Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
i 1~7759 11
1¦ Inventor: RAY M. DOLBY
21 Title: IMPROVEMENTS IN CIRCUIT ARRANGEMENTS
31 FOR MODI~YING DYNAMIC RANGE
4¦ The present invention is concerned in general with
51 circuit arrangements which alter the dynamic range of audio and
61 other signals, namely compressors which compress the dynamic
71 range and expanders which expand the dynamic range. Generally,
81 it relates to improvements in compressors and expanders that
91 comprise series connected circuits and more particularly to
10¦ the cross-coupling of such series circuits.
11¦ Compressors and complementary expanders are often used
12¦ together (a compander system) to effect noise reduction; the
13¦ signal is compressed before transmission or recording and
14¦ expanded after reception or playback from the transmission
15 ¦ channel. However compressors may be used alone to reduce the
16 ¦ dynamic range, e.g. to suit the capacity of a transmission
17 ~ channel, without subsequent expansion when the compressed signal
18 is adequate for the end purpose. In addition, compressors alone
19 are used in certain products, especially audio products which
are intended only to transmit or record compressed broadcasts or
21 pre-recorded signals. Expanders alone are used in certain
22 . products, especially audio products which are intended only to
23 I receive or play back already compressed broadcasts or pre-recorded
24 signals. In certain products, particularly audio recording and
25 I play back products, a single device is often configured for
26 switchable mode operation as a compressor to record signals and
27 as an expander to play back compressed broadcasts or pre-recorded
28 ~ signals.
29 The amount of compression or expansion may be expressed
in dB. For example, 10 dB of compression means that an input
31 dynamic range of N dB is compressed to an output range of (N-10)
32 dB. In a noise reduction system 10 dB of compression followed
1 177~59
1 b 10 dB o~ complementary expansion is said to provide 10 ~B oE
21 noise reduction.
31 The present invention relates in particular to
41 improvements in circuit arrangements for modifying the dynamic
51 range of an input signal, which circuit arrangements comprise
6~ series circuits, each with a bi-linear characteristic ~where
71 "linear" in this context denotes constant gain) composed of:
81 1) a low level linear portion up to a threshold,
91 2) an intermediate level non-linear (changing gain)
10¦ portion, above the threshold and up to a finishing
11¦ point, providing a predetermined maximum compression
12¦ ratio or expansion ratio, and
13¦ 3) a high level linear portion having a gain
14 ¦ different from the gain of the low level portion.
15 ¦ The characteristic is denoted a bi-linear characteristic
16 ~ because there are two portions of substantially constant gain.
17 In practice, the threshold and finishing point are
18l not always well defined "points". The two transition regions
19 where the intermediate level portion merges into the low
level and high level linear portions can each vary in shape
21 from a smooth curve to a sharp curve, depending on the control
22 characteristics of the compressor and expander.
23 It is also pointed out that circuit arrangements with
24 bi-linear characteristics are distinguished from two other known
classes of circuit arrangement, namely:
26 (a) a logarithmic or non-linear circuit arrangement
27 with either a fixed or changing slope and with no linear portion:
28 the gain changes over the whole dynamic range.
29 (b) circuit arrangements with a characteristic
having two or more portions of which only one portion is linear
31~ ("uni-linear").
32 ~
-2-
al7~7~s
1 A circuit arrangement with a bi-linear charaeteristic
2 has particular advantages and is widely used. The threshold can
3 be set above the input noise level or transmission channel noise
4 level in order to exclude the possibility of eontrol of the
circuit by noise. The high level portion of substantially
6 constant gain avoids non-linear treatment of high level signals
7 which would otherwise introduce distortion. Moreover, in the
8 case of an audio signal, for which the circuit must be syllabic,
9 the high level portion provides a region within whieh to deal
with the overshoots whieh oeeur with a syllabie eireuit when the
11 signal level inereases abruptly. The overshoots are suppressed
12 by elipping diodes or similar means. Only bi-linear eharaeter-
13 istics are capable of providing this combination of advantages.
14 Known circuits employing a single stage with a
bi-linear charaeteristie in use today in eonsumer audio products
16 provide 10 dB of compression and expansion, which is adequate
17 for many purposes. However, this leaves some noise audible to
181 some listeners and, for highest fidelity, more compression and
19 expansion is desirable, say 20 dB. A new circuit now also in
use for eonsumer audio products is described in Belgian-PS
21 889,428, Belgian-PS 889,427, Belgian-PS 889,426, Audio, May,
22 , 1981, pp. 20~26 and in paper J-6 and preprint presented at
23 November, 1981 Convention, Audio Engineering Soeiety, New York,
24 New York.
Prior to the above-mentioned circuits, circuits were
26 known and commercially available which provided 20 dB of compres-
27 sion or expansion, and even more, but these were usually constant
28¦ slope logarithmic circuit arrangements in which there is a
29~ constantly changing gain over the whole dynamic range or nearly
30¦ the whole dynamic range. Such circuits suffer from higher
31¦ distortion and signal tracking problems at very low and very
321 high signal levels than the bi-linear circuits in which the
, _3_
1 177~9
l change of gain is restricted to an intermediate portion of the
21 characteristic and overshoot problems are more severe than with
31 bi-linear characteristic arrangements. Known constant slope
41 companders employ compression ratios in the range 1.5:1, 2:1 and
51 3 1~ but 2:1 is most common.
