XMC / Dirac / multi-sub - how to prepare for Dirac.

[KeithL]

OK, if that's "simple", then HOW would you set them?

If you have one sub, you measure it, decide what you're hoping to achieve, and calculate what correction you need to get there - and there's going to be one set of corrections that do the job.

Likewise, if you have two subs, but you're correcting them individually, you do the same thing - once for each (which is our preferred way of doing it).

And, if you have two subs, but you don't want to adjust them individually, you can connect both of them to a splitter, feed then the same exact output, tell the XMC-1 they're "one sub", and still only have one output channel to correct.

HOWEVER, if you have two subs, and you want to "adjust them together", but retain the option of adjusting each one individually, suddenly you have an unlimited number of possible combinations and permutations of adjustments. (Let's say your two subs, together, show a 3 dB notch at 63 Hz; you could turn both of them up 3 dB at that frequency, or you could turn one up 6 dB and leave the other one alone, or you could turn one up 9 dB and one DOWN 3 dB. All of those options will produce the same overall summed result - but in different ways, and with different effects on things like the frequency response at other points in the room. And we haven't even mentioned the possibility of moving one or both of them.) Having two separate subs which you can adjust individually - but as part of a group - does give you a lot more flexibility, but the price you pay is having a lot more options and compromises that you have to figure out. Up that to three or more, and you have an amazing number of possible options and combinations - which, again, you have to navigate to reach a solution that works the way you want it to. Most of the "methods" I've seen basically work out to being ways to simplify all those choices by offering you a procedure to follow - adjust the first one for a certain result, then adjust the second one to modify that result, then adjust the third one to modify that result, or follow the procedure to pick starting points for all three, then adjust each according to a certain formula to get them collectively closer to what you want. Of course, none of them is considering ALL the options; each is just reducing the huge number of possibilities to a manageable number so you have a good chance of ending up someplace where you get a decent result. (And one method may favor the best sound at one listening position, while the other may favor the best response in the entire room, and a third may favor "tight bass" at the expense of absolutely flat response.)

Now I'll be really cynical and remind everyone that simply using EQ to reduce a peak or fill a notch doesn't correct for everything anyway. (If your room has a mode at 64 Hz that makes bass at that frequency sound boomy, you can correct the peak using EQ, but that won't entirely fix the fact that bass around that frequency sounds muddy because it bounces around the room for a long time. You may end up with the right amount of bass, but the only thing that will stop it sounding muddy at that frequency will be to reduce the reverberance of the room at that frequency - with a bass trap, or move the sub somewhere where it avoids exciting that room mode.) 

Mar 10, 2015 10:58:22 GMT -5 nickwin said:

Mar 6, 2015 13:08:40 GMT -5 KeithL said:

Errr.... no.

We believe that you should connect them both to the two sub outputs on the XMC-1, and let Dirac create individual Correction Filters for them (which is what it will do). Then, if they're positioned symmetrically, you should configure them as Stereo Subs; if they're not positioned symmetrically, or you have two dissimilar subs, you should configure them as Dual Mono Subs. In most situations, and in most rooms, doing it that way will work very well - and the individually corrected subs will add together perfectly with few negative interactions. (Doing it the other way is something we would only suggest if you're NOT one of the majority of our customers - who are quite happy with the results they get when they do it the way we recommended ).

Keith, could you explain the logic behind setting up multiple subs this way, as opposed to eqing the summed response? I'm a little confused, it seem to me that especially for people that want to set it and forget it, measuring and eqing the summed response would give a more accurate overall response. It seems like the subs would have to be perfectly time aligned first to get good results with the method you describe, other wise the real (summed) response could be way off. Im sure the set it and forget it crowd won't want to bother with that extra step.

[nickwin]

Let me take a step back here because maybe I'm making a false assumption or over simplifying. Is time allignment not a requirement to achieve any sort of consistency when two or more subs are EQed separately and played together? Your creating a filter that may or may not correlate to what the ear is actually getting post interaction vs creating filters based off what the ear would get at the MLP (measuring the summed response).

[bradford]

It may be Emotiva is operating from this school of thought:

Figure 1c. 1/4 W Placement
Although this isn't spelled out in the CEA recommendation, Dr. Toole references it in his book as a good solution for two subwoofers but suggests additional subwoofers may be needed. In my experience I've had excellent results placing two subs against the front wall at locations of 1/4 the room width. I've had even better results placing two additional subs in a similar manner against the back wall. This configuration can achieve nearly as good frequency response performance as the 4 Corner placement with nearly as much bass gain as well.

