CA2090531A1 - Dual time constant audio compression system - Google Patents
Dual time constant audio compression systemInfo
- Publication number
- CA2090531A1 CA2090531A1 CA 2090531 CA2090531A CA2090531A1 CA 2090531 A1 CA2090531 A1 CA 2090531A1 CA 2090531 CA2090531 CA 2090531 CA 2090531 A CA2090531 A CA 2090531A CA 2090531 A1 CA2090531 A1 CA 2090531A1
- Authority
- CA
- Canada
- Prior art keywords
- signal
- gain
- amplifier
- output
- current
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Abandoned
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
- H04R25/00—Deaf-aid sets, i.e. electro-acoustic or electro-mechanical hearing aids; Electric tinnitus maskers providing an auditory perception
- H04R25/35—Deaf-aid sets, i.e. electro-acoustic or electro-mechanical hearing aids; Electric tinnitus maskers providing an auditory perception using translation techniques
- H04R25/356—Amplitude, e.g. amplitude shift or compression
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03G—CONTROL OF AMPLIFICATION
- H03G7/00—Volume compression or expansion in amplifiers
- H03G7/06—Volume compression or expansion in amplifiers having semiconductor devices
- H03G7/08—Volume compression or expansion in amplifiers having semiconductor devices incorporating negative feedback
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
- H04R25/00—Deaf-aid sets, i.e. electro-acoustic or electro-mechanical hearing aids; Electric tinnitus maskers providing an auditory perception
- H04R25/50—Customised settings for obtaining desired overall acoustical characteristics
- H04R25/502—Customised settings for obtaining desired overall acoustical characteristics using analog signal processing
Abstract
ABSTRACT OF THE DISCLOSURE
An audio compression system for a hearing aid. The audio compression system provides a variable gain amplifier for compensating for loudness recruitment. The audio compression system comprises an amplifier having a feedback circuit which includes a current controlled resistor. The gain of the amplifier is varied by using a gain controller to generate a variable current signal. The gain controller utilizies dual averaging detectors to determine the average energy in the amplifier output and generate a corresponding current signal. The averaging detectors are coupled to the output of the amplifier through a rectifier stage which rectifies the amplifer output. The averaging detectors include a slow averaging detector and a fast averaging detector. The slow averaging detector is normally in control, while the fast averaging signal is activated for loud signal levels. To compensate for non-linearity in the operation of the current controlled resistor, the audio compression system can include a pair of comparators which are coupled to the output of the amplifier. The comparators sense the amplifer output signal and provide a pair of symmetrical thresholds about the normal quiescent point of the amplifier output.
An audio compression system for a hearing aid. The audio compression system provides a variable gain amplifier for compensating for loudness recruitment. The audio compression system comprises an amplifier having a feedback circuit which includes a current controlled resistor. The gain of the amplifier is varied by using a gain controller to generate a variable current signal. The gain controller utilizies dual averaging detectors to determine the average energy in the amplifier output and generate a corresponding current signal. The averaging detectors are coupled to the output of the amplifier through a rectifier stage which rectifies the amplifer output. The averaging detectors include a slow averaging detector and a fast averaging detector. The slow averaging detector is normally in control, while the fast averaging signal is activated for loud signal levels. To compensate for non-linearity in the operation of the current controlled resistor, the audio compression system can include a pair of comparators which are coupled to the output of the amplifier. The comparators sense the amplifer output signal and provide a pair of symmetrical thresholds about the normal quiescent point of the amplifier output.
Description
` ^` ~090~3~
BP File No. 0374-138 Title: Dual Time Constant Audio Compression System FIELD OF THE INVENTION
This invention relates to audio amplifiers, more particularly, it relates to an audio compression circuit for use in hearing aids.
BACKGROUND OF THE INVENTION
Sensori neural hearing impairment is characterized by a loss of sensitivity to quiet or low level sounds. The loss of sensitivity can be across the entire frequency spectrum of the affected individual's hearing, or the loss can be frequency dependent. The degree of sensitivity loss generally varies from one hearing impaired individual to another.
Typically, the hearing impaired individual's loss is not accompanied by an equal increase in loudness discomfort, and frequently no increase is observed at all. Thus, the hearing impaired individual can be viewed as having a reduced dynamic range, i.e. the range of low level signals to high level signals which are audible by the individual.
To compensate for the loss of sensitivity or dynamic range, the audible sounds in the individual's environment can be squeezed into the residual dynamic range of the impaired ear. Known hearing aid systems do this by using some form of automatic gain control circuit. The automatic gain control involves using a variable gain 25 amplifier whose output is fed to a level detector which after some processing produces a gain control signal that is used to modify or control the gain of the amplifier.
In the ideal situation, the most gain would be provided for quiet audio signals or sounds, while the gain would be progressively 30 decreased as the sounds become louder and the impaired hearing more closely matches (in terms of loudness perception) that of a person having normal hearing.
BP File No. 0374-138 Title: Dual Time Constant Audio Compression System FIELD OF THE INVENTION
This invention relates to audio amplifiers, more particularly, it relates to an audio compression circuit for use in hearing aids.
BACKGROUND OF THE INVENTION
Sensori neural hearing impairment is characterized by a loss of sensitivity to quiet or low level sounds. The loss of sensitivity can be across the entire frequency spectrum of the affected individual's hearing, or the loss can be frequency dependent. The degree of sensitivity loss generally varies from one hearing impaired individual to another.
Typically, the hearing impaired individual's loss is not accompanied by an equal increase in loudness discomfort, and frequently no increase is observed at all. Thus, the hearing impaired individual can be viewed as having a reduced dynamic range, i.e. the range of low level signals to high level signals which are audible by the individual.
To compensate for the loss of sensitivity or dynamic range, the audible sounds in the individual's environment can be squeezed into the residual dynamic range of the impaired ear. Known hearing aid systems do this by using some form of automatic gain control circuit. The automatic gain control involves using a variable gain 25 amplifier whose output is fed to a level detector which after some processing produces a gain control signal that is used to modify or control the gain of the amplifier.