61 ~ccording to the staggering aspect of bi-linear
71 circuits described in Belgian-PS 889,428, a first circuit, which
81 has a bi-linear input-output characteristic, is followed by one
91 or more further circuits which also have bi-linear character-
lO¦ istics at any given frequency within a frequency range common to
ll¦ the circuits. The thresholds and dynamic regions of the circuits
12¦ are set to different values so as to stagger the intermediate
13¦ level portions of the characteristics of the circuits to produce
14¦ a change of gain over a wider range of intermediate input levels
15¦ than for any of the circuits individually, and to produce an
16 ¦ increased difference between the gains at low and high input
17~ levels, but with a maximum compression or expansion ratio which
;3 is substantially no greater than the maximum compression ratio
19 of any single circuit, by virtue of the staggering.
In the case of audio circuits, if the circuits have
21 overshoot suppression (limiting) elements, then it is also
22 possible to stagger their thresholds along with the stagger of
23 I the syllabic thresholds. The overshoots of the lower level
24 circuits, or stages, are correspondingly reduced, with minimal
overall overshoot of the several stages. This is in contrast
26 to conventional logarithmic compressors in which large over-
27 shoots are inherently produced.
28 Each of the circuits may introduce an alteration of
29 the spectral content of the signal -- for example, a low level
treble boost in the case of a compressor. Thus each succeeding
31 stage may be actuated by a signal of progressively changing
32 spectral content. In the case of complex signals, this has the
Il -4-
1.
11777~9
1 virtue of spectrally spreading out the chances for error in the
21 decoding function. In the case of a tape recorder with an
31 uneven frequency response characteristic, for example, the
41 spectral shifting tendency reduces the overall dynamic and
frequency response errors of the decoded result.
6 The ability to stagger bi-linear stages provides the
7 designer with an additional way in which to optimize an overall
8 circuit. In so doing, the shapes of the compression character-
9 istics of individual stages can be designed with staggering
specifically in mind. The transient characteristics of the
11 circuits are also taken into account and the opportunity is
12 preferably taken to stagger the overshoot suppression thresholds
13 in audio compressors and expanders so as to result in minimal
14 overall overshoot.
A well known type of circuit, called "sliding band",
16 which can be used for each of the first and second circuits,
17 creates the specified desirable characteristic for the case of
18 high frequency audio compression or expansion by applying high
19 frequency boost (for compression) or cut (for expansion) by way
of a high pass filter with a variable lower corner frequency.
21 As the signal level in the high frequency band increases, the
22 I filter corner frequency slides upwardly so as to narrow the
23 boosted or cut band and exclude the useful signal from the boost
24 or cut. Examples of such circuits are to be found in US-PS Re
28,426, US-PS 3,757,254, US-PS 4,072,914, US-PS 3,934,190 and
26 ~ V~S~ Pc~t~nt /~o,~ql 1,~`7 J ~
27 Accordingly, each of the first and second circuits can
28 be such a "sliding band" circuit. In principle the quiescent
29 corner frequencies of the two sliding band circuits can be
different and use can be made of this to provide a degree of
31 compression or expansion which is higher in one part of the
32~
--5--
1 1777~9
1 ~ treated frequency band than in another. However, in a modifica- ¦
2 ¦ tion the corner frequencies are made substantially identical.
3 ¦ This leads to the advantage of sharper discriminations between
4 ¦ the frequency region where boost or cut is being applied and the
5 ¦ reqion where it is not applied and accordingly a sharper discrimi-
6 1 nation between the region where noise reduction is no longer
7 ¦ taking place, because of the appearance of a significant useful
8 ¦ signal, and the region where noise reduction remains effective.
91 On the other hand, circuits are also well known in
which the frequency spectrum is split into a plurality of bands
11¦ by corresponding band-pass filters and the compression or expan-
12¦ sion is effected in each band by a gain control device (whether
13¦ an automatically responsive, diode type of limiting device or a
14¦ controlled limiting device) in the case of a compressor, with
15¦ some form of reciprocal or complementary circuitry for an
16¦ expander. Examples of such circuits are to be found in US-PS
17¦ 3,846,719. These split band or multi-band circuits have the
181 advantage of independent action in the various frequency bands
19 and, if this property is required, such circuits may be employed
as the first, second, or more stages in the series arrangements.
21 It is known to construct bi-linear compressors and
22 expanders, of both sliding band and split band type, by the use
23 of only a single signal path. However, it is generally preferred
24 to construct such devices by providing a main signal circuit
which is linear with respect to dynamic range, with a combining
26 circuit in the main circuit, and a further circuit which derives
27 ! its input from the input or output of the further circuit and
28~ has its output coupled to the combining circuit. The further
291 circuit includes a limiter (self-acting or controlled) and the
30~ limited further circuit signal boosts the main circuit signal in
31¦ the combining circuit for the case of compression but bucks the
32~ main circuit signal for the case of expansion. The limited
--6--
~ 17775g
1 further path signal is smaller than the main path signal in the
21 upper part of the input dynamic range. The main and further
3 ~ circuits are preferably and most conveniently separately
41 identifiable signal paths.
Such known compressors and expanders are particularly
6¦ advantageous because they enable the desired kind of transfer
7~ characteristic to be established in a precise way without
81 problems of high-level distortion. The low level portion of
91 substantially constant gain is established by giving the further
10¦ path a threshold above the noise level; below this threshold the
11¦ further path is linear. The intermediate level portion is
12 ¦ created by the region over which the further path limiting
13 ¦ action becomes partially effective and the high level portion of
14¦ substantially constant gain arises after the limiter has become
15 ¦ fully effective so that the further path signal ceases to
16 ~ increase and becomes negligible compared to the main path
17~ signal. At the highest part of the input dynamic range, the
18 output of the circuit arrangement is effectively only the signal
19 passed by the linear main path, i.e. linear with respect to
dynamic range. In dual path audio circuits the provision of
21 overshoot suppression is particularly convenient.
