Optimizing subwoofers and integration with mains: multi sub optimizer - Page 13 - AVS Forum | Home Theater Discussions And Reviews
Forum Jump: 
 162Likes
Reply
 
Thread Tools
post #361 of 777 Old 11-29-2016, 09:38 AM
** Man of Leisure **
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 16,232
Mentioned: 205 Post(s)
Tagged: 1 Thread(s)
Quoted: 7562 Post(s)
Liked: 4224
Quote:
Originally Posted by markus767 View Post
Use acoustic timing signal on lets say L (channel 1). Measure all satellites. Then route R to your subs within the 88BM and measure R (now your subs). Keep L as your acoustic timing reference.
Why measure all satellites? Did I miss that part of the Instructions?
AustinJerry is online now  
Sponsored Links
Advertisement
 
post #362 of 777 Old 11-29-2016, 09:50 AM
AVS Forum Special Member
 
andyc56's Avatar
 
Join Date: Feb 2013
Posts: 1,172
Mentioned: 15 Post(s)
Tagged: 0 Thread(s)
Quoted: 17 Post(s)
Liked: 391
Quote:
Originally Posted by AustinJerry View Post
My current set-up consists of a Marantz 8802A feeding a MiniDSP 88BM. On the sub channel I have four subs being integrated using a MiniDSP 2x4. All connections are balanced. Bass management is performed in the 88BM, so all speakers in the 8802A are set to Large. The 2x4 provides the time alignment delays, but has no PEQ. Room correction is, of course, Dirac Live.
Hi Jerry,

I noticed you mentioned that the 2x4 will use no PEQ. This might be a good opportunity to talk about the usage of PEQ in MSO.

Since the 2x4 has only 6 biquads per individual channel, you could say that it's far less powerful than the IIR portion of Dirac in that regard, and you'd be right. But there's more to the story. As soon as different PEQs are used for each sub, their behavior when considered together becomes different from the way PEQs have been traditionally used. Used in this way, they affect the relative amplitude and phase of the individual pressure contributions of the subs at a given listening position. They can also alter the seat-to-seat variation in the bass response, improving it if done correctly via software like MSO.

Having said that, and given that Markus is suggesting a 1-foot cube in which to measure the responses, I'm not sure if there would be an advantage to using individual PEQ for each sub in this application, as this is a pretty tight cluster of listening positions. You could experiment with this using the configurations feature, by having one configuration without individual PEQ and one with.

Quote:
3. Since no bass management is involved, I will use REW HDMI4 (LFE) to generate the test tones. Is that correct? Normally I measure at 90dB. Does that matter?
If I'm understanding correctly, it looks like Markus is recommending that you do use bass management, and that you should also measure the satellites (maybe the center?) in addition to the subs. I think your original plan was to measure only the subs. It's worth mentioning that these two scenarios involve two different choices for MSO configuration type. For optimizing subs only, you'd choose the Config, Add New Sub-Only Configuration menu selection. For optimizing subs and their integration with the satellites, you'd choose Config, Add New Configuration.

Level considerations for the measurements are pretty much the same as you've been discussing in the big REW thread: keep the data out of the noise on the one hand, and out of the distortion and compression on the other.
andyc56 is offline  
post #363 of 777 Old 11-29-2016, 10:08 AM
** Man of Leisure **
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 16,232
Mentioned: 205 Post(s)
Tagged: 1 Thread(s)
Quoted: 7562 Post(s)
Liked: 4224
Quote:
Originally Posted by andyc56 View Post
Hi Jerry,

I noticed you mentioned that the 2x4 will use no PEQ. This might be a good opportunity to talk about the usage of PEQ in MSO.

Since the 2x4 has only 6 biquads per individual channel, you could say that it's far less powerful than the IIR portion of Dirac in that regard, and you'd be right. But there's more to the story. As soon as different PEQs are used for each sub, their behavior when considered together becomes different from the way PEQs have been traditionally used. Used in this way, they affect the relative amplitude and phase of the individual pressure contributions of the subs at a given listening position. They can also alter the seat-to-seat variation in the bass response, improving it if done correctly via software like MSO.

Having said that, and given that Markus is suggesting a 1-foot cube in which to measure the responses, I'm not sure if there would be an advantage to using individual PEQ for each sub in this application, as this is a pretty tight cluster of listening positions. You could experiment with this using the configurations feature, by having one configuration without individual PEQ and one with.