In the ideal situation, the most gain would be provided for quiet audio signals or sounds, while the gain would be progressively 30 decreased as the sounds become louder and the impaired hearing more closely matches (in terms of loudness perception) that of a person having normal hearing.
- 2 ~ S 3 ~ :
In practical terms, there is a wide variation in the results achieved by the automatic gain control circuits found in the prior art.
Known automatic gain control circuits typically use peak detectors to control the gain of the amplifier. If the input signal's instantaneous value 5 is above a predetermined threshold, then a control signal is produced which charges a capacitor that acts to reduce the gain of the amplifier.
When the instantaneous value of the input signal is below the threshold, the capacitor is discharged, usually by a resistor or a current source. In the art, the charging operation is referred to as the "attack time" and can be 10 designed to a variety of values. The discharging operation is termed the "release time" and can also be set to a variety of values. The choice of the attack and release times (i.e. time constants) involves trade-offs between distortion, audible pumping effects, and the effect of occasional impulsive sounds (e.g. a door slamming) which may occur in the hearing 15 impaired individual's environment. Known systems, typically use short attack times in the range 1 to 5 milliseconds to handle impulsive sounds, while release times of 100 milliseconds are common.
Various attempts have been made in the prior art to overcome these compromises in the design of amplifiers for hearing aids.
2 o While the prior art has achieved some success, there are still problems to be overcome.
A known prior art system utilizes a simple parallel resistor-capacitor combination to achieve a "variable release time". The prior art system teaches using a relatively small capacitor in parallel with 25 a series combination of a larger capacitor (e.g. factor of 10) and a high value resistor. The side branch effectively stores the longer term average of the signal which is developed on the small capacitor. Effectively, this system operates to modify the automatic gain feedback loop only after long duration sounds have occurred. Thus, short duration impulsive 30 sounds do not significantly change the gain setting of the amplifier and performance will be compromised.
Another problem with this known system involves 2 ~
steady state distortion. Since a small capacitor is responsible for the instantaneous gain control signal, the associated small time constant can modulate the instantaneous form of lower frequency waveforms which will result in harmonic distortion.
Another problem with the prior art is due to the use of peak detectors for signal level detection in the automatic gain control loop for hearing aid amplifiers. This problem arises from the operation of the peak detector and the effect of the crest factor in the input signal. The crest factor is defined as the "peak to RMS" ratio, where RMS is the 1 o abbreviation for Root Mean Square. The Crest factor can be approximated by the "peak to average" ratio which also provides a measure of the average signal energy, For a continuous sinusoidal waveform, the crest factor is 3.92 dB. For real life signals, the crest factor is usually greater and 15 dependent on the type of signal. For example, the crest factor for music is different from the crest factor for speech. Crest factors in the range 12 to 18 dB are common in real life environments of the hearing impaired individual.
Peak detectors, by the nature of their design, are 2 o responsive to the crest factor. Peak detectors typically utilize some form of comparator to determine if the instantaneous value of the signal is above the pre-determined threshold. If the instantaneous value is above the threshold, then a control signal is produced to lower the gain of the amplifier. The control loop for varying the gain is usually sufficiently 2 5 high to yield a high compression ratio input-output characteristic for the audio amplifier. The high gain of the control loop, however, means that the peaks of the signal effectively determine the operating point. Thus, the amplifier output signal will be substantially constant for input signals above the pre-determined threshold, and the attack and release 30 parameters will result in the tracking of the peaks in the signal. This means that an audio circuit using a peak detector will be responsive to the crest factor, and therefore the circuit will vary the gain of the 2 ~ 3 ~
amplifier irrespective of the signal's average energy.
Since the human ear uses the RMS power of an audio signal to determine loudness, the crest factor becomes a very important parameter in the design of loudness detection circuits. Because the peak 5 detector is sensitive to the peak level of the signal, the detector will see differences as a result of the Crest factor. The human ear, on the other hand, is responsive to the energy level in a signal and not the Crest factor. Therefore, to provide accurate loudness detection and dynamic range compression, ideally an RMS or average energy detector should be 1 o used.
BRIEF SUMMARY OF THE INVENTION
The present invention provides an audio compression system suitable for use in hearing aids. The audio compression system 15 according to the present invention is responsive to the average signal energy in the user's environment. The audio compression system includes two averaging detectors: a slow path detector and a fast path detector. The averaging detectors are used to control the gain of the amplifier. The slow path detector is normally in control of the system 20 gain, but control is switched to the fast path upon occurrence of a fast signal event.
The audio compression circuit according to the present invention is intended to compensate for loudness recruitment, where the highest gain is required for quiet or low level audio signals and 25 progressively less gain is applied to the amplifier as the incoming audio signals increase in loudness.
The audio compression circuit according to the present invention has the advantage that the averaging detectors avoid the over reaction characteristic which is typically exhibited by the peak detec$or 3 0 gain control circuits utilized in known hearing aid designs.
In a first aspect, the present invention provides a hearing aid including microphone means for receiving an audio input signal and ~ 2~9~31 converting the audio input signal into an electrical input signal, a variable gain amplifier coupled to the microphone means, and signal processing m.eans coupled to transducer means for processing the electrical signal and producing an audio output signal which is coupled 5 to the ear of a user, said variable gain amplifier comprising: ( a ) a n amplifier stage having an output port and an input port for receiving the electrical input signal, and including variable gain means for producing a output signal having a gain level at said output port; (b) a rectifier stage coupled to the output port of said amplifier stage and having means for 10 rectifying said output signal to produce a rectified signal; (c) averaging detector means for generating an averaged signal from said rectified signal, said averaging detector means being coupled to said rectifier stage;
(d) gain controller means coupled to said averaging detector means and said amplifier stage, said gain control means being responsive to said 15 averaged signal and having means for producing a gain control signal;
and (e) said variable gain means in said amplifier stage having means . ~
responsive to said gain control signal for varying the gain level of said :
amplifier stage. ::
BRIEF DESCRIPl[lON.OF rE~E DRAWINGS
For a better understanding of the present invention, and to show more clearly how it may be carried into effect, reference will now :;
be made, byway of example, to the accompanying drawings which show 2 5 preferred embodiments of the present invention~
Figure 1 shows in block diagram form a dual time constant audio compression system according to the preferred embodiment of the present invention; : :;
Figure 2 shows the transfer function for the audio 3 o compression system of Figure l; .
