Professional Documents
Culture Documents
AN OVERVIEW
DENNIS A. BOHN
Rane Corporation
Everett, Washington
ABSTRACT
AES6TH INTERNATIONAL
CONFERENCE 371
WHITE ALTEC DAX
ASHLY E-V
PASSIVE ---NARROW-BAND (NOTCH) AUDIOARTSRANE
-- CONSTANT-O
(LCROTARY)I AUDIO
CARVINLOGICWHITE
YAMAHA
L -- 1/5- OCTAVE
--RC --
V WHITE
A
r_ -- PROPORTIONAL--Q
E couRT
KLARK- TEKNIK
SCV
e UREI
C_ r--- CONSTANT-O
I ,JBL/UREI
KLARK - TEKN'IK
Z --CUT ONLY
--
E (NOTCH)
R
S --LC
ALTEC
UREI
DAX
-- ONSTANT-O
ACTIVE -- [_ BOOST/CUT -- PROPORTIONAL- O
CETEC-W1E
ROTARY --
t
LCUT ONLY (L-C)
WHITE
LC
WHITE
WHITE
[_ SLIDERS (GRAPHIC)
JFURMAN SCV
TRANSVERSAL
INDUSTRIAL RESEARCH PRODUCTS
Figure1.Whosaysequalizersdon'tgrowon trees?
(Excludesall microprocessor-controlled
and fulldigitaldesigns.)
Apologiesare madeto manufacturers omittedor incorrectlycategorized.
AES6TH INTERNATIONAL
CONFERENCE 373
a= c0=_at]a_,m,. A_(m) ,, m_(,,) _ ,.. x_:sG:_ Two main categories exist. The first uses active RC
i i i i i ii _ i i i i i iiii -&.l!
.[l.l_i------_---4---_ i i FF-- :- i i [ i i _i _ =. i i i _ ! ': ii i _ _ ! I _ ! _
LS'mi i i i_ !, i, i__,........
i ii i i i [ i _Pi designs as combining, andhow
the terminology
commonplace· Describing well adjacentbecame
bands
L_._[
}_1_ { i i i i.i{
· ; combine is good terminology. However, some varia-
il
ii ii {} ii_i_i i iJ _ !_ ii lng
tionsfilteff' is term,
of this a misnomer,
confuse since what
people. Theisphrase
meant"combin-
is not a
_ i { .....I !i!i'
_ _ i i i i ?'i _
........ filter at all, but rather whether adjacent bands are buf-
.3.ml ...... fered before summing. The other side of this mis-
._._i i i i !i!i
i i i i i i[ i{iliil
i; i}i nomered coin finds the phrase "noncombining filter."
._.m; _ i i ; i i i [ i i i }ii Again, no filter is involved in what is meant. Dropping
._.m'_ ) the word ffilter" helps, but not enough. Referring to an
_ _1_ equalizer as "noncombining" is imprecise. All
Figure 7. Asymmetrical (non-reciprocal) boost/cut equalizers
is how muchcombine
ripple the outputs oftheir filters. The issue
results.
curves.
}_: i [ [ i [ i i { i [ ):_ lk
;' ......
, , __ The second issue involves minimum phase shiftjil-
bandwidths.
ters. There is an implication that some equalizers
Figure 10. Summed response of adjacent filters show- produce less phase shift than others. Again, this does
lng combined response with ripple for constant-Q or not seem to be the case. All 2nd-order bandpass or
proportionaI-Q designs not buffering adjacent filters, band-reject filters (active or passive) shift phase the
re-drawing. ._,_.._..,.
. t
Figure 13. Wheatstone bridge to bridged-T equalizer v,,o_mv_vwv,--,_o_
vjN . OVour
1 Figure
filter section.
14b. Altec-Lansing Acousta-Voice band-reject
Figure 14a. Constant-$ variable band-reiect filter. Figure 15. Series resonant network.
