# Bass arriving out of phase at the listening position

#### ernestcarl

##### Major Contributor
The room is pretty much linear - thee are some non-linearities with adiabetic effects and non-elastic behaviour of surfaces. The room does not cause GD. Just reverb or reflections.

Hmmn… I think I get your point — this might be nothing more than arguing on semantics, but I don’t agree with the statement a room cannot cause increase in GD.

#### Galz

##### Senior Member
The room is pretty much linear - thee are some non-linearities with adiabetic effects and non-elastic behaviour of surfaces. The room does not cause GD. Just reverb or reflections.
I thought the excess GD comes from how room modes are building and as a result get measured differently compared to other frequencies? So the peak could be greatly delayed?

#### NTK

##### Major Contributor
Forum Donor
The model I included in my post #111 uses an absorption coefficient of 0.2 for all boundary surfaces (i.e. walls, floor, ceiling, and the wardrobe partition surfaces).

I suspect there may be a desire to model surfaces with different absorption coefficients to simulate installing acoustic treatments. This can be done by changing the absorption coefficient from a constant into a function that depends x, y, z (and, if you are ambitious, also frequency w = 2*pi*f).

Here are the modifications to the Mathematica notebook. The modification specifies a rectangular patch of the rear wall to have an absorption coefficient of 0.7. (The Neumann boundary condition is only applicable to the boundary surfaces of the simulation volume. The interior of the volume is not affected.) The patch is defined by:
• x ≥ width of the room - 0.01, and
• 0.5 ≤ y ≤ end of the entrance area, and
• 0.3 ≤ z ≤ ceiling - 0.3

Here are the side-by-side comparison between this condition and the earlier one using a uniform absorption coefficient of 0.2.

If we have sufficient confidence in the simulation results, we can use it to target only the offending frequencies by using membrane absorbers that target only those frequencies.

#### hmt

##### Senior Member
You are right. Perhaps I didn't put it clearly: I didn't want to deal with the room mode, I wanted to cut everything below 45 Hz in order to avoid the peak but also to avoid resonances that occur at any volume a bit below that frequency and make the walls and floor of my flat vibrate as if a truck was passing by in my street. That is to remain at peace with my neighbors.

I thought a linear-phase filter would be better in that case?

A minimal-phase filter would rotate the phase a lot all over the spectrum:

View attachment 232485
Not in case of a room mode since the mode itself will rotate the phase inversely. For room modes minimum phase filters should be used.

OP

#### edechamps

##### Addicted to Fun and Learning
Forum Donor
A minimal-phase filter would rotate the phase a lot all over the spectrum

I wouldn't expect that to be audible (as long as you apply the same filter to all channels equally, of course). Feel free to ABX between a minimum-phase and a linear-phase filter if you want to make sure.

The room is pretty much linear - thee are some non-linearities with adiabetic effects and non-elastic behaviour of surfaces. The room does not cause GD. Just reverb or reflections.

I'm not sure I understand what you mean. Reverb/reflections in a room can and do cause group delay, as explained here for example. This can be verified very easily by looking at pretty much any group delay measurement from any speaker in any room.

Here are the side-by-side comparison between this condition and the earlier one using a uniform absorption coefficient of 0.2.

Amazing work. I still haven't found the time to start digging into your simulation code, but I'm quite excited already. The example you showed is already useful to me to figure out the potential effect of placing tuned bass absorbers on the back wall which is on my list of things to consider.

#### SoundGuy

##### Active Member
I'm not sure I understand what you mean. Reverb/reflections in a room can and do cause group delay, as explained here for example. This can be verified very easily by looking at pretty much any group delay measurement from any speaker in any room.
Sorry but the link is not proper science or physics to be precise. It is just nonsense.

#### SoundGuy

##### Active Member
I thought the excess GD comes from how room modes are building and as a result get measured differently compared to other frequencies? So the peak could be greatly delayed?
Yes but this isn’t group delay.