Figure 3 shows in schematic form one implementation of the audio compression system of Figure l; and ' . ', . . " . . "," '' ' , ' "- .~, .' '' .' : ., ' ` ~' 2 ~ 3 ~
Figures 4(a) and 4(b) show in detailed schematic form the averaging detector circuit for the audio compression system of Figure 3.
5 DETAILED DES~RIlrrION OF T~iE PREFERRED EMBODIMENTS
Figure 1 shows a dual time constant audio compression circuit indicated by reference 10 according to the present invention. The audio compression circuit 10 provides a variable gain amplifier stage which is suitable for use in a hearing aid (indicated generally by reference 10 11). The audio compression circuit 10 or variable gain amplifier stage in response to an input signal 12, e.g. an audio signal, produces an output signal 14 having a variable gain.
According to the present invention, the gain of the amplifier stage 10 is a function of the average signal energy in the hearing 15 aid user's environment. While the human ear uses the RMS power to determine loudness, it has been found that the average signal energy can provide excellent performance characteristics in a hearing aid 11. A
simple examination of various speech and music signals has shown that where the Peak to RMS ratios were between 12 to 20 dB, the 20 corresponding RMS to Average ratios were in the range 1 to 3 db Furthermore, when a compression ratio of 2:1 is used, this error is reduced by a factor of 2 (i.e. 0.5 to 1.5 dB), which makes the error arising from average energy level almost inaudible. Therefore, the average signal energy can produce a good approximation for the RMS energy 2 5 level.
In a typical hearing aid 11, the audio compression circuit (or variable gain amplifier stage) 10 is coupled to an input transducer 16 such as a microphone through a coup!ing capacitor 18. The transducer 16 converts an audio input signal into the electrical input signal 12. The 3 0 output of the variable gain amplifier stage 10 is coupled to further signal processing circuits, indicated by reference 20. The signal processing circuit 20 performs further processing of the signal 14 prior to being fed to an `::. . `
output transducer 22, for example, the receiver in a hearing aid 11. The gain of the audio compression circuit 10 is controlled by a gain controller 13, as will be discussed in detail below.
For a hearing aid application, the purpose of the variable 5 gain amplifier 10 is to provide a transfer function 24 as shown in Figure 2. Referring to Figure 2, the transfer function 24 for the variable gain amplifier 10 has three distinct segments 26,28,30. The first segment 26 provides the most gain and is active for quiet sounds, that is, low level input signals 12. As the sound level of the user's environment increases, 1 o the gain provided by the amplifier 10 progressively decreases as shown by the second and third segments 2~,30. In the third segment 30, the gain is unity which corresponds to loud sound level environments, where as will be appreciated by one skilled in the art, the impaired hearing of a user more closely matches that of normal users in terms of loudness 1 5 perception.
In practical hearing aid design, the gain of the amplifier 10 is made constant below a loudness threshold level of 40 to 50 dBspl. As will be understood by one skilled in the art, the constant gain is necessary so that the amplifier 10 does not unduly amplify the noise which can 2o originate from the microphone 16. Above the loudness threshold level, the gain is selected to provide a compression ratio of 2:1, thus, a 10 dB
increase in the input signal 12 is reduced to a 5 dB increase in the output signal 14. For high level input signals 12, e.g. 90 dBspl where loudness growth becomes more normal, the gain of the amplifier 10 becomes 2 5 linear and unity, as indicated by third segment 30 in Figure 2.
To produce the transfer function 24 shown in Figure 2, the audio compression circuit 10 uses two averaging detectors to set the gain. The first detector is usually in control and is characterized by long time constants (i.e. attack and release times) so that the detector reacts 30 slowly and smoothly to changing input signals 12. The second detector, on the other hand, is characterized by relatively fast time constants, e.g. 20 times. The short time constants enable the second detector to react to 2 ~ 3 :1~
abnormal changes in the input signal 12, such as would occur if a door was slammed in the user's environment.
As will be understood by one skilled in the art of hearing aid amplifier design, an amplifier circuit which uses averaging detectors 5 to modify the gain more closely resembles the hearing mechanism in an ear. The ear uses the RMS power to determine the loudness of a signal, and not the peak amplitude of the signal. Therefore, in designing a hearing aid circuit to compensate for loudness, the circuit should determine the loudness of signals in the user's environment using the 1 o RMS value of the signals, which as explained above can be approximated by averaging the signal.
Referring back to Figure 1, the audio compression amplifier stage or variable gain amplifier 10 comprises an amplifier stage 32 and the gain controller 13. The amplifier stage 32 has an input port for 1 5 the input signal 12 which is coupled to the microphone 16 through the capacitor 18 and an output port which couples the output signal 14 to the signal processing circuits 20. The output of the amplifier stage 32 is also coupled to the gain controller 13. As will be explained in detail, the gain controller 13 produces a gain control signal 33, which varies the gain of 2 0 the amplifier stage 32, in order to compress the dynamic range presented to the user's ear.
Referring still to Figure 1, the gain controller 13 includes a slow averaging detector 34 and a fast averaging detector 36. The function of the two averaging detectors 34,36 is to determine the loudness 25 level in the user's environment and generate the appropriate control signal 33 to modify the gain of the amplifier stage 32. The control signal 33 is generated by a gain control stage 38 which is coupled to each of the averaging detectors 34,36. The averaging detectors 34,36 are coupled to the output of the amplifier stage 32 through a rectifier circuit 40. The gain 3 0 controller 13 also includes a comparator stage 42 which is coupled to the fast averaging detector 36.
The operation and structure of each functional block in 3 ~
g Figure 1 will now be explained in detail with reference to Figure 3. In Figure 3, the corresponding elements from Figure 1 are depicted using broken line outline and like reference numbers.
