[net.audio] more on hafler circuit

rzdz@fluke.UUCP (Rick Chinn) (01/11/84)

re: hafler circuit

I tried the Hafler circuit years ago when it first hit audio (sort of like
quad did). I seem to remember that it more or less did what it claimed. The
big minus (seemed to me) was that any ambience was reproduced in MONO.

Then I built a primitive quad decoder, using the matrix coefficients
promulgated by Electro-Voice. No level-shifting or any funny business like
that, just matrixing. I was unable to find any records encoded with this
process, so I tried a standard stereo recording. What a surprise! With
levels adjusted properly, so you could just *barely* hear the rear speakers,
the effect was quite pleasing. It's hard to describe it in subjective terms.
What I percieved was the ability to "fill" the room, without having to
resort to excessive volume (although I still like to do that). If you turned
off the rear speakers, the room suddenly felt "emptier". The usual reaction
was: "Whatever you just did, Undo! it."

The EV decoder used fixed matrix coefficients. It wasn't anything super-
complicated, just cross matrixing of out of phase signals.

Later, I built another version of the same thing into my system's remote
control. This time I made the matrix coefficients variable. Now I can adjust
to one degree or another, the amount of ambience present in the rear
channels. For most pop recordings (multitrack stuff is what I listen to,
mostly), the ambience is any reverberation that was added, and any cross
mixed delays. Center channel information drops out. On recordings made with
"classical" technique (spaced omnis, xy, ortf, blumlein, mid-side), the
effect is still present, and pleasing.

You could describe what I did mathematically as:

	 LeftFront = Left
	 RightFront= Right
	 LeftRear = Left - k * Right
	 RightRear = Right - k * Left

	 where 0 > k < 1

I've been living with 4 channel ambience recovery now for about 9 years. I
feel really strongly that it does make a worthy difference. I've tried
delaying the rear channels, but haven't had time to really closely evaluate
the results.

Rick Chinn
John Fluke Mfg. Co MS 232E
PO Box C9090 Everett WA 98043
{uw-beaver,decvax!microsof,ucbvax!lbl-csam,allegra,ssc-vax}!fluke!rzdz
(206) 356-5232