If I'm understanding correctly, it looks like Markus is recommending that you do use bass management, and that you should also measure the satellites (maybe the center?) in addition to the subs. I think your original plan was to measure only the subs. It's worth mentioning that these two scenarios involve two different choices for MSO configuration type. For optimizing subs only, you'd choose the Config, Add New Sub-Only Configuration menu selection. For optimizing subs and their integration with the satellites, you'd choose Config, Add New Configuration.

Level considerations for the measurements are pretty much the same as you've been discussing in the big REW thread: keep the data out of the noise on the one hand, and out of the distortion and compression on the other.
Well, the measurements are finished. I measured the subs full-range with no bass management. Not sure why, but I measured each satellite speaker in the same way--full range, no bass management, Dirac off.

Now that I am on to the next step, loading the measurements into MSO and experimenting with the options, I am sure I will be back with more questions. BTW, I didn't say the 2x4 will not use PEQ, I said I have not used the 2x4 in the current configuration. The 2x4 is used only to consolidate the four sub signals into the 88BM, and to apply time alignment among the four subs.
AustinJerry is online now  
 
post #364 of 777 Old 11-29-2016, 10:14 AM
** Man of Leisure **
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 16,232
Mentioned: 205 Post(s)
Tagged: 1 Thread(s)
Quoted: 7562 Post(s)
Liked: 4224
FWIW, here are the "All Subs" baseline measurements for 6 positions:

AustinJerry is online now  
post #365 of 777 Old 11-29-2016, 10:17 AM
AVS Forum Special Member
 
markus767's Avatar
 
Join Date: Apr 2009
Posts: 9,873
Mentioned: 111 Post(s)
Tagged: 0 Thread(s)
Quoted: 4189 Post(s)
Liked: 1577
Quote:
Originally Posted by AustinJerry View Post
Why measure all satellites? Did I miss that part of the Instructions?
It would give you more options using MSO. Furthermore it's always good to have more data. There's always the chance of the occasional blown tweeter

Markus

"In science, contrary evidence causes one to question a theory. In religion, contrary evidence causes one to question the evidence." - Floyd Toole
markus767 is offline  
post #366 of 777 Old 11-29-2016, 10:19 AM
** Man of Leisure **
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 16,232
Mentioned: 205 Post(s)
Tagged: 1 Thread(s)
Quoted: 7562 Post(s)
Liked: 4224
And FWIW, here is the current sub response that MSO needs to improve upon, measured at the MLP with Dirac and BM on:

AustinJerry is online now  
post #367 of 777 Old 11-29-2016, 10:19 AM
AVS Forum Special Member
 
markus767's Avatar
 
Join Date: Apr 2009
Posts: 9,873
Mentioned: 111 Post(s)
Tagged: 0 Thread(s)
Quoted: 4189 Post(s)
Liked: 1577
Quote:
Originally Posted by andyc56 View Post
Having said that, and given that Markus is suggesting a 1-foot cube in which to measure the responses, I'm not sure if there would be an advantage to using individual PEQ for each sub in this application, as this is a pretty tight cluster of listening positions.
He could do a second set of wider spaced mic positions. What spacing would you recommend?

Markus

"In science, contrary evidence causes one to question a theory. In religion, contrary evidence causes one to question the evidence." - Floyd Toole
markus767 is offline  
post #368 of 777 Old 11-29-2016, 10:23 AM
AVS Forum Special Member
 
markus767's Avatar
 
Join Date: Apr 2009
Posts: 9,873
Mentioned: 111 Post(s)
Tagged: 0 Thread(s)
Quoted: 4189 Post(s)
Liked: 1577
Quote:
Originally Posted by AustinJerry View Post
FWIW, here are the "All Subs" baseline measurements for 6 positions:

[...]
There's virtually no room for improvement up to 100Hz.

Markus

"In science, contrary evidence causes one to question a theory. In religion, contrary evidence causes one to question the evidence." - Floyd Toole
markus767 is offline  
post #369 of 777 Old 11-29-2016, 10:27 AM
** Man of Leisure **
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 16,232
Mentioned: 205 Post(s)
Tagged: 1 Thread(s)
Quoted: 7562 Post(s)
Liked: 4224
Quote:
Originally Posted by markus767 View Post
There's virtually no room for improvement up to 100Hz.
That has been my assumption all along, which is why I didn't dive into MSO immediately. Regardless, I will continue the analysis just to see how it turns out.
AustinJerry is online now  
post #370 of 777 Old 11-29-2016, 11:29 AM
AVS Forum Special Member
 
andyc56's Avatar
 
Join Date: Feb 2013
Posts: 1,172
Mentioned: 15 Post(s)
Tagged: 0 Thread(s)
Quoted: 17 Post(s)
Liked: 391
Quote:
Originally Posted by markus767 View Post
There's virtually no room for improvement up to 100Hz.
I definitely agree with that.

