Best Room Response

[QMuse]

Ok, so it seems that we actually agree that, in the 400-800Hz range, although the room may have a significant influence on the steady-state response, it is only the speaker's (anechoic) response that should be corrected.

I understood you to be arguing earlier that room correction should be implemented in the 400-800Hz range. As I understand it now, though, what you're actually saying is that only the speaker should be corrected in this range (with which I agree) - but that one way to indirectly measure the speaker is by averaging out a large number of in-room measurements.

Did I understand correctly?

Not really. Here I will be trying to demonstrate if my method of measurement is accurate enough in the 400-800Hz range. As I cannot control the room I will induce a resonance to the speaker and I will try to guess and correct that resonance based on the eventual difference between my measurements (pre and post inducing a resonance).

Assuming this will work, in the real life scenario I wouldn't know if the resonance peak is coming from the speaker or from the room as in this area both can be true: some room modes definitely still exists in the 400-800Hz range and speaker can have resonance anywhere, so it may be there as well.

Your original point (supported by @Willem) was that the resonance in this range (coming from either source, room or speaker) cannot be measured and thus effectively corrected because of short wavelength, so it doesn't make sense to apply any EQ in this range.

Let's now see how this went..

This is the initial measurement of my left speaker taken from my LP. I used MMM RTA pink noise and I took 59 samples over the horizontal area of 70cm x 40cm at my ear height. I applied 1/6 smoothing.

Then I created this filet in rePhase and introduced it into the convolution chain. As you can see the filter is applied at 650Hz, gain +2dB with Q=4. I hope we'll agree that this resonance is far from spectacular. I'm creating resonance in the speaker because I don't have knowledge nor means to modify my room to introduce such resonance, but in this frequency range it anyhow doesn't matter.

Now, my idea is to measure again and try to identify where resonance was introduced, identify it's parameters and correct it.

This is what I got when I re-measured:

Well hello precious, there you are:

Ok, let's now try to find with REW what we're dealing with here and try to correct it to put it down:

To my best knowledge I would estimate the center at 631Hz, Q at 4 and I would apply -1.5dB of correction to compensate for it.

I'd say my correction landed up pretty close to the real thing, wouldn't you agree?

If I was a construction engineer I would may have knowledge and the means to introduce similar kind of resonance to my room but I see no reason why test would end up differently. Let me once again remind you that your and @Willem 's original point was that it is practically impossible due to short wavelength to identify such resonances and even if identified and corrected it would serve no purpose as "if you move head few inches/step away it will be different".

But it turned out not to be so. Not only the resonance survived extensive averaging via MMM over the 70 x 40cm area with 1/6 smoothing applied but I was also able to identify it with pretty good accuracy and correct it. I think this proves my point.

 

[andreasmaaan]

Your original point (supported by @Willem) was that the resonance in this range (coming from either source, room or speaker) cannot be measured and thus effectively corrected because of short wavelength, so it doesn't make sense to apply any EQ in this range.

Sorry, no. My original point was not that a resonance here can't be measured. It was that the room should not be corrected here.

I never said anything about whether or not a change in the speaker's response could be measured by the means you're now suggesting (although I think you'll find that it can)

I'm interested to see what you find though.

 

[QMuse]

Sorry, no. My original point was not that a resonance here can't be measured. It was that the room should not be corrected here.

I never said anything about whether or not a change in the speaker's response could be measured by the means you're now suggesting (although I think you'll find that it can)

I'm interested to see what you find though.

Ok. But why wouldn't you correct the room but would correct the speaker for the same resonance in 400-800 Hz range? How would you even know what you're correcting? In the 20-400Hz range you also don't now if the peak you're correcting is coming from a speaker or from the room. And why would you care, you simply push it down.

 

[andreasmaaan]

Ok. But why wouldn't you correct the room but would correct the speaker for the same resonance in 400-800 Hz range?

IMHO, the two reasons are:

Firstly, below 300-400Hz, the room tends to be perceptually dominant, whereas in the transition region, although the room's effects will still show up in the smoothed/averaged frequency response, perceptually this will be less important than the direct sound from the speakers.

Secondly, room corrections in the transition region will be valid only in a narrow sweet spot (if at all) as the wavelengths are short relative to the normal movements of humans.

