Gene,
Great article, Gene. (As usual!!!)
If you don't mind, some thoughts, etc.:
1. For a slightly different approach, I have often found it useful to calibrate an HT at no fewer than 4 and as many as 10 different seating locations. An analyzer that allows a running average of the sound level to be taken is most useful for this since the measurement can be paused while the meter/mic is moved to the next location. This removes the guesswork involved in taking sound level readings at multiple points. If you can get the average response of all the different listening positions to come out like you want it, all the different compromises you talk about can get that much easier. Just be sure to choose as many left-side positions as right-side positions to average out the effects of proximity to side and rear channel loudspeakers.
2. Your article highlighted (for me anyway) the importance of size of room and subsequent treatment thereof. A larger room has less variation since the relative spacing of seating rows stays the same, but the distance to the sources increases. (Hence the reason a large movie theater is relatively simple to calibrate.) Moral: Build it as big as you can. A larger room is also relatively easier to treat. Since most listening positions are more firmly planted in the reverberant field, there will be fewer intracacies involved when treating with absorption or diffusion or both. In fact, the argument for diffusion gets better since overabsorbing could lead to highlighting echo problems. Of course, all this is moot unless you can get the distance between listener and side/rear source to be greater than about 8 feet. (And you have the appropriate ceiling height and front-to-seats spacing as shown in your diagram.)
3. The change in SPL vs. distance issue is noteworthy. What is also worth mentioning is the frequency dependence of this. It's related to the distance from the source. The 3 dB/halving you mentioned will only apply to frequencies whose wavelengths correspond to 1-2 times the distance from the source. In other words, you'll only observe the halving from your rear speakers - about 4 feet away - above about 300-500 Hz. Below that range, very little decrease versus distance. Highlighting once again the importance of good LF control.
(It's also worth noting that a room treated primarily with diffusion in the rear sound field will experience almost no variation in sound level vs. distance over most of the frequency range.)
Some other minor things:
- IMO, microphone orientation doesn't matter if the mic is omnidirectional. (There's a good discussion of this topic
here.) You could point it at the floor and get the same results. What you said about placing it above the back of the seats is very important though. In fact, be cognizant of diffraction effects over hard seatbacks if you're placing the microphone right behind a seat. The mid/high frequency sound levels could measure higher than when the mic is placed in front of the seat for sound coming from the front channels. (Or vice versa if the sound is coming from behind.)
- This is nitpicky, but - IMO - there's really not much point in reporting sound levels to the 1/10th of a decibel. Rounding to the nearest decibel is perfectly adequate since most people cannot hear changes of less than 1 dB. (I know that some would argue; research has certainly shown changes in 0.1 dB are detectable to some people. However, most of that research has involved the use of tones as the noise source. I know of very, very few people who can detect differences less than 1 dB when the source material is
Batman Begins or
Two Against Nature.)