Quote:
Originally Posted by markus767 View Post
He could do a second set of wider spaced mic positions. What spacing would you recommend?
I don't know! This will probably come across as evasive, but I've never actually thought of the problem in those terms. I've always considered it in terms of putting the mic at the actual position of each listener's head in each seat, and assumed that the seats are far enough apart that there is significant variation in response from seat to seat.

There is the question of how close the positions can get before usage of MSO has negligible benefit, and I don't know the answer to that. Maybe it could be determined using room simulations, running an optimization for a set of points, then repeating, making the points closer and closer together. I don't know how to write a room simulator though.

Quote:
Originally Posted by AustinJerry View Post
That has been my assumption all along, which is why I didn't dive into MSO immediately. Regardless, I will continue the analysis just to see how it turns out.
I noticed that in the other thread too. When you start with near perfection and only one listening position, there just isn't much else to do, other than enjoy it (horrors! ).
andyc56 is offline  
post #371 of 777 Old 11-29-2016, 11:46 AM
** Man of Leisure **
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 16,232
Mentioned: 205 Post(s)
Tagged: 1 Thread(s)
Quoted: 7562 Post(s)
Liked: 4224
Quote:
Originally Posted by andyc56 View Post

I noticed that in the other thread too. When you start with near perfection and only one listening position, there just isn't much else to do, other than enjoy it (horrors! ).
Yes, it is indeed difficult, but I tough it out.
AustinJerry is online now  
post #372 of 777 Old 11-29-2016, 11:48 AM
AVS Forum Special Member
 
markus767's Avatar
 
Join Date: Apr 2009
Posts: 9,873
Mentioned: 111 Post(s)
Tagged: 0 Thread(s)
Quoted: 4189 Post(s)
Liked: 1577
Quote:
Originally Posted by andyc56 View Post
I don't know! This will probably come across as evasive, but I've never actually thought of the problem in those terms. I've always considered it in terms of putting the mic at the actual position of each listener's head in each seat, and assumed that the seats are far enough apart that there is significant variation in response from seat to seat.
By the way, did you or anybody else ever test if the variance at points other than the mic locations was also reduced using MSO?

Markus

"In science, contrary evidence causes one to question a theory. In religion, contrary evidence causes one to question the evidence." - Floyd Toole
markus767 is offline  
post #373 of 777 Old 11-29-2016, 01:05 PM
AVS Forum Special Member
 
andyc56's Avatar
 
Join Date: Feb 2013
Posts: 1,172
Mentioned: 15 Post(s)
Tagged: 0 Thread(s)
Quoted: 17 Post(s)
Liked: 391
Quote:
Originally Posted by markus767 View Post
By the way, did you or anybody else ever test if the variance at points other than the mic locations was also reduced using MSO?
I have not done so.
andyc56 is offline  
post #374 of 777 Old 11-29-2016, 05:47 PM
AVS Forum Special Member
 
artur9's Avatar
 
Join Date: Mar 2013
Location: near philly
Posts: 1,848
Mentioned: 7 Post(s)
Tagged: 0 Thread(s)
Quoted: 671 Post(s)
Liked: 237
Speaking from a position of almost total ignorance regarding MSO....

I have a Xilica that I use to manage my 4 subs. My (minimal) understanding of MSO is that you want to use REW to measure the impulse response of each individual subwoofer so that it can calculate the proper time alignment.

How would I do that since the Xilica is one input and 4 outputs?
artur9 is offline  
post #375 of 777 Old 11-29-2016, 06:06 PM
** Man of Leisure **
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 16,232
Mentioned: 205 Post(s)
Tagged: 1 Thread(s)
Quoted: 7562 Post(s)
Liked: 4224
Quote:
Originally Posted by artur9 View Post
Speaking from a position of almost total ignorance regarding MSO....

