EP0052144A1 - Diotic position recovery circuits - Google Patents

Abstract

A high fidelity reproduction system using at least one (for mono) or a plurality of speakers in conjunction with, for example, a conventional sound production system (such as stereo, headphones, hearing aids, and three speaker diagrams) and an image correction circuit which ensures better reproduction by correcting the angular position of the images. The system comprises for example one, two or three transducers in various geometric positions such as two loudspeakers placed in front of the person listening (eg at angles of 45 from a line of sight - Fig. 3 , 4) or behind the person listening (as in Fig. 16). Three geometric speaker arrangements can also be used (Figs. 18 and 24). The preferred embodiment also provides that the circuit can be used with for example a stereo amplifier (Fig. 1 and 14) when it is used with conventional stereo speakers / headphones or that it can be used like a hearing aid. The circuit improves the separation of the individual components of sound production (such as drums, a singer, a trumpet, a violin and others) while taking into account the "shadow of head" effect.

Description

DIOTIC POSITION RECOVERY CIRCUITS

Technical Field

The present invention relates to systems for reporducing sound, and more particularly relates to a system for the reconstruction of psycho-acoustic images in stereophonic, headphone and hearing aid sound reporduction systems.

Even more particularly, the present invention relates to a system for reconstructing psycho-acoustic images by changing the relative amplitude of signals, for example, between two speakers.

Background Art

Numerous prior publications and patents have disclosed systems and devices which have the objective of distortion free reporduction of recorded sonic materials typically in the home environment. These systems usually introduce some distortion.

One type of distortion is imaging distortion, typically manifested as a "flattening" of the sound stage image, this is, a tendency to hear sounds as if they emanate from a single plane which when recorded were in widely scattered locations throughout the recording area, or stage sound sources. Even with the finest stero system, one can sense only general depth.

SSJBST C M _ that is .a vague "closer" or "farther" perception, regardless of the disposition of the original sound sources. Because stereo systems use only two sound sources, the resultant imaging is necessarily different from randomly placed multiple sound sources.

One proposed, but impractical "ideal" playback system uses one speaker for reproducing the sound of each sources, each speaker being placed in the exact location originally occupied by the respective source with respect to the listener.

A more practical but less satisfactory approach is the quadro-phonic system, in which separate information channels drive each of four speakers placed as to generally surround the listener. The conventional stereophonic system is simpler yet, but represents a probably less satisfactory solution to the sound imaging problem.

The activity of the outer ear relative to distance perception was demonstrated in applicant's copending application, U.S. Serial No. 895,309 entitled "Audio Image Recovery System", filed April 11, 1978-

The "Audio Image Recovery System" patent used a signal controller and the geometrical placement of four loudspeakers to perform these outer ear functions in the following way.

This image recovery-type method took into consideration one aspect of hearing, in which the outer ears contribute to our ability to discern location in space. All sounds which strike the ear drums, or tympanic membrane, have first been partially supplemented by the wing of the outer ear, or auricle as well as the presence of the listener's head. The listener's head and auricle function as a reflector and/or barrier which modify a portion of the sound

entering the ear canal, or external auditory meatus, according to the direction and distance of the sound from the listener.

For a given sound originating in front of the listener, a certain portion of the sound ("direct-in" or D.I. signal) will enter the ear canal opening directly. Due to losses by multiple reflections as the sound travels through the ear canal, higher and lower frequencies will be attenuated relative to the mid- range frequencies according to the angle at which the sound subtends the opening. This attenuation is an addition to the characteristic attenuation of these frequencies due to the distance from the listener.

The other, usually smaller portion of the sound reaching the tympanic membrane ("apparent-on-axis" or A.O.A. , signal) will be reflected off the auricle, more or less directly into the ear canal by that section of the auricle directly adjacent to the ear canal opening and above the ear lobe. Because this sound component undergoes essentially a single reflection, it does not suffer the frequency aberrations of the D.I. signal.

The subjective ratio of A.O.A. signal to D.I. signal is dependent upon both distance and angle of the sound source relative to the facing direction of the listener. The subjective level of A.O.A. signal increases relative to direct in (D.I.) signal as distance is reduced because of 1) an absolute level increase; and 2) the high frequency and low frequency levels increase relative to the mid-range level due to the Fletcher-Munson effect.

The A.O.A. component also increases relative to the D.I. signal as the sound source moves closer to being in front of the listener (angular change) . The D.I. component enters the meatus less directly and is

E SHΞET therefore attenuated more before reaching the eardrum than is the reflected A.O.A. components.

As the A.O.A./D.I. ratio changes, the brain discriminates between angle and distance changes on the basis of the sound information of both ears. Simultaneous similar A.O.A. /D.I. changes indicate change distance while simultaneous dissimilar A.O.A./D.I. changes indicate change in the angular location of the source. An approximate "head shadow" effect was also included.

The previous patent uses four speakers (two in front of the listener for D.I. sound, and two to the sides of ^"the listener for A.O.A. sound, plus a controller to vary the ratios of signal levels between the four speakers) . In this way, the original distance from the sound source to the microphone may be accurately perceived by the listener.