Referring to Figure 3, the amplifier stage 32 comprises an 5 operational amplifier 44 in a negative feedback configuration. The inverting input of the operational amplifier 44 is connected to the coupling capacitor 18 through an input resistor 46. The non-inverting input of the operational amplifier 44 can be connected to a voltage reference VQ. In a typical hearing aid application, the output of the 1 0 operational amplifier 44 is coupled to additional signal processing circuits which indicated generally by reference 20 (in Figure 1). The signal processing circuits 20 in known manner process the output signal 14 from the amplifier 10 (i.e. output of the operational amplifier 44) before the signal is fed to the output transducer 22.
1 5 As shown in Figure 3, negative feedback for the amplifier 44 is provided through a feedback resistor 48. The inverting input of the operational amplifier 44 is connected to the output of the operational amplifier 44 through the feedback resistor 48. In known manner, the gain of the operational amplifier 44 is given by the quotient of the resistance 2 o value of the feedback resistor 48 to that of the input resistor 46 (in sum with any output impedance associated with the microphone 16). Changes in the value of the feedback resistor 48 will cause corresponding changes in the gain of the operational amplifier 44, and therefore changes in the gain of the audio compression or amplifer circuit 10.
In the preferred embodiment of the present invention, the feedback resistor 48 comprises a current controlled resistor. The resistance value of a current controlled resistor 48 is a function of the gain control signal 33 which as shown in Figure 3 comprises a gain control current IGAIN- AS will be understood by one skilled in the art, 3 0 varying the control current IGAIN will change the resistance value of the current controlled resistor 48. In the present invention, the gain of the ., . .... . ~. - - ? ' ' ', -2~iS~ -~S3~, amplifier stage 10 is modified by adjusting the gain control current IGAIN
to vary the resistance of the current controlled resistor 48.
As shown in Figure 3, the gain control current IGAIN is fed to the current controlled resistor 48 by the gain control stage 38. The gain 5 control stage comprises a current summer 50 having three inputs and an output. The first input is connected to a threshold current reference 52, the second input is connected to a first variable current reference 54, and the third input is connected to a second variable current reference 56.
The threshold current reference 52 produces a reference 10 current Ithl. The current reference 52 comprises a current sink which in known manner is designed to sink the current Ithl. The constant gain of the amplifier stage 10 is achieved by the use of the current reference 52.
The constant gain is made a function of the magnitude of the current I
by designing the first variable current reference 54 and the second 15 variable current reference 56 to be zero below the loudness threshold level, e.g. 40 to 50 dBspl, as will be explained below. Therefore, the choice of the magnitude of the threshold current Ithl determines the maximum gain of the amplifier stage 10.
The rectifier stage 40, as shown in Figure 3, comprises a 2 o rectifier circuit 58, and first and second current sources 60,62 respectively.
The rectifier circuit 58 can be implemented in known manner using solid state switches such as diodes (not shown). Furthermore, the rectifier circuit 58 can be a full-wave or a half-wave design. This design choice will not result in substantial differences in performance of the amplifier stage 2 5 10, except for the factor of 2 difference in the average signal energy, as will be understood by one skilled in the art. This scale factor of 2 can be easily accommodated by the appropriate choice of other parameters in the gain controller 13.
The output of the rectifier 58 is coupled to both current 3 o sources 60,62. The two current sources 60,62 are identical and implemented, in known manner, as voltage controlled current sources.
~ 3 ~
The first current source 60 produces a first rectified output current IRECTI-The second current source 62 produces a second rectified output current IRECT2 which is equal to the current IRECTI- Therefore, the instantaneous values of the rectified currents IRECTI and IRECT2 are proportional to the 5 rectified (i.e. absolute) instantaneous voltage level of the input signal 12.
In known systems which use peak detector circuits, there is more sensitivity to half-wave rectification. This occurs because peaks in the input signal will not always be symmetrical about zero. Hence, it is possible that a peak in the input signal 12 will be opposite in phase to the 1 o sensitivity of peak detector and be missed by the detector. In contrast, thevariable amplifier 10 according to the present invention obviates the problem by using averaging detectors 34,36 to determine the average energy in the input signal 12. Moreover, the coupling capacitor 18 acts to make the average energy in the input signal 12 symmetrical.
A feature of the amplifier stage 10 is the use of dual averaging detectors 34,36 to generate the control signal 33 (i.e. IGAIN) for varying the gain of the amplifier stage 10 according to the average energy (i.e. loudness level) of the input signal 12. To explain how the control signal 33 is generated, the operation of the slow averaging detector 34 is 2 o first considered because it is the detector which is usually in operation.
The rectified output current IRECTI is applied to the slow averaging circuit 34. The averaging operation is achieved by feeding the current IRECTI into the combination of a capacitor 61, a resistor 63 and an operational amplifer 65 shown in Figure 4(a). It will be understood by one 2 5 skilled in the art that the current flowing in the resistor 63 is representative ~f the average of the cùrrent IRECT1. This current is sensed using known techniques and replicated by two current sources 66,68, which produce identical averaging output currents ISLOWI and ISLOW2-Reference is now made to Figure 4(b), which shows the 3 0 details of the slow averaging circuit 64 of Figure 4(a). The operation of theslow averaging circuit 64 is evident from Figures 4(a) and (b) and within ~ .
2~ 3~3~
the understancling of one skilled in the art. The circuit for the fast averaging circuit 74 is virtually identical to the one shown in Figure 4.
The only difference is found in the resulting time constants. In the preferred embodiment, the slow averaging circuit 64 produces a time constant of 200 milliseconds and the fast averaging circuit 74 produces a time of constant of 10 milliseconds (i.e. a factor of 20). The selection of the values for respective circuit elements shown in Figure 4 to produce these respective time constants is within the capability of one skilled in the art.
Referring back to Figure 3, the averaging output current 1 O ISLOWI is compared to a second threshold current Ith2 which is producedby a current source 70 in the first variable current reference 54. The current source 70 is coupled to a current mirror 72 which is formed from transistors Ql and Q2. The difference between the averaging current ISLOW1 and the threshold reference current Ithl is produced or mirrored at the collector of transistor Q2, i.e. the output of the current mirror 72. The design of the current mirror 72 formed from transistors Ql and Q2 is within the capability of one skilled in the art and has the characteristic of only being able to sink current at the output, i.e. the collector of transistor 2 o Referring still to Figure 2, if the averaging current ISLOW1 is less than the threshold current Ith2, then transistor Ql will not conduct and according to the operation of the current mirror 72, there will be no collector current in transistor Q2. As the averaging current ISLOW1 increases, in response to an increasing input signal 12 (i.e. louder user 2 5 environment), it will eventually exceed the value of the current Ith2 and this difference in the currents will be mirrored as the collector current of transistor Q2. The collector current of transistor Q2 is added to the first threshold current Ithl by the current summer 50 to produce the gain control current IGAIN which sets the gain of the operational amplifier 44 3 0 through the current controlled resistor 48.
