SBIR Matters for Subs Too!

SBIR Matters for Subs Too!

By Matthew J Poes

This is a quick write-up to show the effect of SBIR on low frequencies, highlighting something I found by accident, and how I addressed it. Recently, I was testing a large 18” subwoofer in my theater and temporarily had it placed in front of my screen, which is 30” from the actual front wall. This created a scenario where I had a subwoofer out into the room (a situation some believe leads to better sub performance). I decided to use this as a case study of why subwoofers should probably not be placed out into a room and should always be placed against at least one wall.


Speaker Boundary Interference Response
What is SBIR? SBIR stands for speaker boundary interference response and refers to the interference caused by low frequency sound waves bouncing off boundaries (e.g. walls, ceilings, and floors) and reflecting against the other waves from the source, causing either peaks or dips. Because of these responses, one unique attribute of SBIR (as compared to typical modal interference) is that resulting peaks and dips will always remain constant regardless of the listener’s position. The only thing that modifies the peaks and dips are moving the speaker.

The cause of the interference has to do with the way that acoustic waves interact with each other when their phase aligns or not. Look at this graphic to help understand how interference works in perfect 0-degree and 180-degree phase to each other:

Figure 1: Interference graphic

As can be seen, when two waves are perfectly in phase (0 degrees or 360 degrees) the result is perfect constructive interference. In other words, the amplitude increases. When the two waves are completely out of phase (180 degrees) then the amplitude is decreased and the resulting sound is cancelled. Since a reflected wave travels a further distance, the phase of the wave will be different from the wave coming directly from the subwoofer. When the reflected and direct waves meet up they interfere with each other, sometimes constructively and sometimes destructively. Because low frequencies are steady state in small rooms by the time they hit our ears, we hear the bass with all the room reflections. This cancelation effect isn’t delayed, it’s a part of the fundamental signal we hear.


What does SBIR look like?
Let me first describe the setup in my room that caused this measurement. I turned off all correction, all EQ, and measured the 18” subwoofer along with my two 12” B&C Bandpass subs and my single 12” Dayton reference sub. When I saw the cancelation illustrated in Figure 2 (below), I assumed all four subs were interfering with each other and phase had been reversed. Because my front wall is a false wall, I forgot how far the subwoofer was from the front wall. I proceeded to switch each sub’s phase, singularly and in pairs, to see if that effected output measurements. While the response changed, a noticeable notch at 32hz never went away. Then I turned off the new subwoofer and the notch was gone. Of course! The 18” subwoofer measures roughly 28” long, and because of its in-room position, was actually sitting nearly 60” away from my room’s actual physical wall.

Figure 2: Uncorrected system response subwoofer front and center

Now you may be asking yourself: how did I know there was nothing wrong with the subwoofer? How did I know this was SBIR? The answer: simple calculations (but we need to keep a few things in mind). This interference effect depends on the acoustic distance to the walls, which can differ from the actual distance for a few reasons. One is that materials in the room can have an impact on the exact position of these interferences, but more commonly, the wall in our room may not be the hard barrier causing the reflection. It may be both a wall’s surface in a normal 2x4 wall (drywall on both sides) or it may be the cement outer wall in a basement room. In my case the cement front wall of the basement foundation is about 24 inches from the theater room’s inner wall. Due to some pipes my side walls are actually 36” from the cement foundation walls. The rear wall is a double stud wall with 4 layers of drywall that measures about 14” thick. All of this impacts the room’s acoustic dimensions, as compared to its internal actual dimensions. Remember, low frequency waves can travel through materials like drywall as a hot knife through butter.

Figure 3: Room simulation of SBIR

In Figure 3, I have created a simulation that seems to somewhat closely match the response I measured in my actual room. This is with the subwoofer placed dead center of the screen and with the LF source roughly 50-60”from the interior front wall. However, with a dip at 32Hz, it suggests the subwoofer is much farther from the walls, as if the room is much larger. I used the dimensions of the foundation walls instead and had a very close match. The proximity of the sub to each barrier caused a bunch of dips to all fall very close to each other. This compounded the dip and made it both very wide and very deep. Room modes caused the peak in that measurement, specifically the 70hz mode is a length axial mode in the room.