In summary, the earlier referenced copending patent application, U.S. Serial No. 895,309 illustrates the direct functional connection between distance perception, the outer ear shape, and the Fletcher- Munson effect by reproducing the physical reflecting activity of the outer ear. A listener with an "average" Fletcher- unson loudness perception characteristic would not need the presence of their own outer ear in order for the circuit and speaker arrangement shown in copending application, U.S. Serial No. 895,309 to work.

The embodiment of the present invention does not duplicate the physical reflective function of the average outer ear. Instead, the embodiment of the present invention is based on the fundamental relationship between human angular perception and simultaneous volume difference heard between each ear.

OM

SUBSTITUTE SHEET The volume changes made by the embodiment of the present invention are made in such a way as to re¬ create the two-ear volume differences heard from real sound sources placed in corresponding positions relative to the listener. That is, this invention corrects for the physical placement of the sound source and head, not the outer ear physiological response to that placement.

These volume differences also create subjective frequency changes perceived as a result of the presence of the listener's outer ear - Fletcher-Munson effect relationship. This secondary effect permits distance perception of images using two loudspeakers, if the presence of the head (head shadow) is acounted for. Further, if the program material was recorded with a cardioid microphone (which nearly duplicates the volume changes heard by a listener with respect to vertical placement of a sound source), and if the listener's shoulders are present, (critical to vertical localization) accurate vertical image placement can be discerned as well.

For example: A sound source moves from a position directly in ..front of a listener to a new position to the left of the listener: A) The volume in the left ear increases.

B) The volume in the right ear decreases.

C) The rate of increase to decrease is influenced by the presence of the head.

The above changes are provided by the embodiment of the present invention.

Because of the Fletcher-Munson effect:

D) High and low frequencies increase in volume relative to mid-range frequencies in the left ear.

E) High and low frequencies decrease in volume relative to mid-range frequencies in the right ear.

Both of the above amplitude changes are modified in the usual way by the listener's outer ears.

Because angular localization depends on

^•3if fβ.l- ϋ¹??-.?. heard between ears:

F) Differences between high and low frequencies heard between both ears are greater than differences heard at mid-range frequencies.

G) These increased differences are perceived as additional cues agreeing with a change in the angular position of the sound source relative to the listener. Because distance perception is dependent on frequency change:

H) These simultaneous amplitude and subjective frequency changes allow the listener to perceive the accurate original angle and distance to the image, if the presence of the head is accounted for.

I) If the program material was recorded with a cardioid microphone, vertical position information can be retrieved as well, if the acoustical contributory effect of the shoulders is taken into account.

_*

If a sound source moves vertically with respect to a listener's ears, the volume level decreases on either side of an imaginary approximately 45, (see Figure 7) inclined line passing through the listener's ear canals. The volume changes with respect to angle approximately agree with the vertical pick-up pattern of a cardioid microphone (see Figure 8) .

This pick-up pattern of the ear is modified in

volume in the lower forward and side quadrants by the presence of the shoulders. All of these changes are analyzed relative to other changes occurring simultaneously in the opposite ear. For vertical localization the listener's shoulder must be present during playback, because of a suspected acoustic interaction between the listener's shoulder and part of the listener's corresponding outer ear. The vertical cardioid microphone pattern provides a head shadow (upper) and body shadow (lower) effect, but not a reflective shoulder function.

It is further believed that this reflective shoulder interaction only occurs at a volume level corresponding to a real sound source placed at a given location, such as the embodiment of the present invention.

With the other patent, the appropriate volume change is not present - it is not designed to do that. Therefore, vertical image recovery cannot occur with the first patent.

With the embodiment of the present invention, it is necessary for the listener's outer ears, head and shoulders to be present for these localization effects to occur. This is because the sound is being created, in this example, by transducers in a free field. (For hearing aids and headphones, frequency tailoring must be used to add the outer ear and head shadow effects of angular change and also must be used to compensate for the actual locations of the transducers.) The invention described in this embodiment suggests that the degree of difference heard between two ears is another fundamental hearing feature to consider for totally accurate localization or focusing of an image. Further it indicates that volume

ITUTE SHEET differences, and consequently frequency differences are reliant on the presence, shape and size of the head and shoulders. Consequently, my prior patent application, U.S. Serial No. 895,309 in some respects fails to produce a totally focused image with respect to angle and distance (the head shadow effect is not as well defined) , and it cannot reproduce as accurately vertical image information.

Conventional stereophonic systems, and other prior art playback methods do not take into account diotic amplitude correction for physical location of a sound source, or sophisticated diotic localizing functions of human hearing involving the physiological shapes of the body. Although some localizing information is provided by level differences between two speakers, the illusion of location is imperfect and at best the sound stage images are fuzzy, vary from listener to listener, and offers no. vertical image position information. In other prior art, as will be later discussed, other methods are used to create random (artificial) image locations from existing recordings, or use special recordings to locate images in their approximate original locations.

When recordings are played through stereophonic speakers which have been "toed-in" (angled inward) center stage images become over emphasized. Toe-in .of front speakers creates overall incorrect image location (i.e., not according to what was recorded), and blurred sound information from sound sources located at center stage. The circuit of the present invention (particularly the active circuit) has the ability to compensate for this effect by subtracting monaural (center stage) information from the signal.

