Geddes Approach on Multi-sub Bass Optimization: Video Presentation

Geddes Approach on Multi-Sub Bass Optimization: Video Presentation

​

AV NIRVANA is happy to release one of our first video presentations addressing technical topics. This video is on the multi-sub approach developed by Earl Geddes, a man I consider an important mentor and influencer of my own views on audio. This is a presentation I gave at a friends Home Theater GTG discussing low frequencies in small rooms and the use of Geddes multi-sub approach. Please put any comments or questions below.

Also, please take a moment to Like the video and Subscribe to AV NIRVANA on YouTube!

 

[tripplej]

Wow, very nice. Thanks for sharing.

 

[Matthew J Poes]

Wow, very nice. Thanks for sharing.

Thanks! I have one more coming. YouTube is just taking it’s good ol’ time loading it. I’ll hopefully have a link posted soon.

 

[Todd Anderson]

Great presentation!

 

[Matthew J Poes]

Great presentation!

Thanks Todd! I’ve done the REW talk so many times now that this was actually the more fun talk. While I’ve had casual conversations or forum exchanges around it, this was the first time I ever turned it into a presentation.

One of the non-intuitive aspects of the approach that I wanted to make a point of discussing also became a tense moment with Dr. Geddes. In discussing this talk I used the word crossover, which I did out of habit in the talk as well. By this I meant a low or high pass filter but actually the term means a low and high pass filter together. So the sound crosses over from one speaker to the next. Geddes doesn’t advocate this and wanted to be very clear that I not imply he allows the use of highpass filters in his approach. I had already identified this as an important point most people miss and even I really confused the terms a bit. Hopefully my explanation in the talk is still clear.

 

[Michael Boeker]

Your presentation is very well done, but I disagree with Geddes on several points. The main point I have issue with is at 17:42 where he/you says, "Precise Time Alignment not important - only steady state response. "

Geddes says that it takes several periods of sound for the ear to recognize "pitch" and at 100 Hz the ear hasn't recognized the tone until 30 ms. However, this has nothing to do with the change that occurs to the sound prior to us hearing it. The sound can change SPL at a given frequency with as little as 1 ms of change in the time alignment.

Use the REW Room Simulation I can show that 1 ms can a make difference of about 20 dB over a narrow Q and 10 ms can make a 38 dB difference over a wider span of frequencies. If I play a sine wave at the affected peak, I can easily hear the SPL difference during the 10 ms change.

I'm building a new house with a new home theater. I've been doing some modeling with the REW Room Simulation. The first picture is a single subwoofer at the front of the room. Then I added another subwoofer at the rear of the room. Finally I reversed phase of the rear subwoofer and time aligned it (less than 30 ms) with the front subwoofer. Precise Time Alignment is all about the phase relationship of the subwoofers to each other.

In pro audio if you are doing a concert outdoors and are aligning the subs and tops, you use ~12.5 ms of delay on the tops to match the subs if they are at the same plane plus any delay caused by the DSP. This is to allow for precise time alignment at an 80 Hz crossover. If you were to take this system into a small room, why would the time alignment no longer matter? I contend that it does and that it makes a huge difference in not only the tactile and impact feel of the system, but also the sound quality.

I owned the 215RT's you heard at AXPONA. I sold them about a month ago and am getting 215RM's for the new theater. Since they are tuned to 17 Hz, the three 215RT's could be considered subwoofers for testing purposes. I put a low pass filter on the left and right "sub" at 100 Hz with a 24/dB per octave rolloff. I played Hotel California while changing the delay. It doesn't take much delay (~10 ms) to muddy the mid-bass and make an audible difference.

At 23:44 it says, "Geddes suggests using EQ to eliminate the peaks in the response caused by modal abnormalities. The only peaks that should be dropped are those that show up in all or most listening positions." I completely agree with this. If you watch the amroc room mode calculator you showed in the video while moving subwoofers, microphones, or listening position throughout the room, you see that the amroc room mode calculator sits there and shows the same room modes. That is because room modes are dependent on room size and have no correlation with subwoofer placement (I realize you know this).

I had my last house for 15 years and had 7 to 8 subwoofer systems in the room including up to 4 sealed subwoofers. My room modes stayed the same regardless of subwoofer brand or position. However, it seems to me that a calibration workflow and strategy to deal with the room modes is rarely implemented.

 

[Matthew J Poes]

Your presentation is very well done, but I disagree with Geddes on several points. The main point I have issue with is at 17:42 where he/you says, "Precise Time Alignment not important - only steady state response. "

Geddes says that it takes several periods of sound for the ear to recognize "pitch" and at 100 Hz the ear hasn't recognized the tone until 30 ms. However, this has nothing to do with the change that occurs to the sound prior to us hearing it. The sound can change SPL at a given frequency with as little as 1 ms of change in the time alignment.

Use the REW Room Simulation I can show that 1 ms can a make difference of about 20 dB over a narrow Q and 10 ms can make a 38 dB difference over a wider span of frequencies. If I play a sine wave at the affected peak, I can easily hear the SPL difference during the 10 ms change.