22 Examples of these known circuits are to be found in
23 US-PS 3,846,719, US-PS 3,9n3,485 and US-PS Re 28,426. There are
24 also known analogous circuits which achieve like results but
25 wherein the further path has characteristics inverse to limiter
26 characteristics and the further path output bucks the main path
27 signal for compression and boosts the main path signal for
28 expansion (US-PS 3,828, 2ao and US-PS 3,875,537).
29 Any of these known bi-linear circuits may accordingly
be employed as the first and second circuits of the series
311 circuit arrangements embodying the invention.
32 ///
--7--
1~777~9
1 ¦ As mentioned previously, it is not essential to create
2 ¦ the desired form of bi-linear characteristic by such "dual path"
3 ¦ techniaues. Alternatives exist, operating with single paths, as
4 1 described in US-P~ 3,757,254, US-P5 3,967,219, US-PS 4,072,914,
5 ¦ US-PS 3,909,733 and ~F_Ke~e~PatentA ~ , for
6~ example. Although these alternative circuits usually are not
71 capable of producing such good results as dual path circuits, or
81 may be less convenient and thereby less economical, they can
9 ¦ produce generally equivalent results. Accordingly, these known
10¦ circuits can also be used as one or more of the circuits of a
11¦ series circuit arrangement embodying the invention. If desired,
12¦ one of the first and second circuits can be a dual path circuit
13¦ and the other a single path circuit.
14¦ In the above described staggered series bi-linear
15¦ circuit arrangements, the series processors operate independently
16 ¦ of each other. In accordance with the teachings of the present
17¦ invention, means for cross-coupling staggered series bi-linear
18¦ circuits are provided, including the coupling of signal path
19 ¦ signal components in one circuit to a signal path in another
20 ¦ circuitl the coupling of control circuit signal components in
21 ¦ one circuit to a control circuit in another circuit and cross-
22 ¦ coupling by way of employing a common control circuit for series
23 ¦ connected devices. Cross-coupling may a) assist in immunizing
24 ¦ the system to control by undesired signals, b) help reduce noise
25 1 modulation effects, c) suppress spurious responses, d) allow
26 ¦ relaxation of individual circuit requirements, e) increase
27 compression or expansion without side effects, f) reduce overall
28 circuit complexity and cost, etc.
29 ¦ A characteristic of staggered series bi-linear circuits
30¦ is that they employ different operating thresholds and, usually,
31 different overshoot thresholds (at least in the case of audio
32 devices~. Consequently, the c rcuits respond differently.
l 1~775~
Cross-coupling of AC signals between (or among) the series
circuits provides the system designer with an additional design
parameter that may be useful in optimizing the operation of the
system.
For example, in the case of dual path bi-linear cir-
cuits, the signal output of the noise reduction path of a
higher threshold level circuit ean be fed to the lower threshold
level circuit and injected, with suitable frequeney response and
phase modifications, into the fixed and variable filter signal
eireuits, thereby to ereate aetive filter aetions that enhance
the sliding band aetion and subsequent noise reduetion effeet.
Similarly, a eross-eoupling ean be effeeted between
the eontrol signals in series bi-linear circuits to improve the
response to rapidly changing signal amplitudes. In one embodi-
ment, a control signal component from the low level eircuit
(which has the first and largest response to incoming trans-
ients, even though it is preferably the second device, because
its threshold is reached first when an input signal rises in
level) is fed via a coupling network to the eontrol cireuit of
the high level proeessor. In this way, the high level proees-
sor is provided with a timely warning of impending signal amp-
litude ehanges. Sueh control signal cross-coupling is particu-
larly useful in minimizing overshoots and distortion in series
bi-linear audio cireuits.
In addition, eircuit complexity and cost can be re-
dueed by the use of a single control circuit for two or more
stages of series bi-linear circuits, in whieh ease the respee-
tive inputs and outputs of the single eontrol eireuit are eross-
eoupled between (or among) the series stages.
In summary the present invention provides a eircuit
for modifying the dynamie range of an input signal comprising:
_g_
~7'77S9
first circuit means having a characterlstic in which only a
portion thereof has a changing gain between low and high input
levels, at least one second circuit means which also has such
a characteristic within a frequency range common to the circuit
means, in series with the first circuit means, the changing
gain portions of the characteristics of the circuit means being
staggered within the frequency range common to the circuit
means such as to provide a change of gain over a wider range
of input levels than for any one of the circuit means individu-
ally, and an increased difference between the gains at lowand high input levels, but with a maximum overall compression
or expansion ratio which is substantially no greater than that
of any single circuit means, by virtue of the staggering, and
at least one coupling circuit means coupling signal components
from one of the circuit means to another of the circuit means
for modifying the action of the said other circuit means in
response to coupled signal components from said one circui-t
means.
The invention will be described in more detail, by
way of example, with reference to the accompanying drawings,
in which:
Figure 1 is an exemplary set of curves showing comple-
mentary bi-linear compression and expansion characteristics.
-9a-
1177~9
l Figure 2 i5 a block diagram showing series bi-linear
2 ¦ devices in general terms.
3 ~ Figure 3 is a schematic circuit diagram of a prior art
41 sliding band compressor.
51 Figure 4 is a schematic circuit diagram of a prior art
6 ¦ sliding band expander.
71 Figure 5 is a schematic circuit diagram of a modifi-
81 cation to Figures 3 and 4.