One attempt to solve the audio image problem in

OMPI stereophonic systems is disclosed in Sorkin Patent No. 3,478,167, entitled "Three Speaker Stereophonic Audio System". That . system comprises a conventional stereophonic audio system with an additional third speaker. This third speaker is placed directly in front of the listener and is fed with both left and right channel information, thus insuring that sound recorded from central sources will be played back through a centrally located speaker. The remaining two speakers are placed on either side of the listener and are fed with the respective left and right channel information as well as the inverted sound information of the opposing channel. The inventor claimed that the inverted signal cancelled the oppositely phased identical audio signal emanating from the opposite pair of speakers, to enhance in an undisclosed manner the stereophonic system.

However, this system does not supply appropriate volume level differences or head shadow effect and thus, does not create the true three-dimensional realism which the present invention does. The Sorkin system at best wraps the flattened sound stage image around the listener in a 180 degree arc. The Sorkin system creates further sound stage confusion because of varying amounts of opposite channel cancellation arising from the sound shadowing of the listener's head and the directionality of the ears.

Other prior art attempts at proper sound image placement have used time delay, frequency tailoring, or phase shift.

Time delay is a method of signal processing which involves delaying the arrival of a signal to a listener. The time delay may occur with only one of the listener's ears (such as in U.S. Patent No.

cr- 4,188,559) so as to deliver a hearing cue to the listener that the sound source is at some angle relative to the listener.

Or, the time delay may follow a signal by both ears to simulate a sound reflection from a distant surface (creating a false spatial or concert hall effect such as used in Patent No. 4,053,711 issued to

R. E. DeFreitas, et al.

The time arrival cue as used in the U.S. Patent No. 4,188,599 is a legitimate application, although it caters to a minor (by comparison to amplitude) hearing localization mechanism.

Time delay to create simulated reverb will create a sensation of depth (not specific distance) and has the ability to blur or mask subtle details in the recorded signals by smearing the arrival time.

Add-on time dela . systems such as U.S. Patent No. 4,053,711 (above) can yield a subjective improvement in clarity, but this is a virtue of the side speaker placement used with add-on time delay type systems. Side speaker placement allows sound to enter the ear canal axis unobstructed.

Frequency tailoring is a method of signal processing whereby specific frequencies in the audio spectrum are increased or attenuated relative to the other frequencies in the spectrum.

Frequency tailoring (equalization) is usually applied in stereo systems to compensate for subjective frequency losses at lower listening levels (loudness compensation) , or non-linearities in associated equipment and psycho-acoustic devices (such as U.S. Patent No. 4,188,599 above), to change the frequency response (in addition to phase shift and time arrival) in relation to human head shadow and outer ear caused

OM

SUBSTIT WIP changes accompanying the angular position of a sound source, i.e., simulate the frequency changes occurring in binaural recordings with loudspeakers. Frequency changes will have psycho-acoustic "effects" because of the Fletcher-Munson effect.

Analog equalization (as opposed to digital applied equalization) is a distortion producing process which will also introduce varying degrees of phase shift. Digitally applied equalization is currently too expensive to be practical. For the sake of fidelity, a signal should be processed minimally.

Phase shift is a method of signal processing, akin to time delay, in which a waveform is shifted backward or forward in phase relative to its original position. Phase shift has been used in many audio devices to create an array of psycho-acoustic effects. Phase shift is less critical to image location than amplitude, and can create both usable or confusing, clues for the listener's hearing mechanisms. Phase shifting also produces distortion.

Discussion of Invention

It is possible to use the activity of the listener's own angular perception mechanism in a way which allows virtual sound sources to be created beyond the limits of the left and right loudspeakers. This is done by manipulating the amplitudes of signals which the ears hear from respective left and right loudspeakers. The amplitude changes are made simply to conform to the physical location left or right of the listener where the sound source is supposed to be, rather than the location of where the sounds are really coming from (two speakers) . For example, as a sound source moves to the right, the volume level will iiicrease in the listener's right ear and decrease in the left through corresponding volume changes from the two loudspeakers. These volume changes are in addition to the normal left and right volume changes which occur in normal stereo recordings. By correcting the. sound levels for the physical location of the sound source, all of the listener's other localization abilities, necessary to deduce location of the image, function automatically. (That is, the listener's localization mechanisms perfectly match the amplitude differences which are present.) For this reason, this design approach (for stereo playback) will work with all subjects. Accounting for head shadows should be included. (The head acts as a sound blocking barrier between the two ears. Hence, the term head shadow.)

For high fidelity playback, this system requires two loudspeakers, which can connect between one stereo amp and one stereo pre-amp. The system may be optimized for a particular stereo speaker geometry, but it is "forgiving" in the area of listener placement, i.e., the effect can be heard to a lesser degree from many locations throughout the listening area. For practical applications, no adjustments are required by user. The two-transducer format permits modification to applications such as hearing aids and headphones.