From the foregoing discussion, it can be seen that there are two regions of operation for the slow averaging stage 34. In the first region of operation, the average energy of the input signal 12 is below the threshold level (i.e. the current ISLOWI is less than the reference current 5 Ith2) and the gain control current IGAIN is equal to the first threshold current Ithl- In the second region of operation, the average energy of the input signal 12 is above the threshold level. This means that the averaging current ISLOWI exceeds the current Ith2 and the difference between these two currents is mirrored in the current mirror 72 and 10 added to the current Ithl to produce the gain control current IGAIN- The gain control current IGAIN can be determined according to the following equation:
IGAIN = Ithl + (ISLOWI - Ith2) = ISLOWI ; given Ithl = Ith2 Since the threshold currents Ithl and Ith2 are designed to be identical, the gain control current IGAIN is equal to the averaging output current ISLOWI
and therefore there is a linear relationship between the control current 2 0 IGAIN and the averaging output current ISLOW1-In the preferred embodiment of the present invention,the resistance value of the current controlled resistor 48 is inversely proportional to the magnitude of the gain control current IGAIN, and above the threshold loudness level, the compression ratio is 2:1. The 2:1 25 compression ratio can be illustrated as follows. If the output signal 14 increases by a factor of 2 (or 6 dB), the gain control current IGAIN which is a function of the averaging current ISLOWI will also increase by the same factor, as shown in the above equation. Since the value of the resistor 48 is inversely proportional to the gain control current IGAlN, the input 30 signal 12 must have increased by a factor of 4 (or 12 dB). This follows from the gain equation of the inverting operational amplifier 44. In other ~, -9 ~ 3 ~ ~:
words, if the output signal 14 increases by 6 dB, then the gain will decrease by 6 dB, and therefore, the input signal 12 must have increased by 12 dB, which provides the desired 2:1 compression ratio.
Referring back to Figure 3, the operation of the fast 5 averaging stage 36 is considered. The current IRECT2 produced by current source 62 in the rectifier stage 40 is identical to current IRECT1. The current IRECT2 is applied to the fast averaging detector 36. The operation of the fast averaging detector 36 is very similar to the operation of the slow averaging detector 34 with the primary difference being the time 10 constants employed in the design of the fast averaging circuit 74. The time constants of the fast averaging circuit 74 are designed to be much shorter than those of the slow averaging circuit 64. In addition, the faster averaging detector 36 only requires one current source which generates `~
an averaging output current IFASTI ` ~
As shown in Figure 3, the averaging output current IFASTI `
is compared to the output of current source 68 which produces a current ISLOW~- The current IFAST1 is compared to current ISLOW2 through a current mirror 76 formed from two transistors Q3 and Q4. By choosing the emitter area ratios of transistor Q4 to transistor Q3 as N:1, the dynamic 2o threshold can be set to determine the amount that the fast occurring event (i.e. IFASTI) must exceed the slower moving averaging current ~ ~ ~
ISLOW2 in order to assume gain control of the amplifier stage 32. The ~ ;;dynamic thxeshold can be determined according to the following -expression:
dynamic threshold = 20 loglo (N) ;- ~
According to the above expression, if N = 2, then the fast averaging .-~.
current IFASTI must exceed the slow averaging current ISLOW2 by 6 dB.
The difference between the fast averaging current IFAST1 and the scaled slow averaging current ISLOW2 is mirrored by a current -2 ~
mirror 78 which is formed from two transistors Q5 and Q6. The difference is reproduced or mirrored as the collector current of transistor Q6. The collector of transistor Q6 provides the third input to the current summer 50.
Similar to the operation of the current mirror 72 formed from transistors Q1 and Q2, the current mirror 78 will produce zero output current (i.e. collector current of transistor Q6) if the fast averaging current IFAST1 is less than N times the slow averaging current ISLOW2- If the fast averaging current IFAST1 exceeds the scaled slow averaging current 1 O ISLOW2~ then difference is reproduced as the collector current of transistorQ6- The current summer 50 adds the collector current of transistor Q6 to the first threshold current Ithl to produce the gain control current IGAIN
which reduces the gain of the amplifier stage 32 as explained above.
In practical systems, the current controlled resistor 48 will 15 have limitations in terms of linear operation. This non-linearity typically manifests itself as distortion at very high level input signals 12. The function of the comparator stage 42 is to compensate for this non-linearity.
As shown in Figure 3, the comparator stage 42 comprises 20 first and second comparators 80 and 82. Each comparator 80,82 can be implemented using an open loop operational amplifier and known design techniques. The first comparator 80 is configured to process positive going signals 14 with a voltage reference 84 connected to the non-inverting input of the comparator 80. The output of the comparator is 25 connected to the base of a drive transistor Q7. The emitter of the drive transistor Q7 is tied to a po-vver supply rail 90. The collector (output) of thetransistor Q7 is tied to the input of the fast averaging detector 36 which is also connected to the output of the current source 62. In a similar fashion, the second comparator 82 is coupled to a second voltage reference 86 and 30 a second drive transistor 92. The collector of the second drive transistor 92 is also tied to the input of the fast averaging detector 36. By connecting the second voltage reference 86 to the inverting input of the comparator 82, the second comparator 82 is configured to process negative going output signals 14 from the operational amplifier 44.
The function of the comparator stage 42 is to provide 5 symmetrical thresholds about the normal quiescent point of the output from the operational amplifier 44. The use of high gain comparators 80,82 connected to the fast averaging detector also effectively produces a peak detecting high compression ratio automatic gain control circuit with the accompanying short time constants (from the fast averaging circuit 74). In 10 practical terms, this means that the gain controller 13 according to the present invention produces a quick response time which lowers distortion at high signal levels while providing very fast recovery due to the short time constants in the fast averaging circuit 74. The voltage references 84,86 are selected so that they are seldom exceeded while still 15 obtaining optimum distortion performance from the amplifier as will be understood by one skilled in the art.