I have a Xilica that I use to manage my 4 subs. My (minimal) understanding of MSO is that you want to use REW to measure the impulse response of each individual subwoofer so that it can calculate the proper time alignment.

How would I do that since the Xilica is one input and 4 outputs?
How is your Xilica different than my 2x4, which has one input and four outputs? As input to MSO, you measure each sub independently. I am assuming the Xilica allows for each output to be muted, correct?
AustinJerry is online now  
post #376 of 777 Old 11-29-2016, 06:32 PM
AVS Forum Special Member
 
artur9's Avatar
 
Join Date: Mar 2013
Location: near philly
Posts: 1,848
Mentioned: 7 Post(s)
Tagged: 0 Thread(s)
Quoted: 671 Post(s)
Liked: 237
Quote:
Originally Posted by AustinJerry View Post
How is your Xilica different than my 2x4, which has one input and four outputs? As input to MSO, you measure each sub independently. I am assuming the Xilica allows for each output to be muted, correct?
In this regard the Xilica isn't all that different than the 2x4 (in others, a great deal, link in case you're curious)

so you use 2 channels, the timing reference and the one sub, muting all the others in turn? Doesn't that mess up the timing measurements?
artur9 is offline  
post #377 of 777 Old 11-29-2016, 06:52 PM
** Man of Leisure **
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 16,232
Mentioned: 205 Post(s)
Tagged: 1 Thread(s)
Quoted: 7562 Post(s)
Liked: 4224
Quote:
Originally Posted by artur9 View Post
In this regard the Xilica isn't all that different than the 2x4 (in others, a great deal, link in case you're curious)

so you use 2 channels, the timing reference and the one sub, muting all the others in turn? Doesn't that mess up the timing measurements?
Not sure what your concerns are. For input to MSO, you want raw sub response measurements, with room correction off, all delays zeroed out, trims set to zero. Then you measure one sub at a time, independent of the other subs.
AustinJerry is online now  
post #378 of 777 Old 11-29-2016, 07:15 PM
AVS Forum Special Member
 
andyc56's Avatar
 
Join Date: Feb 2013
Posts: 1,172
Mentioned: 15 Post(s)
Tagged: 0 Thread(s)
Quoted: 17 Post(s)
Liked: 391
Quote:
Originally Posted by artur9 View Post
In this regard the Xilica isn't all that different than the 2x4 (in others, a great deal, link in case you're curious)
Looks like a nice unit. I'll have to ask @JohnPM about how the PEQ is specified in that one. I might have to modify MSO's filter report to make it compatible.

Quote:
so you use 2 channels, the timing reference and the one sub, muting all the others in turn? Doesn't that mess up the timing measurements?
As long as you always use the same speaker as the timing reference for all measurements like people have been doing in the big REW thread, the timing will be okay.
andyc56 is offline  
post #379 of 777 Old 11-30-2016, 08:31 AM
** Man of Leisure **
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 16,232
Mentioned: 205 Post(s)
Tagged: 1 Thread(s)
Quoted: 7562 Post(s)
Liked: 4224
I have imported the REW measurements, configured MSO, and run several trial optimizations. The setup was straightforward due to the very good documentation. I have several questions before proceeding to the next steps:

The first optimization (5 minutes) produced delay blocks that exceeded the MiniDSP 2x4 7.5ms delay limit. I went back into Tools/Application Options and set a 7.5ms maximum delay value.



I then ran a second optimization (5 minutes), but still received gain delay blocks in excess of 7.5ms. How can I limit the delay results?

Next question, the results between the two optimizations are quite different, even though the only changed setting was the maximum delay value. Is this normal, and why would the results be so different? Here are the results:



And finally, the optimization produced no Polarity Inversion blocks. REW measurements show that the front sub pair is 180 degrees out of phase with the rear pair, and my current configuration has a polarity inversion for both front subs. I am curious why MSO is not making the same recommendation.

Next steps (correct me if I am wrong):

1. Export the bi-quads for each sub as a text file, and import the file into each of the four 2x4 output channels.
2. Import the gain and delay settings (once I figure out what is wrong with the delays).
3. Measure the results using REW.
4. Run a fresh Dirac calibration.
5. Measure the final results using REW and compare to the baseline results.
AustinJerry is online now  
post #380 of 777 Old 11-30-2016, 10:48 AM
AVS Forum Special Member
 
andyc56's Avatar
 
Join Date: Feb 2013
Posts: 1,172
Mentioned: 15 Post(s)
Tagged: 0 Thread(s)
Quoted: 17 Post(s)
Liked: 391
Quote:
Originally Posted by AustinJerry View Post
The first optimization (5 minutes) produced delay blocks that exceeded the MiniDSP 2x4 7.5ms delay limit. I went back into Tools/Application Options and set a 7.5ms maximum delay value.

