Multi Subs- How to time align with different distances

you must have something like minidsp

I'm also interested in this as I have a MiniDSP.
Is there a trick one can use to do the adjustment?
Thanks

Same here. I think it is something you have to adjust and remeasure with each change till the fr is where you want it. Im no expert by any means though

For my multi-sub arrangemet, 2 of my 3 are same R/L and front from MLP, while 3rd is different located on LH wall.

Since I don't have mini dsp, but a Denon 4520 CI with discrete 2 sub, my plan was put both RH/LH main subs on sub1, and 3rd on sub 2, and use the distance to time align them.
(actually after I use PEQ for the basic flat freq to run Audyssey XT32)

You'll need room measurements and the ability to adjust delay. This would be the phase dial on some plate amps. Distance setting in the avr. Or delay with a minidsp or other dsp device.

Unless your installation is outdoors don't worry about it. For one thing you can only perfectly time align the direct waves in one spot in the room. More important, room reflections will effectively defeat any attempts at time aligning the direct waves.

I disagree.

Just as it's clearly obvious how important the alignment of a sub system, to the remaining speakers in a bass managed/crossover system are, it's similarly important to align each sub source within that group. It works.

Yes, I agree it better to work outdoors, without the infinitely complex interference patterns and modal influences of a typical listening room. But it is beneficial to individually adjust the time/signal alignment of each sub, if response smoothing is the goal. The discussion over multiple listening positions is an ancillary topic in my opinion, but even then, it's still worthwhile.

Whether one uses the ubiquitous Behringer DCX, the miniDSP, or up to the pro gear like the Xilica offerings, a multi-sub system is the path to bass nirvana, and the tool that helps facilitate that is loudspeaker processing DSP. Also, whatever model one uses, the Geddes approach, the Harman, Welti/DeVantier method, your own home brew technique, it largely doesn't matter as typically practicality intervenes and often influence sub placement. It is what it is.

With measurement software, be mindful of as much acoustic summation occurs as broadband as possible (with weighted concern in the upper octaves), EQ globally, as a group. Tend to the overall characteristic blending between the mains (wrt level and time), and this sub group at the crossover freq(s). Fine tune subsequent response with small adjustments in both time alignment/delay, and to a lesser extent adjust low pass filtration of each sub.

Rinse repeat, rinse repeat .... each step interacts with the final response somewhat differently. Establish the baseline (max acoustic summation) and experiment. It's effective.

EQ globally, time align individually.

The FR follows the time domain, the time domain is where the gold is.

---

After that time domain is addressed, examine the room decay characteristics (bass trapping) .... as tidying up thusly allows for some superb bass detail.

I love seeing all these highly capable sytems, but actual optimization efforts is a "system multiplier" the likes of which many don't fully understand. I'm not referring to Bill Fitzmaurice, he likely knows more than I'll ever know, we simply disagree on whether the aforementioned is worthwhile.

It's all about the room, ... the room's all about the time domain.


Thanks


(wrt the "system multiplier", a good acoustic environment/optimization is akin to a kick-ass offensive line, as that way both your run game and passing game is improved)

And that's my point. Unless the room is anechoic every boundary represents another source. Multiply three subs by the number of boundaries present and trying to time align is a nightmare of epic proportion. But the good news is that for the same reasons that subs aren't directionally locatable and group delays of even 30ms aren't audible neither is the lack of time align in the sub frequencies of any major consequence.

By optimizing for smoothest frequency response.

If you are using the Welti/Devatier method to attempt to reduce FR differences across several seats, you do not use individual delays, as to cancel modes, all the subs need the same coherent signal. If you are trying to use SFM principles, you do adjust delays and EQ on each individual sub, and the process is incredibly iterative, which is why JBL made the BassEQ module.

JSS

^ ^ ^

Good point, and perhaps I mis-spoke.


I posted;


I intended, that whatever you set out to do, whatever model you may intend to emulate, oftentimes the actual locations used are dictated by influences other than the purely technical ones that you originally set out with, ie., practicality/aesthetics, etc.

So personally, I recomend as much dedicated control as possible. Individual circuits for each location, thus facilitating differing adjustments for each sub. Regarding the Welti-DeVantier positioning, ... if one's subs are all mid-wall, or some up front and some in the back, due to a variety of residential factors, it's entirely possible you may be significantly closer to one group or another. From there, if individual control is called SFM, then that's cool too. It's audio, individuals have been time/signal aligning multiple sources for years. I've never read SFM with any amount of real attentiveness. A re-read is in order before I continue off ill advisedly. I get it, it seems they all want to stake claim of their specific "process", everything has to have a name.


I've not really paid close attention to other multi-sub techniques, and actual actual methodologies. For HT, it seems there's so few actual documented in use implementations, with measurments, data, and solid results.