#### NTK

##### Major Contributor
Forum Donor
Sorry but the link is not proper science or physics to be precise. It is just nonsense.
Are you serious? Do you know you are in effect saying John Mulcahy doesn't know room acoustics

#### daftcombo

##### Major Contributor
Forum Donor
I wouldn't expect that to be audible (as long as you apply the same filter to all channels equally, of course).
That's what I think (after having made countless ABX of the same kind a few years ago). That's why I'm not convinced by the advice to replace the linear-phase filter by a minimal-phase filter!

OP

#### edechamps

##### Addicted to Fun and Learning
Forum Donor
I'm not convinced by the advice to replace the linear-phase filter by a minimal-phase filter

To be fair, you are going at it a bit backwards compared to most people. Most people start with a minimum-phase filter, which is usually simpler to implement (IIR vs. FIR), more efficient and has zero latency. Then they wonder if they should use a more sophisticated linear-phase filter instead. You're looking at it the other way around which is unusual.

#### SoundGuy

##### Active Member
To be fair, you are going at it a bit backwards compared to most people. Most people start with a minimum-phase filter, which is usually simpler to implement (IIR vs. FIR), more efficient and has zero latency. Then they wonder if they should use a more sophisticated linear-phase filter instead. You're looking at it the other way around which is unusual.
It depends on your understanding of acoustic physics. Minimum phase is best avoided if you care about high fidelity. Phase distortion is bad. Linear phase is always preferred. If you simplify everything to a flat or target frequency response and totally ignore phase then you won’t ever achieve high fidelity. Minimum phase is powerful but best used on individual tracks in the mix stage to clean things up - it really has no place in playback where all sound is already mixed together. That DACs offer minimum phase as an option simply shows how ignorant the industry is about acoustic physics.

OP

#### edechamps

##### Addicted to Fun and Learning
Forum Donor
It depends on your understanding of acoustic physics. Minimum phase is best avoided if you care about high fidelity. Phase distortion is bad. Linear phase is always preferred.

I don't think that's true when trying to fix issues that are themselves minimum phase (e.g. room modes). My understanding is that the best way to address minimum phase frequency response deviations is with minimum phase filters, because the original phase response of the system and the phase response of the filter cancel themselves out, and the resulting system (with the filter included) ends up back in linear phase. IIRC Toole said as much in his book as well.

Maybe that's actually what you meant by your post, in which case I agree.

If you simplify everything to a flat or target frequency response and totally ignore phase then you won’t ever achieve high fidelity.

Please provide references to serious, peer-reviewed, properly controlled studies showing that (realistic) phase response deviations are audible with real listeners in a real system with real content in a real room. So far I haven't seen any, which is why I've always been skeptical of claims like these.

Last edited:

#### SoundGuy

##### Active Member
I don't think that's true when dealing trying to fix issues that are themselves minimum phase (e.g. room modes). My understanding is that the best way to address minimum phase frequency response deviations is with minimum phase filters, because the original phase response of the system and the phase response of the filter cancel themselves out, and the resulting system (with the filter included) ends up back in linear phase. IIRC Toole said as much in his book as well.

Maybe that's actually what you meant by your post, in which case I agree.

Please provide references to serious, peer-reviewed, properly controlled studies showing that (realistic) phase response deviations are audible with real listeners in a real system with real content in a real room. So far I haven't seen any, which is why I've always been skeptical of claims like these.
It may be listener dependent. Just listen to MQA yourself or a minimum phase DAC filter. If you can’t hear the difference minimum phase filtering makes on your setup (vs linear phase) then you don’t need to worry. I have given advice. You have ignored it so far. It is your loss.

#### gnarly

##### Senior Member
Interesting thread. Super room mode calculator, NTK.

If I've gathered correctly, the bass cancellation has been entirely with regard to the main speakers, and no sub was involved (?)
But since the issue is is the 90 Hz zipcode, I'd like to think of the speakers as subs for a moment, and offer a few thoughts.
So here's a few ideas/thoughts that have helped me try to figure out rooms, and what's going on with bass.

