Mono Subwoofer timing signal without an AV receiver?

[Daverich4]

He doesn't have an avr from what I can tell, so he's simply dealing with pre-amp outputs in stereo to start. A lot of lower bass is already mono otoh. He needs something more than an old 2ch pre-amp....

It’s a traditional 2 channel stereo. I’ve received extensive assistance on this forum as far as room treatment & speaker and listening position. It’s been suggested a number of times that a pair of subwoofers should be able to correct several room modes. If it doesn’t work out with what I have I’ll sell them.

 

[Flaesh]

Okay, measure it and see what happens. And if you place subs in the middle of the room height, the first vertical mode also won't be exited!))

Btw, what svs3000 did you buy? Real life amplitude test always is interesting.

 

[Keith_W]

It’s a traditional 2 channel stereo. I’ve received extensive assistance on this forum as far as room treatment & speaker and listening position. It’s been suggested a number of times that a pair of subwoofers should be able to correct several room modes. If it doesn’t work out with what I have I’ll sell them.

It's actually a pair of subwoofers WITH DSP that will correct room modes. You may or may not get away without DSP, but it will be much more difficult. And BTW, DSP means that you should ideally have individual control of each sub and speaker, so for a 2.2 system you will need either a 4 channel hardware DSP or a 4 channel DAC if you are planning to do it manually.

 

[Flaesh]

Found it in another thread https://www.audiosciencereview.com/...ds/main-listening-position.70502/post-2563378

IMO integration with unfiltered speakers may be problematic, and a few ms of delay don't make much difference in general case.

ps: But noticeable improvements, including in the time domain, are possible. At the cost of increasing the bass level. Theoretical exercise (simplified!!):

 

[janbth]

There are a number of topics that are discussed repeatedly here on ASR which is inflicted with - AHEM! - «low quality information». Room modes and using (multiple) subs to fix them is one such topic. It’s clear to me after spending years here on ASR that many people don’t understand room modes. I don’t understand why. It’s a solved problem, and it’s really not that difficult.

An excellent source for reading about room modes and how they can be dealt with is Floyd Toole’s Sound Reproduction. I take the liberty of posting a photo from the 4th edition, Chapter 14.5.4, where he includes an informative graphic on a multi-sub technique to reduce them. I recommend getting a copy of Sound Reproduction if you are serious about this hobby.

 

[janbth]

First, stereo subs: L and R subs must be deliberately placed on the left and right of the listening position. During playback, they are sent L and R music signal.

Mono subs: the subs are deployed wherever they produce the flattest response. You must find these positions by experimentation (i.e. subwoofer crawl). During playback, the subs are sent a summed mono signal which your AVR should do internally, but each subwoofer will have an individual DSP profile. Meaning, each sub will have its own delays, phase adjustments, PEQ's, etc. So they aren't actually playing the same signal - they are "mono" in the sense that the input signal is the same, but the output to each sub is individualised. Meaning that you CAN NOT use a physical connector on your AVR outputs to sum the subs together to create a mono signal. If you did that, you wouldn't be able to send individual signal to each sub.

I agree with what you say about stereo subs, but what you say about mono subs is wrong. Do not EQ subs separately. This may defeat the purpose of using multiple subs to reduce modal peaks.

To see this, consider two subs - one located at each wall. When played separately, they produce a peak in the frequency response of the lowest lying mode at the listening position. When played together, same mono signal, in phase, they will reduce the peak. Now, if you EQ each sub separately, by applying a PEQ filter with negative gain, you will actually produce a dip when the subs play together. Which they always will, when you set them up in mono!

 

[andyc56]

I agree with what you say about stereo subs, but what you say about mono subs is wrong. Do not EQ subs separately. This may defeat the purpose of using multiple subs to reduce modal peaks.

Well, you don't want to EQ each sub to flat, that's true. But in Floyd Toole's book, section 14.5.8, "Manipulate Modes Using Multiple Subwoofers and Computer Optimized Signal Processing—All Rooms", applying EQ to each sub individually (and also delays, and possibly all-pass filters) is exactly what is done. This applies to Harman SFM, Dirac DLBC and MSO among others. These systems don't EQ each sub flat, but rather apply separate EQ, delay etc. to the subs to try to minimize seat-to-seat variation. Then they apply global EQ to all subs at once to flatten them (or reach a desired curve).

 

[janbth]

Well, you don't want to EQ each sub to flat, that's true. But in Floyd Toole's book, section 14.5.8, "Manipulate Modes Using Multiple Subwoofers and Computer Optimized Signal Processing—All Rooms", applying EQ to each sub individually (and also delays, and possibly all-pass filters) is exactly what is done. This applies to Harman SFM, Dirac DLBC and MSO among others. These systems don't EQ each sub flat, but rather apply separate EQ, delay etc. to the subs to try to minimize seat-to-seat variation. Then they apply global EQ to all subs at once to flatten them (or reach a desired curve).