I then ran a second optimization (5 minutes), but still received gain delay blocks in excess of 7.5ms. How can I limit the delay results?
The Tools/Application options only applies to the parameter limits of newly-created filters. For existing filters, you'll need to modify the minimum and maximum properties of the parameters via the Properties window on the right. See this screenshot of the main window.

More information can be found in "Adding Filters to Filter Channels" in the tutorial.

Edit: The above links may have some problems on and off today 11/30/2016 as my web provider is switching over to new servers.

Quote:
Next question, the results between the two optimizations are quite different, even though the only changed setting was the maximum delay value. Is this normal, and why would the results be so different?
The optimization minimizes an aggregate measure of the response flatness error (a single number that's ideally zero) over multiple listening positions. It's sometimes the case that several different combinations of parameters can produce an aggregate error that's very similar.

Imagine that you were only adjusting two parameters: sub distance and level. You'd have an error that's a function of two variables. Mathematically, that looks like:

z = f(x, y)

This is the equation of a surface in three dimensions. Now imagine that the surface looked like this:

https://en.wikipedia.org/wiki/Rastri...n_function.png

That's called a "multi-modal objective function". This means the error function has many local minima. It can be a challenge to create an optimizer that will find the best minimum in the presence of the many local minima. If you were varying 20 parameters, that "surface" would be a "hyper-surface" in 21 dimensions. It can't be visualized, but the concept is the same. These local minima are a challenge for all optimizers. If you're using a number of PEQs, you'll want to run MSO for a pretty long period of time, like 30 minutes, to allow it to explore the solution space and avoid these local minima.

Quote:
And finally, the optimization produced no Polarity Inversion blocks. REW measurements show that the front sub pair is 180 degrees out of phase with the rear pair, and my current configuration has a polarity inversion for both front subs. I am curious why MSO is not making the same recommendation.
MSO only changes the parameters of existing filters. It does not add filters. So if you want to use a polarity inversion, you'll need to add it. Did you do the measurements with the polarity inversions in place? If so, no change is necessary. If not, you'll need to add them manually.

One quirk of MSO is that it only varies parameters that have a continuous range of adjustment. Introducing a polarity inversion on some iterations of the optimizer but not others would mess up the convergence of the optimizer. That's why it doesn't add inversion blocks. Unfortunately, exploring that option requires using a distinct configuration for each discrete combination of inversion blocks to be explored.

Optimizers that allow discrete parameter values are called "combinatorial optimizers". MSO does not use such an optimizer. I chose the optimizer on the basis of its ability to avoid local minima of the objective function as described above.

Unless your subs are really out of polarity with one another (as yours are), adding a polarity inversion to individual subs is generally not a good idea, as it increases the likelihood of non-minimum-phase behavior, giving the potential of degraded time-domain response. But since you're aware of the issue, you're equipped to handle it per the above advice.

Quote:
Next steps (correct me if I am wrong):

1. Export the bi-quads for each sub as a text file, and import the file into each of the four 2x4 output channels.
2. Import the gain and delay settings (once I figure out what is wrong with the delays).
3. Measure the results using REW.
4. Run a fresh Dirac calibration.
5. Measure the final results using REW and compare to the baseline results.
That looks right.

You'll probably want to explore negative delays. You'll have N-1 delays for N subs, assuming a sub-only configuration. I'd do this:

1. Allow the N-1 delays to have a range of -7.5 msec to 7.5 msec
2. Run the optimization
3. Do a "normalize delays" from the "Subwoofer Channels" node of the Config View.
4. This will convert all delays to positive delays or no delay and move some delay blocks around.
5. Some of these delays may be more than 7.5 msec as a result of normalization
6. Change the limits of all delays to 0 and 7.5 msec for min and max.
7. Re-run the optimization

This is just a roundabout way of figuring out which sub needs no delay.

Also, you may want to decrease the minimum allowable center frequency for the PEQs to 20 Hz. I should probably make this the default.
Svenibaer likes this.