In addition to room modes which I think of as independent of speaker/sub placement, I like to picture the idea of "virtual subs" .....which are created from mirror reflections with each and every room boundary that a sub's (presumably omni) radiation hits.
It helps me visualize 1/4 wave cancellations that occur between sub and 'virtual subs', and 1/2 wave reinforcements.
Helps separate one piece of the puzzle,..... is it a mode, or a reflection.

Another piece to remember is there will be 1/2 wave cancellations between two real subs, at every odd half cycle. (1/2, 3/2, 5/2 WL distance between ...for same signal)

A couple of 'sometimes hard to implement' measurement techniques:

The first is I make sure there's good summation between a single speaker and sub. If at all possible, i measure and do the xover work outdoors using the physical separation as planned indoors. Then I know once indoors, problems are either modes or reflections.

Hopefully the speaker and sub can be located within 1/4 WL of each other, for easier in-phase summation.
Not always possible though, or always desirable.
If subs can be located anywhere, another technique is the reciprocity principle...... put the sub in the listening position and move the mic around the room.
Which will show what the subs' frequency response will be at listening position, if it placed where the mic is.
Saves a lot of sub placement experimentation (if you can get the dang sub into the listening position Lol )
If you have a real-time dual-channel transfer measurement setup, this test really gets nice. Even better, if it has automatic delay tracking (removing any constant delay & ToF in real time) you can quickly see what phase at listening position will be too..

#### theSuede

##### New Member
I've never trusted the straight output of REW when it comes to in room phase (long term, close to steady state amplitude-to-phase calculations).
Understand me correctly here, I don't think the continuous amplitude-phase theorem is wrong! I just think that when you trust amplitude sweeps that are slow enough to include LARGE amounts of room or resonator influence, large enough to dominate the driver output direct radiation SPL, then trying to solve a null or a constructive resonance by shifting phase at the source - sound origin - is almost totally meaningless...

If you verify that two speakers/drivers are phased correctly to each other when measured really near-field, but they're wildly off when measured in far-field at equi-distance, then the only real difference between the measurements is how much of the room resonances/reflections they include in the amplitude (and hence phase) estimation.

Think of it like this: The sound SOURCES have a certain (identical!) phase response, and if you were in a perfectly anechoic room or out on a field you would measure the SAME phase response curves for each source at 1m, 2m ,3m distance (after subtracting for signal delay of course...)
The room modes on the other hand don't in any way care where the measurement point is, it they just care about the excitation points (sources). A source at point X1,Y1,Z1 in a room might excite a room resonance that gives a signal reflected that's +90 degs off compared to the source when measured from a fixed measurement point somewhere else in the room. Another source at point X2,Y2,Z2 might excite the same resonance, but with a realtive phase of -90 degs when inspected from the SAME measuring point. And since the room mode might be 70-80% of the steady state sound pressure in a mode in a room, the result is phase cancellation, even if the sources are perfectly in-phase.

#### theSuede

##### New Member
Judging by the pictures of the room, I'd be willing to bet that most, almost all of the L/R phasing problem for mid-bass would go away if you didn't have that hallway in the entrance to the room. The "left" speaker sees a much longer diagonal primary than the "right" speaker does, even if they're roughly symmetrically placed with regards to their closest back/side reflection surfaces.
But if music sounds ok, then I think trying to phase match L/R with respect to steady-state excitation (in stead of direct radiation/near field reference) could even be counterproductive. A large 90Hz tuned resonator/absorber in the rear left part of the room might maybe do some good....

#### daftcombo

##### Major Contributor
Forum Donor
I think I have not expressed myself precisely enough.

I don't have problems at Listening Position. There are room modes, and peaks in the FR, but a bit of EQ minimal-phase deasl with them gently.

But I listen mostly to electronic music, and when I stand in the corridor of my flat, I can hear a lot of bass. I put the microphone at that place and, indeed, there is something like +20dB (!!) at 40 Hz there.

I am sure my neighbours can hear it too and be disturbed, so I want to kill everything under 45 Hz, even if it doesn't sounds as good at LP.

That(s why I am using a HP filter, 96dB/oct, linear-phase, at 46 Hz.

But the bass doesn't only sounds lacking, it sounds weird on some tracks. And that echo to what @ernestcarl wrote in a previous post in which he said that a 48dB/slope sounds better. But why?