The invention has now been explained with reference to specific embodiments. Other embodiments will be apparent to those of ordinary skill in art in view of this description. For example, the feedback 20 circuit for the amplifier has been described as including a current controlled resistor. It would be possible to design the system in the voltage domain by using a voltage controlled resistive device in the feedback loop and generating a voltage based gain control signal. It is therefore not intended that the invention be limited except as indicated 25 bytheappendedclaims.
In practical terms, there is a wide variation in the results achieved by the automatic gain control circuits found in the prior art.
Known automatic gain control circuits typically use peak detectors to control the gain of the amplifier. If the input signal's instantaneous value 5 is above a predetermined threshold, then a control signal is produced which charges a capacitor that acts to reduce the gain of the amplifier.
When the instantaneous value of the input signal is below the threshold, the capacitor is discharged, usually by a resistor or a current source. In the art, the charging operation is referred to as the "attack time" and can be 10 designed to a variety of values. The discharging operation is termed the "release time" and can also be set to a variety of values. The choice of the attack and release times (i.e. time constants) involves trade-offs between distortion, audible pumping effects, and the effect of occasional impulsive sounds (e.g. a door slamming) which may occur in the hearing 15 impaired individual's environment. Known systems, typically use short attack times in the range 1 to 5 milliseconds to handle impulsive sounds, while release times of 100 milliseconds are common.
Various attempts have been made in the prior art to overcome these compromises in the design of amplifiers for hearing aids.
2 o While the prior art has achieved some success, there are still problems to be overcome.
A known prior art system utilizes a simple parallel resistor-capacitor combination to achieve a "variable release time". The prior art system teaches using a relatively small capacitor in parallel with 25 a series combination of a larger capacitor (e.g. factor of 10) and a high value resistor. The side branch effectively stores the longer term average of the signal which is developed on the small capacitor. Effectively, this system operates to modify the automatic gain feedback loop only after long duration sounds have occurred. Thus, short duration impulsive 30 sounds do not significantly change the gain setting of the amplifier and performance will be compromised.
Another problem with this known system involves 2 ~
steady state distortion. Since a small capacitor is responsible for the instantaneous gain control signal, the associated small time constant can modulate the instantaneous form of lower frequency waveforms which will result in harmonic distortion.
Another problem with the prior art is due to the use of peak detectors for signal level detection in the automatic gain control loop for hearing aid amplifiers. This problem arises from the operation of the peak detector and the effect of the crest factor in the input signal. The crest factor is defined as the "peak to RMS" ratio, where RMS is the 1 o abbreviation for Root Mean Square. The Crest factor can be approximated by the "peak to average" ratio which also provides a measure of the average signal energy, For a continuous sinusoidal waveform, the crest factor is 3.92 dB. For real life signals, the crest factor is usually greater and 15 dependent on the type of signal. For example, the crest factor for music is different from the crest factor for speech. Crest factors in the range 12 to 18 dB are common in real life environments of the hearing impaired individual.
Peak detectors, by the nature of their design, are 2 o responsive to the crest factor. Peak detectors typically utilize some form of comparator to determine if the instantaneous value of the signal is above the pre-determined threshold. If the instantaneous value is above the threshold, then a control signal is produced to lower the gain of the amplifier. The control loop for varying the gain is usually sufficiently 2 5 high to yield a high compression ratio input-output characteristic for the audio amplifier. The high gain of the control loop, however, means that the peaks of the signal effectively determine the operating point. Thus, the amplifier output signal will be substantially constant for input signals above the pre-determined threshold, and the attack and release 30 parameters will result in the tracking of the peaks in the signal. This means that an audio circuit using a peak detector will be responsive to the crest factor, and therefore the circuit will vary the gain of the 2 ~ 3 ~
amplifier irrespective of the signal's average energy.
Since the human ear uses the RMS power of an audio signal to determine loudness, the crest factor becomes a very important parameter in the design of loudness detection circuits. Because the peak 5 detector is sensitive to the peak level of the signal, the detector will see differences as a result of the Crest factor. The human ear, on the other hand, is responsive to the energy level in a signal and not the Crest factor. Therefore, to provide accurate loudness detection and dynamic range compression, ideally an RMS or average energy detector should be 1 o used.
BRIEF SUMMARY OF THE INVENTION
The present invention provides an audio compression system suitable for use in hearing aids. The audio compression system 15 according to the present invention is responsive to the average signal energy in the user's environment. The audio compression system includes two averaging detectors: a slow path detector and a fast path detector. The averaging detectors are used to control the gain of the amplifier. The slow path detector is normally in control of the system 20 gain, but control is switched to the fast path upon occurrence of a fast signal event.
The audio compression circuit according to the present invention is intended to compensate for loudness recruitment, where the highest gain is required for quiet or low level audio signals and 25 progressively less gain is applied to the amplifier as the incoming audio signals increase in loudness.
The audio compression circuit according to the present invention has the advantage that the averaging detectors avoid the over reaction characteristic which is typically exhibited by the peak detec$or 3 0 gain control circuits utilized in known hearing aid designs.
In a first aspect, the present invention provides a hearing aid including microphone means for receiving an audio input signal and ~ 2~9~31 converting the audio input signal into an electrical input signal, a variable gain amplifier coupled to the microphone means, and signal processing m.eans coupled to transducer means for processing the electrical signal and producing an audio output signal which is coupled 5 to the ear of a user, said variable gain amplifier comprising: ( a ) a n amplifier stage having an output port and an input port for receiving the electrical input signal, and including variable gain means for producing a output signal having a gain level at said output port; (b) a rectifier stage coupled to the output port of said amplifier stage and having means for 10 rectifying said output signal to produce a rectified signal; (c) averaging detector means for generating an averaged signal from said rectified signal, said averaging detector means being coupled to said rectifier stage;
(d) gain controller means coupled to said averaging detector means and said amplifier stage, said gain control means being responsive to said 15 averaged signal and having means for producing a gain control signal;
and (e) said variable gain means in said amplifier stage having means . ~
responsive to said gain control signal for varying the gain level of said :
amplifier stage. ::
BRIEF DESCRIPl[lON.OF rE~E DRAWINGS
For a better understanding of the present invention, and to show more clearly how it may be carried into effect, reference will now :;
be made, byway of example, to the accompanying drawings which show 2 5 preferred embodiments of the present invention~
Figure 1 shows in block diagram form a dual time constant audio compression system according to the preferred embodiment of the present invention; : :;
Figure 2 shows the transfer function for the audio 3 o compression system of Figure l; .
