I wasn't able to read the article yesterday, but I got through some of it tonight - honestly, I think a fair amount of it is not directly pertinent, so I skipped through a little. There seems to be two sections that directly relate to this conversation.
First, section III.: Technique for Creating Decorrelated Signals,
Second, section IV. Part A: Elimination of the Perception of Destructive and Constructive Interference
Unfortunately, the description of the techniques for creating decorrelated signals is cursory at best. It gives us a few details: the convolution is accomplished via FIR filters; the filters should not change frequency response (duh!); phase changes should be constructed from combinations of random number sequences.
This paper is, as explained in the introduction, intended to provide access to unpublished ideas that require further work before publishing, so it seems like we shouldn't expect it to explain the fundamentals of the convolution - but I sure wish it did. There don't seem to be any references from this section that could lead to more informative reading.
Section IV, Part A describes the application of the decorrelation to sound reinforcement (the use of multiple loudspeakers in an array to distribute sound over a larger area). Honestly, I was surprised to read this first paragraph and find that it expresses pretty exactly what I was trying to convey yesterday - though my use of terms doesn't match theirs exactly.
"Multiple loudspeakers create interference patterns that can be heard especially clearly when the listener is moving in relationship to the loudspeakers." "The composite magnitude spectrum will exhibit spectral peaks and notches that results form the constructive and destructive interference of the acoustic waves. The frequency of these peaks and notches is dependent on the difference in arrival times of the acoustic signals at the measurement position."
The bits later in that section describe the important characteristics of the interactions between the decorrelated signals.
"Both coloring and combing can be eliminated when the delayed signals are decorrelated from the leading signal. When the decorrelated signal has random phase changes spaced more closely than critical bands, the resulting composite magnitude spectrum will exhibit spectral peaks and notches which are narrower than a critical band and the smoothed spectral envelope is much more likely to retain its original shape. Combing itself is impossible with decorrelated signals because the decorrelated signal is smeared in time and the temporal periodicity between the original and delayed signal varies with frequency. It is interesting to note that the constructive and destructive interference is still present, but the perceptual effects are eliminated."
So here's what I've come to:
- Simple filters and delays are not the ideal treatment for an additional side surround loudspeaker signal
- An optimal solution is a digital FIR filter based on a random number sequence, which adjusts phase and delay based on frequency.
What I'm unclear about
- Is there some pre-programmed filter that can be set up based on positions and dimensions and distances in a listening room, such that the calibrator need only know what measurements to make and how to adjust the filter within the device, or is this done in a more iterative or intuitive/organic way by the calibrator, by applying a number of different filters and measuring and listening to the result?
Much of this is new to me, but I've become more-or-less familiar with most of the technical jargon by now, so I think I've got it straight. I'd be happy for anyone's feedback about my reading and conclusions.