How did I fix the SBIR effect?
I was in fact able to largely fix this problem without moving the subwoofer behind the screen. I moved the subwoofer to the right side wall and oriented it to face the front wall. Why did this help? By moving the sub to a barrier, the distance between surfaces is no longer equal and the interference effects are spread out. This reduces the significance of their effect. Additionally, some of the interference effect is raised above the operating range by placing them against a wall. Turning the subwoofer toward the front wall placed it closer to the front wall, again moving the interference to a higher frequency. This is how it looked after I moved the subwoofer. To be clear, I did not change any settings in the DSP, nor did I adjust anything in the electronics. I only moved the subwoofer to the side of the room, and turned it around.

Figure 4: Uncorrected Systems Response with sub on right side wall

As you can see the response is much flatter. The dips and peaks are no longer on top of each other and the previously measured deep valley in the response is removed. Instead, the bass response is relatively flat and smooth.

Now let’s look at what a little EQ does:

Figure 5: Corrected Systems Response with sub on right side wall

As you can see, once some EQ is applied, the bass response can be made much flatter. The response falls within a +/- 5db window at nearly all frequencies relative to my intended house curve. It’s important to note that no amount of EQ could have filled in that dip in the bass response. That dip was a wide cancelation effect and the only solution was to move the subwoofer to a more optimal position. What was that more optimal position? It was pushing the subwoofer against the side wall and closer to the front wall.


Conclusion
Two things have been demonstrated in this exercise. First, subwoofers should not be placed far away from room boundaries. Second, the distance between a low frequency source and surrounding boundary surfaces should not be equal or multiples of each other. These compound cancelation effects, resulting in a large measurable trough in bass response. Hopefully this also helps to explain why full range speakers (and their typical position in a room) are perpetually suboptimal low frequency sources. The ideal placement of main speakers is rarely if ever the best location for bass reproduction. This is why I frequently say, subwoofers are audiophile. It’s a lack of subwoofers (separate LF sources which can be optimally placed in a room) which is not audiophile.

 

[Matthew J Poes]

I went ahead and asked GIK if he suggests Full or with the Limiter for the front wall. He said Full to better extend and tighten the bass.
To get a better idea say for example. I use Audyssey DEQ and as you know the bass gets boosted at levels below reference levels. Having said that the bass boost sometimes depending on content can be overwhelming to others (not for me though lol). If I use a Range Limiter it will absorb even more of the lower frequencies vs if I use Full?

I'm not really sure I follow GIK's logic that the full range would extend and tighten bass over the limiter, I would argue the opposite is true. Did you talk to Glenn? Maybe I'll email him to see what he is thinking and why?

Audyssey doesn't apply the loudness compensation in the EQ, its actually in one of the selectable options, Dynamic EQ (DEQ as you call it). If it bothers people, you can turn that off while still getting the benefits of the room correction (You may already know this, but I feel best not to assume).

The application of absorption like this doesn't change the volume of bass. Because it's called an absorber that is a really common misconception. On the forums, I often read "bass heads" who won't use bass traps because they want the bass to be as loud as possible. That isn't what bass damping does, it won't make it any less loud. It has no effect on the direct sound or initial bass, only on the reflections. As such, the overall bass level stays about the same. Because it has the same effect on the peaks and nulls, the average level around the room doesn't change. It is possible that it would be quieter at your listening position, but only because you were previously sitting in a peak. As such, the use of the range limiter won't change the volume really. What it would do is increase the damping efficiency in the critical low frequencies. SBIR is only of concern below about 300hz, and in fact, its biggest problems tend to be quite a bit lower, below 150hz.

 

[Asere]

I'm not really sure I follow GIK's logic that the full range would extend and tighten bass over the limiter, I would argue the opposite is true. Did you talk to Glenn? Maybe I'll email him to see what he is thinking and why?

Audyssey doesn't apply the loudness compensation in the EQ, its actually in one of the selectable options, Dynamic EQ (DEQ as you call it). If it bothers people, you can turn that off while still getting the benefits of the room correction (You may already know this, but I feel best not to assume).