Figure 3 shows in schematic form one implementation of the audio compression system of Figure l; and ' . ', . . " . . "," '' ' , ' "- .~, .' '' .' : ., ' ` ~' 2 ~ 3 ~
Figures 4(a) and 4(b) show in detailed schematic form the averaging detector circuit for the audio compression system of Figure 3.
5 DETAILED DES~RIlrrION OF T~iE PREFERRED EMBODIMENTS
Figure 1 shows a dual time constant audio compression circuit indicated by reference 10 according to the present invention. The audio compression circuit 10 provides a variable gain amplifier stage which is suitable for use in a hearing aid (indicated generally by reference 10 11). The audio compression circuit 10 or variable gain amplifier stage in response to an input signal 12, e.g. an audio signal, produces an output signal 14 having a variable gain.
According to the present invention, the gain of the amplifier stage 10 is a function of the average signal energy in the hearing 15 aid user's environment. While the human ear uses the RMS power to determine loudness, it has been found that the average signal energy can provide excellent performance characteristics in a hearing aid 11. A
simple examination of various speech and music signals has shown that where the Peak to RMS ratios were between 12 to 20 dB, the 20 corresponding RMS to Average ratios were in the range 1 to 3 db Furthermore, when a compression ratio of 2:1 is used, this error is reduced by a factor of 2 (i.e. 0.5 to 1.5 dB), which makes the error arising from average energy level almost inaudible. Therefore, the average signal energy can produce a good approximation for the RMS energy 2 5 level.
In a typical hearing aid 11, the audio compression circuit (or variable gain amplifier stage) 10 is coupled to an input transducer 16 such as a microphone through a coup!ing capacitor 18. The transducer 16 converts an audio input signal into the electrical input signal 12. The 3 0 output of the variable gain amplifier stage 10 is coupled to further signal processing circuits, indicated by reference 20. The signal processing circuit 20 performs further processing of the signal 14 prior to being fed to an `::. . `
output transducer 22, for example, the receiver in a hearing aid 11. The gain of the audio compression circuit 10 is controlled by a gain controller 13, as will be discussed in detail below.
For a hearing aid application, the purpose of the variable 5 gain amplifier 10 is to provide a transfer function 24 as shown in Figure 2. Referring to Figure 2, the transfer function 24 for the variable gain amplifier 10 has three distinct segments 26,28,30. The first segment 26 provides the most gain and is active for quiet sounds, that is, low level input signals 12. As the sound level of the user's environment increases, 1 o the gain provided by the amplifier 10 progressively decreases as shown by the second and third segments 2~,30. In the third segment 30, the gain is unity which corresponds to loud sound level environments, where as will be appreciated by one skilled in the art, the impaired hearing of a user more closely matches that of normal users in terms of loudness 1 5 perception.
In practical hearing aid design, the gain of the amplifier 10 is made constant below a loudness threshold level of 40 to 50 dBspl. As will be understood by one skilled in the art, the constant gain is necessary so that the amplifier 10 does not unduly amplify the noise which can 2o originate from the microphone 16. Above the loudness threshold level, the gain is selected to provide a compression ratio of 2:1, thus, a 10 dB
increase in the input signal 12 is reduced to a 5 dB increase in the output signal 14. For high level input signals 12, e.g. 90 dBspl where loudness growth becomes more normal, the gain of the amplifier 10 becomes 2 5 linear and unity, as indicated by third segment 30 in Figure 2.
To produce the transfer function 24 shown in Figure 2, the audio compression circuit 10 uses two averaging detectors to set the gain. The first detector is usually in control and is characterized by long time constants (i.e. attack and release times) so that the detector reacts 30 slowly and smoothly to changing input signals 12. The second detector, on the other hand, is characterized by relatively fast time constants, e.g. 20 times. The short time constants enable the second detector to react to 2 ~ 3 :1~
abnormal changes in the input signal 12, such as would occur if a door was slammed in the user's environment.
As will be understood by one skilled in the art of hearing aid amplifier design, an amplifier circuit which uses averaging detectors 5 to modify the gain more closely resembles the hearing mechanism in an ear. The ear uses the RMS power to determine the loudness of a signal, and not the peak amplitude of the signal. Therefore, in designing a hearing aid circuit to compensate for loudness, the circuit should determine the loudness of signals in the user's environment using the 1 o RMS value of the signals, which as explained above can be approximated by averaging the signal.
Referring back to Figure 1, the audio compression amplifier stage or variable gain amplifier 10 comprises an amplifier stage 32 and the gain controller 13. The amplifier stage 32 has an input port for 1 5 the input signal 12 which is coupled to the microphone 16 through the capacitor 18 and an output port which couples the output signal 14 to the signal processing circuits 20. The output of the amplifier stage 32 is also coupled to the gain controller 13. As will be explained in detail, the gain controller 13 produces a gain control signal 33, which varies the gain of 2 0 the amplifier stage 32, in order to compress the dynamic range presented to the user's ear.
Referring still to Figure 1, the gain controller 13 includes a slow averaging detector 34 and a fast averaging detector 36. The function of the two averaging detectors 34,36 is to determine the loudness 25 level in the user's environment and generate the appropriate control signal 33 to modify the gain of the amplifier stage 32. The control signal 33 is generated by a gain control stage 38 which is coupled to each of the averaging detectors 34,36. The averaging detectors 34,36 are coupled to the output of the amplifier stage 32 through a rectifier circuit 40. The gain 3 0 controller 13 also includes a comparator stage 42 which is coupled to the fast averaging detector 36.