AES6TH INTERNATIONAL
CONFERENCE 377
approach makes use of gyrators [28], [29]. This syn-
Active LC thetic inductor replaces the series resonant circuit as
Active LC designs commonly use the simpler series shown in Fig. 17. This is the most common proportion-
resonant network (Fig. 15) over the more complex aI-Q design and perhaps a dozen different manufac-
bridged-T configuration. A popular topology, based on turers use it. This is the simplest gyrator form; many
Peter Baxandall's famous negative feedback tone con- others exist.
trol circuit [27] appears as Fig. 16. The LCR series APl, Audio Products, Inc. developed a unique
resonant circuit creates a bandpass filter function. The proportionaI-Q approach that uses the bridged-T RC
slider routes the bandpass filter either to the input for filter section shown in Fig. 18 as the variable building
boosting or to the output for cutting. This design is in- block. Many such buffered sections string together in
dicative of approaches used byWhite [21] and others; series. Although drawn as single elements in Fig. 18,
the capacitors are really a bank of capacitors selected
I -'- AA^_--_ /
, ¢
Figure 16. Active LC equalizer based on Baxandall
-/
negative feedback tone control circuit [27]. Figure 18. Bridged-T RC section used by APl in ac-
tive proportionaI-Q equalizer.
. YIN
iL crC>
L'J I Vou_
Figure 17. Active LC circuit showing gyrator substitu- Figure 20a. Passive Wien-bridge.
tion for inductor.
VIN_
VOlT
ViN
Your(VCVS)
Figure 21. Voltage-controlled voltage source
Figure 20c. Active Wien-bridge bandpass filter.
AES6THINTERNATIONAL
CONFERENCE 379
useasaevarabe
Carefully designed state-variable topology allows corn- v,,o- --
pletely independent control over frequency and
bandwidth without changing the amplitude. Relegating
the amplitude control function outside of the state-vari-
able filter then completes a true parametric equalizer.
Any of Fig.s 17, 19, or 23 work as parametrics with the I ...
bandpass function being replaced with the state-vari-
able design of Fig. 24. --- uT
FR£0
!
/ Figure 25. Transversal filter graphic equalizer.
Figure 24. State-variable non-inverting bandpass filter Figure 26. Simple all-pass filter delay block.
section.
Transversal filter equalizers are constant-Q designs So, there you have it. Fifteen categories to choose
based on a tapped delay line as shown in Fig. 25. Each from. To sum up, as the great London autioneer Mr.
tap roughly represents an area of the frequency Christie said, in 1770, 'q'he whole of which is truly neat."
response affected. Scaling each of these outputs by a This many categories exist primarily due to simple
'lap weight" (constants al, a2, etc.) and summing the historical evolution. As technology evolved, so did
results, produces any desired frequency response. Ac- equalizer design. A natural course of events. Transistor
tive filters can be designed either in the frequency or and integrated circuit developments led to active
time domain with the same results.Frequency andtime designs, invention of gyrators created a new category.
are inexorably linked by physics. Transversal filters Proliferation of modern active RCfilter designs created
take advantage of this knowledge by modifying the fre- new ways of doing old tricks, and old ways to do new
quency response using time delay (alsothe foundation tricks. And, today, digital technology propels us into a
for all digital filters), whole new generation of equalizers.
Analog transversal filter designs require using either My personal favorite is the parametric. It allows you
analog delay lines (bucket-brigade devices) or all-pass to go anywhere and do anything. Yet, there are those
active filters. The simplest all-pass filter appears in Fig. who claim the best parametric will not sound as good
26. It has the properties of producing a fiat amplitude as old passive bridged-T designs. Perhaps, but that
response with changing phase shift. (Interchanging cannot be objectively proven. Tightly controlled A-B
the positions of the non-inverting input resis- testing demonstrates that all equalizers designs, creat-
tor/capacitor network produces either phase-lead or ingthesameexactfrequencycurve (important --it must
phase-lag characteristics). This circuit starts with zero be identical) are indistinguishable, it does not matter
degrees at DC, yields 90 degrees at the design frequen- whether they are passive or active, proportionaI-Q or
cy and ends up with 180 degrees at high frequencies, constant-Q, LC or RC, fixed band or parametric, or
Since time is nothing more than phase shift divided by operate in the frequency or time domain. With
frequency, you can use a string of phase shifters to apologies to Gertrude Stein, a transfer function is a
create time delay (although it is frequency-dependent transfer function is a transfer function.