Last edited by andyc56; 11-30-2016 at 01:16 PM.
andyc56 is offline  
post #381 of 777 Old 11-30-2016, 12:38 PM
Senior Member
 
AV_mike's Avatar
 
Join Date: May 2009
Posts: 425
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 115 Post(s)
Liked: 105
Hi Jerry, good to see you here.
Obviously Andy (and Markus) are the real experts around here, but I would like to offer some advise.

When using MSO to optimise multiple location subs, over multiple seating positions - the optimiser tends to balance out the individual subs contributions in order to achieve the lowest seat to seat variation - and with a little PEQ thrown in to each sub channel to manage modal peaks, etc, the results can be very good.

However, when optimising multiple subs for a single position (or closely spaced mic positions) MSO will often simply reduce the gain of uneven channels, and leave the most even channels doing all the work - as this does give the flattest response - but does not make full use of the available subs.

Better results will be obtained by giving MSO some PEQs on each sub channel so that the basic unevenness of individual subs can be addressed - allowing all the channels to contribute to the final result (giving more headroom to the combined channel).

Also, as the number of variables increases - make sure you give MSO a good amount of time, I often leave an optimisation running overnight!!!!!!

Give it a go - we're all watching - no pressure,

Regards, Mike.
dgage, andyc56 and Svenibaer like this.
AV_mike is offline  
post #382 of 777 Old 11-30-2016, 01:02 PM
** Man of Leisure **
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 16,232
Mentioned: 205 Post(s)
Tagged: 1 Thread(s)
Quoted: 7562 Post(s)
Liked: 4224
Thanks for the feedback, Andy and Mike. I am going to work on it a bit more this afternoon. And I have no issue giving MSO plenty of time to crunch the data.
AustinJerry is online now  
post #383 of 777 Old 12-01-2016, 07:26 AM
** Man of Leisure **
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 16,232
Mentioned: 205 Post(s)
Tagged: 1 Thread(s)
Quoted: 7562 Post(s)
Liked: 4224
Quote:
Originally Posted by andyc56 View Post

MSO only changes the parameters of existing filters. It does not add filters. So if you want to use a polarity inversion, you'll need to add it. Did you do the measurements with the polarity inversions in place? If so, no change is necessary. If not, you'll need to add them manually.

One quirk of MSO is that it only varies parameters that have a continuous range of adjustment. Introducing a polarity inversion on some iterations of the optimizer but not others would mess up the convergence of the optimizer. That's why it doesn't add inversion blocks. Unfortunately, exploring that option requires using a distinct configuration for each discrete combination of inversion blocks to be explored.

Optimizers that allow discrete parameter values are called "combinatorial optimizers". MSO does not use such an optimizer. I chose the optimizer on the basis of its ability to avoid local minima of the objective function as described above.

Unless your subs are really out of polarity with one another (as yours are), adding a polarity inversion to individual subs is generally not a good idea, as it increases the likelihood of non-minimum-phase behavior, giving the potential of degraded time-domain response. But since you're aware of the issue, you're equipped to handle it per the above advice.

Sorry, but I am still not fully understanding the Polarity Inversion subject.

The pre-MSO sub configuration has the polarity inverted on the two front subs. REW measurements were taken with the polarity inversion turned off. Is this correct?

MSO now has a polarity inversion block configured for the two front subs, and I have re-run the optimization. After transferring the results to the 2x4, do I reverse polarity for the two front subs manually?


Second area of confusion (sorry, my learning curve seems a bit steep). The MSO optimization is resulting in negative delays. The 2x4, of course, cannot have negative delays. I set the delay max value at 7.5, should I have set the min delay to zero?
AustinJerry is online now  
post #384 of 777 Old 12-01-2016, 10:33 AM
Senior Member
 
AV_mike's Avatar
 
Join Date: May 2009
Posts: 425
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 115 Post(s)
Liked: 105
Hi Jerry,

I would configure all delays to minimum of 0ms and maximum of 7.5ms - one for each sub channel.
This will prevent delays being chosen that you can not actually set.

The better method is to set the max to 7.5ms - and set the minimum to a value that your AVR can set by adding distance to the LFE channel. When distance is added - negative delays are effectively possible - as the AVR will actually delay all the other channels with respect to the LFE. Using this method I can set +7.5ms to -17.6ms (17.6ms is approx. 20 feet - the maximum difference between channels allowed in my Denon 4520 - and I always take my initial sweeps with the sub channel distance set to the same value as the nearest speaker)

Hope that makes sense?