OK, you have a point. I don’t deny that an additional step of adding separate filters to the subs may improve the response after an initial baseline has been established, but the starting point should be common filters. I assume this is what the algorithms you mention do. But who knows? They are proprietary.

Do you agree that my example in post #26 shows that common filters are necessary - at least to set a baseline?

 

[john61ct]

I assume this is what the algorithms you mention do. But who knows? They are proprietary.

That is the creator of MSO.

> If you're a two-channel user not using an AVR, this places the burden on you to understand the concept of bass management and implement it as used in AVRs. **That means summing bass to mono, adding sub EQ that's common to all subs, then splitting the sub signal into as many paths as you have subs, with EQ and delay for each one.** The approach of "rolling your own bass management" with dedicated DSP hardware is harder than using an AVR, because an AVR does this work for you.

Multi-Sub Optimizer (MSO): Lessons Learned, Tips & Tricks

I am in the process of making some tweaks to my 2.2 system, and wanted to use MSO. I've got an Okto Pro 8 on the way and want to run MSO for sub alignment. Is there an idiot's guide somewhere? Some basic questions: Do you measure both subs and mains or just subs (there seems some differing...

www.audiosciencereview.com

...

obviously less relevant to using Stereo subs

 

[andyc56]

OK, you have a point. I don’t deny that an additional step of adding separate filters to the subs may improve the response after an initial baseline has been established, but the starting point should be common filters.

In SFM, the starting point is not established by applying common filters first, but after the seat-to-seat variation has first been minimized using individual sub EQ, delay, and attenuation. The common filters are applied afterwards. The algorithm, though it's patented, is described in detail in Todd Welti's 2006 AES article, Low-Frequency Optimization Using Multiple Subwoofers (PDF).

In MSO (my software), it's all done at once within the same optimization. This helps ensure that any attempts to minimize seat-to-seat variation don't render the final frequency response too difficult to equalize properly using common EQ. Though DLBC is proprietary, they do reveal some information about its algorithm in some of their white papers. I describe some of that here.

I assume this is what the algorithms you mention do. But who knows? They are proprietary.

Do you agree that my example in post #26 shows that common filters are necessary - at least to set a baseline?

I was having trouble figuring out what you were trying to say in post 26. It might be helpful to refer to Figure 14.17 (b) and (c) in section 14.5.4 of Floyd Toole's book. In 14.17 (b), that is an odd-order mode (1st-order), so identical drive to identical subs against the two walls will result in mode cancellation. The two lobes have opposite signs, + and -. In figure 14.17 (c), that is an even-order mode (2nd-order). That means identical drive to identical subs against the two walls will result in mode boosting. The two lobes have the same signs (both +). That is also stated in that figure's accompanying text, where it says, "The second-order mode has been amplified because the subwoofers are in lobes having the same polarity." Because of this complexity, trying to come up with hard and fast rules about what happens to the overall response is harder than it seems. When seat-to-seat variation is of interest, that's harder yet.

 

[janbth]

I assume this is what the algorithms you mention do. But who knows? They are proprietary.

In SFM, the starting point is not established by applying common filters first, but after the seat-to-seat variation has first been minimized using individual sub EQ, delay, and attenuation. The common filters are applied afterwards. The algorithm, though it's patented, is described in detail in Todd Welti's 2006 AES article, Low-Frequency Optimization Using Multiple Subwoofers (PDF).

In MSO (my software), it's all done at once within the same optimization. This helps ensure that any attempts to minimize seat-to-seat variation don't render the final frequency response too difficult to equalize properly using common EQ. Though DLBC is proprietary, they do reveal some information about its algorithm in some of their white papers. I describe some of that here.

Ah, you are ‘andyc’ - the creator or MSO. Sorry I didn’t catch that. You obviously know your own software. I haven’t tried MSO myself for the only reason that I’m on Mac, and MSO is Windows only.

Anyway, fine, let me backtrack once again: Fully automated systems/algorithms for correcting room modes - SFM, DLBC, MSO, etc. - may well optimize subs separately, with success, as part of the algorithm, in order to produce a smooth response in mono.

The point that is of concern to me, and that I have made a few times here on ASR, is what practical guidelines should someone follow if he/she were to set up a system with sub(s) by hand. This last part about ‘by hand’ is important. I assume no automated systems. Just level/delay of the sub(s), possibly with some DSP (i.e. PEQ filters) available, possibly a miniDSP device. It is in this scenario that I say: EQ the subs together first. Optionally add tweaks separately to see if any improvement can be found.

