Best Room Response

[QMuse]

Thank you for the data.

The difference is not spectacular.

The difference is in margin of error. RT60 is not dependent of amplitude correction.

 

[QMuse]

Well, I prefer having a flat frequency response in the listening window of my speakers, calibrated by the manufacturer in anechoic conditions, than to try to blindly guess what may be close using measurements from the listening position.

Speaker that has flat LW response in anechoic conditions would have constantly falling flat response in a room, like famous Harman target curve. Your measured in room response is nothing like that. Don't get me wrong here, If that is what you like it is certainly fine with me.

I don't know if this is better. I think that this is more like "having the musicians playing in the room" while equalizing from the listening position would rather be like "being transported into the concert hall".

I'm not a fan of classical music and I don't go to concert hall, so I can't really imagine what you mean,

 

[Pio2001]

Speaker that has flat LW response in anechoic conditions would have constantly falling flat response in a room, like famous Harman target curve. Your measured in room response is nothing like that.

Above 800 Hz, my measured response IS the response of a flat speaker (KH-120) in a room (except for a small 20 kHz bump, and the treble trim set to -1, active from 8 kHz (0 dB) to 20 kHz (-1.5 dB)).

 

[andreasmaaan]

3-how to know if local variations in the frequency response caused by the acoustics (variations in wall absorptions leading to irregular RT60 across the spectrum, SBIR...) should be corrected through increasing the total energy (filling the gaps), decreasing the total energy (removing resonances), or keeping the total energy constant ?... which relates to question 1, but here the question is about the general shape of the ideal target curve, rather than about the general slope.

I'm not sure I understand the question. Assuming that clipping in your electronics is avoided, are all three approaches not functionally equivalent?

Say we want to make the following frequency response flat. We can introduce either a +6dB low shelf or a -6dB high shelf:

So long as clipping is avoided and SNR is kept within acceptable limits, what's the difference?

 

[Pio2001]

I'm not sure I understand the question. Assuming that clipping in your electronics is avoided, are all three approaches not functionally equivalent?

Say we want to make the following frequency response flat. We can introduce either a +6dB low shelf or a -6dB high shelf:

That's right. In this example, the two options are equivalent.

But let's consider a more ambiguous example :

Here, let's say that the measured curve is the light blue one.

The question is to know if the right correction is :
-the blue one, an extreme example where any peak is completely flattened
-the green one, we consider the general slope of the light blue curve above 300 Hz as a reference, and we decide that this reference is supposed to be valid down to 100 Hz
-the red one, we keep the total energy unchanged, filling dips as much as cutting peaks.

Which is the same question as "what should be the shape of the target" ?

And this example is an easy one. In real life, it's difficult to see something consistent between 300 and 1000 Hz on which we can rely in order to draw our target curve between 50 and 200 Hz, like here.

 

[andreasmaaan]

That's right. In this example, the two options are equivalent.

But let's consider a more ambiguous example :

View attachment 60666

Here, let's say that the measured curve is the light blue one.

The question is to know if the right correction is :
-the blue one, an extreme example where any peak is completely flattened
-the green one, we consider the general slope of the light blue curve above 300 Hz as a reference, and we decide that this reference is supposed to be valid down to 100 Hz
-the red one, we keep the total energy unchanged, filling dips as much as cutting peaks.

Which is the same question as "what should be the shape of the target" ?

And this example is an easy one. In real life, it's difficult to see something consistent between 300 and 1000 Hz on which we can rely in order to draw our target curve between 50 and 200 Hz, like here.

Ah thanks, I see now.

FWIW, I think there are two important points to consider here:

1) Perceptually, the speaker/room system is predominant in the room's modal region/below 200Hz, as opposed to the anechoic response of the speaker.

2) The research tends to suggest that the modal region is one where personal preferences tend to diverge (in contrast to the midrange/treble where preference tends to be far more convergent).

Due to (1), IMO the general slope of the unequalised speaker in the room is not a useful guide/starting point. So I wouldn't focus on the question whether you're adding to or subracting from the unequalised response, and instead focus on your in-room target.

Due to (2), it's at least arguable that there is no "correct" (ofc this comment cannot apply to the production side; I speak here only of reproduction). So again I'd say: focus on smoothness as a universal goal, and then tilt to personal taste (and in some cases program).

 

[QMuse]

I don't know if this is better. I think that this is more like "having the musicians playing in the room" while equalizing from the listening position would rather be like "being transported into the concert hall".

So, with this target curve you're saying you're having "musicians in the room"?

Your filters affects mostly LF (as it should) and LF doesn't really contain any spatial information so I'm not sure how would this, or any other room EQ made only in the 20-400Hz region, bring "musicians in the room".

 

[Pio2001]

So, with this target curve you're saying you're having "musicians in the room"?

Your filters affects mostly LF (as it should) and LF doesn't really contain any spatial information so I'm not sure how would this, or any other room EQ made only in the 20-400Hz region, bring "musicians in the room".