time delay; frequency independent time delay requires Differences do exist, but they are in areas other than
bucket-brigade devices or digital techniques). An all- those described above. Secondary considerations
pass filter approach produced the first transversal such as noise performance, dynamic range, and tran-
equalizer by IRP [13] in 1984. sient stability all enter into explaining perceived sonic
attributes.
380 AES
6TH
INTERNATIONAL
CONFERENCE
Manydeslgns are decades old, while others are but 15. T. Thomas, "Digital Processing for the Digital
a few years. The latest is not necessarily the best, al- Age,"Roland UsersGroup,vol. 6, pp. 60-62(Jan. 1988).
though, we like to think so. Each new development is 16. "AES Recommended Practice for Digital Audio
embraced as the ultimate -- for a while. Then, we tend Engineering -- Serial Transmission Format for Linearly
to migrate back to proven ways that are comfortable Represented Digital Audio Data (AES3-1985 & ANSI
and known, if for no other reason. This, too, is not al- S4.40-1985)," J. Audio Eng. Soc., vol. 33, pp. 975-984
ways best. Ours is a human industry, with human (Dec. 1985).
quirks. 17. D. Bohn, "Bandpass Filter Design,"StudioSound,
The decision as to which is best is a personal one. vol. 25, pp. 36-37 (Jan. 1983).
Manysubjectivethingsenterintotheselectionprocess. 18. T. Pennington, 'q'he Rane GE 30 Interpolating
There are those who swear by one design over another Constant-Q Equalizer," Rane Note 117, Rane Corp.
and will never be convinced otherwise. Nothing can be (1987).
done about this, nor should we try. Objectively, much 19. IEEE Standard Dictionary of Electrical and
could be written regarding the performance virtues of ElectronicsTerms(ANSI/IEEEStdlO0-1984),3rd ed., p.
eachdesign. Nevertheless, suffice itto say, applications 548 (IEEE, New York, 1984).
exists for all these designs. Eventually, the market 20. H. Blinchikoff and A. Zverev, Filteringin the Time
determines lasting favorites. For now, viveladifference, and FrequencyDomains, pp. 89-91 (Wiley, New York,
REFERENCES 1976).
21. C. Van Ryswyk, "Filtersfor Equalization: Active or
1.T. Uzzle,"Boost vs. Cut," Altec-Lansing Corp., Ap- Passive?," presented at the 55th Convention of the
plication NoteAN-6 (1981). Audio Engineering Society, J. Audio Eng. Soc.
2. G. Ballou, Ed., Handbook for Sound Engineers: (Abstracts),vol. 24, p. 851 (Dec. 1976), preprint 1177.
TheNewAudio Cyclopedia (H.W. Sams, Indianapolis, 22. R.A.Greinerand M. Schoessow, "Design Aspects
1987). of Graphic Equalizers," presented at the 69th Conven-
3. D. Davis and C. Davis, Sound SystemEngineer- tion of the Audio Engineering Society, J. Audio Eng.
lng, 2nd. Ed., (H.W. Sams, Indianapolis, 1987). Soc. (Abstracts), vol. 29, p. 556 (July/Aug. 198t),
4. H. Tremaine, Audio Cyclopedia, 2nd. Ed., (H.W. preprint 1767.
Sams, Indianapolis, 1973). 23. T. Pennington, "Constant-Q," Studio Sound, vol.