I was having trouble figuring out what you were trying to say in post 26. It might be helpful to refer to Figure 14.17 (b) and (c) in section 14.5.4 of Floyd Toole's book. In 14.17 (b), that is an odd-order mode (1st-order), so identical drive to identical subs against the two walls will result in mode cancellation. The two lobes have opposite signs, + and -. In figure 14.17 (c), that is an even-order mode (2nd-order). That means identical drive to identical subs against the two walls will result in mode boosting. The two lobes have the same signs (both +). That is also stated in that figure's accompanying text, where it says, "The second-order mode has been amplified because the subwoofers are in lobes having the same polarity." Because of this complexity, trying to come up with hard and fast rules about what happens to the overall response is harder than it seems. When seat-to-seat variation is of interest, that's harder yet.

Yes, it can be complex, but let me try again. Let’s simplify.
- First, consider one listener position only - the seat-to-seat variation problem is for another day.
- Second, let’s focus on the lowest order mode only, Figure 14.17 (b). What happens further up or down the frequency range is not relevant to this argument. (For example, the second order mode will indeed be boosted when subs play mono, so that’s a tough problem. Maybe we need to compromise on that mode.)
- Third, the listener is somewhere close to a wall, or at least away from the null in the middle.

So, in the scenario of Figure 14.17 (b), before we apply any filters, if we play either sub separately, that will lead to a frequency response with a peak at the frequency of the mode. Do we agree on that?

If the two subs play together, exact same signal, in phase, this will produce a near smooth response, i.e. a significantly reduced peak at the frequency of the mode. A cancellation effect is in action here. This corresponds to the SPL line with small dots in the figure. Do we agree?

Then EQ each sub separately with a suitable PEQ filter. These would be notch filters. Then play the subs in mono. You will then observe a dip at the frequency of the mode. Agreed?

The conclusion I draw from this argument is that separate EQ is a flawed procedure.

 

[sqrl]

SVS auto eq is now on their app

 

[RPG]

SVS auto eq is now on their app

But only for the R Evolution series subs. If you have anything older than that, you're back to manual EQ or using your AVR if you have one.

 

[andyc56]

Fully automated systems/algorithms for correcting room modes - SFM, DLBC, MSO, etc. - may well optimize subs separately, with success, as part of the algorithm, in order to produce a smooth response in mono.

These systems do not optimize the subs separately. They apply individual sub EQ and use the Superposition Principle to optimize the summation, taking phase into account, of all the subs together at each listening position. Because the optimization criteria are applied not to individual subs, but to their summation, they are optimized together. One can do this because the system is linear, so that the Superposition Principle applies.

The point that is of concern to me, and that I have made a few times here on ASR, is what practical guidelines should someone follow if he/she were to set up a system with sub(s) by hand. This last part about ‘by hand’ is important. I assume no automated systems. Just level/delay of the sub(s), possibly with some DSP (i.e. PEQ filters) available, possibly a miniDSP device.

Yes, you do not want to try such an optimization by hand because it is not practical. That is different from saying it is not valid though. For instance, a typical comprehensive MSO optimization might be 40,000 iterations of the optimizer. Each iteration is 100 evaluations of the summed responses ("together") at each position. So that's 4 million evaluations of the combined responses multiplied by the number of listening positions. That's 4 million times the number of subs times the number of listening positions. Obviously that isn't practical, as when doing this manually, each of those evaluations would need to be its own measurement. Superposition allows you to do these computations with the number of measurements being the number of subs times the number of listening positions measured.

It is in this scenario that I say: EQ the subs together first. Optionally add tweaks separately to see if any improvement can be found.

Even this is questionable. Suppose the only EQ you have available is the common EQ, but you have adjustable delays available for each sub. What response do you equalize? You don't know until you adjust the delays first. The relative delays have a strong effect on the frequency response of the combined subs. So you adjust the delays first (maybe to maximize output at the MLP, say, with the REW alignment tool), then apply EQ.

- Second, let’s focus on the lowest order mode only, Figure 14.17 (b). What happens further up or down the frequency range is not relevant to this argument.

Okay, the frequency response (that is, "what happens further up or down the frequency range") is not relevant to this argument.

So, in the scenario of Figure 14.17 (b), before we apply any filters, if we play either sub separately, that will lead to a frequency response with a peak at the frequency of the mode.

Now the frequency response is relevant to the argument?

If the two subs play together, exact same signal, in phase, this will produce a near smooth response, i.e. a significantly reduced peak at the frequency of the mode.
[...]
Then EQ each sub separately with a suitable PEQ filter. These would be notch filters. Then play the subs in mono. You will then observe a dip at the frequency of the mode.