Regards, Mike.
AV_mike is offline  
post #385 of 777 Old 12-01-2016, 10:43 AM
** Man of Leisure **
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 16,232
Mentioned: 205 Post(s)
Tagged: 1 Thread(s)
Quoted: 7562 Post(s)
Liked: 4224
Quote:
Originally Posted by AV_mike View Post
Hi Jerry,

I would configure all delays to minimum of 0ms and maximum of 7.5ms - one for each sub channel.
This will prevent delays being chosen that you can not actually set.

The better method is to set the max to 7.5ms - and set the minimum to a value that your AVR can set by adding distance to the LFE channel. When distance is added - negative delays are effectively possible - as the AVR will actually delay all the other channels with respect to the LFE. Using this method I can set +7.5ms to -17.6ms (17.6ms is approx. 20 feet - the maximum difference between channels allowed in my Denon 4520 - and I always take my initial sweeps with the sub channel distance set to the same value as the nearest speaker)

Hope that makes sense?

Regards, Mike.
Can't follow the recommendation for the AVR, Mike. There is no bass management being done in the AVR (all speakers set to large). Any changes to the delay parameter in the AVR affects only the LFE channel, not the re-directed bass. I went ahead and set the min delay in MSO to zero and am in the process now of measuring and evaluating the results.
AustinJerry is online now  
post #386 of 777 Old 12-01-2016, 12:45 PM
** Man of Leisure **
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 16,232
Mentioned: 205 Post(s)
Tagged: 1 Thread(s)
Quoted: 7562 Post(s)
Liked: 4224
OK, making some progress here. I have re-run the MSO Optimization a number of times. Currently, max delays are set t0 7.5, min delays set to zero. I have changed the Center Frequency minimum value to 20Hz per Andy's recommendation. And I changed the target MLP level from 75dB to 80dB. Here are the results showing all subs at Pos 1 (the MLP). Dirac and bass management are off. The post-MSO measurement includes a polarity inversion for the front two subs.



Below 100Hz, there is not much difference, except for a slight boost below 20Hz. My usual crossover is set to 100Hz. Above 100Hz, there is some improvement, but not sure if this is important or not. As Mike noted earlier, the trims on the two front subs are -15dB, so for someone who is used to religiously gain-matching all four subs, this will take some getting used to. The MSO delays seem reasonable.



FWIW, here is a comparison of the post-MSO All Subs at MLP, with polarity normal and polarity inverted. As you can see, polarity inverted is the correct configuration.



Next steps? If the consensus is that it is worth it, I will now run the Dirac Live calibration, re-enable bass management, and take final REW measurements.
AustinJerry is online now  
post #387 of 777 Old 12-01-2016, 01:34 PM
AVS Forum Special Member
 
andyc56's Avatar
 
Join Date: Feb 2013
Posts: 1,172
Mentioned: 15 Post(s)
Tagged: 0 Thread(s)
Quoted: 17 Post(s)
Liked: 391
Quote:
Originally Posted by AustinJerry View Post
Sorry, but I am still not fully understanding the Polarity Inversion subject.

The pre-MSO sub configuration has the polarity inverted on the two front subs. REW measurements were taken with the polarity inversion turned off. Is this correct?
It sounds right, but let me try to explain it in a different way.
1) The filter, gain and delay blocks in MSO always represent a change from the "as-measured" condition
2) You've determined by another means that you have to invert the polarity of your two front subs
3) If you've done your measurements with these inversions in place, there is no need to add polarity inversion in MSO, as the "as-measured" and required conditions are the same.
4) If you've measured without these inversions, you'll need to add them in MSO to model the required change from the "as-measured" condition.

Quote:
MSO now has a polarity inversion block configured for the two front subs, and I have re-run the optimization. After transferring the results to the 2x4, do I reverse polarity for the two front subs manually?
Yes, however you were doing it before.

Quote:
Second area of confusion (sorry, my learning curve seems a bit steep). The MSO optimization is resulting in negative delays. The 2x4, of course, cannot have negative delays. I set the delay max value at 7.5, should I have set the min delay to zero?
I may have been confused in my previous post. I thought you were using a sub-only configuration, and the technique I described of using "normalize delays" was based on that assumption. With a sub-only configuration you can always get rid of negative delays using "Normalize delays". With "sub+mains" configurations, you use "Rearrange delays" to get rid of all individual per-sub negative delays. This lumps them all into a shared negative delay. When using an AVR, one then implements this shared negative delay by increasing the sub distance. I don't know how this would be handled with the 88A though.