TBH, I really think this is mostly a semantic debate here To me, by definition "room correction" means a deviation from flat anechoic response to compensate for the effects of the room.

If you're using room measurements to try to build an indirect picture of the speaker's anechoic response, and then correcting that, in my book you're attempting to do speaker correction. Although not ideal, IMHO may be your best option if you don't have anechoic data on your speaker.

 

[QMuse]

IMHO, the two reasons are:

Firstly, below 300-400Hz, the room tends to be perceptually dominant, whereas in the transition region, although the room's effects will still show up in the smoothed/averaged frequency response, perceptually this will be less important than the direct sound from the speakers.

Don't get me wrong, but is this your personal opinion or there exists a proper scientific confirmation for this?
Btw, the resonance I introduced may easilly be audible, and if it is it would be equally audible if it was caused by the room.

IMHO, the two reasons are:
Secondly, room corrections in the transition region will be valid only in a narrow sweet spot (if at all) as the wavelengths are short relative to the normal movements of humans.

I have shown that correction I made for the resonance I introduced in speaker was valid throughout my single seat LP area (70 x 40 cm), Why would it be any different if that same resonance (so the same short wavelengths!) was coming from room and not the speaker (in which case the same correction would also be applied)?

As I understand it, the whole point of transition region is that response there is equally affected by room as it is by the speaker. How I see it, the difference between your position and mine is that you're trying to separate room and the speaker and I'm trying ot put equal sign between them. But again, IMHO same resonance would cause same issue in that region, no matter if caused by room or by speaker. And same correction would improve it, and I have demonstrated it can be done.

 

[andreasmaaan]

Don't get me wrong, but is this your personal opinion or there exists a proper scientific confirmation for this?

It's a good question, and the answer is ofc science But it's been a long time since I've delved into it, so I'll have to get back to you with some references (hopefully the simplified model I've used to stored this info in my head turns out not to be at fault).

 

[QMuse]

It's a good question, and the answer is ofc science But it's been a long time since I've delved into it, so I'll have to get back to you with some references (hopefully the simplified model I've used to stored this info in my head turns out not to be at fault).

It better be a good one, as I don't expect it is easy to explain why room resonance peak at 250Hz is offending to our ears but at 650Hz it wouldn't be.

Btw, as we cannot distuinguish if room resonance at 250Hz is coming from room or speakers I'm really sceptic toward the idea that we would be able to do that at 650Hz.

 

[andreasmaaan]

It better be a good one, as I don't expect it is easy to explain why room resonance peak at 250Hz is offending to our ears but at 650Hz it wouldn't be.

Btw, as we cannot distuinguish if room resonance at 250Hz is coming from room or speakers I'm really sceptic toward the idea that we would be able to do that at 650Hz.

Are you sceptical of the idea that our brains can distinguish between direct and reflected sounds in a room at say 6500Hz?

 

[QMuse]

Are you sceptical of the idea that our brains can distinguish between direct and reflected sounds in a room at say 6500Hz?

I am sceptical that we can distinguish resonance resulting in a standing wave caused by speaker vs the one caused by room. At any frequency.

And no, I don't think room can have mode at 6500Hz. Maybe at 650Hz, but at 6500Hz certainly not.

Btw, why do you think we would perceive room mode at 300Hz differently than at 600Hz? But plz don't say it's about wavelength because if that was the case I wouldn't be able to catch that resonance at 650Hz as I demonstrated.

 

[Igor Kirkwood]

Is it possible that an un-optimized spectrum is the preferred response for your room? I've experimented with maybe 2 dozen custom targets and corresponding filters with Dirac.
I keep coming back to the un-optimized response being my preferred.

If you prefer your room without EQ it's because your EQ is not perfect.

Notice it's more difficult to have better with an EQ if your room have without EQ a good sound .

In my room (without my 4 subs) and without any EQ I have this:

But with an EQ I get that

The sound is bettet with EQ even if i's not a big earing difference between without EQ

Especialy voices are smaller , and with a better color. Without EQ gives voices too warm.

