Speaker tuning to room.

[Liftro]

Hi, I have a question on my mind and I'd like to know if my reasoning makes sense. I'm not an expert, so please bear with me. The question is, if a room has a dip in a specific frequency range, and the speaker we're going to put in that room has a peak in that same frequency range, is it possible that the peak will fill the dip, and in that specific room, that specific speaker will sound flat in that specific range? Or will the result be a disaster anyway?

 

[voodooless]

The question is, if a room has a dip in a specific frequency range, and the speaker we're going to put in that room has a peak in that same frequency range, is it possible that the peak will fill the dip, and in that specific room, that specific speaker will sound flat in that specific range? Or will the result be a disaster anyway?

A dip is a null, meaning that whatever the speaker produces, the sound will be canceled. So raising the output at that frequency will not work. Now, not all nulls are created equal, and some are not perfect cancellations. Those might be fixable. But not by buying a speaker that happens to have a peak. You use EQ for that. Much more precise, and much more convenient.

 

[Arindal]

A dip is a null, meaning that whatever the speaker produces, the sound will be canceled. So raising the output at that frequency will not work.

That is the case with effects like room modes causing the null, often to be observed in smaller or medium-sized rooms with solid walls. Typical for this scenario is a very narrow-banded dip at a certain position and particularly frequency, while other positions show a peak at the same frequency (I tend to do some more measurements exactly 1/4 wavelength away from the null maximum position and oftentimes encounter peaks there). I agree that in this case filling the null or applying narrow-banded filters to increase level, will most likely result in a disaster, as the peaks also increase.

Now, not all nulls are created equal, and some are not perfect cancellations. Those might be fixable.

Such effects do exist, for example as a result of light walls acting as a parasitic resonator, i.e. unwanted bass trap. Most of such cases show a rather broad dip which stays more or less the same over several positions.

In this case, applying parametric filters usually works well. While target frequency and quality factor can easily be extradited from the measurements, the exact level of boost should be judged and adjusted by ear.

There are cases in between these two extremes, like nulls resulting from a single reflection of what appears to be a more or less even wavefront hitting a solid surface (typically found with speakers/subwoofers close to each other, placed in only moderate distance to the rear wall, like 40-90cm). In this case, only experiments would show if EQ might help.

Or will the result be a disaster anyway?

Except for the pronounced room mode scenario (which can often be tackled by changing the speaker´s positions relative to the room), EQ in most cases can help with nulls. It should be noted, though, that usually where there are nulls, there are also peaks at different positions/frequencies, and these are much more audible/annoying/non-EQable.

The one thing I have learned from setting up numerous systems and tuning by DSP, is that one should never aim for a flat target curve for its own sake. It does not translate to a balanced bass, as decay and subjective energy of different frequency bands vary.

 

[Newman]

if a room has a dip in a specific frequency range, and the speaker we're going to put in that room has a peak in that same frequency range, is it possible that the peak will fill the dip, and in that specific room, that specific speaker will sound flat in that specific range?

No, because we perceive the direct sound from the speaker distinctly from the summed sound, and it's the direct sound that we most value when it is smooth, flat and extended in frequency response.

 

[Arindal]

we perceive the direct sound from the speaker distinctly from the summed sound, and it's the direct sound that we most value when it is smooth, flat and extended in frequency response.

That is in theory correct for shorter wavelengths which translate to frequencies like 300Hz upwards.

In practice, most of nulls or narrow-banded cancellation issues I have seen, were below 300Hz, oftentimes below 150Hz. In these bands, our brain is incapable of distinguishing direct from indirect sound, as a single period is already too long.

In some cases, dips might appear in mid/high bands as well, but they are oftentimes a result of reflections not correctly summing up with direct sound (aka comb-filtering effects or alike). I agree these should not be filled using EQ filters, but for a different reason. They do come early enough to contribute to perceived early/direct sound tonality, but wavelengths are so short that the situation might change when moving microphone or ear by only 5cm.

I don´t really understand what you mean by ´smooth direct sound´, but indirect sound plays a crucial role for tonality as well, particularly when coming in early and discretely after the first wavefront.

 

[Liftro]

That is in theory correct for shorter wavelengths which translate to frequencies like 300Hz upwards.

In practice, most of nulls or narrow-banded cancellation issues I have seen, were below 300Hz, oftentimes below 150Hz. In these bands, our brain is incapable of distinguishing direct from indirect sound, as a single period is already too long.

In some cases, dips might appear in mid/high bands as well, but they are oftentimes a result of reflections not correctly summing up with direct sound (aka comb-filtering effects or alike). I agree these should not be filled using EQ filters, but for a different reason. They do come early enough to contribute to perceived early/direct sound tonality, but wavelengths are so short that the situation might change when moving microphone or ear by only 5cm.

I don´t really understand what you mean by ´smooth direct sound´, but indirect sound plays a crucial role for tonality as well, particularly when coming in early and discretely after the first wavefront.

I think he was referring to on axis response.

 

[Newman]

I think he was referring to on axis response.

Yes.....but......the reason I prefer to call it the direct sound (direct from the speaker to the listener) is that the listener is not obliged to point the speaker straight at him or her self. If the listener chooses to point the speakers straight down the room, or crossed in front of or behind the listener, then the on-axis anechoic response is NOT the one that is most critical. It is the anechoic response at whatever angle is subtended from the speaker to the listener. So I call it the direct sound.

cheers

 

[Arindal]

It is the anechoic response at whatever angle is subtended from the speaker to the listener. So I call it the direct sound.

Agreed, but I still do not understand what ´smooth´ in this regard meant. Linear?

It is clear that direct sound is important for tonality and might change depending on toe-in angle, but I want to highlight that indirect sound tonality can also be important. Ideally, both have to be linear.

 

[Newman]

Agreed, but I still do not understand what ´smooth´ in this regard meant. Linear?

Not jagged?

Basically I mean the same as the author of this post means by 'smooth', link. IMHO the word is established in the vernacular.

It is clear that direct sound is important for tonality and might change depending on toe-in angle, but I want to highlight that indirect sound tonality can also be important. Ideally, both have to be linear.

Agreed, I didn't imply otherwise. But I was answering the question in post #1: if the direct and indirect sound are non-flat in ways that 'cancel out', is everything fixed?

cheers