The approach I posted, was forged out of both discussions w/pros and audio enthusiasts, and my interest, background, and reasearch in pro audio. In the pro world, the blending of multiple sources, time alignment of delay towers, cardioid sub arrays and signal manipulation, wavefront shaping and main lobe broadening so that you project as much LF into the primary LP and away from the performance mics, all is well vetted and firmly entrenched in use. There exists a great deal of well documented examples in whitepapers, AES submitted work, and the like.


I'm not truly well versed in much of it, and I've not used all the techiques either, not even close. But I have tried to read and experiment with as much as I can. However taking all that I'd learned in that regard, and adding forum readings and discussions w/MSeaton, I recomend the approach I layed out above.


Control and manipulate as many elements as possible. I mean it only makes sense, as long as the signal processing is transparent, the more control the better. We strive for a smooth, fully time/phase coherent response. When the output range our subs cover, benefits from such efforts to aid in evening out the response, both measurably and subjectively, then clearly this becomes worthwhile.


I just thought of an analogous example;
The entire process seems akin to internet/LAN packet switching, ie., the original signal is fractionalized, spread across a paralleled network that's optimized for robustness, then fully re-correlated or re-constituted for end use within the primary listening area.
It's funny because I just thought of this, my son and I were recently watching/discussing an ARPANET piece, and Kleinrock, the dude that wrote the book, and built the first packet swiching gear connecting UCLA, and Stanford, I believe.


I sat down at a Harman SFM demo while attending Cedia last year, and was perplexed at some of the things the presenter said to me, what a joke. He was dumbing it down for mainstream consumption, and I listened to the entire thing. I had some very specific questions but he insisted not going too deep. I understand, but at that time, it was just two of us listening to his presentation. My only guess was he was well versed in the demo, but nothing more. I stayed polite and didn't push.

Kevin Voecks, Harman big shot, was in the booth too, but he was wrapped up smoozing, etc. I wanted someone to address my question but I decided to move on. The presenter (I'll have to review my notes, .. Todd something, not Welti) sugested I go accross the street to the big Harman demo at the neighboring Westin. He tuned the $400k total, $270k audio, SFM multi-sub, M. Levinson powered, JBL Synthesis 7.1 system. This guy explained to me the best area of couple seats, as that's were he tuned the room from.

So later that day, I went to check out the demo. I got one of the two prefferred seats, quite excited and unfortunately, it was all for naught .... it was terribly underwhelming. I posted details shortly after the show in some AVS discussions. It was not what I thought would be representative of a Harman $400k tour de force, especially in context of the Harman brass overseeing the entire affair, .. oh well.

Here's my post wrt the SFM demo; http://www.avsforum.com/t/1413241/room-correction-vs-digital-signal-processors/30#post_22388617

I mean come on, a decent room, JBL beryilium drivered / Levinson powered, four, corner loaded 18s, ... good mid-bass but it ended there. I don't have a clue ... right upstairs in a private room was a superb system set up by Dan Francis, and the gentlemen from XTX. It, had it all. The Harman/JBL demo, ... junk. It was a shame too, I'm a fan of JBL and all the R&D they put up. I wanted to try to take in the demo again, but they made you get a ticket at the booth (diff building than the demo), I never came back, too much other good sound to be experienced....like Steinway Lyngdorf. That's a multi-paragrapgh post by itself,...best loudness, dynamics, transient attack, I've ever experienced ...anywhere, period. All set up by Lyngdorf, I got two front row center demos, and discussion w/Lyngdorf. All that a short walk from the JBL demo,...no contest. The JBL room, ain't nobody got time for that. Even better, the remix.



All the best

That's not necessarily an absolute. You can use delays creatively to shift the peaks and nulls of multiple subs around reducing the amount of EQ needed at the end to flatten it out the rest of the way. Adding a delay to one of the subs is not really any different than moving it further away.

More an algorithm than a process. SFM uses measurements from multiple seats (typically 8, but up to 12) to compute the best combination of 3 parameters (level, delay/phase, single band PEQ) for each subwoofer, with the goal being to get the maximum possible consistency from seat to seat. It's been in use since April 2011 as part of ARCOS room correction on JBL Synthesis systems.

You can re-read the SFM paper here.

The algorithm itself is capable of helping with placement as well. For example: if you have 2 subs and 6 locations where you can put them in the room, you can measure one of the subs in all 6 locations and SFM will tell you which of 15 possible placement combinations will result in the greatest seat to seat consistency. The algorithm can also be set to solve for other results, like the placement combination that yields the flattest response at the main listening position or placement combination that yields the maximum output. This can be done for up to 4 subwoofer and up to 8 placement locations.