I'm building a new house with a new home theater. I've been doing some modeling with the REW Room Simulation. The first picture is a single subwoofer at the front of the room. Then I added another subwoofer at the rear of the room. Finally I reversed phase of the rear subwoofer and time aligned it (less than 30 ms) with the front subwoofer. Precise Time Alignment is all about the phase relationship of the subwoofers to each other.

View attachment 9818 View attachment 9819 View attachment 9820

In pro audio if you are doing a concert outdoors and are aligning the subs and tops, you use ~12.5 ms of delay on the tops to match the subs if they are at the same plane plus any delay caused by the DSP. This is to allow for precise time alignment at an 80 Hz crossover. If you were to take this system into a small room, why would the time alignment no longer matter? I contend that it does and that it makes a huge difference in not only the tactile and impact feel of the system, but also the sound quality.

I owned the 215RT's you heard at AXPONA. I sold them about a month ago and am getting 215RM's for the new theater. Since they are tuned to 17 Hz, the three 215RT's could be considered subwoofers for testing purposes. I put a low pass filter on the left and right "sub" at 100 Hz with a 24/dB per octave rolloff. I played Hotel California while changing the delay. It doesn't take much delay (~10 ms) to muddy the mid-bass and make an audible difference.

At 23:44 it says, "Geddes suggests using EQ to eliminate the peaks in the response caused by modal abnormalities. The only peaks that should be dropped are those that show up in all or most listening positions." I completely agree with this. If you watch the amroc room mode calculator you showed in the video while moving subwoofers, microphones, or listening position throughout the room, you see that the amroc room mode calculator sits there and shows the same room modes. That is because room modes are dependent on room size and have no correlation with subwoofer placement (I realize you know this).

I had my last house for 15 years and had 7 to 8 subwoofer systems in the room including up to 4 sealed subwoofers. My room modes stayed the same regardless of subwoofer brand or position. However, it seems to me that a calibration workflow and strategy to deal with the room modes is rarely implemented.

Hi Michael,

Thanks for watching and thanks for the feedback.

I actually think the point you make is the point Geddes made with the only difference being the relative importance of time alignment to achieve that goal. Geddes isn’t saying that a flat response is unimportant. He believe that a flat response is the only thing that matters. You can achieve a flat response without precise time alignment because at such low frequencies “in phase” behavior across the measurement locations can still be fairly far off in time alignment.

Keep in mind those simulations are far from perfect and often massively overstate the changes you will see in real life. They assume a perfect cuboid room with rigid walls. We virtually never have that. My own experience has been that I can achieve a flat response without precise time alignment of the subs, in fact, often the flattest response is achieved when the subs are not time aligned.

We can’t use pro audio as a standard for comparison here. A pro audio venue is a large acoustic space, it may even be an open outdoor acoustic space. Such large spaces no longer have he huge room effects so it should be obvious why time alignment matters more. It doesn’t as much in small rooms because the response your ear hears and the mic picks up already contains reflections from around the room. The energy you hear and measure is already quite delayed. Even if all of the sources of group delay at LF’s could be eliminated you would see that the peak energy time at Lf’s Is wider than at mid and high frequencies when measured in room.

I actually have some indoor and outdoor measurements of a subwoofer that helps shownthis effect but I am waiting on permission to publicly share. The measurements are of a product that hasn’t come to market yet. In the mean time, take a look at the shape of the wavelet I showed at the end of the video. While you can’t see what it looks like inside and both suffer group delay, you can also still see that both are cone shaped. That’s the point here.

 

[jgmixstudio]

Hello Every one.
I was wondering if any of you have ever heard or used an Active sub in the back of your room to control room modes in the sweet spot of a Studio?

 

[Matthew J Poes]

Hello Every one.
I was wondering if any of you have ever heard or used an Active sub in the back of your room to control room modes in the sweet spot of a Studio?

It’s conceptualky related. It’s about adding additional LF sources to the room to interact differently with the modes.

I can’t speak for all, but what I’ve read about this makes me think it may come from some confusion over how the concept works. Some say that it cancels the modes at LF’s. That isn’t correct. This particular approach can’t cancel the modes. It actually excites the modes more evenly. That is why one additional LF source is nice, but not adequate. You need 3 or more.

I believe the confusion comes from a technically correct but confusing way of talking about mode cancelation. If you excite all the modes that are positive and negative evenly the net effect is 0. A neutral response. To me, cancelation means stoping the reflection that causes the mode in the first place. For example, canceling the reflection or absorbing it. That’s isn’t how the active sub method works, it’s an ineffective approach if you try it. Instead, it excites more of the modes at a given listening position. With sufficient excitation, enough of the positive and negative excitation of modes leads to an even response. Hence why one sub is not enough. It helps but is not sufficient.

Actual cancelation of the reflections through active subs is only possible through a large number of carefully placed subs which generate a Planar wave at low frequencies and then cancels the Planar wave with an equal antiphase Planar wave at the other end of the room.