Brief Description of Drawings

For a further understanding of the nature and objects of the present invention, reference should be had to the following detailed description, taken in conjunction with the accompanying drawings, in which like parts are given like reference numerals and wherein:

Figure 1 is a schematic connection diagram illustrating the preferred embodiment of the apparatus of the present invention as connection to a conventional amplifier and left and right transducer;

Figure 2 is a schematic plan view of a typical geometric transducer layout with respect to a listener as used with the preferred embodiment of the apparatus of the present invention, the actual placement of the original sound generating instruments being illustrated in phantom lines. Figure 3 is a schematic plan view of a. conventional stereo transducer's layout with respect of a listener- illustrating, with graphical representation, the sound separation as experienced with conventional stereo systems; Figure 4 is a schematic plan view of the preferred embodiment of the apparatus of the present invention showing a typical geometric representation of sound separation as achieved by the present invention being shown as distributed to each transducer; Figures 5 and 6 are schematic plan views of monaural transducer placement illustrating the preferred embodiment of the apparatus of the present invention as used during monaural playback;

Figure 7 is a schematic representation of the invention of shoulder/ear of a listener;

Figure 8 is a schematic representation of a cardioid microphone vertical response pattern as compared to the vertical response pattern of a listener; Figures 9 and 10 are circuit diagrams of the preferred embodiment of the apparatus of the present invention as used in a passive stereo enhancement circuit; ^" Figure 11 is a circuit diagram of an active image recovery circuit as used with the preferred embodiment of the apparatus of the present invention;

Figures 12 and 13 are graphical representations which illustrate simulated head shadow affect as achieved by the circuit of Figure 11;

Figure 14 is a connection diagram for the two- speaker active circuit of Figure 11;

Figure 15 is a plan view geometrical speaker placement diagram as used with the circuits of Figures 9-11;

Figure 16 is another plan view geometrical speaker placement diagram as used with the circuits of Figure 11 with equalization; Figure 17 is a graphical representation of an equalization curve as used with a two-speaker (or three speaker) circuit with side loudspeakers placed behind the listener at 135, with respect to the listener's line of sight; Figure 18 is a plan geometrical view speaker placement diagram as used with the image recovery three speaker circuit of Figure 20;

Figure 19 is a plan view geometrical speaker placement diagram as used with the image recovery three speaker circuit of Figure 23;

Figure- 20 is a three speaker passive circuit diagram of another embodiment of the apparatus of the present invention;

Figure 21 is a circuit diagram of the preferred embodiment of the apparatus of the present invention using four speakers;

Figure 22 is a connection diagram for the four speaker passive circuit of Figure 21;

Figure 23 is a circuit diagram of the preferred

O "- - embodiment of the apparatus of the present invention as used in a three speaker active recovery circuit;

Figure 24 is a geometrical speaker placement diagram as used with the circuit of Figure 23; Figure 25 is an equalization curve for the center loud speaker as used with the circuit of Figure 23 and the geometric speaker placement diagram of Figure 24;

Figure 26 is a connection diagram for the three speaker active circuit of Figure 23; Figure 27 is a connection diagram for the passive three speaker circuit of Figure 20; and

Figure 28 is a single channel variation of the active image recovery circuit of Figure 11 with balanced inputs and outputs.

Best Mode for Carrying Out the Invention

In Figure 1, there can be seen a schematic view of the preferred embodiment of the apparatus of the present invention designated by the numeral 10. A pair of speakers 12, 14 are shown as connected to a conventional stereophonic audio generating source 11 known in the art as an amplifier.

^■ The circuit designated by the numeral 10C is illustrated as it would be normally connected with the amplifier 11 and speakers 12, 14.

In Figure 2, there can "be 'seen a schematic representation and plan view of a listener L and a pair of transducers 12, 14 which are, for example, left and right stereophonic speakers. Also shown schematically in Figure 2 are sound generating instruments 15, 16, 17 which are shown in phantom lines. In the schematic illustration of Figure 2, a horn 15, drum 16, and violin 17 are shown. This view merely schematically illustrates with respect to a

listener a typical speaker geometric placement as well as the original location with respect to the listener of the sound generating instrument doing, for example, a live performance. The listener is now hearing the live performance through speakers 12, 14 and with conventional stereophonic playback, the sound comes from the two speakers, 12, 14 and appears (See Figure 3) to be psycho-acoustically placed generally between speakers 12, 14 rather than to be coming from their actual location as originally generated during the live playback which positions are schematically shown in Figure 2.

In Figure 3, there is seen also listener L as well as speakers 12, 14 with sound generating instruments 15 - 17 shown in phantom lines generally between speaker 12 and speaker 14. In conventional stereophonic playback this would be the psycho-acoustic position of horn 15, drum 16, and violin 17, or somewhat in between three speakers which were playing back the recorded signals recorded at the live playback. A graphical illustration is shown which approximates an exemplary breakdown of the amplitude of each sound generating instrument 15 - 17 as proportionately divided between the two speakers. Thus, horn 15 shows a certain percentage of amplitude generated as represented by horn 15L while a lesser percentage is generated from speaker 14 represented graphically as 15R. Similarly, a drum 16 is graphically represented as being generated from speaker 12 as in part shown at 16L and a lesser percentage of amplitude graphically illustrated at 16R. In a like manner, violin 17 which during the live performance was at a position graphically illustrated in Figure 2 as being somewhat right of transducer 14, the graphical representation of relative amplitude now

^" fUH

SUBSTITU^" —r.__T OM

.__. i places a certain larger percentage of violin 17 as graphically illustrated by violin 17R while a lesser percentage of amplitude is graphically illustrated as 17L. The above graphical representation merely shows that somewhat of an improper psycho-acoustic effect is produced by conventional stereophonic loudspeaker playback.