Reference> www.audioholics.com/subwoofer-setup/multiple-subwoofer-setup-calibration-1

[KeithL]

Yes, time alignment is important if you want to be able to calibrate each sub individually, and then have the sum of all of their responses add up to what you expect when you play them all together.

Your speakers are being time aligned as part of the calibration process. The distance setting is an "across the board" time alignment, then the individual Correction Filters apply frequency dependent time alignment to each individual speaker if necessary; this includes all the speakers, including the subs.

Your first statement is actually more or less accurate... although it's overstated. If your subs are not time aligned, then you will almost certainly experience all sorts of annoying cancellation effects which will make it much more difficult to convince them to produce a flat overall frequency response. And, in order for two or more subs to be time aligned, then they must either be the same distance from the listening position, or you must use time delay to correct for the difference so that they are "acoustically" the same distance. Note that, for two subs, all of the spots in the entire room that meet this requirement fall along a single straight line; and, for three subs, they are a single point. So, for three subs, there is precisely one spot in the room where they can all be properly time aligned (with any given set of delay settings). Oh, and don't forget that they also all have to be time aligned with your main speakers at the crossover frequency to avoid cancellation effects between them around the crossover point.

Hmmmmf... but we're not done yet. For any given size and shape of room there are going to be certain frequencies that have wavelengths that happen to be multiples of the dimensions of the room. At those frequencies, the room will tend to collect energy from any sub located at the wrong spots, which will then cause a null or a peak in the frequency response. Where these occur will depend on the location of each sub, the dimensions of the room, and the frequency you consider (different frequencies will have different room modes at different locations in the room). Your goal is to find locations for all your subs such that the combination of all these problems cancels out in such a way that you end up with a reasonably flat frequency response at your listening position. A given room will typically have several "big" room modes at specific frequencies, and specific spots that are especially bad for aggravating each one - and these are spots where you want to avoid placing any of your subs.

The general process is this (excluding changing the room itself):

1) Find locations for your subs that avoid aggravating any known room modes
2) Adjust the delay applied to each sub so they end up being time aligned at your listening position
3) Take some measurements (from your listening position) to make sure you haven't discovered any new and exciting room modes or cancellations
4) If you care about a wider listening area, then take more measurements
4) If you find anything nasty, go back to step #1 and try some different locations
5) Assuming we're OK so far, NOW use room correction to iron out any small inconsistencies in frequency response
 
Incidentally, to answer your question directly..... Yes, you are most likely to get the best results in general if your subs are time aligned, and there shouldn't be major interactions between multiple subs at your listening position if they are time aligned at your listening position. (As far as that listening position is concerned, those time aligned subs should all act like one big sub... so, if the filter for each one is working as it should, then the result should add more or less exactly.) To say it another way: As long as your subs are time aligned, you should get pretty much the same result whether you measure them all at once, or you measure them one at a time and simply add the results together. 

Now, so far we've been assuming that you have two or more IDENTICAL subs. If not, then all bets are off... because the phase shift of every speaker varies with frequency (we're talking about hundreds of degrees here - not just a few), and the way it varies is different for every speaker. So, if your two subs of different models are time aligned at your listening position at 20 Hz, it's a virtual certainty that they won't be even close at 32 Hz, and, if you get them time aligned at 32 Hz, then they won't match at 20 Hz or at 60 Hz. (To say that another way: There is no simple adjustment you can make that will allow you to time align subs with different phase characteristics.)

Interestingly, this is one area where Dirac can help you - since its mixed-mode filters can do quite a bit towards correcting this inconsistency. (So, if you let Dirac adjust them, those two subs will probably add together more cleanly than if you time aligned them with some program that only lets you do it for a single frequency, because Dirac will at least try to time align them at ALL frequencies..)
 
Incidentally, now that we've located our two or three subs at their perfect spots, all equidistant from your listening position (or corrected to seem that way), I'm going to ruin your day... because sound from each of those subs isn't just going straight to your listening position; a lot of it is bouncing off walls and arriving a bit later, so it can add and mix with the other sound in newer and even more interesting ways.