5. W. Rudmose, "Equalization of Sound Systems," 27, pp. 82-85 (Oct. 1985).
Noise Control,vol. 24 (Jul. 1958). 24. H.R. Kimball, MotionPictureSound Engineering
6. C.P. Boner, H. Wayne [Rudmose] Jones and W.J. (Van Nostrand, New York, 1938).
Cunningham, "Indoor and Outdoor Response of an Ex- 25. F. Bies, "Attenuation Equalizers," J. Audio Eng.
ponential Horn," J. Acoust. Soc. Am., vol. 10, p. 180 Soc., vol. 1, pp. 125-136 (Jan. 1953).
(1939). 26. B. Soloman and C. Broneer, "Constant-S
7. C.P. Boner, "Sound Reinforcement Systems in Equalizers,"J. AudioEng. Soc., vol. 6, pp. 210-215 (Oct.
Reverberant Churches," presented at the 67th Meeting 1958).
of the Acoustical Society of America, New York, May 8, 27. P. Baxandall, "Negative Feedback Tone Control -
1964. - Independent Variation of Bass and Treble Without
8.C.P. Boner and C.R. Boner, "Minimizing Feedback Switches," WirelessWorld,vol. 58, p. 402 (Oct. 1952).
inSound Systems and Room-Ring Modes With Passive 28. R. Riordan, "Simulated Inductors Using Differen-
Networks," J. Acoust.Soc. Am., vol. 37, p. 131 (Jan. rial Amplifiers," Electron. LEE., vol. 3, pp. 50-51 (Feb.
1965). 1967).
9.- ....... , "A Procedure for Controlling Room-Ring 29. T.H. Ishimoto, "Applications of Gyrators in
Medes and Feedback Modes in Sound Systems with Graphic Equalizers," presented at the 63rd Convention
Narrow-Band Filters," J. Audio Eng. Soc., vol. 13, pp. of the Audio Engineering Society, J. Audio Eng. Soc.
297-299 (Oct. 1965). (Abstracts),vol. 27, p. 598 (July/Aug. 1979) preprint
10. -....... , "Behavior of Sound System Response 1501.
Immediately Below Feedback," J. AudioEng.Soc., vol. 30. G.R.Thurmond, "New Devices for Equalization,"
14, pp. 200-203 (Jul. 1966). presented at the 52nd Convention of the Audio En-
11. W. Conner, "Theoretical and Practical Con- gineering Society, J. Audio Eng. Soc. (Abstracts),vol.
siderations in the Equalization of Sound Systems," J. 23, p. 827 (Dec. 1975) preprint 1076.
Audio Eng.Soc., vol. 15, pp. 194-198 (Apr. 1967). 31. D. Bohn, "Constant-Q Graphic Equalizers," J.
12. D. Davis, "A 1/3-Octave Band Variable Notch Fil- Audio Eng. Soc., vol. 34, pp. 611-626 (Sep. 1986).
ter set for Providing Broadband Equalization of Sound 32. W. Kerwin and L. Huelsman, 'q'he Design of High
Systems," presented at the 33rd Convention of the Performance Active RC Bandpass Filters," IEEE Int.
Audio Engineering Society, J. Audio Eng. Soc. Conv.Rec., vol. 14, pt. 10, pp. 74-80 (1960).
(Abstracts),vol. 16, p. 84 (Jan. 1968). 33. R. Sallen and E. Key, "A Practical Method of
13. "Transversal Equalizer DG-4017," Industrial Designing RCActive Filters," IRETrans.CircuitTheory,
Research Products, Inc., data sheet (1984). vol. CT-2, pp. 74-85 (Mar. 1955).
14. "Digital Equalizer DEQ7," Yamaha, data sheet 34. K. Gundry, "Adjustable Equalizers Useable in
(1987). Audio Spectrum," U.S. Patent 3,921,104 (Nov. 1975).
AES6THINTERNATIONAL
CONFERENCE 381