You seem to imply that without room EQ, the musicians would not be in the room. Why not ?

The speakers are dead neutral. The musicians are in the room without EQ. All the stuff that is visible on the curve would affect the sound coming from a guitar, a violin or a human voice exactly as it does with the sound coming from the speakers.

If a double bass player was standing where my speakers are standing, the building would collapse as soon as he touches the strings of his instrument, because of the +18 dB room mode

Also, the details of the curve can vary. Here is an MMM measurement of my current correction, taken in november 2018 :

And another one of the exact same correction, taken in april 2019 :

It's always like that. There are differences that depends on the exact position of the furniture in the room, the position of the computer screen, if the doors are close or open, if there are pillows on the sofa etc.

And these are MMM, which is already a simplification of reality. Here are some multipoint measurements around one seat only, with an older correction, that show the variation across +/- 30 cm of the sweet spot, that are smoothed out in the above curves.
Each individual curve is the frequency response at a different fixed point. The sound that reaches my ears is described by one of these curves, and it changes from one curve to the next everytime I move my head.

 

[QMuse]

You seem to imply that without room EQ, the musicians would not be in the room. Why not ?

Not at all, I thought you said your room EQ brought them into your room. As this thread is about room response and room EQ I thought you said that in that context.

Also, the details of the curve can vary. Here is an MMM measurement of my current correction, taken in november 2018 :

View attachment 61047

And another one of the exact same correction, taken in april 2019 :

View attachment 61046

It's always like that. There are differences that depends on the exact position of the furniture in the room, the position of the computer screen, if the doors are close or open, if there are pillows on the sofa etc.

Frankly, i don't believe in oddly shaped response curves like yours. As Toole explained in his book I think most people would prefere smoothly falling curve and not the shape like yours. Besides, once you eliminate room modes linear speaker would have smoothly sloped in-room response, definitely not similar to yours.
But ok, if this is what rocks your boat I'm of course fine with that.

 

[thewas]

The speakers are dead neutral.

Sorry if I missed it, but what loudspeakers are they?

 

[Igor Kirkwood]

Frankly, i don't believe in oddly shaped response curves like yours. As Toole explained in his book I think most people would prefere smoothly falling curve and not the shape like yours. Besides, once you eliminate room modes linear speaker would have smoothly sloped in-room response, definitely not similar to yours.
But ok, if this is what rocks your boat I'm of course fine with that.[/QUOTE]

Not at all, I thought you said your room EQ brought them into your room. As this thread is about room response and room EQ I thought you said that in that context.




Frankly, i don't believe in oddly shaped response curves like yours. As Toole explained in his book I think most people would prefere smoothly falling curve and not the shape like yours. Besides, once you eliminate room modes linear speaker would have smoothly sloped in-room response, definitely not similar to yours.
But ok, if this is what rocks your boat I'm of course fine with that.

Do you agree this curve QMuse ?

 

[QMuse]

Frankly, i don't believe in oddly shaped response curves like yours. As Toole explained in his book I think most people would prefere smoothly falling curve and not the shape like yours. Besides, once you eliminate room modes linear speaker would have smoothly sloped in-room response, definitely not similar to yours.
But ok, if this is what rocks your boat I'm of course fine with that.

Do you agree this curve QMuse ?View attachment 61099[/QUOTE]

This looks very good! I would probably try to push up a little (3-4 dB) that wide dip at L channel in the 100-180Hz range, but it also looks good to me like this.

 

[Pio2001]

Sorry if I missed it, but what loudspeakers are they?

It's a pair of Neumann KH-120 A. Not yet tested by Amir.
Here is a third-party measurement : https://www.kenrockwell.com/audio/neumann/kh-120-a.htm

 

[Igor Kirkwood]

Do you agree this curve QMuse ?View attachment 61099

This looks very good! I would probably try to push up a little (3-4 dB) that wide dip at L channel in the 100-180Hz range, but it also looks good to me like this.[/QUOTE]
Thank You.

About 3-4 dB around 100/180 Hz perhaps . But notice than the EQ is the same on Left side and Right side, the result is better , if we try to change we lose this point of quality.

This curve is the best curve for jazz because the 4 subs are mono. For Classical Music a "Griesinger curve" with more L-R on subs sound better , but its not so nice !

For both curves (jazz and classical music ) thE EQ of loudspeakers is the same ( L=R) . Change only the bass management on subs.
You can notice than the positive EQs are smalls (+1 or +2 dB maxi +3,5 dB at 420 Hz).

 

[Pio2001]

Besides, once you eliminate room modes linear speaker would have smoothly sloped in-room response, definitely not similar to yours.

In a well-treated room, with optimal speaker placement, maybe. But my room and speaker placement are far from optimal. The main problem being a large window just behind the speakers.
The acoustic absorption of glass is even inferior to concrete's. 99% of the sound bounces on the window, and because of this, and also because of the very small size of the speakers that makes them omnidirectional at higher frequencies than usual, the SBIR interferences are high, with a wide low part around 280 Hz (drivers and surface are 80 cm away, that's the highest recommended value by the user manual). It can't be completely corrected because the frequencies are already too high for an accurate compensation.