The present invention solves these prior art problems and shortcomings by utilizing a six-resistor separation enhancement circuit (see Figure 9) in the preferred embodiment which uses preferably two speakers in the circuit IOC with the psycho-acoustic localization abilities of the listener being manipulated by changing the relative amplitude of the signals between the two loudspeakers 12, 14. This increased difference in amplitude directly stimulates the subjective change in the lateralization caused by the degree of sound which is common to both ears. Figure 9 shows a circuit diagram of the preferred embodiment of the apparatus of the present invention and more specifically the circuit portion thereof designated by the numeral IOC. In Figure 9, 26-29^" indicate respectively left amp ground, right amp ground, left amp output and right amp output. Circuit IOC is provided with six resistors 20 - 25 and two speakers 12, 14. The function of circuit 10C is to allow the dominant channel to suppress the volume of the loudspeaker attached to the opposite channel. The weaker signal in the opposite channel also increases the output of the dominant channel by allowing the weaker hot terminal of amplifier 11 to act progressively or like an additional grounding terminal. In figure 9, the value shown for resistors 20 - 25 are exemplary values as used with an exemplary pair of speakers 12, 14 being B and W DM-7 loudspeakers. These values of resistors 20 - 25 would change if loudspeakers with different impedance characteristics were used. For these values in the circuit to apply, the loudspeakers must be placed relative to the listener .generally as shown in Figure 2. The speakers 12, 14 should be facing directly forward and -should geometrically be placed from the listener at about forty-five (45) degrees from a line of sight of listener L or from forward center.

The effect of the circuit 10C of Figure 9 is to allow acoustic images to be psycho-acoustically placed beyond the left and the right loudspeakers 12, 14 respectively for those acoustic images which are left and right of center to a degree. Note is taken here that in this case, the values of the circuit resistors 20 - 25 and the layout of the circuit attempt to cancel phase inversion in opposite loudspeakers 12, 14 when a stereo phonograph record with average channel separation is used as a signal source. If resistor values are chosen to allow phase inversion in the weaker channel and allow greater amplitude in the stronger channel (hence keeping the difference heard between channels unchanged) the image placement will still be correct but blurred, and the sound quality will become increasingly "grainy" with increasing amounts of phase inversion.

Ideal lack of phase inversion may not be possible for this circuit when used with close speaker placement or wide dispersion loudspeakers. The increased acoustical crosstalk from these factors would necessitate increasing the degree of gain added to the predominant channel via the differential stages. This would also cause the out-of-phase signal in the weaker channel.

It is expected that this "grainy" character is caused from the out-of-phase signal acoustically cancelling at the weaker signal ear the signal from the opposite channel loudspeaker. This effect is helpful under some circumstances (close placement of speakers, etc.) but the cancelling wave is somewhat out of step with the signal produced by the predominant channel because of the small difference in path length between the car and the two loudspeakers. Consequently, the acoustic cancellation is not perfect, and parts of the signal component are still heard in the opposite or far ear, hence the graininess. If left channel 12 is stronger than right channel 14 with respect to a particular sound generating instrument, the stronger signal passes through isolation resistor 20. The signal is then divided between the left loudspeaker 12 and crossfeed resistor 22. After the signal passes through the left loudspeaker 12, it is divided between resistor 24 which goes to the positive side of right speaker 14. By virtue of the value of resistor 21, a signal passing through resistor 23 is partially fed through the positive side of right speaker 14. This counteracts the tendency of the signal passing through resistor 22 to cause a phase inverted left channel signal in the right loudspeaker 14. The choice of resistor 24 determines to a degree the strength of signal fed through resistor 23. The same circuit can be placed anywhere in the recording of playback chain, and it could be done actively as well as passively as will be described more fully hereinafter.

The circuit of Figure 9 is primarily for use with an amplifier 11 and between amplifier 11 and two loudspeakers 12, 14 as shown in figure 1.

To obtain position recovery using this basic format, there are a couple of additional considerations. One is compensation for the distance that the speakers 12, 14 are spaced. Another is some degree of compensation for whether a recording had a small amount or a large amount of separation in it and it would be desirable to add a "head shadow effect" as will be described more, fully hereinafter.

In the passive circuit IOC of Figure 9 adjustment for separation in recording and the separation of the loudspeakers can automatically be accomplished by choosing the resistance values. Alternately, a separate additional adjustment for separation in recordings could be attempted by the addition of isolation resistors (shown in Figure 10) (variable isolation resistors 31, 32) which allow a variable bridging resistor (either 33 or 34) to add a desirable amount of crosstalk between channels for a stronger monaural image, and to reduce phase inversion at the weaker channel output.

