So I read more about what the Adjust+ tool is actually measuring (this info isn't so easy to find!), and I was initially mistaken. It's measuring the phase of voltage induced in the adjacent channel relative to the original signal. I thought the "phase angle" was somehow trying to correlate absolute phase difference between L+R to crosstalk. The language he uses in many places is rather ambiguous, IMO. But this makes more sense, and I take back what I said above about wild phase variations within 0.5 degrees. I don't know how likely this is, as I have not paid close attention to it on my scope. (I don't consider it a first order problem)
As a first order, we should always care about lowest crosstalk. Reducing crosstalk reduces distortion, and it also reduces the distortion caused by an out-of-phase crosstalk signal! We shouldn't trade off higher crosstalk for improved crosstalk phase. In the Adjust+ "Compendium", Chris has one example showing this, but that example is misleading. You can clearly see that despite the crossover points of phase and crosstalk not aligning, the MEAN crosstalk is indeed lowest at the crossover of the L->R and R->L crosstalk phase angles. So, he's not actually accepting higher average crosstalk in that example. And it does look reasonable to leave the cart at the crossover point of the phase angles in that particular case. So really, where Adjust+ can come in handy, then, is where you have a mean crosstalk curve that bottoms out and flatlines across several azimuth angle settings. He also shows such examples of this in the compendium as well. This a fortunate situation as it now allows you to minimize both crosstalk and crosstalk phase offsets by choosing the point on the curves where both are lowest. That's useful.
Why would we care about crosstalk phase offset: because it causes distortion. Just keep in mind, it is very negligible distortion. In the scope trace shown in post #7, the crosstalk is out of phase by roughly 170 degrees. You can see it visually. The absolute mean channel separation would have needed to worsen by 3dB in order to make the phases aligned when I took those measurements. There is no tradeoff to be made here IMO - I want the additional 3 dB of separation. 3dB is significant. The harder it is to hear the crosstalk, the better. I would never make it 3dB more audible simply to align its phase to the fundamental.
As a first order, we should always care about lowest crosstalk. Reducing crosstalk reduces distortion, and it also reduces the distortion caused by an out-of-phase crosstalk signal! We shouldn't trade off higher crosstalk for improved crosstalk phase. In the Adjust+ "Compendium", Chris has one example showing this, but that example is misleading. You can clearly see that despite the crossover points of phase and crosstalk not aligning, the MEAN crosstalk is indeed lowest at the crossover of the L->R and R->L crosstalk phase angles. So, he's not actually accepting higher average crosstalk in that example. And it does look reasonable to leave the cart at the crossover point of the phase angles in that particular case. So really, where Adjust+ can come in handy, then, is where you have a mean crosstalk curve that bottoms out and flatlines across several azimuth angle settings. He also shows such examples of this in the compendium as well. This a fortunate situation as it now allows you to minimize both crosstalk and crosstalk phase offsets by choosing the point on the curves where both are lowest. That's useful.
Why would we care about crosstalk phase offset: because it causes distortion. Just keep in mind, it is very negligible distortion. In the scope trace shown in post #7, the crosstalk is out of phase by roughly 170 degrees. You can see it visually. The absolute mean channel separation would have needed to worsen by 3dB in order to make the phases aligned when I took those measurements. There is no tradeoff to be made here IMO - I want the additional 3 dB of separation. 3dB is significant. The harder it is to hear the crosstalk, the better. I would never make it 3dB more audible simply to align its phase to the fundamental.
Last edited: