[net.audio] equalizers and equalization

rzdz@fluke.UUCP (05/15/84)

re: equalizers, rooms, equalization, and music

I won't get into a discussion of which musical form constitutes music or
which form doesn't. That isn't my purpose here. This is a synopsis of a
piece that I wrote for Audio Control Corp. If you really are interested, you
can see it in their catalog. Since there has been more than a little bit
said over the last couple of weeks on equalization and equalizers, I will
add my two cents.

On equalizers:

There are several popular circuit topologies in use today:

	 O    A series resonant circuit that is inserted into the input
	      circuit of an amplifier, or into the negative-feedback circuit
	      of the same amplifier.

	 O    A bandpass filter that is inserted in parallel with the input
	      build-out resistor of an inverting amplifier, or in parallel
	      with the feedback resistor of another inverting amplifier.

	 O    A parallel bank of bandpass filters whose Q and passband gain
	      are calculated to produce flat response when they are summed
	      together equally.

The first circuit is probably the most widely used today. It's used by Audio
Control, Soundcraftsmen, ADC, TAPCO, Bi-Amp, etc. The method by which the
designer created the series-resonant circuit varies (some use real LC's,
other's use an op-amp inductor-simulator). The circuit is capable of *very*
flat response *if* the designer used center tapped pots for the equalizer
controls. If not, then the response is as flat as the linearity of the
control and the manufacturing tolerances of same allow. (if the pots are
center-tapped, {and they used the center-taps}, then the resonant circuits
are grounded out when the pots are mechanically centered.)

The second circuit is characteristic of many parametric equalizers and to
some degree, the Crown EQ-2 (i think that's the number). Again, if the pots
are center-tapped, they are capable of *very* flat response.

The third circuit was used until the designer got smart (my opinion) and
discovered the first two topologies. To my recollection, it was only used by
Advent (remember the FBC), Blonder-Tongue (the Audio Baton), and Sunn
Musical (prehistoric Coliseum series Graphic Equalizer). This topology
doesn't work real well. The flatness is *very* dependent upon the passband
gain of the bandpass filters and the linearity of the pots. As I recall, my
old FBC was good for +/- 2 dB for mechanically centering the pots, and maybe
+/- 1 dB if I did it with a square wave.

Conclusion: If your equalizer uses either of the first two topologies, and
	    uses reasonably recent op-amps (if it does at all!), then it
	    should be inaudible in your signal chain. (Inaudible with regard
	    to induced distortion, not noise, which is installation
	    dependent)

On equalizing:

The important thing to remember is that you are probably not listening to
the recording on the same speakers that the recording engineer did. If
you've ever shopped for speakers before, I'm sure that you noticed that they
all sound different. (so do microphones!) Thus, one of the first thing
selection criteria for a speaker is: 

	 How does it sound on my particular type of music? 

More specifically: 

	 does the speaker's particular coloration (read equaliztion) impart
	 a pleasing quality to my sort of music?

What I'm trying to say here is that the speakers coloration acts just like
an equalizer. Since most speakers are not flat (more on that later), you can
view their response curve as a sort of fixed equalizer that you subject all
of your music to. I don't think that anyone will dispute that it's the
frequency response that we first notice in any speaker. (if you do, fine,
but don't flame at me about it.) Most recordings are equalized to one degree
or another by the time they reach our ears. If you listen to popular music
(as opposed to the more classical forms), then this is *almost* (99.9%)
universal. The goal of equalization is to make an end product that is more
sonically satisfying to our ears, i.e. to remove the rough edges from a
sound, or to (perhaps) enhance/diminish some particular nuance that the
producer/engineer/artist/other found good/notgood. 

At best, equalization is a highly subjective process, involving taste (or
lack of), judgement, and a pair of loudspeakers. Since these speakers posess
coloration to one degree or another, this coloration is imparted in a
reverse fashion to the finished product. If the monitor speakers have a
peaky mid-range, recordings equalized on them will tend to have their
mid-range depressed. If the monitors are boomy, the finished product will
probably be bass shy. 

	 Note: A good engineer can learn to compensate for this, but it's
	 tricky at best.

