L/R channel Room Correction with subwoofer using REW

[slartibartfast]

This is the REW "Overlays" view, using the "Phase" tab - useful in this instance as it allows you to compare phase responses of multiple measurements.

OK so I can see the phase difference between 100-200Hz. How do I find the cause? At first I thought maybe reflections from a TV screen just outside the right speaker but covering the TV with a blanket made no difference at all.

 

[canard]

so observe the group delay can be more explicit ;-)

 

[slartibartfast]

so observe the group delay can be more explicit ;-)

This is with no smoothing

 

[canard]

1/12 or 1/6...
+-50ms or less scale max ,40/400hz
clean at 40hz (25ms)
?

your scales don't really allow reading
;-)
(your phase slope looks pretty normal in your previous image)

 

[dominikz]

At first I thought maybe reflections from a TV screen just outside the right speaker but covering the TV with a blanket made no difference at all.

Unfortunately at those frequencies the blanket won't have any effect at all because the wavelength is around 2m - the sound will simply pass through the blanket unaffected.

OK so I can see the phase difference between 100-200Hz. How do I find the cause?

I'd suggest to move one of the loudspeakers around and measure how the phase changes for each position.
Second thing I'd try is to measure the speakers full-range, without the sub.
Third thing you could try, if possible, is to at least temporarily set the speakers symmetrically in the room - with identical distances to adjoining walls and major obstacles.
In each case you should measure the responses and see how the phase is affected - that might in the end give you some clues as to what exactly might be causing the phase mismatch at your main speaker placement.

 

[slartibartfast]

Third thing you could try, if possible, is to at least temporarily set the speakers symmetrically in the room - with identical distances to adjoining walls and major obstacles.

Easier said than done . It would be relatively easy to temporarily remove the TV completely. I could also swap the speakers around to make sure it is room related. Main issue is I can only experiment for a short time when the house has gone to bed

 

[slartibartfast]

Unfortunately at those frequencies the blanket won't have any effect at all because the wavelength is around 2m - the sound will simply pass through the blanket unaffected.

I'd suggest to move one of the loudspeakers around and measure how the phase changes for each position.
Second thing I'd try is to measure the speakers full-range, without the sub.
Third thing you could try, if possible, is to at least temporarily set the speakers symmetrically in the room - with identical distances to adjoining walls and major obstacles.
In each case you should measure the responses and see how the phase is affected - that might in the end give you some clues as to what exactly might be causing the phase mismatch at your main speaker placement.

I made phase measurements then swapped the speaker positions to compare the results.

You can see the 180 degree phase difference around 150Hz


After swapping the speakers it looks the same so the phase change is definitely down to the room


I made two measurements with the speakers in each position and comparing the two sweeps of a speaker on the left side I see this. What could cause the 360 degree phase change between around 350-600Hz?

 

[dominikz]

I made phase measurements then swapped the speaker positions to compare the results.

You can see the 180 degree phase difference around 150Hz

After swapping the speakers it looks the same so the phase change is definitely down to the room

Good job! I agree, it appears to be down to placement of speakers in room.

I made two measurements with the speakers in each position and comparing the two sweeps of a speaker on the left side I see this. What could cause the 360 degree phase change between around 350-600Hz?

Given the fact that it appears to be exactly 360° in the entire range, and the shape of the discontinuities (especially at ~350Hz) I suspect it is an artefact of the REW phase unwrapping function.