' As can be seen from Figure 4 and the inspection of circuit 30, the amount of signal contributed through resistors 33 or 34 will vary slightly with channel balance. This is caused by the hot amplifier terminal behaving as a variable alternative path to ground for the signal passing across either resistors 33 or 34. Resistors 33, 34 are adjusted for different types of recordings according to the amount of channel separation in the recording. The contribution of this bridge, either resistor 33 or 34, could be limited to a very small amount (i.e., the resistors would preferably be small compared with the other resistances in the circuit) .

Numeral 36 indicates a two-position selector switch which allows to switch between two bridging resistors 33 and 34. Also shown in Figure 10 are isolation resistors 38 and 41, cross-feed resistors 39 and 42 and resistors 40, 43.

If the contribution of the optional bridge is too great, the enhancement of separation will be received or negated. The resistance values for the circuit (of Figure 9) for a successful passive adaptation of the circuit require that the circuit would have a high insertion loss when it was hooked into the system and for that reason (along with not having the "head shadow effect") it is more desirable to use an active circuit which can perform the functions that this circuit (Figure 9) could perform to a more pronounced degree adding the necessary "head shadow effect".

Figure 11 is a circuit diagram of the active imaging circuit 50 as used with the preferred embodiment of the apparatus of the present invention. Figure 11 as shown already includes the desired separation characteristics by virtue of resistor selection. Note also Figure 14 where a schematic illustration is provided of a connection diagram suitable for connecting circuit 50 between a conventional stereophonic audio generating source 11 such as amplifier 11 and pre-amp 52. Returning to Figure 11 there can be seen the connections which are respectively left pre-amp output 52, right pre-amp output 53, left circuit output 96, and right circuit output 97. These would be connected as shown in Figure 14. The circuit of Figure 11 will be described more fully at this time.

An active circuit 50 is shown in Figure 11. In Figure 11, resistors 52, 54 56, 58, 94 and 95 are impedance matching resistors and 55, 57, 91 and 93 are DC blocking capacitors could be provided as useful in maintaining electrical stability of the circuit. Resistors 90, 92 are conventional grounding resistors. The first two amplifiers 60, 62 shown in the circuit as non-inverting voltage followers which isolate the left and right inputs from the rest of the circuit. The outputs of 60 and 62 are each sent to both a summing amplifier (either 84 or 85) as shown in Figure 11 via variable bypass resistors 64 and 66 and resistors 70 and 71. Non-inverting input grounding resistors are provided and indicated in Figure 11 as numerals 78 - 79. Feedback resistors are provided in circuit 50, as indicated by the numerals 75, 76, 86 and 88. The signals are also sent to inverting inputs of differential amplifiers 73 or 74 through the variable differential resistors or controls 68 or 69. Additionally, signals are sent through resistors 70 or 71 to non-inverting differential amplifier inputs.

The stereo source fed into the inputs of circuit 50 will in practical use never have ideal separation. That is, there will always be a signal" in both channels of the program source even for sound sources placed to the left side or the right side of the listener. Ideally, this characteristic of program sources allows the circuit of the type in Figure 11 to be used without introducing out-of-phase information at the outputs of either channel. The introduction of out-of-phase information should be avoided because phase inversion by itself destroys imaging. If, as previously

OM

,-» τ * I

SUBSTIT i k- t H i :ET ^ A^•fy_j. WI

" mentioned, the values of channels are changed so that interchannel differences remain the same and phase inversion does occur, the image location will be correct, but the sound will become increasingly grainy. The purpose of differential amplifiers 73 and 74 is to provide the necessary increase in volume (by adding a signal) or decrease in volume (by subtracting a signal) through resistors 80 and 82.

Whether or not the signal is going to be increased or decreased is dictated by the channel balance of the program source coming to the inputs of the circuit. The ratio of increase to decrease is determined by adjusting controls 68 and 69. This ratio has to be adjusted subjectively by the listener. The 2.255 k values shown in Figure 11 are exemplary as suitable values for speaker placement of forty-five (45) degrees left and right respectively, with respect to the listener's line of sight.

The outputs of amplifiers 73, 74 are then added to corresponding left and right channel signals. The ratio between the original volume signal and the volume modifying signal is controlled by variable resistors 64 and 66. Again, exemplary values are given for 45 degree speaker placement above described. For instance, if the left channel input is the stronger incoming signal by some amount (eg. the maximum difference that can exist in the source material) then this signal is dominant at the non-inverting input of differential amplifier 74. (The degree of dominance being determined by the ratio of the resistor values of 69 and 70 and the strength of the signal present in the right channel.) The non-inverted signal is then added back to the original left channel signal at the input of summing amplifier 84. The result at the output of

SHEET summing amplifier 84 is that the volume of the left channel signal is also present at the inverting input of differential amplifier 73 after it has been attenuated by variable resistor 68. The resistance of resistors 68 and 71 and the signal strength present in the right channel determine the amplitude of the inverted signal present at the output o,f differential amplifier 73. The inverted output of amplifier 73 is subtracted from the original right channel signal at the input summing amplifier 85. Ideally, resistance values of 68 and 66 are adjusted so that the inverted output of 73 is not greater than the right channel signal it is subtracted from so that no phase inversion occurs. (As previously noted, however, this is not a necessary limitation.) The result is that the output of summing amplifier 85 is reduced.