Unfortunately, Harman only seems interested in using it for calibrating/tuning purposes (i.e., subwoofers are already in place, ARCOS certified calibrator shows up and uses SFM to make the best of it). Shame it doesn't come in a separate box, since it would be a good pre-equalization step for any room correction system (Audyssey, manual PEQ, etc). Once the user gets the response as similar as possible across the seating area, it's a matter of using global EQ to tame peaks and dips (treat the multiple subs as a single output). Todd Packer

So where does one purchase one of these SFM devises?

Thanks for the added clarification.

As I now understand it, ARCOS is the piece of gear one would buy (this is what was demoed at CEDIA) , SFM is the predictive sim for multiple seats, multiple potential sub locations, developed by Welti/DeVantier. Is this correct Sanjay? And HATS, was the result, over time evolving into ARCOS?

The predictive sim software, as cool as it is, isn't the realm of this discussion. That'd be for Yates, Erskine, Grimani, et al, in the design stage of a megabuck build, ... typically not the area of primary interest of this DIY forum. I know the Yates (perhaps the finest of the high end designers, in my opinion) group, has their own proprietary computer algorithm software for such work. But I did stumble across an article in FOH online, stating Yates was one of four people attending the initial ARCOS, master level training course, a couple years ago.

Like any Synthesis branded gear, a Harman/JBL luxury dealer would be a source.

But, as I suggested in my first post, just get a processor, ... and roll your own. In this DIY forum, sub locations are way more often determined by practicality, not by sims. SFM, ARCOS, ... I get it, and it is very cool and a good read.

---

My contention, back OT, is individual sub manipulation/control is beneficial, resulting in tangible benefits, both empirical and subjective. Individuals have been manipulating the time alignment of multiple, spaced sources, long before any of this had acronyms, etc.

A bit more clarification if you're willing. The only piece of gear one would buy is one of the SDEC units. The rest is all software.

Harman Audio Test System (HATS) is the unified measurement system used across all Harman facilities around the world. Prior to that, their speaker builders were using one measurement suite and room correction designers were using something else. Now they're all on the same page and can easily exchange data.

Adaptive Room Correction & Optimization System (ARCOS) is a user friendly interface/skin for HATS, similar to how the early versions of Windows used to ride on top of DOS to make it easier to use. Otherwise, ARCOS is HATS.

ARCOS calibration has three steps. The first is Sound Field Management (SFM), which tweaks multiple subs so that their combined interaction minimizes spatial variation across the seating area. The intent isn't to improve frequency response, just get it as similar as possible from seat to seat. Once that is done, subwoofers are treated as a single sub for the rest of the calibration. The second step is equalization, which conforms the frequency response of all speakers and "subwoofer" to the target curve, prioritizing the most audible problems (peaks & dips) first.

The final ARCOS step is subwoofer to speaker blending. HATS has a feature called AutoCurveSum that Harman's speaker designers use when designing crossovers between various drivers. Someone at Harman must have figured that if it could optimize tweeter to woofer blend on a loudspeaker, it might also be able to optimize speaker to subwoofer blend during system calibration. The blending is done using all pass filters, which allow all frequencies to pass, affecting only the phase (which is how they smoothen out the blend). Any unused filters left over from the second step (equalization) can also be used for the sub/speaker blending if the allpass filters alone aren't able to make the blend smooth enough. Understood, just wanted to be thorough in explaining SFM, considering the post I was replying to. Is anyone arguing against the manipulation of individual subs? The only thing folks sometimes don't consider is that listeners will be hearing the combined interaction of the subwoofers. I've known people that painstaking dialed in each of their subs (even though they would never ever be listening to them that way), only to find that effort evaporate the moment all the subs were playing together.

Figuring out how to manipulate individual subs in the service of their interaction isn't easy. With that in mind, SFM isn't merely an acronym for the type of subwoofer manipulation that people had been doing all along, otherwise it wouldn't have earned a patent (prior art).

Ok, so I have denon avr3311 with Audessy multEQ xt, and a mini dsp 8x8 capable of calculating biquad filters sent directly from REW. Two differently located IB manifolds on seperate amp channels. Do I independently measure, EQ with the mini dsp, then use audessy with the subs combined? I never have enough time to try it several different ways unfortunately. Thanks

True, very true. I didn't mean to say that you must not mess with delays, I like FOH's approach, have as much flexibility as you can. Many have found impressive results just by tweaking delays in addition to multiple sub positions (Unicron-WMD's 8 Dayton Reference 15" HF build is a perfect example). The original theory was to use multisubs to cancel or not excite the most modes, which relies on a signal coherence to each sub. Another thing to remember for everyone else reading: Welti/Devantier theory does not say much about the flatness of the frequency response attained, only that the difference between seats will be minimized. This is very important. Toole addresses this in his book, and states that while very uniform seat to seat, some subwoofer placements result in frequency responses that are impossible to EQ without tons of headroom.