The manner in which the subs combine is very much influenced both by the relative magnitude and phase of the signals driving them. Having different EQ for each sub can actually make the seat-to-seat variation worse than driving them identically. It can also make it better. It's possible to conjure up both scenarios. Of course, if one makes as a precondition the stipulation, "consider one listener position only - the seat-to-seat variation problem is for another day.", that conveniently removes the main reason that multiple-subwoofer optimization (with its attendant separate EQ) exists in the first place.

The conclusion I draw from this argument is that separate EQ is a flawed procedure.

That is because you have stipulated, as a precondition for your argument, that seat-to-seat variation does not matter. But minimizing seat-to-seat variation is the very reason for having different EQ for each sub in multi-sub optimization. In 14.1 of Floyd Toole's book, Todd Welti lists that as the number one objective.

The first one is the most difficult to achieve and is the primary focus of this chapter

 

[john61ct]

Note also that even for "just one" MLP, the thickness of your skull + the distances involved in changing posture, shifting your weight / body within the seat, turning & tilting your head etc

all add up to non-trivial variations (huge understatement there!)

In an overall workflow – no matter what tools are used – only keep to a single precise mic position when that is REQUIRED for valid results - as with some MSO procedures.

Otherwise, best to use MMM or similar over an (in my case ~18" minimum) sphere.

And ofc that is still for a single LP, leaving multiple listeners "for another day".

 

[janbth]

These systems do not optimize the subs separately. They apply individual sub EQ and use the Superposition Principle to optimize the summation, taking phase into account, of all the subs together at each listening position. Because the optimization criteria are applied not to individual subs, but to their summation, they are optimized together. One can do this because the system is linear, so that the Superposition Principle applies.

Yes, this is exactly the point.

Even this is questionable. Suppose the only EQ you have available is the common EQ, but you have adjustable delays available for each sub. What response do you equalize? You don't know until you adjust the delays first. The relative delays have a strong effect on the frequency response of the combined subs. So you adjust the delays first (maybe to maximize output at the MLP, say, with the REW alignment tool), then apply EQ.

If we were in the scenario of Figure 14.17 (b), the relative delay of the subs would need to be zero. I guess it might be appropriate in some scenarios to apply a delay first. It would depend on the standing wave patters of the modes, and the location of the subs and listener. However, it would never be appropriate to apply separate PEQ filters first.

Okay, the frequency response (that is, "what happens further up or down the frequency range") is not relevant to this argument.
…
Now the frequency response is relevant to the argument?

The frequency response further up or down is not relevant to this argument. For the second-order mode we get a peak in the response at the frequency of the mode, which can be handled with other tools if you do it by hand, like applying a common PEQ filter.

That is because you have stipulated, as a precondition for your argument, that seat-to-seat variation does not matter. But minimizing seat-to-seat variation is the very reason for having different EQ for each sub in multi-sub optimization. In 14.1 of Floyd Toole's book, Todd Welti lists that as the number one objective.

Yes, as I said, I’m addressing a specific scenario that is typical in discussions here on ASR, that of a person with a stereo system, with a sub or two, a single listening position, some bothersome room modes, and who is not using an automated room correction system. The seat-to-seat variation problem is important and worthy, but it is a different problem

 

[Flaesh]

The specific scenario in this thread now is two towers without any processing including HPF absence and two SVSs with some kind of DSP capabilities. For one LP.

 

[janbth]

The specific scenario in this thread now is two towers without any processing including HPF absence and two SVSs with some kind of DSP capabilities. For one LP.

Yes, that’s true. This scenario is a bit more unpredictable than a traditional bass managed one. However, as far as the EQing procedure goes, the underlying principles are the same: the system should be EQed as a collective unit, not each sub/speaker separately.

If EQing can only be achieved from DSP onboard the subs (so no EQing the mains), then the OP will have to take what he has and make the best out of it.

 

[Daverich4]

Okay, measure it and see what happens. And if you place subs in the middle of the room height, the first vertical mode also won't be exited!))

Btw, what svs3000 did you buy? Real life amplitude test always is interesting.
View attachment 548937

The new Revolution 3000’s

 

[Daverich4]

It's actually a pair of subwoofers WITH DSP that will correct room modes. You may or may not get away without DSP, but it will be much more difficult. And BTW, DSP means that you should ideally have individual control of each sub and speaker, so for a 2.2 system you will need either a 4 channel hardware DSP or a 4 channel DAC if you are planning to do it manually.

I’m going to start with what DSP is built into the subwoofers. (SVS Revolution 3000’s) You’ve posted several times on how to do that so I’ll see if I can get that to work out. If not, I’ll probably look into additional hardware. Or maybe not. With your help and that of others on this forum it sounds pretty good as it is. This is mostly just out curiosity for me, I keep reading about the benefits of adding 2 or more subwoofers to a system and thought I’d try it.