The values of 68 and 69 also determine the difference in volume change rate which produces the head shadow effect. Figures 12 and 13 show relative volume changes for^' one channel illustrating the head shadow effect. In this case both drawings show just the right channel 110 and show the relative change in volume for that channel according to channel balance at the inputs of circuit 50. In Figure 12 there is a solid line indicating the function of the circuit electrically for some resistance values of resistors 68, 69, 64 and 66. If we reduce the resistances of 68 and 69 in the circuit then examine the right channel output, the resistance changes affect it as shown by the dashed line in Figure 12. That is, there is little change in volume level in the right channel 110 when only the right channel is active, but when both channels 112 are operating the volume of the right channel is decreased. When just the left channel 111

HEET ^OMP is on more subtraction occurs in the left channel 111.

The dotted line of Figure 13 schematically shows changes in volume level when the resistance of the bypass resistors 64 and 66 are reduced. This change can be used to reduce the phase inverted signal present in the left channel while increasing the output in monaural arc in the extreme right channel position.

Figures 12 and 13 both show a change in the slope of the volume level curve. The rate of change in volume in the right channel 110 when the signal moves from the middle 112 to the right channel 110 is different from the rate of volume change in the right channel 110 when the incoming signal moves from the middle position to the left channel 111. This volume rate change is the equivalent of providing the head shadow effect, and can be adjusted by changing the resistance values of resistors 68 and 69. Once the head shadow effect is adjusted, the resistance values of variable resistors 64 and 66 could be adjusted so that there is minimum phase inversion occurring at the outputs even under extreme imbalanced conditions (as shown in Figure 13) .

Also the problem with speaker toe-in can be compensated by the circuit in Figure 11, the resistive values of 68, 69, 64 and 66 can be changed so that in "monaural", the differential amplifiers 73, 74 are still in a subtractive mode. The very strong central image which is emphasized by toeing-in loudspeakers can be subtracted from and consequently reduced. These adjustments, of course, would have to be made subjectively by the person who is setting up the playback system. It is possible to adjust these resistances to permit improved monaural playback. As is discussed concerning Figure 2, it is also possible

SUBSTITUTE SHEET to add equalization to the circuit (at point labelled "eg" in Figure 11) which permit compensation for speakers placed to the rear of the listener. Such a speaker arrangement is shown in Figure 16. If the loudspeakers are placed behind the listener and forty- five (45,} degrees to either side of the listener's head, the equalization curve shown in Figure 17 can be used to correct for the non-linear frequency response caused by the location of the speakers relative to the listener.

In Figure 16, there is seen a view of two speakers 12, 14 roughly behind the listener. These situations can be handled by adding an equalization circuit EQ at the positions EQ shown in Figure 11. The equalizations EQ suitable for use in Figure 11 are seen in Figure 17.

Thus, for different speaker angular placement values when they are behind the listener, the equalization as shown in Figure 17 with its position in

Figure 11 being also shown. In Figures 18 and 19 there are seen two geometric speaker layout plans with three speakers in front of the listener as seen in Figure 18, and in Figure 19 one speaker in front of the listener and two speakers behind. With regard to Figure 18, there can be seen listener L as well as three transducers 12, 13, and 14. In the preferred embodiment for layout as shown in Figure 18, a common radius is shown. It should be understood, however, that variations in radius thus varying the distance of each transducer 12-14 from a listener L could be used. In this case, for satisfactory audio results of the volume changes with respect to each transducer 12-14 would be varied to suit the listener.

SUBSTITUTE SHEET j - ^OIP In Figure 27 is shown the^" wiring or connection schematic which teaches the acquiring of circuit 200 between pre-amp 52 and amplifier 11 and circuit 200. Transducers 12-14 are also seen in Figure 27. Referring back to Figure 20 in the drawings, there is seen a circuit 200 which is passive for use with three speaker layouts as shown in Figures 18, 19 and 27.

In Figure 23 there is seen a circuit diagram of the active circuit 250 used with the three speaker embodiments of Figures 18, 19 and 24.

In Figures 18, 19 and 24, three speakers 12-14 are shown in each figure at various angular positions with respect to the listener L.

Equalization is necessary for the embodiments of Figures 19 and 24 but not for the embodiments of Figure 18.

Referring now to Figure 23, there is seen circuit

250, a description of which follows hereinafter. Left and right inputs 253, 255 are seen as well as left, center and right outputs 293, 291 and 295. At numbers

252, 254 are input resistors for summing amp 260, where

256 is a feedback resistor and 258 is the resistor from the non-inverting^' input to ground. The output of 260 is added to the non-inverted output signal of summing amp 270, which is connected to the outputs of 50 (the circuit of Figure 11) via input resistors 262 and

264. 268 is the feedback resistor and 266 is the resistor from the non-inverting input to ground.

The output levels of 270 and 260 are adjusted so that in monaural, a monaural signal component is present at the output of summing amp 280, where 272 and 274 are input resistors, 276 is a feedback resistor, and 278 is the resistor from the non-inverting input to ground.

SUBSTITUTE SHE! This signal (mostly present when monaural signals are present) is then passed through an inverting amplifier 290, via input resistor 282, where 284 is the feedback resistor, and 286 is the resistor from the non-inverting input to ground. (This preserves proper phase between channels.)