You might ask: Why not just leave well enough alone and leave it all flat?
Good question. The answer is simple...the record company must be reasonably
sure that the product will sound good on 80% of all reproducing equipment in
the marketplace. That means (especially in the case of pop material) that
it's gotta sound good on your walkman, your kid's ghetto-blaster, and on
your lunatic-fringe-polypropylene-capacitors-only stereo. (If the knob is in
front of you, it's hard to resist...let's see what 2 dB more 10k will
do?...)

The way to do this is to listen on a *small* sample of representative
systems, apply the empirical sop (seat-of-pants) equalization algorithm
(modified by the amount of hype that the producer insists be added) and go
for it! In most studios, this means listening on the ubiquitous Auratones (a
4 inch speaker in an equally small box), the studio reference monitors
(whatever the designer built into the wall), and the engineer's (or
producer's) pet bookshelf speakers. In an extreme case, I saw the producer's
ghetto-blaster become the reference standard during a mixdown. I won't say
what the product sounded like.

With all this going on, it's definitely not flat when you get it.

On room equalization:

Some writers to this newsgroup have intimated that an equalizer can do
wonders to a bad room. While this may be partially true, generally, it's
not. To my way of thinking, the only sure-cure for a bad room is dynamite
and a subsequent building permit.

True, an equalizer can help (underline that) to tame a bad resonance. It
can't deal with a bad case of flutter echo caused by an architects
insistence on parallel walls, ceilings and floors, and 90 degree angles
everywhere (not to mention wall dimensions that are integer multiples of
each other). A fat throw rug or a tapestry will do a lot more than an
equalizer will in that case.

On real-time-analyzers

I can't recall one writer to this newsgroup who hasn't made the rash
generalization about equalizing a room (to use a misnomer) with a
real-time-analyzer:

	 adjust the equalizer controls until all the lights are in a
	 straight line.

If you follow that line of thinking, I'd be willing to bet that you listen
to it once, then park the equalizer in the closet because it made your
system sound horrible. 

Even in recording studios, where the speakers are almost always 1/3 octave
equalized, the final curve at the engineer's position is *not* flat. the
usual trend is for a 3dB/octave rolloff beginning at maybe 8 khz. The amount
and breakpoint of this rolloff varies according to the practicioner. The
reasons for this are varied; they include:

	 O    Home systems aren't flat, must use rolloff to ensure that
	      finished product is bright enough (high enough in treble
	      content).

	 O    Boner Preference Curve (after C. R. Boner, a noted acoustician)

	 O    High-frequency absorption of air/vs distance.

If you attempt to modify your loudspeaker/room interface with an equalizer,
you should:

	 O    use the controls on the speakers first, to get a reasonable
	      first approximation.

	 O    ignore the big dips in the curve, concentrating on the big
	      peaks first (they're lots more audible). Dips can be caused by
	      diaphragmatic absorption within the room, trying to fix them
	      by adding more energy to the room at the absorption frequency
	      could turn out to be a losing (for your amplifier and speaker)
	      proposition.

	 O    try to get a relatively smooth curve (note that I didn't say
	      *flat*), incorporating the usual HF rolloff to your own taste.

Conclusions:

What the equalizer does best is to give you a second chance at
second-guessing what the recording engineer tried to do in the first place:
make it sound good on your speakers. If you don't particularly care for you
speakers, but can't afford to replace them, perhaps re-voicing (changing
their EQ curve) via an equalizer will buy you enough time to save the bucks
to buy speakers that are pre-voiced the way you like them.

I welcome any discussion on this topic, however please direct all flames to
/dev/null or net.flame (which I don't read).

Rick Chinn
John Fluke Mfg. Co MS 232E
PO Box C9090 Everett WA 98206

ihnp4!uw-beaver----\
decvax!microsof     \
ucbvax!lbl-csam      \ 
                      +====!fluke!rzdz
sun                  /
sb1!allegra         /
ssc-vax------------/ 

(206) 356-5232
-- 
Rick Chinn
John Fluke Mfg. Co MS 232E
PO Box C9090 Everett WA 98206

ihnp4!uw-beaver----\
decvax!microsof     \
ucbvax!lbl-csam      \ 
                      +====!fluke!rzdz
sun                  /
sb1!allegra         /
ssc-vax------------/ 

(206) 356-5232