Having troubles integrating subwoofer with REW

[stigger]

So, I've been planning to optimize my setup with REW for a while, have read a ton of tutorials and watched many YouTube guides. I kind of have the basic understanding of the theory by now, but once I started actually trying to do something, I immediately encountered a problem while integrating a sub, which I could not resolve myself. I'll try to show my a bit simplified process.

First, my initial measurements.

To me this looks like it's OK to do a crossover anywhere between 40 and 100 Hz, so I'll pick the standard 80 Hz for this experiment. I don't really have a feeling for different crossover types, so I'll somewhat arbitrary pick 4th order Linkwitz-Riley.

I don't see anything suspicious here, so I move on to the time/phase alignment.

Looking at these graphs, I decide to add a 39.5 ms delay to the loudspeaker. Next, phase alignment.

At 80 Hz the phase difference is ~21.2 degrees, 21.2 / 360 / 80 * 1000 = 0.736 ms, resulting in a total delay of 40.236ms.

Let's do the final measurement and look closer at the result.

After aligning the sub I expect a better frequency response, so the new dips at 68 and 91 Hz make me think that I'm missing something or doing something wrong. I tried a bunch of other crossover frequencies as well, but the result is always the same: new dips appear as soon as I add the delay. Or maybe they are expected, so I should just redo the whole process for a bunch of different crossover frequencies and choose the dips I like the most?

Here are all the measurements, if anyone wants to take a closer look. Any guidance is appreciated.

 

[Purité Audio]

40 millisecond delay?
How far away are the subs from the mains 40feet?
Keith

 

[ppataki]

So, I've been planning to optimize my setup with REW for a while, have read a ton of tutorials and watched many YouTube guides. I kind of have the basic understanding of the theory by now, but once I started actually trying to do something, I immediately encountered a problem while integrating a sub, which I could not resolve myself. I'll try to show my a bit simplified process.

First, my initial measurements.

View attachment 425709

To me this looks like it's OK to do a crossover anywhere between 40 and 100 Hz, so I'll pick the standard 80 Hz for this experiment. I don't really have a feeling for different crossover types, so I'll somewhat arbitrary pick 4th order Linkwitz-Riley.

View attachment 425710

I don't see anything suspicious here, so I move on to the time/phase alignment.

View attachment 425711
View attachment 425712

Looking at these graphs, I decide to add a 39.5 ms delay to the loudspeaker. Next, phase alignment.


View attachment 425713

At 80 Hz the phase difference is ~21.2 degrees, 21.2 / 360 / 80 * 1000 = 0.736 ms, resulting in a total delay of 40.236ms.

View attachment 425714

Let's do the final measurement and look closer at the result.

View attachment 425716

After aligning the sub I expect a better frequency response, so the new dips at 68 and 91 Hz make me think that I'm missing something or doing something wrong. I tried a bunch of other crossover frequencies as well, but the result is always the same: new dips appear as soon as I add the delay. Or maybe they are expected, so I should just redo the whole process for a bunch of different crossover frequencies and choose the dips I like the most?

Here are all the measurements, if anyone wants to take a closer look. Any guidance is appreciated.

Just use acoustic timing reference measurements in REW, let's say your front left speaker being the reference
Then do a Left->Left, Left->Right, Left->Sub measurement, take a note on the Delay value under the Info button and then calculate the delay by subtracting the biggest number from each measured delay value (this is all without crossover)
Then engage your crossover, LR 24dB/octave at 80Hz and measure with the sub having normal polarity and then with the sub having inverted polarity
This is what I always do (even when I have multiple subs) and for me it always works like a charm

 

[ZolaIII]

Flip the phase, adresa the first peak (44 Hz room mode) with PEQ by hand and mesure with it in place. Now you can see natural decay in the room to do orders of crossover slopes and do it to Butterwort model.

 

[stigger]

Thanks for the replies. The sub stands right next to the speaker and I was using the acoustic time reference, so I also was surprised by the large group delay, but didn't know what to do about it. I don't know what exactly was causing it, but after reducing the 44 Hz peak and re-matching the levels I don't see it anymore. I ended up with the 1.622 ms delay, and the result looks as expected even with the PEQ removed. Now the measurements make sense, so I can continue experimenting. Thanks again!

Haven't tried reversing the polarity yet, by the way.

 

[stigger]

I ended up with the crossover at 82hz and FIR filters made according to this guide (1/A of a minimal phase version, up to 225hz). Left channel seems good, the right is quite messy, but I don't think anything can be done about that due to the room geometry and lack of options for speaker positioning. The much flatter bass sounds, well, flat, so I'm experimenting with some house curves, but since I don't know what else could be done (aside moving to another apartment), I mostly consider the process to be over.

 

[ppataki]

Try a low shelf filter at 75Hz, Q=1
Gain to taste, but min. 3dB

 

[stigger]

Try a low shelf filter at 75Hz, Q=1
Gain to taste, but min. 3dB

Why 75Hz specifically? Is this a more or less arbitrary frequency, or was it deliberately chosen to be below the crossover frequency?

 

[ppataki]

Why 75Hz specifically? Is this a more or less arbitrary frequency, or was it deliberately chosen to be below the crossover frequency?

It simply sounds good
But feel free to play around with it anywhere between 100Hz and 50Hz
I find that if I set it closer to 100Hz bass tends to get boomy whereas if I set it lower I miss the 'juice'