The application of absorption like this doesn't change the volume of bass. Because it's called an absorber that is a really common misconception. On the forums, I often read "bass heads" who won't use bass traps because they want the bass to be as loud as possible. That isn't what bass damping does, it won't make it any less loud. It has no effect on the direct sound or initial bass, only on the reflections. As such, the overall bass level stays about the same. Because it has the same effect on the peaks and nulls, the average level around the room doesn't change. It is possible that it would be quieter at your listening position, but only because you were previously sitting in a peak. As such, the use of the range limiter won't change the volume really. What it would do is increase the damping efficiency in the critical low frequencies. SBIR is only of concern below about 300hz, and in fact, its biggest problems tend to be quite a bit lower, below 150hz.

Thanks, I have been communicating via email with Nick from GIK. He is one of their associates.
Yes I already knew about turning DEQ off if its bothersome but thanks for mentioning the option as you are correct some people may not know.
I am learning here because I am one that thought bass traps lower the output and or loudness. What does increase in damping efficiency mean to better grasp this? I apologize in advance for my ignorance and Thank you.

 

[Matthew J Poes]

Thanks, I have been communicating via email with Nick from GIK. He is one of their associates.
Yes I already knew about turning DEQ off if its bothersome but thanks for mentioning the option as you are correct some people may not know.
I am learning here because I am one that thought bass traps lower the output and or loudness. What does increase in damping efficiency mean to better grasp this? I apologize in advance for my ignorance and Thank you.

It's not a problem, it's a very misunderstood topic.

Let's start with an example, Imagine a suspension for a moment and think about what damping does to a suspension. The damping resists motion and this resistance effectively converts motion into heat. In space, for example, if we took a car and bounced its suspension with a spring and no damper, it would spring up and down forever, perpetually. There would be nothing to dissipate that energy that you input and so nothing to stop it. On earth, we have lots of opposing forces, but they are too slow, so we add a damper and that keeps the car from bouncing around endlessly. It doesn't limit or inhibit the up-down motion of the suspension over the bump, only the amount of bounce afterward (I'm simplifying, a very highly damped suspension would, in fact, resist initial motion and make the ride very harsh).

If you go back to my article you will see the impulse responses that I did at 1/3 octave bands. These band limited impulses let us look at the "bouncing" motion of the sound wave in the room at a given frequency. The bouncing over time is seen in the tail of the impulse, and this is what we want to damp. You can even test the truth of what I'm saying by looking at impulses bandwidth limited after LF damping has been applied. You will see that on average around the room, the impulse peak doesn't change, just the length of the tail. That bouncing motion is, in this case, the amount of energy contained in the wave as it continues to bounce around the room. If a room had a medium for sound to transfer in that did not have any damping property at all, the sound would bounce around forever. In real rooms, this isn't what happens. First, some of the sound simply passes through the walls. Especially bass. Some of the sound reflects. This is the sound, from a mode or SBIR standpoint that we are concerned with. The sound that escapes is sound energy that is now gone so while our neighbors and loved ones may care, we do not. For the sound that reflects, we want a scenario where the initial sound (direct sound) is preserved, but the reflected sound is diminished. We don't actually want to remove it completely as this would cause an artificial sound, but we do want to minimize the effect of these reflections to the extent possible. The direct sound is the initial bump that a car suspension feels. The reflections are the residual bouncing, and we want to damp them by lessening them and dissipating them as heat. That is what LF damping does.

Efficiency, in this case, refers to the amount of LF energy that can be dissipated to heat for a given square foot of material. The more energy that is dissipated per square foot the more efficient it is. Because bass is so energetic and the waves so long, absorbing it efficiently is critical. We can't cover every inch of our room in very thick insulation, so efficiency is helpful.

I personally think (and have found in my room measurements) that most domestic spaces have enough absorption in the mid and high frequencies. Not always true for sure, and I would rather add it in manually myself, but the truth is, most rooms have curtains, carpet, couches, etc. which all absorb at mid and high frequencies. That mixed with their small acoustical size means the RT60 of most rooms is under .5 seconds and this is low enough for the most part. The problem with most rooms is that the RT60 is not flat, it rises dramatically to over 1 second (sometimes over 3 seconds) by 100hz or so. A look at the slope of the waterfall (or the Schroeder integral I showed above) shows that the bass decay is very shallow compared to higher frequencies. This is normal, but it's not desirable. You want to increase the decay of bass as much as possible. This is why I advocate range limited. The range limiter doesn't eliminate the mid and high-frequency absorption, just reduces it. Since rooms need more bass, this helps tilt things in that direction.

