Originally Posted by audioguy
You are way more "anal" than I am. One dead center at the MLP. One about 6" to 12" on either side in the same seat. And then one dead center in each of the two adjacent seats. I have a forth seat that is far to the left that I ignore.
My own approach is probably closer to Berland's. All are at ear level. Fortunately all critical listening adults in our house are my ear height. Although I might do an experiment next time and do an MLP at my six year old's ear height to see if he can tell the difference on an A/B/X test (LOL):
1) MLP at ear level
2) Four (4) measurements in a square about 6" around my ears, using MLP as the center
3) Measurements at the center of the seats to the left and right of MLP (2)
4) Three (3) measurements about 6" from the back of the sofa directly behind left, center, and right seats
5) Three (3) measurements about 6" from the front of the sofa directly in front of left, center and right seats
I use this data for two separate calibrations: a "selfish" one for MLP (60% MLP, 40% at the square coordinates), and an "egalitarian" one for all three seats, putting 70% of the weight on MLP and the square, and 30% on the other two seats. With one row of seats I use measurement approaches 4 and 5 more for confirming sub placement, MSO settings and PEQ for the subs rather than have them for the full calibration, but that's just me. If I had a multi-row setting I'd feel differently, of course. I used to have a separate preset for when the six year old hogs MLP and demanded I sit on the floor below him, but he's come around to letting that position go to our 13 year old Maltese who doesn't participate on AVS (the year old one thinks she's a listener and wants the left seat to MLP)
This may be overkill, but it gives you the flexibility to explore using lower and higher resolution octave acoustic correction, which is especially valuable now that we can have separate FIR and IIR settings with the most recent room EQ. More data isn't a bad thing, even if you don't wind up using it all.