The signals for left, mono, or right are then passed through the appropriate equalization sections 294, 292, and 296 in order to provide the appropriate equalization for particular speaker placement.

Volumes between channels are adjusted by potentiometers 297, 298 and 299. For example, if the speaker arrangement of Figure 24 is used, the equalization curve of Figure 25 would be used for the monaural channel (292 of Figure 23) , and the equalization applied by blocks 294 and 296 of Figure 23 would be according to curve of Figure 17.

The circuit of Figure 20 can be considered similar to the circuit of Figure 9 with a monaural stage added. The resistors 202, 208, 212, 214, 216 and 218 perform substantially the same functions as in the Figure 9 resistors 20, 21, 22, 23, 24 and 25 respectively. In addition, resistors 202 and 208 are used as adjustable controls for determining the side loudspeaker signal level relative to the center speaker level. Numbers 201 and 203 represent left and right amplifier grounds respectively, positions 205 and 207 show left and right amplifier outputs.

Variable resistors 204 and 206 supply left and right signals through volume control resistor 210 to center loudspeaker 13. It can be seen that the resistance bridge betwen hot amplifier terminals formed by 204 and 206 allow the weaker hot amplifier terminal to behave as a virtual path to ground which varies with the signal in that channel. This allows the volume of the center speaker to reduce when left and right channel levels are not the same.

Referring to Figure 21 which is. the resistance values for a standard mixed recording shown for this circuit are shown for four B&W DM7 loudspeakers arranged geometrically as shown in Figure 8 of ray copending patent application, U.S. Serial No. 895,309. In Figure 21, 315-318 indicate respectively left amp ground, right amp ground, left amp output and right amp output.

The function of this circuit is basically the same as in the copending patent, but with a more articulate head shadow effect. In Figure 21, the "ratio" bridge formed by resistors 301 and 302 is joined commonly to two additional resistors 303 and 304. This output from the ratio bridge 301 and 302 varies with channel balance because the weaker amplifier terminal behaves as a "virtual" path to ground. Maximum signal at the joining point is when both channels are reproducing a common signal.

The left channel signal is directed to left rear speaker 311 through resistor 305 and to the left front speaker 312 through resistor 306.

The right channel signal is directed to right rear speaker 314 through resistor 308 and to the right front speaker 313 through resistor 307. The ratio of resistors 305 to 306 and of resistors 308 to 307 determine A.O.A.-D.I. ratio.

The amount of mixing between channels for the front speakers as provided by resistor 309, is greater than the amount of mixing between the side speakers, as provided by resistor 310. This difference simulates

SUBSTITUTE SHE the presence of the head shadow.

As can be seen from this arrangement, as a signal becomes predominant in one channel, the output of the side speaker on the weaker channel is reduced faster than the output of the front speaker of the weaker channel, thereby reducing the weaker D.I. component substantially. In this way, the sound blocking effect of the head is more articulately simulated. Consequently, sharper and more accurate image placement is possible.

Referring to Figure 11, for a speaker angle of 30, to left and right sides of a listener, and the speakers are facing straightforward (not "toed-in"), and assuming average crosstalk in the signal source (a phonograph record) the following more important resistance values when inserted into the circuit of figure 11 will produce optimum results:

Inverting differential input resistors 68+69=10.008K ohm Non-inverting differential input resistors 70+71=15K phm

Differential feedback resistors 75+76-10K ohm Differential output resistors 80+82=10K ohm Bypass output resistors 64+66=3.478K ohm Summing amp feedback resistors 86+88=8.87K ohm

Differential input grounding resistors 78+79=10K ohm

The other resistances shown in Figure 11 are less critical to the functioning of the device, and may be chosen to accommodate the input or output impedences and configurations of associated electrical equipment. The circuit of Figure 28 is a variation of the circuit of Figure 11. It incorporates balanced, rather than unbalanced inputs and outputs, plus a switch which allows the differential amplifier stages to be bypassed. Only one channel of the circuit is shown.

Because many varying and different embodiments may be made within the scope of the inventive concept herein taught, and because many modifications may be made in the embodiment herein detailed in accordance with the descriptive requirement of the law, it is to be understood that the details herein are to be interpreted as illustrative and not in a limiting sense. What is claimed as invention is:

Claims

1. An image recovery circuit comprising: a. first transducer mens for generating an acoustic image; b. second transducer means for generating an acoustic image; c. amplifier means for selectively transmitting left and right stereophonic signals, each having a particular amplitude value, to respond to said left and right transducer means; d. image reconstruction means associated- with said left and right transducer means and said amplifier for changing the amplitude of said left and right signals relative to one another.

2. A method of reconstructing sound image placement in audio systems, comprising the steps of: a. providing a pair of transducers; . b. placing the transducers in substantially equi¬ distant spaced positions from the listener at angular positions to the left and right of the listener's line of sight in substantially equiangular positions; c. transmitting left and right stereophonic signals respectively to^' the left and right stereophonic transducers; d. varying the relative amplitude of the left and right stereophonic signals prior to their transmission to the left and right transducers.