 

[Asere]

It's not a problem, it's a very misunderstood topic.

Let's start with an example, Imagine a suspension for a moment and think about what damping does to a suspension. The damping resists motion and this resistance effectively converts motion into heat. In space, for example, if we took a car and bounced its suspension with a spring and no damper, it would spring up and down forever, perpetually. There would be nothing to dissipate that energy that you input and so nothing to stop it. On earth, we have lots of opposing forces, but they are too slow, so we add a damper and that keeps the car from bouncing around endlessly. It doesn't limit or inhibit the up-down motion of the suspension over the bump, only the amount of bounce afterward (I'm simplifying, a very highly damped suspension would, in fact, resist initial motion and make the ride very harsh).

If you go back to my article you will see the impulse responses that I did at 1/3 octave bands. These band limited impulses let us look at the "bouncing" motion of the sound wave in the room at a given frequency. The bouncing over time is seen in the tail of the impulse, and this is what we want to damp. You can even test the truth of what I'm saying by looking at impulses bandwidth limited after LF damping has been applied. You will see that on average around the room, the impulse peak doesn't change, just the length of the tail. That bouncing motion is, in this case, the amount of energy contained in the wave as it continues to bounce around the room. If a room had a medium for sound to transfer in that did not have any damping property at all, the sound would bounce around forever. In real rooms, this isn't what happens. First, some of the sound simply passes through the walls. Especially bass. Some of the sound reflects. This is the sound, from a mode or SBIR standpoint that we are concerned with. The sound that escapes is sound energy that is now gone so while our neighbors and loved ones may care, we do not. For the sound that reflects, we want a scenario where the initial sound (direct sound) is preserved, but the reflected sound is diminished. We don't actually want to remove it completely as this would cause an artificial sound, but we do want to minimize the effect of these reflections to the extent possible. The direct sound is the initial bump that a car suspension feels. The reflections are the residual bouncing, and we want to damp them by lessening them and dissipating them as heat. That is what LF damping does.

Efficiency, in this case, refers to the amount of LF energy that can be dissipated to heat for a given square foot of material. The more energy that is dissipated per square foot the more efficient it is. Because bass is so energetic and the waves so long, absorbing it efficiently is critical. We can't cover every inch of our room in very thick insulation, so efficiency is helpful.

I personally think (and have found in my room measurements) that most domestic spaces have enough absorption in the mid and high frequencies. Not always true for sure, and I would rather add it in manually myself, but the truth is, most rooms have curtains, carpet, couches, etc. which all absorb at mid and high frequencies. That mixed with their small acoustical size means the RT60 of most rooms is under .5 seconds and this is low enough for the most part. The problem with most rooms is that the RT60 is not flat, it rises dramatically to over 1 second (sometimes over 3 seconds) by 100hz or so. A look at the slope of the waterfall (or the Schroeder integral I showed above) shows that the bass decay is very shallow compared to higher frequencies. This is normal, but it's not desirable. You want to increase the decay of bass as much as possible. This is why I advocate range limited. The range limiter doesn't eliminate the mid and high-frequency absorption, just reduces it. Since rooms need more bass, this helps tilt things in that direction.

It's more clear now. Thanks for explaining. I think I'll call Glenn and ask why his rep is suggesting Full vs Limiter bass trap.

 

[Asere]

I have another question. With measurements you can see what the room is doing but does it matter if you can't hear the difference? If you are in a hall or a room without much furniture I am sure you can tell but let's say you can't tell like in a family room that has furniture. Since the traps don't change the sound and only damper how will the person be able to tell other then with measurements? And if measurements call for traps what would be the purpose of placing traps when you can't tell and sound does not change? I hope I make sense in what I am asking and I apologize if I cause confusion with my questions.

 

[Eric SVL]

I have another question. With measurements you can see what the room is doing but does it matter if you can't hear the difference? If you are in a hall or a room without much furniture I am sure you can tell but let's say you can't tell like in a family room that has furniture. Since the traps don't change the sound and only damper how will the person be able to tell other then with measurements? And if measurements call for traps what would be the purpose of placing traps when you can't tell and sound does not change? I hope I make sense in what I am asking and I apologize if I cause confusion with my questions.