The concept of negative delay seems strange, but in the context of MSO it means "reduce the delay relative to the as-measured condition" (see item 1 in the list above). With an AVR, "reduce the delay of all subs relative to the as-measured condition" corresponds to "increase the sub distance relative to the as-measured condition".

Last edited by andyc56; 12-01-2016 at 01:55 PM.
andyc56 is offline  
post #388 of 777 Old 12-01-2016, 01:59 PM
** Man of Leisure **
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 16,232
Mentioned: 205 Post(s)
Tagged: 1 Thread(s)
Quoted: 7562 Post(s)
Liked: 4224
Quote:
Originally Posted by andyc56 View Post
I may have been confused in my previous post. I thought you were using a sub-only configuration, and the technique I described of using "normalize delays" was based on that assumption. With a sub-only configuration you can always get rid of negative delays using "Normalize delays". "Normalize delays" with "sub+mains" configuration gets rid of all individual per-sub negative delays, and lumps them all into a shared negative delay. When using an AVR, one then implements this shared negative delay by increasing the sub distance. I don't know how this would be handled with the 88A though.

The concept of negative delay seems strange, but in the context of MSO it means "reduce the delay relative to the as-measured condition" (see item 1 in the list above). With an AVR, "reduce the delay of all subs relative to the as-measured condition" corresponds to "increase the sub distance relative to the as-measured condition".
Thanks for the revised description of the polarity inversion block. I am pretty sure I understand it now.

I suspect the shared negative delay would be introduced on the 88BM Outputs tab for Output Channel 3 (combined subs going to the 2x4).

Am I good to proceed with the Dirac calibration, or do I still have corrections to make to MSO?
AustinJerry is online now  
post #389 of 777 Old 12-01-2016, 02:29 PM
AVS Forum Special Member
 
andyc56's Avatar
 
Join Date: Feb 2013
Posts: 1,172
Mentioned: 15 Post(s)
Tagged: 0 Thread(s)
Quoted: 17 Post(s)
Liked: 391
Quote:
Originally Posted by AustinJerry View Post
Below 100Hz, there is not much difference, except for a slight boost below 20Hz. My usual crossover is set to 100Hz. Above 100Hz, there is some improvement, but not sure if this is important or not. As Mike noted earlier, the trims on the two front subs are -15dB, so for someone who is used to religiously gain-matching all four subs, this will take some getting used to. The MSO delays seem reasonable.
With the two front subs turned down so much, it looks like you're experiencing what Mike has described, repeated below:

Quote:
Originally Posted by AV_mike View Post
When using MSO to optimise multiple location subs, over multiple seating positions - the optimiser tends to balance out the individual subs contributions in order to achieve the lowest seat to seat variation - and with a little PEQ thrown in to each sub channel to manage modal peaks, etc, the results can be very good.

However, when optimising multiple subs for a single position (or closely spaced mic positions) MSO will often simply reduce the gain of uneven channels, and leave the most even channels doing all the work - as this does give the flattest response - but does not make full use of the available subs.
Quote:
Originally Posted by AustinJerry View Post
Am I good to proceed with the Dirac calibration, or do I still have corrections to make to MSO?
I don't like that -15 dB gain change on the front subs at all. I'd suggest constraining the sub gain blocks in MSO to not allow so much attenuation. Or in the extreme case, you could get your gain matching by not having individual gain blocks at all and only use a shared gain block, letting the individual PEQs and delays do the work.

How many PEQs per sub are you using?

Last edited by andyc56; 12-01-2016 at 02:34 PM.
andyc56 is offline  
post #390 of 777 Old 12-01-2016, 02:45 PM
** Man of Leisure **
 
AustinJerry's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 16,232
Mentioned: 205 Post(s)
Tagged: 1 Thread(s)
Quoted: 7562 Post(s)
Liked: 4224
Quote:
Originally Posted by andyc56 View Post

How many PEQs per sub are you using?
I have been following the MSO guide that was posted on the MiniDSP forum. The guide recommended setting up three PEQ's, so that is what I did.
AustinJerry is online now  
Sponsored Links
Advertisement
 
Reply Subwoofers, Bass, and Transducers

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off