 

[jgmixstudio]

It’s conceptualky related. It’s about adding additional LF sources to the room to interact differently with the modes.

I can’t speak for all, but what I’ve read about this makes me think it may come from some confusion over how the concept works. Some say that it cancels the modes at LF’s. That isn’t correct. This particular approach can’t cancel the modes. It actually excites the modes more evenly. That is why one additional LF source is nice, but not adequate. You need 3 or more.

I believe the confusion comes from a technically correct but confusing way of talking about mode cancelation. If you excite all the modes that are positive and negative evenly the net effect is 0. A neutral response. To me, cancelation means stoping the reflection that causes the mode in the first place. For example, canceling the reflection or absorbing it. That’s isn’t how the active sub method works, it’s an ineffective approach if you try it. Instead, it excites more of the modes at a given listening position. With sufficient excitation, enough of the positive and negative excitation of modes leads to an even response. Hence why one sub is not enough. It helps but is not sufficient.

Actual cancelation of the reflections through active subs is only possible through a large number of carefully placed subs which generate a Planar wave at low frequencies and then cancels the Planar wave with an equal antiphase Planar wave at the other end of the room.

I read about it here
https://amcoustics.com/articles/roommodes
and found it extremely interesting, I don't need multiple listening positions, I just need a nice flat response from my monitors in the sweet spot so that I can Mix, I've had many small to medium size working spaces and always struggled with room modes, since I didn't have the capability of building from the ground Up.
Also I have to mention that I never liked mixing with subs, found it distracting, but using a sub as an absorption tool sounds like a really cool idea. And if it works, maybe there is a solution for smaller scale studios like mine.
All this I ask for opinions before investing in equipment.
I don't know if it works but my Idea was.
Radiate from my monitors and cancel in the back with the sub. maybe feeding the sub with a digital processor so that you can change polarity and phase.
What do you think? would it be worth investing in +500 bucks between sub and processor?

 

[Matthew J Poes]

I read about it here
https://amcoustics.com/articles/roommodes
and found it extremely interesting, I don't need multiple listening positions, I just need a nice flat response from my monitors in the sweet spot so that I can Mix, I've had many small to medium size working spaces and always struggled with room modes, since I didn't have the capability of building from the ground Up.
Also I have to mention that I never liked mixing with subs, found it distracting, but using a sub as an absorption tool sounds like a really cool idea. And if it works, maybe there is a solution for smaller scale studios like mine.
All this I ask for opinions before investing in equipment.
I don't know if it works but my Idea was.
Radiate from my monitors and cancel in the back with the sub. maybe feeding the sub with a digital processor so that you can change polarity and phase.
What do you think? would it be worth investing in +500 bucks between sub and processor?

It doesn’t really work that way. The sub will contribute to the LF’s as a source even if you set it up to cancel the reflection causing the mode.

The only way to cancel just the reflection is with an active bass trap. That works reactively and is a different beast.

 

[Matthew J Poes]

But I should add, you should still try it. I think if you found it distracting it is possible it wasn’t optimally set up. Optimally set up it should not be a distraction.

 

[jgmixstudio]

But I should add, you should still try it. I think if you found it distracting it is possible it wasn’t optimally set up. Optimally set up it should not be a distraction.

Thanks I will borrow some equipment and post my results.
All the Best Keep on Subin

 

[Trdat]

Some of the comments are similar to what I am considering as an option of helping room modes and SBIR. Apparently SBIR and room modes should have different methods to tackle. Although, the above active sub technique doesn't actually cancel modes it excites room modes evenly according to what has been said that is essentially what Geddes is suggesting if I have understood correctly. I am not sure how different the following approach I will suggest is.

What about the flanking sub technique where you place two subs on the back wall but in a different low frequency range? I mean, did I understand right that the 3 different subwoofers that you suggested can be at different crossovers points?

My example set up is to have the low frequency from the AVR Pre Processor into an active crossover where I will send 70hz and under to the low out of the active crossover at the front and the high out to subs from 70hz to about 180hz.

I am trying to add a more low frequencies to the set up in particular to the back to help with envelopment. Also, I get to DIY a small sealed cabinet with an efficient driver giving me some punch and slam.

Would this work? Any advantages or disadvantages?

The main reason I am not just adding the same subs that I have at the front to the back is purely cause there pretty big. I suppose ideally that is what you would to or the method mentioned in the video randomly spacing them asymmetrically.

PS> I find when it comes to fast bass, meaning music with fast repetitive bass drum EQing corrupts the transients of my sub woofers, obviously if the toll off is sealed and you flatten till 30hz then your going from a roll off that is better in transients to a roll off that is ultimately worse. With slow bass music, this works but how to extend the low frequency response without ruining the natural roll off of my sub woofer is what intrigues me especially when EQ is so revered. Essentially your loosing the whole transient capability of the sub.

 

[Anand Raman]

A descriptive discussion of my room, low frequency setup a la Geddes, etc…here.

Best,
Anand.