9¦ Figure 6 is a block diagram of a dual path bi-linear
lO¦ sliding band compressor such as that described in connection
ll¦ with Figure 3 or Figure 3 with the modification of Figure 5.
12 ¦ Figures 7 and 8 are block diagrams showing a prior
13¦ art fixed band compressor and expander.
14¦ Figure 9 is a block diagram showing the present
lS¦ invention in general terms.
16 ¦ Figure 10 is a block diagram showing the present
17¦ invention embodied in a dual stage bi-linear compressor and
18¦ expander.
19¦ Figure 11 is a more detailed block diagram of the
20 ¦ embodiment of Figure 10.
21 ¦ Figure 12 is a schematic diagram showing an exemplary
22 ¦ cross coupling network for use in the embodiment of Figure 11.
23 ¦ Figure 13 is a block diagram showing a further
24 embodiment of the invention.
25 ¦ Figure 14 is a schematic diagram showing an exemplary
26 ¦ cross-coupling network for use in the embodiment of Figure 13.
27 ¦ Figure 15 is a block diagram showing the invention
2~ ¦ embodied in an arrangement for providing a common control
29 circuit to series connected sliding band bi-linear devices.
Figure 16 is a block diagram showing the invention
311¦ embodied in an arrangement for providing a common control
32 circuit to series connected fixed band bi-linear devices.
-10-
1,'
1177759
1 Exemplary bi-linear complementary compression and
2 expansion transfer characteristics (at a particular frequency)
3 are shown in Figure 1, indicating (for the compression charac~
4 teristic) the low level portion of substantially constant
gain, the threshold, the portion where dynamic action occurs,
6 the finishing point, and the high level portion of substantially
7 constant gain.
8 Figure 2 shows series bi-linear devices in general
9 terms: a first bi-linear compressor 2 receives the input infor-
mation and applies its output to a second bi-linear compressor 4
11 connected in series, which has its output applied to a noisy
12 information carrying channel N. A pair of series connected
13 bi-linear expanders 6 and 8 receive the input from channel N at
14 expander 6 and provide a noise reduction system output at the
output of expander 8. The areas of dynamic action of the series
16 devices are separated or staggered with respect to each other
17 within the freauency range that is common to the devices.
18 Although the figure shows two devices on each side of the
19 information channel N, two or more can be employed: the inven-
tion contemplates the cross-coupling of two or more series
21 bi-linear compressors or expanders as explained further herein-
22 after. When configured as a complementary noise reduction
23 system, like numbers of series bi-linear compressors and expanders
24 are provided.
25 ¦ The order of stages having particular characteristics
26 ¦ in the compressor is reversed in the expander. For example,
27 ¦ the last staqe of the expander is complementary to the first
28 ~ stage of the compressor in all respects--steady state and time
29 dependent dynamic response (frequency, phase and transient
30 I response under all signal level and dynamic conditions).
31 ¦ As mentioned earlier, it is usually preferable for
32 the hiqh level stage to be first in a compressor series and the
I
l l
117775g
1 low-level stage to be last. However, a reversed arrangement
2 ¦ is also possible. In the reversed case the control amplifier
3 ¦ of the first stage needs a high gain in order to achieve the
4 ¦ required low threshold. This low threshold then applies even in
the presence of high level signals, which in the case of sliding
61 band systems ~nown in the prior art usually leads to poor noise
71 modulation performance of the overall system. In this reversed
81 arrangement each stage must provide sufficient control amplifier
91 gain to achieve the threshold required of that stage. Moreover,
10¦ each threshold is essentially fixed and independent of the opera- ¦
11¦ tion of the other staaes. This is a consequence of the fact that
12 the signal gain of each earlier stage has fallen substantially to
13 unity when the threshold is reached for the corresponding suc-
14 ceeding stage.
In contrast to the reversed situation, in the preferred
16 arrangement (in which the high level stage is first in the
17 compressor chain, and the low-level stage is last), there is a
18 useful interaction between the stage gains and the thresholds.
19 The thresholds of the downstream stages are partly determined by
the signal gains of the preceding stages. Thus in a 2 stage
21 system with 10 dB of low level gain per stage, the control ampli- ¦
22 fier gain requirement of the second stage is reduced by 10 dB, by
23 virtue of the low-level siqnal gain of the first stage. When a
24 high level sianal appears, the 10 ds gain of the first stage is
eliminated and the threshold of the low level stage is effectively
26 raised by 10 dB. With sliding band companders this improves the
27 noise modulation performance of the noise reduction action.
28 In the preferred arrangement the qains of all preceding
29 stages are fully effective up to the threshold of any particular
succeeding stage. Thus, in contrast with the reversed order
31 system described above, the preferred arrangement takes best
32 ///
11777~9
1 advantage of the prevailing signal gains of the individual
2 ¦ staqes. Namely:
3 ¦ 1. Under very low level (sub-threshold) signal
4 ¦ conditions the control amplifier gain requirement
5 ¦ of each staqe is reduced by an amount equal to the
61 cumulative signal gains of all preceding stages.
7¦ 2. A signal dependent variable threshold effect is
81 achieved, whereby with sliding band stages, noise
91 modulation effects are reduced. The effective
10¦ thresholds of the low level stages are progressively
11¦ raised with increasing signal level at a particular
12¦ frequency. At high signal levels (on the hiqh level
13¦ linear portion of the transfer characteristic) the
14¦ effective threshold of the lowest level stage is
15¦ raised by a level equal to all the low-level (sub-
16¦ threshold) stage gains up to that point.
