Effect of Room modes on Distortion - Is there research on active distortion reduction using destructive interference?

[Plompudu]

I have been wondering about the following:

If we assume a system with a perfectly flat magnitude response after correction based on in room measurements - will the distortions still show the distortion caused by room modes (standing wave causing peak/dip) that were previously corrected and is there a way to reduce them? (I've got the idea after reading about some people saying that they could localize subwoofers with a 80Hz crossover or lower)

As an example this is the response in my room with dual SVS SB-2000 (blue: time aligned and with a -5dB LS at 14Hz to reduce the room gain; orange: PEQed response):

(measurement done with a UMIK-I so the response below 10Hz isn't as accurate)

Would the Distortion with Harmonics at ~32Hz and ~65Hz be amplified by ~10dB?

A direct reduction would only be possible by using a lower distortion driver or reducing the level of the fundamental.
Are there any ways this could be corrected other then active or passive room treatment? If so, has there been research on active distortion reduction at the listening position using destructive interference? (if not, would I consider it as a topic for my master thesis in a few years)

 

[Digital_Thor]

The room will always be the room... Unless you change the properties of that room. All EQ can do - no matter the type - is to change the signal that you sent to the driver. Having multiple drivers can smooth out the response, but only the room can change reflections and how quickly they decay.
Multiple subwoofers can definitely smooth the response as a sum, but still only the level, not for how long it takes that level to die out.
When you measure, you also have to decide where you measure, how and if you want a smaller or bigger listening area.
This might be of interest.

 

[DVDdoug]

Would the Distortion with Harmonics at ~32Hz and ~65Hz be amplified by ~10dB?

Yes. The summation or subtraction of soundwaves (the effect of room modes) or EQ are linear. If you reduce the signal at 65Hz, the distortion components at 65Hz are reduced proportionally. Higher frequency harmonics won't be affected.

Except that speakers aren't necessarily linear and you'll generally have higher percentage distortion at higher levels so the distortion could be worse at higher levels and better at lower levels.

 

[jmdesignz2]

More people need to do acoustic room treatments. After nearly 40 years of buying this and that I finally ran REW and made some 6" thick panels and a massive corner bass trap for less than $500 total.

Basically cured any bass boom, vastly better decay and essentially cured my SBIR issues
INSANE ROI of music quality to my ears and significantly measurable difference in REW

 

[KSTR]

Would the Distortion with Harmonics at ~32Hz and ~65Hz be amplified by ~10dB?

Do you mean the harmonics of lower tones (like 11Hz) that land on 32Hz and 65Hz, or the harmonics of tones at 32Hz and 65Hz?
For the former, nothing happens as the level of the hamonics only depends on the level of the fundamental. EQ at the frequency of the harmonics has no effect.
For the latter, the distortion product of tones at 32Hz and 65Hz will be lower after room EQ, for the same reason. Excitation level of the fundamental has been reduced.

 

[NotInYourHead]

More people need to do acoustic room treatments. After nearly 40 years of buying this and that I finally ran REW and made some 6" thick panels and a massive corner bass trap for less than $500 total.

Basically cured any bass boom, vastly better decay and essentially cured my SBIR issues
INSANE ROI of music quality to my ears and significantly measurable difference in REW

How do you place those treatments?

 

[Plompudu]

Do you mean the harmonics of lower tones (like 11Hz) that land on 32Hz and 65Hz, or the harmonics of tones at 32Hz and 65Hz?
For the former, nothing happens as the level of the hamonics only depends on the level of the fundamental. EQ at the frequency of the harmonics has no effect.
For the latter, the distortion product of tones at 32Hz and 65Hz will be lower after room EQ, for the same reason. Excitation level of the fundamental has been reduced.

I'm talking about the first case where the Second/Third/... Harmonic land at the frequency of the room mode exciting a standing wave at the MLP.

Why would it only depend on the fundamental?
I would have expected the harmonic distortion at the MLP to be equally impacted by the room effects. Is this a wrong assumption and if so, why?

Also, yes I agree that EQing the fundamental down is one way to reduce Harmonic distortion and that EQing at the harmonics won't impact it.

 

[KSTR]

Why would it only depend on the fundamental?
I would have expected the harmonic distortion at the MLP to be equally impacted by the room effects. Is this a wrong assumption and if so, why?

You are completely right, please excuse my early morning brain fart ;-)
The acoustic boost applies to any output of the driver.

 

[MRC01]

Another point to keep in mind is how FR affects the distortion vs. frequency curve. If you have a dip or null at frequency F, then distortion will show a bump in distortion at F. The reason is simple: distortion is the ratio of harmonics to fundamental. If the fundamental measures smaller and the harmonics are not affected, that ratio increases. This same FR dip at F will also show a dip in 2nd harmonic at F/2, and in 3rd harmonic at F/3, for the same reasons.

It's useful to compare the FR and distortion curves to see where this is happening to avoid chasing imaginary problems.