Last edited:

#### hmt

##### Senior Member
Did a 48 db slope sound better or was that only an assumption?

#### ernestcarl

##### Major Contributor
But the bass doesn't only sounds lacking, it sounds weird on some tracks. And that echo to what @ernestcarl wrote in a previous post in which he said that a 48dB/slope sounds better. But why?

Sorry, I really dunno why... though, I am a just describing what I hear with my own listening tests, too. Some of the artifacts can be visibly seen in the measurements way before something anomalous or weird becomes evidently audible in the test tracks -- using my own flawed and limited hearing apparatus (i.e. ears).

Also, when I say "worse" it's kind of subjective really -- you would have to listen loud enough, and pay attention very critically to hear it... the sound effect/artifact or whatever you want to call it is as if (perhaps) the impulse has been interrupted and there's sort of "echo" like a "whop" -- which can be exaggerated with extreme filters just for the sake of testing -- but, often it's super short and quiet enough to be practically drowned out/masked/cancelled when summed eventually anyway with an additional sub(s) and/or other speakers in a multichannel system.

*Also, there is actually a sound difference when I apply the stacked all-pass filters vs without when critically A/B listening, but in that example, I cannot say with any certainty that one is worse than the other -- only that there is indeed a small audible difference.

Last edited:

#### gnarly

##### Senior Member
I've never trusted the straight output of REW when it comes to in room phase (long term, close to steady state amplitude-to-phase calculations).
Understand me correctly here, I don't think the continuous amplitude-phase theorem is wrong! I just think that when you trust amplitude sweeps that are slow enough to include LARGE amounts of room or resonator influence, large enough to dominate the driver output direct radiation SPL, then trying to solve a null or a constructive resonance by shifting phase at the source - sound origin - is almost totally meaningless...

If you verify that two speakers/drivers are phased correctly to each other when measured really near-field, but they're wildly off when measured in far-field at equi-distance, then the only real difference between the measurements is how much of the room resonances/reflections they include in the amplitude (and hence phase) estimation.

Think of it like this: The sound SOURCES have a certain (identical!) phase response, and if you were in a perfectly anechoic room or out on a field you would measure the SAME phase response curves for each source at 1m, 2m ,3m distance (after subtracting for signal delay of course...)
The room modes on the other hand don't in any way care where the measurement point is, it they just care about the excitation points (sources). A source at point X1,Y1,Z1 in a room might excite a room resonance that gives a signal reflected that's +90 degs off compared to the source when measured from a fixed measurement point somewhere else in the room. Another source at point X2,Y2,Z2 might excite the same resonance, but with a realtive phase of -90 degs when inspected from the SAME measuring point. And since the room mode might be 70-80% of the steady state sound pressure in a mode in a room, the result is phase cancellation, even if the sources are perfectly in-phase.

For best in-room tuning, i think room modes need to be differentiated from speaker placement reflections, which need to differentiated from speaker anechoic response.
Any in-room measurement has to lump them all together.

My experience is getting anechoic right first (though speaker purchase selection or DIY), and then playing with a combination or speaker/sub placement and acoustic treatments has given the most pleasing results.
The only "room-EQ" i've used is to tame room modes.

That said, a number of folks find room correction DSP to help at their specific listening position, especially for low frequencies.
So to the extent one is willing to correct to a spot, or perhaps narrowly defined region, i think phase correction of the lumped modes/reflection/anechoic response has potential merit.

I think whether the correction to the listening position is minimum, linear, or maximum phase really doesn't matter....because all of them are bogus for anywhere other than the spot they were used to correct. All that matters imo, is did it work at the spot.

If i were edechamps, i'd first try some of the physical things i mentioned earlier, then I'd try to correct each speaker and sub individually at listening position with whatever filter complement works. Iow, DIY room correction. My next play around step too, i think. Flat mag and phase rock !!!!!

Replies
13
Views
639
Replies
3
Views
257
Replies
28
Views
2K
Replies
12
Views
603
Replies
36
Views
2K