Now that you're really depressed, there is some GOOD news as well. First off, we aren't especially sensitive to very narrow cancellation notches, and all that reflected sound is steadily decreasing in level, so it's not likely to completely cancel anything out. Second off, at 20 Hz, the wavelength is about 50 feet, which tends to reduce the number of even multiples of a wavelength that you're likely to find in a normal room (for example, the ceiling isn't high enough to allow a "vertical resonance mode" to exist, and even a big room probably only has one or two in other directions. Third, since the wavelengths are so big, many of the weird effects are somewhat spread out. (You aren't likely to hear a huge difference when you move your head an inch to the left; it's more likely that you'd have to move it a few feet to notice a major change.)
 
So, in summary, yes you DO want all your subs to be time aligned at the listening position.... and Dirac Live will do a very good job of doing this for you (as long as you only have two subs). And, in general, as long as those subs are time aligned, the response of both WILL add together to give you pretty much the sum of both - as you would hope.

Is it possible that with a lot of trial and error, and a lot of measurements, you could do it better in some situations?
Maybe (but you can't time align your subs across a range of frequencies like Dirac can, which gives it a "home court advantage").

Is it possible that some of the methods we've seen discussed that use multiple subs and various measurements and adjustments performed in a certain order might do better?
Sure, it's quite possible.
Is it likely?
In a "normal situation" it's going to be hard to beat the job Dirac Live does, but, if you have a situation where Dirac doesn't do a great job, it couldn't hurt to try another method.

Is there a simple way to do a better job (or even an equal one)?
Errrrr.... no!

Mar 11, 2015 12:46:43 GMT -5 nickwin said:

Let me take a step back here because maybe I'm making a false assumption or over simplifying. Is time allignment not a requirement to achieve any sort of consistency when two or more subs are EQed separately and played together? Your creating a filter that may or may not correlate to what the ear is actually getting post interaction vs creating filters based off what the ear would get at the MLP (measuring the summed response).

[Mike Ronesia]

When we get the ability to store two Dirac profiles on the XMC-1 it will be much easier to dial the subs in with a little trial and error. For now what sounds the best to me in my room is using a splitter and setting my 2 subs up as mono on the XMC. It helps that they are identical and positioned between my fronts.



Cheers
Mark

[teaman]

Mar 11, 2015 19:45:07 GMT -5 Mike Ronesia said:

When we get the ability to store two Dirac profiles on the XMC-1 it will be much easier to dial the subs in with a little trial and error. For now what sounds the best to me in my room is using a splitter and setting my 2 subs up as mono on the XMC. It helps that they are identical and positioned between my fronts.



Cheers
Mark

Beautiful set up Mark

[Mike Ronesia]

Thanks, It's always sounded good, but it now sounds amazing with the XMC-1 and Dirac dialed in. The tv is 50" and I have a 65" on the way which should make it look a little more balanced.

Cheers
Mark

[nickwin]

Keith, maybe I am misinterpreting, but what I am gathering from all this is that perfect time alignment is difficult if not impossible to achieve across the entire sub frequency range. It sounds like Dirac does a better job at his than most ARC systems and maybe (I think probably) you can do better manually with extra software but thats not a sure thing. With this in mind, wouldn't we generally achieve better results at the MLP measuring and creating filters for the summed bass response? Like I said before the summed bass IS what our ears get in the end, after the subs interact with each other and with the room itself.

Mark Seaton suggests measuring the summed response to create an EQ filter:
"There are rare cases of large narrow peaks in one sub (particularly a rear sub) and not the other above 30Hz that I've applied 1-2 PEQ notches to 1 units and not the other, but unless you can demonstrate a good reason to do so, it's a better approach to EQ as a whole."
www.avsforum.com/forum/113-subwoofers-bass-transducers/1633729-six-6-seaton-submersive-install-exceptional-service-mark-seaton-3.html

Its how MiniDSP recommends doing it with there own DDRC-88A with Dirac:
"Note that all subs are connected to a single channel of the DDRC-88A. The behavior of low-frequency sound in a listening room is very complex, but this is a relatively straightforward way of obtaining a good result together with Dirac Live, as we will see in the following sections."
www.minidsp.com/applications/home-theater-tuning/multi-sub-optimization-with-the-ddrc-88a

And it is also how Audioholics recommends doing it:
"I tried to setup Audyssey using two subwoofer outputs of my AVP-A1HDCI where I grouped the front subs as "Sub1" and the side/back subs as "Sub2". This made the most sense since the subs that made up "Sub1" were equidistant from the listening area and the subs that represented "Sub2" were equidistant from the listening area. Theory was great but reality was terrible as can be seen in the graph below."
"When I first setup the Audioholics Showcase Home Theater room, Audyssey's Chief Tech Officer Chris Kyriakakis paid me a visit. We spent countless hours calibrating measuring and calibrating my system. It was at that time we realized the best way to apply EQ would be to apply a simultaneous single correction curve to all of the subs."
www.audioholics.com/subwoofer-setup/multiple-subwoofer-setup-calibration-1/multiple-subwoofer-setup-listening-and-adjusting

It just seems like in most cases this will provide the most accurate results, especially if you don't perfectly phase align each sub first in which case the post eq summed response could potentially be worse than the same subs with no EQ.

