Originally Posted by audioguy
What prompted my question was a thread on Emotiva's site where some guy was suggesting the it was important to have 1 measurement at the MLP and then the other 8 below (approximately 9 inches) and above (also approximately 9 inches) that point in various positions around the MLP. That won't work for surround since just a few inches below my ears block the rear channels and low behind the seats .
What I did find out through a few trials was two things (not doing what the Emotiva guy suggested but my version of it so as not to block any
channels: (1) the measured response of the LFE channels was a good bit worse doing it this way BUT sound good and (2) the occasional[/LIST]edginess that could be heard (primarily in vocals) was greatly reduced or eliminated !!
I will try Carl's approach next the see what it sounds (and measures) like.
Maybe we should start a whole new thread on this topic since Dirac Live now exists on the the Dataset products as well as Theta, and now Arcam, miniDSP, Audio Control and Emotiva. We can't be the only folks who have had questions on this particular subject. While the platforms all may be different, it is still the same filter calculation algorithms that are being utilized.
I looked on the Dirac web site hoping to find something but did not. That would be the ideal place to have these discussions.
The issue comes down to having a good recommended practice for most to get decent results vs. finding the best option or different techniques for the best result in your system. One will rarely perfectly align with the other.
We have to first consider if we are correcting the entire range or just below some frequency which Dirac does allow. When we include the full range of high frequencies we have to always consider the dispersion of the speakers and make sure our measurement locations are not too far out of the coverage window of the speakers. If you get too far out of the high frequency window, Dirac will see the response drooping at some locations which could cause the primary positions to be overly boosted in this range. IF the high frequency loss off axis is relatively smooth, you can ultimately compensate for this by ear with the target curve, but it's important to understand why you might want to. This can be useful if you are interested in correcting the mid-lower frequencies for a wider listening area.
My recommendation is to pipe some pink noise into each speaker individually while quickly moving your head around the potential listening/measurement window. You will quickly hear if going too high, low, left, right or backward gets into a dramatically different speaker response. This is also the stage at which you may find you want to adjust aiming of speakers, realizing that there will be some balance and compromise between the best imaging/soundstaging vs coverage.
I find having multiple measurement locations vertically and fore-aft very useful so any correction effort doesn't over-correct for problems that only exist at one listening height. That height difference might only be +/-4" to 12", but that's important to make the listening bubble much more comfortable and natural. In many rooms you will have seats with greatly elevated bass response near the back of the room. In such cases that subwoofer placement and quantity isn't optimal for the rear rows, I will take a few measurements over the front row but in line of sight with the rear seats. This gives you a general correction of the direct arriving upper frequencies, while not mucking up the mid-lower frequencies at the main seats.
The take away from the rambling above is to check with your ears or a microphone the information you are giving the correction system to work with and adjust accordingly. Inevitably you will get hit with a dose of reality that no system/room is as ideal as we'd like it to be, and that's where some creativity comes in.