MMM (Jean-Luc Ohl) mesurement distance =3,10 meter from speakers , distance between them = 2,0 metter

 

[tuga]

IMHO, the two reasons are:

Firstly, below 300-400Hz, the room tends to be perceptually dominant, whereas in the transition region, although the room's effects will still show up in the smoothed/averaged frequency response, perceptually this will be less important than the direct sound from the speakers.

Secondly, room corrections in the transition region will be valid only in a narrow sweet spot (if at all) as the wavelengths are short relative to the normal movements of humans.

TBH, I really think this is mostly a semantic debate here To me, by definition "room correction" means a deviation from flat anechoic response to compensate for the effects of the room.

If you're using room measurements to try to build an indirect picture of the speaker's anechoic response, and then correcting that, in my book you're attempting to do speaker correction. Although not ideal, IMHO may be your best option if you don't have anechoic data on your speaker.

Above which frequency can a gated measurement (@1m on optimal axis) of one's speaker in the middle of the room using a short sweep be considered accurate enough?

 

[andreasmaaan]

I am sceptical that we can distinguish resonance resulting in a standing wave caused by speaker vs the one caused by room.

But we're not talking here about a "resonance resulting in a standing wave caused by the speaker". We're talking about a resonance in the loudspeaker's anechoic response vs. a resonance caused by a standing wave. The former has nothing to do with standing waves.

And no, I don't think room can have mode at 6500Hz.

You don't see these patterns of peaks and dips in your highly averaged and/or smoothed graphs because they are narrow in bandwidth. However, there's no fundamental difference between what we call "comb filtering" at high frequencies and "modes" at low frequencies.

The difference is how we deal with them (or not), which simply comes down to (1) how audible they are and (2) how fixed the listening position is.

For head-in-a-vice listening, it might make sense to equalise to compensate for modes whose period is only 10 or 20cm. I don't believe this makes sense for normal listening, even for a single listener.

Anyway I still need to get back to you on the audibility question, which I will do.


Having said that though, that the problem concerning short period of the mode would be enough for me not to correct the room above 300 or 400Hz, regardless of the question of audibility (as I would always want at least 20cm of back-forward/side-side movement available for comfortable listening).

Btw, why do you think we would perceive room mode at 300Hz differently than at 600Hz? But plz don't say it's about wavelength because if that was the case I wouldn't be able to catch that resonance at 650Hz as I demonstrated.

The question is: what are you actually catching.

In theory at least, if you averaged out enough measurements across the whole room, then compensated for the room's loss and absorption, you could derive the speaker's power response.

I don't know if you have enough data points to draw any conclusions, but certainly possible that what you're seeing in that graph is not a "room mode", but a nonlinearity in the loudspeaker itself.

If that's the case, for the reasons I mentioned above, your EQ is not room correction.

 

[andreasmaaan]

Above which frequency can a gated measurement (@1m on optimal axis) of one's speaker in the middle of the room using a short sweep be considered accurate enough?

This involves a complex calculation, but essentially has a straightforward answer.

First, determine the resolution of your gated measurement based on the path length difference between direct sound and sound from the first reflection point. In a typical small room you might get a value for this of say a couple of hundred Hz, depending on your speaker height and measurement distance.

Second, look at experimentally-derived audibility thresholds for resonances. Olive & Toole's studies (1988 and 1997) are the standard on this AFAIK. This is where the complexity comes in, because audibility is highly dependent not only on frequency, Q, and level, but also program (and is marginally dependent on a few other factors, but I think they can be safely disregarded for present purposes).

Finally, use these data to determine how likely it is that an audible resonance could escape detection at a given frequency. Obviously, the greater the resolution, the less likely a resonance is to escape detection.

Of course, to be 100% certain that no possible audibly resonance could escape detection, you'd need an extremely high-resolution measurement (and accordingly, an unrealistically large path-length difference between direct and reflected sound). Having said that, it's clear that, as frequency increases, the likelihood of missing an audible resonance decreases.

The question in any given case is: at what frequency does the degree of potential error from a gated measurement become greater than the degree of potential error of whatever low-frequency measurement technique is available (e.g. compensated nearfield, groundplane).

