system. What I would pay good money for—feel free to disagree—is to extend my reach without introducing any textural otherness to my already existing bandwidth. What I would pay good money for is to heighten the sense of a recorded acoustic that overlays my own with a stronger sense of different space at already low SPL. Likewise for improving my system's overall intelligibility by injecting less distortion from down below than my passive speakers are capable of. Ideally I'd want it all in a small form factor; in a color other than black; with all adjustments controllable by remote; then add purely acoustical room correction. For the latter trick I already committed to the above RiPol DSUB 15 for its directional radiation pattern with lateral nulls and reduced front-wall involvement. That this must get physically bigger than a KEF micro sub is the obvious price to pay. Avoiding the room-pressurization trap by acting as a velocity converter instead is the worthwhile gain.
• when not to sub. If our speakers already manage -3dB at 33Hz to possibly tickle 25Hz at -10dB, justification for a music sub relative to extra bandwidth is highly questionable. In most all cases, we won't have any success with just a low-pass filter that simply won't go low enough. Now it's down to us to double cross then decide whether a sub still improves the quality of our bass since adding quantity is academic. If the low bass of our speakers beds in just so with our room, a sub could well end up being an unnecessary complication and expense. Also, if our speakers are 3-ways or beyond, high-pass advantages shrink. Our midrange driver already sees no low bass. Bigger advantages come with two-ways whose mid/woofers run wide open on the bottom. It's restricting those from seeing low bass that makes a very audible difference. To get the very most from a subwoofer thus will want to mate it to more compact 2-ways. Not only doesn't that throw away bass which bigger floorstanders make on their own and which we already paid (dearly?) for. It also shrinks the mains which now go easier on eye and wallet and routinely image more spookily.
I'll add to this article as I learn more. In our high-end 2-channel space, subbing is poorly covered. Now it requires lots of personal trials and errors. If some of the above points help you sidestep some of your own, so much the better. If you're already well along this tangent, feel free to share some of your own findings so we can learn from your experience!
PS: more on where to cross. When Bel Canto's e1X DAC/preamp landed, I suddenly had 40-120Hz filter options in 10Hz increments. They simply applied 2nd-order Butterworth slopes to cross shallower than my usual 4th-order Linkwitz-Riley so produced more overlap. I also still had Grimm's SB1 on hand for one day of play with the e1X. This sorted whether—stereo imaging assured by two subs—crossing high had advantages. Perhaps because Alain Pratali's 4th-order bandpass woofer loading in our Aurai Lieutenant speakers is already very effective, blending it out at 120Hz even 80Hz didn't add anything desirable. It just fattened up the sound a bit which I didn't fancy. It also punched up dynamics. This soon struck me as a bit discontinuous like how many AMT 2-ways lead with their pleated tweeters. Their dynamic response is twitchier than that of their cone mid/woofers. Since more dynamics translate to getting temporarily louder on peaks, I've often found such AMT integrations to sound top heavy or forward. Blending in stereo subs well into the upper bass had a similar forward effect. It's why working my way back down to 40Hz really blended most seamlessly also in the dynamic domain. That's how I ended up with the most lucid quick overall response. On our single Dynaudio S18 and smaller mains upstairs, I ended up with 50Hz for the same reason. 120Hz also became directional. My ears knew that the sub sat off to the left. My original instinct of crossing low thus did find itself validated with our speakers.
The moral of this finding is to assess seamlessness not just in the textural realm of damping and subjective speed vs. substance; not just in the steady-state frequency domain for a linear tonal balance. Consider too the real-time domain of constantly changing dynamics. If the bass or midbass get dynamically forward, you've still got a bump even if it might seem impressive. Set any of it as you please of course. If true continuousness or virtual subwoofer invisibility is your ultimate aim, just don't forget to also track dynamic linearity across the sub's bandwidth. A good test to know what the bass dynamics of your recordings are (i.e. should be also with speakers) is to use superior headphones. They will eliminate your room's EQ on the sound and the added delay from long reflective path lengths which is nonlinear across the audible band.
Or as the old saw has it, one really shouldn't perceive a subwoofer as anything added. One should just notice when it turns off and something vital vanishes.
PS 2. On September 20th, Aavik announced their new C280/580 control amplifiers. Relevant to this page, there's an optional €4'000 crossover module with hi/lo-pass filters and independent gain controls for either filter leg. The press release added that the same circuitry may be used for surgical room compensation instead. This suggests an unusual degree of adjustment flexibility. What's more, the filter builds into the preamp to avoid extra cabling; and operates in the pure analog domain to create no latency. Voilà, a new option in the still very much underground movement of doing hifi 2.1 in proper double-crossing mode.
Would you like to download the current image?
Would you like to download the current image?