[socketman]

Nick I have a DSpeaker AntiMode 8033s II and it will do stereo subs but they too recommend using a y cable and EQ both subs together. This is how I have the subs in my bedroom setup and EQ'd .

[mcluvin]

Why can't this be an option added to the other bass management options currently available? It's obvious that there are a number of references to the method as nickwin pointed out a couple of posts above. Why would this be hard to add? I am looking to order an XMC-1 and luckily I have a mini-dsp to handle this but it would be nice to leave it out of the chain if the XMC-1 would just have the option available.

[ansat]

Mar 30, 2015 22:38:28 GMT -5 mcluvin said:

Why can't this be an option added to the other bass management options currently available? It's obvious that there are a number of references to the method as nickwin pointed out a couple of posts above. Why would this be hard to add? I am looking to order an XMC-1 and luckily I have a mini-dsp to handle this but it would be nice to leave it out of the chain if the XMC-1 would just have the option available.

Hey mcluvin. I believe the issue is an order of operations issue and would require a major rework. Since I have not liked any processors handling of multiple subwoofers, I am ok with needing an external controller to handle speaker arrays. In the end, I think to make this work with the XMC, Dirac would need to give us the ability to do this which would require a rework on their end.

Tony

[mcluvin]

Mar 30, 2015 22:46:06 GMT -5 ansat said:

Mar 30, 2015 22:38:28 GMT -5 mcluvin said:

Why can't this be an option added to the other bass management options currently available? It's obvious that there are a number of references to the method as nickwin pointed out a couple of posts above. Why would this be hard to add? I am looking to order an XMC-1 and luckily I have a mini-dsp to handle this but it would be nice to leave it out of the chain if the XMC-1 would just have the option available.

Hey mcluvin. I believe the issue is an order of operations issue and would require a major rework. Since I have not liked any processors handling of multiple subwoofers, I am ok with needing an external controller to handle speaker arrays. In the end, I think to make this work with the XMC, Dirac would need to give us the ability to do this which would require a rework on their end.

Tony

Thanks Tony.... If it would require a major rework then I would understand their reluctance in making it an option. Sounds like they just missed the boat in thinking through multiple sub integration but hopefully they will put this on a future road-map since it's obvious that what we are asking for is nothing out of the ordinary.

[pittsburghjoe]

Mar 31, 2015 8:05:21 GMT -5 mcluvin said:

Mar 30, 2015 22:46:06 GMT -5 ansat said:

Hey mcluvin. I believe the issue is an order of operations issue and would require a major rework. Since I have not liked any processors handling of multiple subwoofers, I am ok with needing an external controller to handle speaker arrays. In the end, I think to make this work with the XMC, Dirac would need to give us the ability to do this which would require a rework on their end.

Tony

Thanks Tony.... If it would require a major rework then I would understand their reluctance in making it an option. Sounds like they just missed the boat in thinking through multiple sub integration but hopefully they will put this on a future road-map since it's obvious that what we are asking for is nothing out of the ordinary.

Help with XMC-1, bass management and Subwoofers.

I am picking up my new XMC-1 Pre-pro and XPA-5 amp today at UPS.

I am replacing an old Pioneer Elite VSX-56TXi(no HDMI).
I have a 5.1 system and am more interested in great musical sound quality than theater effects.

My dilemma is which way is best set up the subwoofers.

My current setup is as follows:

(2) QUAD 63 Electrostatics as my Front Left and Right Speakers - 50-100 w (20-30 Vrms).
I currently use an Electrocompaniet Ampliwire IIA, Class A, 50w/channel to drive these speakers.

(3) Magnepan MGMC1s (modified with ribbon tweeters)for the 2 rear surrounds and center channel (4 ohm imp - 40-200 w @ 8 ohms)
I currently use the Pioneer receiver to drive these 3 speakers.

(I plan on using the XPA-5 to drive these five speakers.)