 

[QMuse]

But we're not talking here about a "resonance resulting in a standing wave caused by the speaker". We're talking about a resonance in the loudspeaker's anechoic response vs. a resonance caused by a standing wave. The former has nothing to do with standing waves.



You don't see these patterns of peaks and dips in your highly averaged and/or smoothed graphs because they are narrow in bandwidth. However, there's no fundamental difference between what we call "comb filtering" at high frequencies and "modes" at low frequencies.

The difference is how we deal with them (or not), which simply comes down to (1) how audible they are and (2) how fixed the listening position is.

For head-in-a-vice listening, it might make sense to equalise to compensate for modes whose period is only 10 or 20cm. I don't believe this makes sense for normal listening, even for a single listener.

Anyway I still need to get back to you on the audibility question, which I will do.


Having said that though, that the problem concerning short period of the mode would be enough for me not to correct the room above 300 or 400Hz, regardless of the question of audibility (as I would always want at least 20cm of back-forward/side-side movement available for comfortable listening).



The question is: what are you actually catching.

In theory at least, if you averaged out enough measurements across the whole room, then compensated for the room's loss and absorption, you could derive the speaker's power response.

I don't know if you have enough data points to draw any conclusions, but certainly possible that what you're seeing in that graph is not a "room mode", but a nonlinearity in the loudspeaker itself.

If that's the case, for the reasons I mentioned above, your EQ is not room correction.

Huh, don't get me wrong but you wrote many words and yet you didn't explain why I was able to catch quite precisely the resonance I introduced in speaker's response and you offered no argument why room wouldn't be able to have the same responance in the 400-800Hz region. Room resonances are usually quite high in the 20-400Hz region but room doesn't magically give up at 300 or 400 Hz but it's ability to resonate diminishes continually as frequency raises, so it is reasonable to expect them, but with lower magnitude.

Once again, If speaker would be linear in the 400-800Hz but instead room would be causing same resonance I couldn't really be able to distinguish those two and would corrected it in the equal way but that certainly cannot be called "speaker's" correction. Yes, room resonancce is caused by speaker, but they all are - room doesn't really resonate by itself.

I think I demonstrated that room and/or speaker's resonances can be measured in the 400-800Hz with enough precision to be corrected.

Btw, you promised to quote and explain why speaker's resonance like the one I introduced would be audible but same resonance caused by the room wouldn't, but I wasn't able to find it in your 2 posts. You mentioned Toole & Olive work on audibility of the resonances but that work doesn't really differentiate between room and speaker's resonance. We are not talking here if resoancne at 650 Hz of +2dB and Q=4 would be audible or not as I could easilly made it +3dB and Q=3 which would make it even easier to be detected and corrected. We are here talking if the same resonance would be audible if it was caused by the room. Below Schroeder room resonances are certainly audible and you offered no quotation why they wouldn't be in the transition region.

Let me also comment on this what you said:

"I don't know if you have enough data points to draw any conclusions, but certainly possible that what you're seeing in that graph is not a "room mode", but a nonlinearity in the loudspeaker itself. "

I'm not sure you understood the point of my experiment. I initially posted measured reponse of my speaker as it is. Then I introduced a resonance in speaker's reposnse of +2dB at 650Hz with Q=4 via DSP and measured again. As introduced resonance was clearly visible as a difference between 2 measurements it certainl isn't "non-linearity of the speaker itself" as it didn't exist in the initial measurement but only in the measurement after I introduced a resoance.

I'm affraid we're running in circles here..

 

[Igor Kirkwood]

One precicion on my mesurement without EQ

On my active Yamaha NS 1000x mesurements are not as good as a passive Yamaha NS-1000x .
Because on the active filter QSC we have any correction , (especialy on médium loudspeaker) when we mesure without EQ.

The QSC FIR EQ work only on mesurement "with EQ"

 

[andreasmaaan]

Once again, If speaker would be linear in the 400-800Hz but instead room would be causing same resonance I couldn't really be able to distinguish those two and would corrected it in the equal way but that certainly cannot be called "speaker's" correction.

That's right, you wouldn't - because you don't have good anechoic data on your speakers! If you did, you could simply EQ them to flat (or leave them un-EQ'd if they're already flat) in the 400-800Hz region and call it a day.