Your question makes sense to me - you want to make sure you're making decisions and spending money on real, tangible improvements.

 

[Matthew J Poes]

I have another question. With measurements you can see what the room is doing but does it matter if you can't hear the difference? If you are in a hall or a room without much furniture I am sure you can tell but let's say you can't tell like in a family room that has furniture. Since the traps don't change the sound and only damper how will the person be able to tell other then with measurements? And if measurements call for traps what would be the purpose of placing traps when you can't tell and sound does not change? I hope I make sense in what I am asking and I apologize if I cause confusion with my questions.

I totally missed this.

It is a fair question. However the issue is not so much can you hear it. IF you can’t hear it, then no it doesn’t matter. Now just because you can’t hear it doesn’t mean everyone can’t. Some things require training and experience to recognize (and like a small stain on your white shirt, once you hear it you can’t ignore it).

I think you have to be careful about interpreting a steady state measurement as what you hear. Just because you can’t measure it doesn’t mean it can’t be measured nor that you can’t hear it. A mic doesn’t work like our ears nor are measurements a great approximation of what we hear or how we perceive it. That doesn’t in any way invalidate measurements, you just have to know how to use them and how to best replicate what we hear. We do hear in the amplitude and time domains, yet a steady state measurement is capturing essentially one point in space and one point in time. That’s why we need to use impulse responses to examine temporal variation (like wavelets) and spatial measurements to capture more of what our ears are really hearing.

 

[RichardDolan]

Thanks a lor for infobmation,guys

 

[bsinger]

with large drivers and/or ports pointing towards typical walls, how close is usually too close?

 

[Matthew J Poes]

with large drivers and/or ports pointing towards typical walls, how close is usually too close?

So what happens is that at some point the proximity to the wall will start to load the driver and port and that will change the response. My opinion is that 2” is a minimum and maybe 6” is a maximum. By 6” it won’t load it in a meaningful way.

The loading isn’t a bad thing necessarily, but it can cause some resonances and possibly some bad noises. It’s all a balancing act. Loading could also lower distortion and increase output.

If you can measure, you could always place the mic near the sub and wall and do a sweep. That should help you see what’s going on.

You don’t have to point the sub at the wall. You can point it along the length of the wall.

 

[Mike-48]

Matthew, A very informative article, which I'm reading years later, but still it's relevant.

I never thought about SBIR as a consequence of the sub's cabinet dimension and orientation. Mine's woofer is about 16" from the wall, which puts the SBIR frequency at about 211 Hz. I'll have to turn it 90 degrees (which will put the center at about 8" from the wall), remeasure, and see if I detect a difference in FR.

Thanks again for this!

Added 10 mins later: Since the SBIR frequency is over an octave above the crossover frequency, I don't expect SBIR to have practical consequences in my case! Obviously, it's time for bed for me.

 

[tamas juhasz]

I very much agree with you. It is my opinion that when folks indicate finding the bass to integrate poorly or lack the quality they get from just a full range speaker, that this is caused by one of two things:

1. A hole in the bass response caused by SBIR giving the false perception of cleaner/tighter bass
2. Poor integration due to improper setup

What I never agree with is the notion that even the best full range speakers somehow have superior bass to that of separate subs. The biggest factor effecting bass sound quality is location and the optimal location for mains is not the optimal location for subs, that location imparts a significant amount of distortion. This also means that it makes no sense for someone to operate main speakers down very low and use subs only for the lowest frequencies and LFE channel. The most important range for a separate subwoofer to operate is actually between 50hz and ~150hz (room FS). Operating a sub below 50hz only and using the mains above that point is a bad idea (and this is not my opinion but something that is fully supported by science).

Hi Matthew . What do you think about open baffle subwoofers . Do they have a similar or different SBIF then enclosed subwoofer . They are recommended to be pulled away from walls .

 

[natty]

Hi Matthew . What do you think about open baffle subwoofers . Do they have a similar or different SBIF then enclosed subwoofer . They are recommended to be pulled away from walls .

Why would they not have the same SBIR issues? They are producing sound and the boundaries are still present.