Why does miniDSP has a 0dB peak with Q=0.7 every time the frequency doubles?

[yllanos]

So after a couple months of having my equipment stored, I was finally able to set it up again. I decided to start from scratch and when I logged into miniDSP console, I noticed the PEQ was flat but enabled by default. Like this:

So I turned it off and went from there. However, I do have a crossover with LPF/HPF at 70Hz between my speakers and my sub and by ear I noticed there was something missing from the low end.

I did not know that in reality, miniDSP is using by default a 10-band PEQ that starts at 20Hz with a peak type filter, a Q=0.7 and a 0dB gain. Then, every time the frequency doubles, the same EQ is applied, like this: 20Hz => 40Hz => 80Hz => 160Hz => 320Hz => 640Hz => 1280Hz => 2560Hz => 5120 Hz => 10240Hz

I re-enabled the first three PEQ bands like this:

And then this somehow gave the effect I am used to hear in the low end.

So I have some questions:

1. How come a 0dB increase can have a perceivable effect?
2. Should it still be considered as "flat" EQ?
3. Why would they include this by default?
4. I won't complain, I like this effect, but I wonder what is this based on? (specially the frequency doubling part)
5. Why specifically Q=0.7?

Thanks all.

 

[staticV3]

More data required, honestly.

Does this peak with gain=0dB actually exist, or is it just placebo?

 

[JustJones]

0 gain means no boost or cut so the filter "sharpness" Q = 0.7 isn't doing anything.

From Minidsp manual

A filter has no effect if its gain is set to 0 dB.

 

[kyuu]

Yeah, a 0dB gain filter should do nothing. Besides potential placebo, is it possible some sort of processing is enabled or configured differently when the EQ is enabled even if the filters aren't doing anything?

 

[MAB]

Then, every time the frequency doubles, the same EQ is applied, like this: 20Hz => 40Hz => 80Hz => 160Hz => 320Hz => 640Hz => 1280Hz => 2560Hz => 5120 Hz => 10240Hz

Those are the default settings, all set to zero, spaced one octave apart.
And when set to zero, no effect.

Did you get yourself confused, and lead yourself to think that you can hear the effect of them when set to zero?
If so, get your microphone and prove to yourself that at 0dB they are inaudible, immeasurable, and have no effect on the sound.

 

[Ruhled]

Your ears deceive you. I've done enough measurements of a 2x4hd while developing another DSP solution that I'm sure there's zero difference between disabling and enabling with 0db. I no longer own the product so you'll have to forgive my lack of actual evidence.

 

[yllanos]

Those are the default settings, all set to zero, spaced one octave apart.
And when set to zero, no effect.

Did you get yourself confused, and lead yourself to think that you can hear the effect of them when set to zero?
If so, get your microphone and prove to yourself that at 0dB they are inaudible, immeasurable, and have no effect on the sound.

The effect is definitely there, I can hear it. But I agree I should not.

The mic idea is good, but it can be weeks or months until I could test that, since I live in apartment right next to a busy street and live with my family including small children, so it is almost impossible to get reliable results.

But even if I could, I'm not sure how should I proceed to test. What do you guys suggest me to do?

I'm just still wondering why would they implement this by default?

 

[MAB]

The effect is definitely there, I can hear it. But I agree I should not.

The mic idea is good, but it can be weeks or months until I could test that, since I live in apartment right next to a busy street and live with my family including small children, so it is almost impossible to get reliable results.

But even if I could, I'm not sure how should I proceed to test. What do you guys suggest me to do?

Surely you can. If it the PEQ set to zero is actually changing the signal then trivial to measure, even in a noisy apartment. You can measure such a change in the frequency response with a microphone. You can also measure the volume at the output of the unit, or at the output of the amp, with no need for a microphone.

You can answer your question trivially if you wish to. The test would be a simple frequency sweep, compare frequency response when inputs are set to 0dB vs. the PEQ defeated. Sweeps take ~10 seconds. Doesn't have to be blasting. Can even have background noise, like when my wife says how irritating the sweeps are while I am making measurements.
You could also get out DeltaWave, and compare down to the sample.

I do wonder why you can't measure, when measuring your speakers is the most fundamental aspect of the MiniDSP. And you now seem overcome with doubt over basic DSP, signal processing, and data. On a device that is basically designed to answer the odd question you ask.

Perhaps think about why a computer, when the numerical gain is set to '0', can manifest a change in the data. And to be clear, nothing but data at that point in the signal chain:

I'm just still wondering why would they implement this by default?

Because most people are going to load up PEQ. And zero is really zero.

 

[Joffy1780]

The effect is definitely there, I can hear it. But I agree I should not.

The mic idea is good, but it can be weeks or months until I could test that, since I live in apartment right next to a busy street and live with my family including small children, so it is almost impossible to get reliable results.

But even if I could, I'm not sure how should I proceed to test. What do you guys suggest me to do?

I'm just still wondering why would they implement this by default?

Does your PC/Mac have a microphone? You don't need a separate mic to disprove your theory, just run a sweep in REW with the built in mic.

 

[Emlin]

The effect is definitely there, I can hear it. But I agree I should not.

The mic idea is good, but it can be weeks or months until I could test that, since I live in apartment right next to a busy street and live with my family including small children, so it is almost impossible to get reliable results.

That applies to your ears as much as it does to a mic. Probably more so, as the mic has no emotional involvement.