(2) JANIS W1 Subwoofers
I currently use two Janis Interphase 1 Systems designed by the subwoofer manufacturer to power the them.
They take the Left and Right Front channel signals, and with high and low pass filter sections w/18db per octave slopes and 100 Hz corner frequency split the signals, passing on the high end to their own amps.

QUESTION: Should I even bother with the subwoofer outputs on the XMC-1, just using the left and right front channels outputs as I do now on the Pioneer? Pioneer's room compensation allows me to just call the front speakers "Large".

Or is it worth experimenting with Emotiva's room compensation system by using the sub outs directly to their amps using the subwoofer amps for power and phase compensation only?

Although I know both ways would work, is their an advantage to using the Emotiva's sub outs?

Your thoughts would be greatly appreciated.

[socketman]

Honestly Joe if it was me , I would try both ways and see which works the best to your ears. Originally I had 2 subs one at each end of the room and I have minidsp and tried different things I read on many sites including this one and in the end the best setup was both subs opposite each other. What I am getting at is that every room is so different and we are all individuals I think you will be the best judge. I can tell you that once you get Dirac set up and subs dialed in your will be floored by how great it sounds.

Richard

[ansat]

Fixed the images on the first two pages.

[orangeLollies]

Hi guys, Just got myself an XMC-1 and before I get too deep into my XMC-1/Dirac/Multi-sub adventures, I just wanted to ask if the 'do I need outboard sub management' still a thing?
I upgraded to Dirac Full and have latest firmware 3.1
I have two SVS PB-2000's positioned mid-wall front and rear and had (before XMC-1) been running them as dual-mono using miniDSP.
There's a lot of threads regarding XMC-1 and bass management, but all seem to have gone quiet since 2015...
So, can the XMC-1 and Dirac-Full deal with dual subs correctly now? or is outboard processing still a potential requirement?

I'm vary familiar with REW, so if I have to go there... then I'm okay with that :-)

[geebo]

Aug 28, 2016 18:55:46 GMT -5 orangeLollies said:

Hi guys, Just got myself an XMC-1 and before I get too deep into my XMC-1/Dirac/Multi-sub adventures, I just wanted to ask if the 'do I need outboard sub management' still a thing?
I upgraded to Dirac Full and have latest firmware 3.1
I have two SVS PB-2000's positioned mid-wall front and rear and had (before XMC-1) been running them as dual-mono using miniDSP.
There's a lot of threads regarding XMC-1 and bass management, but all seem to have gone quiet since 2015...
So, can the XMC-1 and Dirac-Full deal with dual subs correctly now? or is outboard processing still a potential requirement?

I'm vary familiar with REW, so if I have to go there... then I'm okay with that :-)

I like to use a balanced MiniDSP with the 2-way advanced plugin and REW to time align my dual ULS-15 MK2 subs' and adjust the combined room response. I then run Dirac as though I have a single subwoofer.

[orangeLollies]

Also, I know we're told to setup speakers before running Dirac, does this include setting the xover point? ...I would have thought Dirac would determine the ideal xover.

Aug 28, 2016 19:10:32 GMT -5 geebo said:

I like to use a balanced MiniDSP with the 2-way advanced plugin and REW to time align my dual ULS-15 MK2 subs' and adjust the combined room response. I then run Dirac as though I have a single subwoofer.

Thanks geebo, are you running firmware v3.1 and Dirac full these days? ...If so, you still perfer outboard processing?

cheers

[mickseymour]

Aug 29, 2016 0:37:39 GMT -5 orangeLollies said:

Also, I know we're told to setup speakers before running Dirac, does this include setting the xover point? ...I would have thought Dirac would determine the ideal xover.

Dirac tests full range so you can set the XO point before or after the run. Whatever you set, the XMC-1 will use.

[geebo]

Aug 29, 2016 0:37:39 GMT -5 orangeLollies said:

Also, I know we're told to setup speakers before running Dirac, does this include setting the xover point? ...I would have thought Dirac would determine the ideal xover.

Aug 28, 2016 19:10:32 GMT -5 geebo said:

I like to use a balanced MiniDSP with the 2-way advanced plugin and REW to time align my dual ULS-15 MK2 subs' and adjust the combined room response. I then run Dirac as though I have a single subwoofer.

Thanks geebo, are you running firmware v3.1 and Dirac full these days? ...If so, you still perfer outboard processing?

cheers

The latest firmware, Dirac Full and a CS calibrated UMIK-1. I still prefer the MiniDSP in conjunction with Dirac.