The odd shape of the curve is thus not a personal choice, but the natural consequence of a suboptimal speaker placement.

The part that is a personal choice is definitely the 35 - 80 Hz range, that is lower than it should, and that is slightly rising.
From 250 to 800 Hz, I made the correction according to the measurements, as long as the result was acceptable. If some parts sounds overcorrected, I just left them uncorrected, and that's why the curve doesn't look completely balanced.

 

[thewas]

You seem to imply that without room EQ, the musicians would not be in the room. Why not ?

The speakers are dead neutral. The musicians are in the room without EQ. All the stuff that is visible on the curve would affect the sound coming from a guitar, a violin or a human voice exactly as it does with the sound coming from the speakers.

If a double bass player was standing where my speakers are standing, the building would collapse as soon as he touches the strings of his instrument, because of the +18 dB room mode

Also, the details of the curve can vary. Here is an MMM measurement of my current correction, taken in november 2018 :

View attachment 61047

And another one of the exact same correction, taken in april 2019 :

View attachment 61046

Its good confirmation to see that the sound power irregularity above 1,5 kHz I mentioned here shows up in your MMM and confirms why the KH120 sound too dark/polite outside of the nearfield (which is ok that, as they are nearfield monitors), see also https://www.hifi-selbstbau.de/component/content/article?id=469

 

[QMuse]

The odd shape of the curve is thus not a personal choice, but the natural consequence of a suboptimal speaker placement.

The part that is a personal choice is definitely the 35 - 80 Hz range, that is lower than it should, and that is slightly rising.
From 250 to 800 Hz, I made the correction according to the measurements, as long as the result was acceptable. If some parts sounds overcorrected, I just left them uncorrected, and that's why the curve doesn't look completely balanced.

I don't have @BYRTT graphical skills but when I pushed your sub for 3dB I got very "classical" curve with 10dB slope:

Did it sound boomy that way or you simply like less bass?

 

[Pio2001]

I don't have @BYRTT graphical skills but when I pushed your sub for 3dB I got very "classical" curve with 10dB slope:

View attachment 61318

Did it sound boomy that way or you simply like less bass?

Good question.
The global balance between the volume of low, medium and high frequencies sound better with your setting. But the bass quality sounds bad. It is disturbing to hear the bad quality of these frequencies all the time.
With my settings, there are less bass, but they sound cleaner.

I suspect that the highest peak, here at 58 Hz, stands out above the rest of the frequency response. It is annoying with techno-trance music, for example, that has a lot of energy down to 45 Hz. I think that the wild variations between 52 and 58 Hz, then 65 Hz then 75 Hz, are unpleasant to hear when they dominate all the frequency response, while they are less audible if the frequency response is lowered in this range.
It is less annoying with recordings of strings and voices, for example, that have little energy in this frequency range.

I once had the chance to compare directly a church organ concert given in my town with a CD recording made on the same organ, right after the concert, on my system.
The sound was very faithful, except that there was more bass in reality, than with my current setting, as the measurement shows. So the lack of bass is audible.
But if I compare the bad quality of the low frequencies in my installation vs the lack of tonal balance if I reduce them, I clearly prefer the second solution.

 

[A800]

I'm not sure I understand the question. Assuming that clipping in your electronics is avoided, are all three approaches not functionally equivalent?

Say we want to make the following frequency response flat. We can introduce either a +6dB low shelf or a -6dB high shelf:

View attachment 60622

So long as clipping is avoided and SNR is kept within acceptable limits, what's the difference?

Amplifier power, SPL, phase, group delay...

 

[QMuse]

Good question.
The global balance between the volume of low, medium and high frequencies sound better with your setting. But the bass quality sounds bad. It is disturbing to hear the bad quality of these frequencies all the time.
With my settings, there are less bass, but they sound cleaner.

I suspect that the highest peak, here at 58 Hz, stands out above the rest of the frequency response. It is annoying with techno-trance music, for example, that has a lot of energy down to 45 Hz. I think that the wild variations between 52 and 58 Hz, then 65 Hz then 75 Hz, are unpleasant to hear when they dominate all the frequency response, while they are less audible if the frequency response is lowered in this range.
It is less annoying with recordings of strings and voices, for example, that have little energy in this frequency range.

I once had the chance to compare directly a church organ concert given in my town with a CD recording made on the same organ, right after the concert, on my system.
The sound was very faithful, except that there was more bass in reality, than with my current setting, as the measurement shows. So the lack of bass is audible.
But if I compare the bad quality of the low frequencies in my installation vs the lack of tonal balance if I reduce them, I clearly prefer the second solution.

Interesting..

As the tonal balance was right but the bass quality was bad my first thought was that maybe your sub is not up to the task regarding sound quality so you compensated with lowering the overal bass level. Which unit are you using?