Hi Greg,
It looks like BDT shuffling is a little different to the width adjustment
process you've described. I had a chance to read most of Gerzon's AES paper on
the train this afternoon and he is makes the following statement on the
necessity of phase shift for BDT shuffling:
"The original application of Blumlein [1] processed input signals differing
mainly in phase, not amplitude, and in his application, as we shall see, phase
shifts in the filters were not merely desirable, but absolutely essential. In
[14] Griesinger quotes the present author as having claimed that elimination of
phase shifts was desirable in this application--something that is simply not
true. In this paper we adopt the term Blumlein shuffling to meanshufflingin
whichphaseshifts in the S signal path relative to the M signal path are used to
convert phase differences in the input stereo signals into amplitude
differences in the output stereo signals. This is to be contrasted with those
kinds of shuffling in which phase shift differences between the S and M signal
paths are desirably minimized, designed to adjust the width at different
frequencies of amplitude-panned stereophony. Griesinger's cited paper [14]
appears to confuse these two quite distinct kinds of shuffling, and rules
applicable to the latter width-adjustment type of shuffling are inap-
propriate to the former Blumlein shuffling, which con- verts phase into
amplitude differences."
The interesting thing is that the time coefficient of the filter is the key
factor in BDT shuffling rather than dB gain of the filter. After a quick read
of the later parts of the paper appears that the frequency of the filter is a
function of the distance between capsules. This is apparently not straight
forward to determine for baffled arrays, hence the Waves disclaimer about
non-standard arrays no doubt.
The paper is well worth hunting down.
cheers
Paul
|