 

[MRC01]

The room will always be the room... Unless you change the properties of that room. All EQ can do - no matter the type - is to change the signal that you sent to the driver.

Yep. This is why EQ is ineffective at correcting dips or nulls. You're just pumping more energy into a self-cancelling black hole. EQ can be effective at flattening peaks, but the room is still resonating at that frequency, you're just pumping less energy into the resonance.

 

[andyc56]

Another point to keep in mind is how FR affects the distortion vs. frequency curve. If you have a dip or null at frequency F, then distortion will show a bump in distortion at F. The reason is simple: distortion is the ratio of harmonics to fundamental. If the fundamental measures smaller and the harmonics are not affected, that ratio increases. This same FR dip at F will also show a dip in 2nd harmonic at F/2, and in 3rd harmonic at F/3, for the same reasons.

It's useful to compare the FR and distortion curves to see where this is happening to avoid chasing imaginary problems.

One useful REW trick is the "Use harmonic frequency as ref" option for distortion measurement. It adjusts the harmonic levels based on the measured system frequency response. This is useful when the system can be accurately represented as a distortion-producing device (such as a speaker or sub), followed by a distortion-free, frequency-dependent filter-like system (such as the response variation introduced by room modes). The MSO documentation has a discussion about that distortion measurement option. It shows two identical subs, driven identically and measured in a location where room modes have an effect on the computed distortion. Without the "Use harmonic frequency as ref" option, the two subs appear to have very different distortion, and with it they look quite similar.

 

[Oddball]

Distortion will rise with SPL - as will decay. You seem to have a relatively good response when PEQ-ed. If you think your distortion is still too high for your preferred listening levels, you would probably need additional subs to get there. Also if you would want to tighten the FQ response range.

Active room treatment such as ART will generally reduce the decay time and also might contribute to a smoother FQ response. It is difficult to understand what that really means, so experiencing a demo is probably the best way to get there. ART works to distribute the load amongst capable support speakers and generally works even better in larger multi-channel systems, although plenty of people are reporting solid gains for 2.1 or 2.2 setups as well.

Passive treatments always worked if implemented effectively, but they will generally not deal with distortion. That is the hardware issue, at least how I understand it.

 

[Digital_Thor]

Yep. This is why EQ is ineffective at correcting dips or nulls. You're just pumping more energy into a self-cancelling black hole. EQ can be effective at flattening peaks, but the room is still resonating at that frequency, you're just pumping less energy into the resonance.

Another reason to carefully analyze the results you are measuring and properly take advantage of multiple subwoofers.
When I first set up my 4 subwoofers, I had to find out which subwoofer that was "stealing" the lowest bass, and which one was creating a dip in my listening position. When that was done, measurements showed both deeper bass and a smoother overall response + more headroom.
As you point out, if I had just EQ'd the f*** out of my system, I would properly overdrive either drivers or amps - well in my case it just sounded weird, since I do have plenty of both. But understanding my specific setup, helped me to get more for less, by simply taking my time, and see that sometimes it does not have to be that complicated.

 

[Digital_Thor]

The room is a "system" a box - just like a speaker is a system/box - so little box in a big box - same shit but bigger. Damping lowers the Q, meaning that you get dips and peaks that are broader and way easier to handle with multiple subwoofers without that much trouble from phase. Basically we want it smoother, so that we need less to get more - easier.

 

[MRC01]

One useful REW trick is the "Use harmonic frequency as ref" option for distortion measurement. It adjusts the harmonic levels based on the measured system frequency response. This is useful when the system can be accurately represented as a distortion-producing device (such as a speaker or sub), followed by a distortion-free, frequency-dependent filter-like system (such as the response variation introduced by room modes). .

Yep. Here's an example - what those distortion graphs look like on my system.

First, REW's default distortion graph. See the big bump around 170 Hz? It's not caused by distortion. It's caused by a null (room mode cancellation). You can also see a small distortion dip at 85 Hz, since the 170 Hz null is the 2nd harmonic of that frequency.

Below is the same graph shown as dbFS absolute level instead of relative to the fundamental. Those peaks & dips are gone, though it did shift the Y scale by about 30 dB. I don't know why, but it's the shape of the curve that matters.

 

[andyc56]

In my previous post, I should have shown the two graphs of the measured distortion of two identical subs driven identically, both with and without the REW "Use harmonic frequency as ref" option. So I'll add them here. More detailed information is at the link in that previous post.

Here's the first plot of the two identical subs, driven identically, without the REW "Use harmonic frequency as ref" option. The measured distortion in percent should be nearly the same, but they are pretty far off because the room mode responses are changing the distortion spectrum in different ways. That's because the subs are in different positions.

Next, the "Use harmonic frequency as ref" option is used. This method backs out the in-room frequency response of each sub at the listening position before calculating the distortion. You can see the measured distortion for the two is now much closer with this option. They should in theory be identical. Not perfect, but much better.