171
1~¦ One known embodiment of series bi-linear processors
19 employs series sliding band devices: the compressors 2 and 4
20 ¦ and the expanders 6 and 8 of Figure 2 are sliding band devices
21 ¦ as set forth in US-PS Re 28,426 with modifications thereof as
22 described in Belgian-PS 889,428. Such modifications include
23 staggering the syllabic and overshoot thresholds and changing
24 the filter corner frequencies.
Details of the basic circuit are set forth in Figures
26 3, 4 and 5 which are the same as Figure 4, 5 and 10 respectively
27 of US-PS Re 28,426 and further details of said circuits, their
28 operation and theory are set forth therein. The following
29 ~ description of Figures 3, 4 and 5 is taken in large part from
30 ¦ US-PS Re 28,426.
31 ¦ The circuit of Figure 3 is specifically designed for
32 incorporation in the record channel of a consumer tape recorder,
j -13-
1 177759
l ¦ two such circuits being required for a stereo recorder. The
21 input signal is applied at terminal 10 to an emitter follower
3 ¦ stage 12 which provides a low impedance signal. This signal is
4¦ applied firstly through a main, straight-through path constituted
51 by a resistor 14 to an output terminal 16 and secondly through a
61 further path the last element of which is a resistor 18 also
71 connected to the terminal 16 The resistors 14 and 18 add the
8¦ outputs of the main and further paths to provide the reauired
91 compression law.
lO¦ The further path consists of a fixed filter 20, a
ll¦ variable cut-off filter 22 including a FET 24 (these constituting ¦
12¦ the filter/limiter), and an amplifier 26 the output of which is
13 ¦ coupled to a double diode limiter or clipper 28 and to the
14¦ resistor 18. The non-linear limiter suppresses overshoots of
15 ¦ the output signal with abruptly increasing input signals. The
16¦ amplifier 26 increases the signal in the further path to a level
17¦ such that the knee in the characteristic of the limiter or
18¦ overshoot suppressor 28, comprising silicon diodes, is effective
19 at the appropriate signal level under transient conditions.
20 ¦ The effective threshold of the overshoot suppressor is somewhat
21 ¦ above that of the syllabic filter/limiter. The resistors 14 and
22 ¦ 18 are so proportioned that the required compensating degree of
23 ¦ attenuation is then provided for the signal in the further path. I
241 The output of the amplifier 26 is also coupled to an ~ -
25 ¦ amplifier 30 the output of which is rectified by a germanium
26 ¦ diode 31 and integrated by a smoothing filter to provide the
27 ¦ control voltaqe for the FET 24.
28 Two simple RC filters are used, though equivalent
29 LC or LCR filters could be used. The fixed filter 20 provides a
cut-off freguency of 1700 Hz (now 1500 Hz), below which dimishing
31 ¦I compression take place. The filter 22 comprises a series
32 1l ///
-14-
11777~9
1 ¦ capacitor 34 and shunt resistor 36 followed by a series resistor
2 ~ 38 and the FET 24, with its source-drain path connected as a
3 shunt resistorO Under quiescent conditions with zero signal on
41 the gate of the FET 24, the FET is pinched off and presents
51 substantially infinite impedance; the presence of the resistor
61 38 can then be ignored. The cut-off frequency of the filter 22
71 is thus 800 Hz (now 750 ~z), which it will be noted is substan-
81 tially below the cut-off frequency of the fixed filter 20.
91 When the signal on the gate increases sufficiently for
10¦ the resistance of the FET to fall to less than say 1 K, the
11¦ resistor 38 effectively shunts the resistor 36 and the cut-off
12 ¦ frequency rises, markedly narrowing the pass band of the filter.
13¦ The rise in cut-off fre~uency is of course a progressive action.
14¦ The use of a FET is convenient because, within a
15¦ suitable restricted range of signal amplitudes, such a device
16¦ acts substantially as a linear resistor (for either polarity
17¦ signal), the value of which is determined by the control voltage
18¦ on the gate. - ¦
19~ The resistor 36 and FET are returned to an adjustable
20 ¦ tap 46 in a potential divider which includes a temperature
21 ¦ compensating germanium diode 48. The tap 46 enables the compres-
22¦ sion threshold of the filter 22 to be adjusted.
23 ¦ The amplifier 26 comprises complementary transistors
24 ¦ giving high input impedance and low output impedance. Since
25 the amplifier drives the diode limiter 28, a finite output
26 impedance is required and is provided by a coupling resistor
27 1 50. The diodes 28 are, as already noted, silicon diodes
28 ~ and have a sharp knee around 1/2 volt.
29 The signal on the limiter and hence on the resistor 18
can be shorted to ground by a switch 53 when it is required to
31 ~ switch the compressor out of action.
32 ~ ///
-15-
1 1177753
l The amplifier 30 is an NPN resistor with an emitter
21 time constant network 52 giving increased gain at high fre-
31 quencie~. Strong hiah frequencies (e.g. a cymbal crash)
4 will therefore lead to rapid narrowing of the band in which
compression takes place, so as to avoid signal distortion.
6 The amplifier is coupled to the smoothing filter 32
7 through the rectifying diode 31. The filter comprises a series
8 resistor 54 and shunt capacitor 56. The resistor 54 is shunted
91 by a silicon diode 58 which allows rapid charging of the
lO¦ capacitor 56 for fast attack, coupled with good smoothing under
ll¦ steady-state conditions. The voltage on capacitor 56 is applied
12¦ directly to the gate of the FET 24.
13 A complete circuit diagram of the complementary
14 expander is provided in Figure 4, but a full description is not
re~uired as substantially as the circuit is identical to Figure
16 3, component values, are therefore not for the most part shown
17 in Figure 4.
