Acourate and the unbalanced stereo setup

[UliBru]

Please send me the measured pulse responses

[jvvita]

I have uploaded two working folders. one of them contains the filters generated without any pre-filtering. The others, the filters using the allpass that is also on the folder. Both folders contain a sub-folder named test, that carries the measurement after the filter applied (in the case of the pre-filter, only a mono measurement up to 300hz to demonstrate the correction).

Filters generated with 4/4 FDW and PRC 2/2.

Thanks for your time and help.

[UliBru]

Hello Jose,

viewing the pulses I'm feeling a bit helpless. They are challenging indeed.
So you can simply load Pulse48L and calculate the excessphase by TD-Functions - Phase Extraction.
The group delay directly shows the strange result (brown curve) as you have posted already


The spikes would require filters of very high quality values with a lot of ringing. This makes no sense anyway.
So I would think about the actual setup of speakers, microphone and the surrounding environment. Are there close-up surfaces? Hard surfaces?
Is it possible to measure the speakers at another position to check if you get the same result.
Anyway it makes sense to change the setup and to check if you can get an improved behaviour.

IMHO it is not possible to improve the actual situation by filters.

[jvvita]

Thank you very much for takin the time to consider the situation. Yes, the room is asymmetric and challenging. I have tried other positions to the speakers, but the excess phase did not change much. Again, the good news is that the filters are working very well, and the setup is sounding excellent. The first correction method, using only one all pass filter solved the big dip in low bass frequency response. Knowing that I did what was technically possible, I can now sit back and enjoy the music.

Thanks again for your support, and congrats for the wonderful toolbox you built.

[jvvita]

After experimenting a lot with positioning, I was able to get better results of the bass response. Good enough so I could apply the second method. To good results.

This was the step response I got with the first method:

Time1.png (22.3 KiB) Viewed 1728 times

Now the new step response:

Time3.png (21.97 KiB) Viewed 1727 times

Any thoughts?

[Ecwl]

Congratulations!
You can see from the step responses the second method is much more coherent than the first as both left & right channels now have the same phase shifts in the low frequencies. This is different than the first method where you’re correcting one channel to realign the out of phase frequencies to the other channel but as a result the step responses are much less coherent.
I’m sure you’ll notice second method gives you much higher IACC scores than the first method too.

[Ecwl]

So Mitch Barnett just posted his first video under the Accurate Sound channel under YouTube.

One interesting thing on the video is that It shows the upcoming version of Acourate for this functionality. But instead of only targeting a specific trough due to phase cancellation between left and right speakers, Mitch recommended matching all the phase shift for better coherence of the two speakers (without specifying up to what frequency). I just tried it briefly last night and psychologically it felt like there was slightly more coherence between 100-160Hz for me. Not sure how Uli feels about this and his opinion on how high we should go to correct for this. Also some of these phase shifts can have very high Q and I’m not sure if those are worth matching. Too bad I don’t have early access like Mitch. Adding those extra phase shifts took a little more time but probably was worth the effort.