As introduced resonance was clearly visible as a difference between 2 measurements it certainl isn't "non-linearity of the speaker itself" as it didn't exist in the initial measurement but only in the measurement after I introduced a resoance.

As I've explained before, without anechoic data, you don't know whether your speaker is linear in the first place, so you don't know whether your introduced resonance is making your speaker more linear or less linear.

The fact that your introduced resonance shows up in averaged room measurements is neither here nor there.

You could introduce a resonance at 5kHz and that too would show up in averaged room measurements. By your logic, this implies we should apply room correction at 5kHz. But that's obviously a bad idea: it's the direct sound that matters here.

Btw, you promised to quote and explain why speaker's resonance like the one I introduced would be audible but same resonance caused by the room wouldn't, but I wasn't able to find it in your 2 posts.

Come on, I need some time here to go through papers, books, etc.

I'm not demanding that you demonstrate why such a thing is audible

Below Schroeder room resonances are certainly audible and you offered no quotation why they wouldn't be in the transition region.

As I already asked, why are so-called comb filter effects not audible above the transition region?

Why are these effects audible in the transition region in your view, but not above it?

 

[QMuse]

That's right, you wouldn't - because you don't have good anechoic data on your speakers! If you did, you could simply EQ them to flat (or leave them un-EQ'd if they're already flat) in the 400-800Hz region and call it a day.



As I've explained before, without anechoic data, you don't know whether your speaker is linear in the first place, so you don't know whether your introduced resonance is making your speaker more linear or less linear.

The fact that your introduced resonance shows up in averaged room measurements is neither here nor there.

You could introduce a resonance at 5kHz and that too would show up in averaged room measurements. By your logic, this implies we should apply room correction at 5kHz. But that's obviously a bad idea: it's the direct sound that matters here.

Again, you're missing the point of my experiment. It is totally irrelevant how the response of my speaker looked at the beginning of the experiment. The only relevant thing is that I introduced a change of +2dB at 650Hz with Q=4 and that same change was visible in the repeated measurement.

It makes no sense to further discuss this if you are not willing to accept the implications of this deomstration.

 

[QMuse]

You could introduce a resonance at 5kHz and that too would show up in averaged room measurements. By your logic, this implies we should apply room correction at 5kHz. But that's obviously a bad idea: it's the direct sound that matters here.

Yes, it would show. But in the case of 5khz we would be certain it is not coming from the room but from the speaker, as at 5kHz room is showing only absorption effects.

At 500Hz we cannot be sure as it may be coming from the room as weell, but frankly, I don't think it is important for the final effect once you EQ-it. My point was to demonstrate 300-800Hz region can be measured and corrected with MMM method in spite of the short wavelengths.

I did it because it is widely accepted that region below 300 should be corrected as it is room dominated. It is also widely accepted that region above app 900Hz should also be corrected, based on anechoic or pseudo anechoic gated measurements) as it is speaker dominated. It seemed to me there was no consensus about what to do with 300-900Hz region as it is some kind of dead-lands or nobody's territory because of mixed room and speaker's influence. What I demonstrated is that tis region can be succesfully measured with MMM using suffiecient number of samples which allowed for precise correction, no matter if the non-linearity in this region is coming from the room or from the speaker.

 

[andreasmaaan]

The only relevant thing is that I introduced a change of +2dB at 650Hz with Q=4 and that same change was visible in the repeated measurement.

But I agree with this 100%.

I also point out that exactly the same would be true of a resonance introduced at 5kHz, or indeed any frequency at all.

What makes 800Hz different from 5kHz in your opinion?

But in the case of 5khz we would be certain it is not coming from the room but from the speaker, as at 5kHz room is showing only absorption effects.

What do you mean? At 5kHz there is absorption, reflection, and transmission. Same as any frequency.

 

[andreasmaaan]

I did it because it is widely accepted that region below 300 should be corrected as it is room dominated. It is also widely accepted that region above app 900Hz should also be corrected, based on anechoic or pseudo anechoic gated measurements) as it is speaker dominated.

This is the point we disagree on.

I argue that below 300Hz the steady-state response should be corrected because it is perceptually dominant.

This is not about whether the speaker or room dominate the steady-state response, but about which dominates our perception.