18 The differences between Figures 3 and 4 are as follows: ¦
19 In Figure 4, the further path derives its input from
the output terminal 16a, the amplifier 26a is inverting,
21 and the signals combined by the resistors 14 and 18 are applied
22 to the input (base) of the emitter follower 12, the output
23 (emitter) of which is coupled to the terminal 16a. To ensure
24 low driving impedance, the input terminal 1Oa is coupled to the
resistor 14 through an emitter follower 60. Suitable measures
26 must be taken to prevent bias getting in the expander.
27 The amplifier 26a is rendered inverting by taking the
28 output from the emitter, instead of the collector, of the second
29 (PNP) transistor. This alteration involves shifting the 10 K
resistor 62 (Figure 3) from the collector to the emitter (Figure
31 3), which automatically gives a suitable output impedance for
32 ///
1177~59
l drivinq the limiter. The resistor 50 is therefore omitted in
2 Figure 4.
3 It should be noted that it is important in aligning
4 a complete noise reduction system to have equal signal levels
on the emitters of the transistors 12 in both compressor
6 and expander. Metering terminals M are shown connected to
7 these emitters.
8 Figure 5 shows a preferred circuit, for replacing the
9 circuit between points A, B and C in Figures 3 and 4. When the
lO ¦ FET 24 is pinched off, the second RC network 22 is inoperative,
ll¦ and the first RC network 20 then determines the response of the
12 ¦ further path. The improved circuit combines the phase advantages
13¦ of having only a single RC section under quiescent conditions
14 with the 12 dB per octave attenuation characteristics of a
15¦ two-section RC filter under signal conditions.
,61 In the practical circuit, using MPF 104 FETIs, the 39 K ¦
17¦ resistor 36a is necessary in order to provide a finite source
18¦ impedance to work into the FET. In this way the compression
19¦ ratio at all frequencies and levels is held to a maximum of
20 ¦ about 2. The 39 K resistor 36a serves the same compression
21 ¦ ratio limiting function in the improved circuit as the resistor
22 ¦ 36 in the circuit of Figure 6 or Figure 7. In addition, this
231 resistor provides a low frequency path for the signal.
241 Certain details of the circuit of Figures 3, 4 and 5
25 ¦ have evolved over the years and more modern forms of the circuit
26 have been published and are well known in the art. Reference to
27 I the specific circuit in US-PS Re 28,426 is made for convenience
28 ¦ in presentation.
29 In Belgian-PS 889,428 modifications to the cir-
cuit just described are disclosed, particularly for the purpose
31 I of operating two such circuits in series. These modifications
32 I ///
-17-
l ll
117775~
1 ¦ include changing the frequencies of filters 20 and 22, the
2 1 changin~ of the overshoot suppression levels, and changin~ the
3 ¦ syllabic threshold of one of the circuits by modifying the
41 control amplifier 30. This is done by altering its pre-emphasis
51 characteristics which are controlled by emitter time constant
61 network 52. Increasing the capacitor value in the emitter
71 network of control amplifier 30 increases the amplifier gain at
8 any given frequency, thus causing the sliding band filter to
9 respond to lower signal levels. As explained above and and in
US-P5 Re 28,426, as the control voltage (from amplifier 30,
ll rectifier 31 and smoothing filter 32) increases, the cut off
12 fre~uency of the variable RC Filter 22 rises. Thus, with larger
13 values of capacitance in network 52 (thus lowering the control
14 amplifier turnover frequency), the variable filter responds by
moving up in frequency from its quiescent value. The threshold
16 of the overshoot suppressor is lowered by the application of
17 suitable DC biases (in the forward direction) to the diodes 28.
18 Alternativelyv the gain of amplifier 26 (Figure 3) can be
19 increased to the required level or the amplifier 26 gain can be
increased to a high level and attenuation used to adjust the
21 signal level to the diodes.
22 Figure 7 shows a block diagram of a fixed band dual-path
23 bi-linear compressor and expander configuration. The fundamental
24 aspects of this system is disclosed in US-PS 3,846,719, US-PS
3,903,485 and in Journal of the Au io Engineering Society, Vol.
26 15, No. 4, October, 1967, pp. 383-388.
27 In the known embodiment of Figure 7, the further path
28 networks 250 provide four bands. Bands 1, 3 and 4 have conven-
29 tional 12 dB/octave input filters: an 80 Hz low pass filter 252
at the input of band 1, a 3 kHz high pass filter 2S4 at the
31 input of band 3 and a 9 kHz high pass filter 256 at the input of
32 ///
-18-
1 1~759
1 ¦ band 4. Each is followed by an emitter follower isolation stage
2 ¦ 258. Band 2 has a frequency response which is complementary to
31 that of bands 1 and 3. Such a response is derived by adding (in
41 adder 260) the outputs of the emitter followers 258 in bands 1
5¦ and 3 and subtracting that sum from the overall input signal (in
61 subtractor 262). The output of emitter follower 258 in each
71 band and the output of subtractor 262 are applied to respective
8¦ except that limiters 264' in bands 1 and 2 have time constants
91 twice those in bands 3 and 4. The outputs of bands 1-4 are
10¦ combined with the main path signal in combiner 265. The com-
11¦ pressor output is applied to a noisy channel for transmission to
12 ¦ the complementary expander in which the output of the identical
13¦ further path networks are subtracted from the input signal to
14¦ provide the complementary expansion characteristic.
