Using REW to verify AL correction

[Iainmc]

The filters created in AL are used in the Jriver convolver and everything is working fine. When I do a measument of left and right chanels combined at the listening position I get this result,

The amplitude resonse is very close to the simulated resonse in AL but the phase response is not as expected, minimum phase. and the group delay is off.

If I apply a frequency dependat window, REW default, I get this result. Changing the window size still gives similar results.

The phase result is now as expected but the freqency response has worsened. The group delay is now as expected.

So the question is, what is the correct processing of the raw measurment to verify that the AL corrections are working as intended?

 

[juicehifi]

I would disregard phase and group delay, and look at the frequency response and the shape of the impulse … and perhaps the step response too.

 

[Iainmc]

Thank you for your reply.
Can you please explain why the phase and group delay measurments should be disregarded? Just trying to understand.

 

[juicehifi]

Phase displays such as the one above does not tell me a lot about the time domain behavior. This is my personal opinion, others may disagree.

I use a combination of group delay and magnitude to spot for problem frequencies, though. Group delay peak and magnitude dip at the same frequency = acoustic trouble, usually not worth trying to time correct.