FOH,

You and I agree on this. Flexibility is a good thing.

OP,

Experiment, Experiment, Experiment......I hope you find a combination that works. It can get very frustrating, and sometimes you need to remember that you may not necessarily notice that 50Hz suckout you are chasing down...unless you measured for it.

JSS

You can use the C input on the DCX to measure the time alignment. It auto corrects but it's always good to check what it's doing. There are quite a few other pro DSPs that have this function

So is this Harmon ARCOS system available to consumers, and if so, is it affordable?

The SDEC 4500 retails for over $15,000. Ouch.

Indeed, they make no mention (zero) of improving frequency response, just minimizing seat to seat variation. You'll see them stick to this approach in their paper on subwoofer locations (placement optimization) and the paper on SFM (electronic optimization).

They're starting from a premise that equalization will be used at some point, whether manual (PEQ) or automatic (Audyssey, etc). Their assumption is understandable, considering how ubiquitous room correction systems are on current receivers and pre-pros,

This is why disclaimers like your's above are important: folks should understand that this is but one approach, not the approach. Lot's of people aren't concerned with more than one seat and/or do not want to insert electronic equalization in the signal path. For those listeners, the Geddes method would be a better fit for their usage than Welti/Devantier.

Been available since April 2011, but only through Synthesis installers/dealers, and set up though an ARCOS certified calibrator (the calibration itself cost several thousand dollars). That's one of the reasons some folks were looking forward to the (now cancelled) pre-pro from Lexicon, since it would have delivered ARCOS (including SFM) in a consumer product. Unfortunately, at the current cost of SDEC units, you're talking ADA/Trinnov or Meridian price range.

I use a "modified" approach, as follows:

Preconditions:
First off, I don't use any automatic software; everything is tuned manually.
Secondly I don't EQ any signals above roughly 300hz (I find that sounds horrible SQ-wise).
Thirdly, the sweet-spot has to sound flawless; other seats are optional by comparison.
Fourthly, I have more than 4 subs, which I find significantly helps get SPL/headroom up and initial seat variance low to start with.
Fifthly, it's a dedicated theater/man-cave, so I can place anything anywhere I want, short of impacting walk space. So that helps.

Step 1) The first thing I do is place all the subs and basstraps where I would like to see them; everything in-phase and flat and widest XO bandpass.
Step 2) I then listen in the central sweet-spot

If that is good, I start sitting in the other seats.
If those are good or at least decent, then great, I'm done.

Otherwise... (which is often the case)
3) I start by measuring the response in the sweetspot, just to see where I'm at.
4) I then I start playing with the phase and timing of the rear subs; and rechecking seats.
5) Afterwhich EQ'ing is done with REW, averaged across all the seats. (EQ'ing MUST be measurement based PERIOD.)

Usually by this point the problem is solved and I'm done. More often than not, I have to lower the low-pass bandwidth in the rears because of localization too.

If that doesn't work I repeat steps 3 and 5 for the front subs (step 6 to 8).
If all that fails, only then do I start physically moving subs out of their ideal spots (after reseting all parameters); starting with the rear and side subs.
Rise and repeat.

I have yet to encounter a room where at this point I'm not getting good sound, in most, if not all seats; even in an L-shaped room. (Knock on wood).

Usually all of this can be achieved with 6 or less main EQ points, and under 3 EQ points per sub (because they are sealed).

If you only have 1 or 2 subs, then it can be a bit of a nightmare; otherwise I don't find it very difficult or all that time consuming.

Via this method I was able to obtain this; and I didn't even attempt to correct the secondary bumps (just primaries).

This was with 9 subs in 7 different locations, all still in their ideal mirror-image locations, in an nasty L-shaped room, with only 6 main EQ points in total (and no sealed-box corrections applied); and only two delay shifts (front and back) and no phase corrections.

Not too shaby!!! I spent more time lifting the subs into their intial positions than configuring.

Exactly.

Part of the understanding gets back to the measurement and realities at low frequencies. Determining arrival time gets fuzzy at low frequencies. High frequency inclusion makes impulse peaks nice and sharp, but an 80Hz low pass gives you a big mound. The reality is that within say +/- 1/4 wavelength/period, fine adjustment simply changes the summation more than anything else.

You want to use both distance measurements and measured group delay to help guide delay adjustments, but the goal is really for constructive and/or smooth combination through the frequency range more than strict impulse alignment. A few overlaid measurements of individual subwoofers at the same drive level and their combined response tell the tale very quickly. If you are prioritizing a full set of 12-16 seats, yes, SMF/ARCOS would be much more expedient. If your primary concern is one row and wanting good results in the other rows, you can do surprisingly well empirically with some invested time and patience.