15¦ Figure 8 shows further details of the limiters 264
16¦ and 264'. Each includes an FET attenuator 270 that operates
17¦ in response to a control signal. The attenuator output is
18¦ amplified by signal amplifier 272, the gain of which is set
19¦ to provide the desired low level signal gain. The outputs
20¦ of all the bands are combined with the main siqnal in such a
21 ¦ way as to produce a low level output from the compressor which
22 ¦ is uniformly 10 dB higher than the input signal up to about
23 ¦ 5 kHz, above which the increase in level rises s~oothly to
24 ¦ I 5 dB at 15 k~z.
25 ¦ The EET attenuator is controlled by a control signal
26¦ sub-circuit that provides a compression threshold of 40 dB below
27~ peak operating level. The control sub-circuit includes control
28 signal amplifier 276 followed by a phase splitter 278 which
29 I drives a full wave rectifier 280. The resulting DC is applied
30 ~ to a smoothing network 282 the output of which is the control
31 ~ signal. Network 282 includes an RC pre-integrator, an emitter
32 j follower and a final RC integrator that operate in conjunction
-19-
1~77759
l with diodes such that both the pre- and final integrators have
2 non-linear characteristics produced by the diodes. Fast,
3 large changes in signal amplitude are passed quickly, whereas
4 small changes are transferred slowly. This dynamic smoothing
action produces optimum results with respect to modulation
6 effects, low fre~uency distortion, and distortion components
7 generated by the control sianal. The circuit achieves both fast
8 recovery and low signal distortion~
9 Figure 9 shows generally the possible cross-coupling
configurations between two series bi-linear devices. If more
ll than two devices are operated in series, the possible cross-
12 coupling configurations increases. Thus, for example, the first
13 device may be cross-coupled to the third device, and so on.
14 Referring to Figure 9, "n" possible cross-couplings from com-
pressor 2 to compressor 4 are shown, having respective transfer
l6 functions f1(s), f2(s) to fn(s). Also, "n" possible cross-
l7 couplings from compressor 4 to compressor 2 are shown, having
l8 respective transfer functions g1(s), g2(s) to gn(s). In
19 the complementary expander arrangement of expanders 6 and 8, the
cross-couplings are reversed such that the g1(s), g2(s) and
21 gn(s) cross-couplings are from expander 6 to expander 8 and
22 the f1(s), g2(s) and gn(s) cross-couplings are from
23 expander 8 to expander 6. Thus, in general, there may be one or
24 more cross-couplings, either forward or backward [f(s) or g(s)]
and the cross-couplina(s) may be in only one direction (e.g.,
26 either the f(s) or g(s) directions are omitted) or, alterna-
27 tively, may be in both directions via a single coupling means.
28 The transfer functions f1(s)~ f2(5), g1(s)~
29 etc., may be implemented by various active or passive devices
which may include freauency and/or level dependent elements.
3l The input and output connections of the cross-coupling paths may
32 ///
-20-
~777t~
1 include suitable points for deriving from or coupling to any of
2 the following: input signals, output signals, main path signals
3 ¦ (in a dual-path bi-linear device), further path signals (in a
41 dual-path bi-linear device) f the AC input signals to the control
51 circuit(s), or the DC control circuit signals themselves (the
61 latter two in bi-linear circuits where compression or expansion
71 is effected by a controllable element responding to a control
8~ signal sub-circuit).
91 In Figure 10, an exemplary cross-coupling arrangement
10¦ is shown between the further paths of dual path bi-linear
11¦ compressors and expanders. The series devices are arranged
12¦ such that the syllabic threshold of the first compressor circuit
13 is at a higher level than the second compressor circuit. For
14 complementarity, the order is reversed in the series expander
circuits. Blocks M1 and N2 denote the further path circuitry.
16 In the arrangement of ~igure 10 the output of the high level
17 stage 280 further path from N1 is applied through coupling
18 circuitry having a transfer function f(s) into the circuitry of
19 the further path circuit N2 of low level stage 282. The
transfer function f(s) may have suitable frequency and phase
21 characteristics so as to enhance the limiting action of the low
22 level compressor stage and conse~uently, the noise reduction
23 effect. If the bi-linear devices are sliding band devices the
24 signal from the high level stage is injected, for example, in
the filter circuitry of the low level stage in order to enhance
26 the sliding band action. In the complementary expander the
27 output from N1 in the high level expander stage is applied
28 through the same transfer characteristic f(s) to the circuitry
29 of the low level stage further path circuit N2.
///
31 ///
32 ///
-21-
1177759
l ¦ A specific embodiment of the general arrangement of
2 ~ Figure 10 is shown in Fiqures 11 and 12. ~igure 11 shows the
3 ¦ series compressor circuits 280 and ~82, indicating the input and
4 ¦ output connection points for the cross-coupling that includes
51 the transfer function f(s). Figure 12 shows the details of the
61 transfer function f(s) network and its connection to the filter
71 circuitry of compressor 282. For ~onvenience, the filter
8¦ circuitry of compressor 282 is taken to be as described in
91 connection with Figure 5.
l0¦ The cross-coupling network of Figure 12 includes a
ll¦ high frequency boost network 284 that provides a 10 dB boost at
12¦ high frequencies and having a corner frec~uency equal to the
13 quiescent filter corner frequency of circuit 280. The network
14 284 output is split into two paths and injected via adjustable
gain means into the fixed filter 20 and the variable filter 22.
16 One end of the 3.3 K resistor in fixed filter 20 is lifted from
17 ground and the signal derived via potentiometer 286 and amplifier
l8 288 is then applied. The end of the 39 K resistor 36a that was
19 formerly connected at the junction of the 0.033 capacitor and
the 3.3 X resistor is lifted from that junction and the signal
21 derived via potentiometer 290 and amplifier 292 is applied
22 thereto. Amplifier 288 has a gain of about 3/4 and amplifier
23 292 has a gain of unity.
2~ In operation, at low levels, the high level circuit
280 would not yet operate. Under this condition voltage V2
26 eauals voltage V4, because voltage V3 contains a low level
27 high frequency boost (resulting from the below threshold gain
28 of high level circuit 280) which is matched by the boost
29 network 284. The signal levels applied to the fixed filter
30l 20 and variable filter 22 may be adjusted to obtain the best
31¦ results. If about 3/4 of the network 284 output sîgnal is fed
32~ to the 3.3 K resistor, its effective resistance becomes about 13
-22-
~17775~
1 K. When the high level circuit reaches its threshold, both
2 filters (20 and 22~ in the low level circuit 282 would then have
3 a sliding band action that would improve the overall noise
4 modulation performance without exacerbating mid-band modulation
effects, i.e. excessive band sliding is minimized.
6 Figure 13 shows generally an arrangement for cross-
7 coupling control signal components from one series device to
8 another. In the example of Figure 13, a control signal component ¦
9 from smoothing network 32 of the low level compressor 232 is
applied via transfer function f(s) block 294 to the smoothing
11 network 32 of the high level stage 22.
12 In operation, the low level circuit 282 has the first
13 and largest response to incoming transients and thus, the
14 cross-coupled signal provides the high level processor with a
timely warning of impending signal amplitude changes.
16 One suitable cross-coupling transfer function f(s) is
17 provided by the network shown in Figure 14, which includes an
18 emitter follower 296 which drives series diodes 298, 300 in
19 parallel with a resistor 302. A 560 K resistor is connected
between the anode of diode 300 and ground. The emitter follower
21 has a 220 K resistor in its base input that is driven from the
22 input of the first smoothing stage of the low level circuit.
23 Thus, the input signal is essentially a pulse from rectifier 31
24 that is not delayed by time constants in the low level smoothing
stage. The pulse is AC coupled by the emitter follower stage,
26 which can provide adequate current dumping, into the output
27 filter section of the high level stage smoothing network, thus
28 providing a fast additional control signal component.
29 In Figure 15, a further embodiment of the invention
for cross-coupling series sliding band dual-path devices is
31 shown. Elements common to the embodiment of Figure 11 retain
32~1 /// ,
~ -23-
11777~9
l the same reference numerals. According to the embodiment of
2 Figure 15, a single control circuit (blocks 30, 31, 32) is fed
3 by adder resistors 304 and 306 from the outputs of the further
4 ¦ paths in each of the series compressors 280 and 282. The
51 control circuit output is applied to the variable filter 22 of
~ j each series compressor through respective level setting means
71 308 and 310 (if reguired). The values of adder resistors 304
81 and 306 may be selected to weight the controlling effect of one
9¦ compressor with respect to another. The arrangement provides an
lO¦ economy in circuit components, while providing performance in
ll¦ large part similar to configurations employing individual
12 ¦ control circuits for each series device. A single control
13 circuit may also be provided for the case of three or more
14 series connected compressors or expanders, in which case each
further path output is fed through a summing resistor to the
16 control circuit input, the output of which is applied through
17 respective level setting means (if required) to the respective
18 variable filters.
19 A single control circuit arrangement is also applicable
20 ¦ to fixed band series connected devices as shown in the embodiment
2l¦ of Figure 16, in which fixed band compressors having a single
22 ¦ further path are shown. The arrangement is also applicable to
23 ¦ series fixed band compressors and expanders each having a
24 I plurality of ~urther paths, in which case the common control
25 j circuit iSf of course, connected between or among the further
26 ¦ paths operating in the same frequency band.
27 ¦ Referrinq to Figure 16, the input signal is split in
28 ¦ compressor 326 into a main path which is applied to a combining
29 network 316 and to a further path that includes a fixed filter
30 ¦ 312 and a voltage controlled amplifier (VCA) 314. The VCA may
31 ¦ be an FET attenuator followed by an amplifier such as described
32 ¦ in connection with Figure 8 (blocks 270 and 272). In the second
~ 1~7'775~
l ¦ compressor 328, the threshold of VCA 314' is staggered with
2 ¦ respect to VCA 314 in the first compressor 326, as explained in
3 ¦ Belgian-PS 889, 428. A common control circuit (blocks 276, 278,
4 ¦ 280 and 282, as described in connection with Figure 8), is fed
5 ¦ by summina resistors 318 and 320 from the further path output in
6 the manner of the Figure 15 embodiment. The control circuit
7 output is applied to the V~A's 314 and 314' through level
81 adjusting means 322 and 324. The arrangement is also applicable
9 ~ to three or more series fixed band devices in the same way as
lO¦ mentioned in connection with the embodiment of Figure 15.
ll¦ The embodiments of Figures 15 and 16 are both based on
12 ¦ the observation that in such series connected compressors and
13 ¦ expanders, the dynamic action occurs primarily in one stage at a
14 ¦ time. For example, starting at low input signal levels, the
l5¦ lowest level stage generates most of the combined control
16¦ signal. As the input signal level rises the lowest level stage
17 ¦ phases out of dynamic action and the next higher threshold level
l3 stage becomes active and contributes most of the combined
19 ¦ control signal.
20 I ///
21 ~
22 I /// I
23 I ///
24 ///
25 ///
26 ///
27 ///
28 ///
29
30 ~
31 ///
32 ///
~1 -25-