The operation and structure of each functional block in 3 ~
g Figure 1 will now be explained in detail with reference to Figure 3. In Figure 3, the corresponding elements from Figure 1 are depicted using broken line outline and like reference numbers.
Referring to Figure 3, the amplifier stage 32 comprises an 5 operational amplifier 44 in a negative feedback configuration. The inverting input of the operational amplifier 44 is connected to the coupling capacitor 18 through an input resistor 46. The non-inverting input of the operational amplifier 44 can be connected to a voltage reference VQ. In a typical hearing aid application, the output of the 1 0 operational amplifier 44 is coupled to additional signal processing circuits which indicated generally by reference 20 (in Figure 1). The signal processing circuits 20 in known manner process the output signal 14 from the amplifier 10 (i.e. output of the operational amplifier 44) before the signal is fed to the output transducer 22.
1 5 As shown in Figure 3, negative feedback for the amplifier 44 is provided through a feedback resistor 48. The inverting input of the operational amplifier 44 is connected to the output of the operational amplifier 44 through the feedback resistor 48. In known manner, the gain of the operational amplifier 44 is given by the quotient of the resistance 2 o value of the feedback resistor 48 to that of the input resistor 46 (in sum with any output impedance associated with the microphone 16). Changes in the value of the feedback resistor 48 will cause corresponding changes in the gain of the operational amplifier 44, and therefore changes in the gain of the audio compression or amplifer circuit 10.
In the preferred embodiment of the present invention, the feedback resistor 48 comprises a current controlled resistor. The resistance value of a current controlled resistor 48 is a function of the gain control signal 33 which as shown in Figure 3 comprises a gain control current IGAIN- AS will be understood by one skilled in the art, 3 0 varying the control current IGAIN will change the resistance value of the current controlled resistor 48. In the present invention, the gain of the ., . .... . ~. - - ? ' ' ', -2~iS~ -~S3~, amplifier stage 10 is modified by adjusting the gain control current IGAIN
to vary the resistance of the current controlled resistor 48.
As shown in Figure 3, the gain control current IGAIN is fed to the current controlled resistor 48 by the gain control stage 38. The gain 5 control stage comprises a current summer 50 having three inputs and an output. The first input is connected to a threshold current reference 52, the second input is connected to a first variable current reference 54, and the third input is connected to a second variable current reference 56.
The threshold current reference 52 produces a reference 10 current Ithl. The current reference 52 comprises a current sink which in known manner is designed to sink the current Ithl. The constant gain of the amplifier stage 10 is achieved by the use of the current reference 52.
The constant gain is made a function of the magnitude of the current I
by designing the first variable current reference 54 and the second 15 variable current reference 56 to be zero below the loudness threshold level, e.g. 40 to 50 dBspl, as will be explained below. Therefore, the choice of the magnitude of the threshold current Ithl determines the maximum gain of the amplifier stage 10.
The rectifier stage 40, as shown in Figure 3, comprises a 2 o rectifier circuit 58, and first and second current sources 60,62 respectively.
The rectifier circuit 58 can be implemented in known manner using solid state switches such as diodes (not shown). Furthermore, the rectifier circuit 58 can be a full-wave or a half-wave design. This design choice will not result in substantial differences in performance of the amplifier stage 2 5 10, except for the factor of 2 difference in the average signal energy, as will be understood by one skilled in the art. This scale factor of 2 can be easily accommodated by the appropriate choice of other parameters in the gain controller 13.
The output of the rectifier 58 is coupled to both current 3 o sources 60,62. The two current sources 60,62 are identical and implemented, in known manner, as voltage controlled current sources.
~ 3 ~
The first current source 60 produces a first rectified output current IRECTI-The second current source 62 produces a second rectified output current IRECT2 which is equal to the current IRECTI- Therefore, the instantaneous values of the rectified currents IRECTI and IRECT2 are proportional to the 5 rectified (i.e. absolute) instantaneous voltage level of the input signal 12.
In known systems which use peak detector circuits, there is more sensitivity to half-wave rectification. This occurs because peaks in the input signal will not always be symmetrical about zero. Hence, it is possible that a peak in the input signal 12 will be opposite in phase to the 1 o sensitivity of peak detector and be missed by the detector. In contrast, thevariable amplifier 10 according to the present invention obviates the problem by using averaging detectors 34,36 to determine the average energy in the input signal 12. Moreover, the coupling capacitor 18 acts to make the average energy in the input signal 12 symmetrical.
A feature of the amplifier stage 10 is the use of dual averaging detectors 34,36 to generate the control signal 33 (i.e. IGAIN) for varying the gain of the amplifier stage 10 according to the average energy (i.e. loudness level) of the input signal 12. To explain how the control signal 33 is generated, the operation of the slow averaging detector 34 is 2 o first considered because it is the detector which is usually in operation.
The rectified output current IRECTI is applied to the slow averaging circuit 34. The averaging operation is achieved by feeding the current IRECTI into the combination of a capacitor 61, a resistor 63 and an operational amplifer 65 shown in Figure 4(a). It will be understood by one 2 5 skilled in the art that the current flowing in the resistor 63 is representative ~f the average of the cùrrent IRECT1. This current is sensed using known techniques and replicated by two current sources 66,68, which produce identical averaging output currents ISLOWI and ISLOW2-Reference is now made to Figure 4(b), which shows the 3 0 details of the slow averaging circuit 64 of Figure 4(a). The operation of theslow averaging circuit 64 is evident from Figures 4(a) and (b) and within ~ .
2~ 3~3~
the understancling of one skilled in the art. The circuit for the fast averaging circuit 74 is virtually identical to the one shown in Figure 4.
The only difference is found in the resulting time constants. In the preferred embodiment, the slow averaging circuit 64 produces a time constant of 200 milliseconds and the fast averaging circuit 74 produces a time of constant of 10 milliseconds (i.e. a factor of 20). The selection of the values for respective circuit elements shown in Figure 4 to produce these respective time constants is within the capability of one skilled in the art.
Referring back to Figure 3, the averaging output current 1 O ISLOWI is compared to a second threshold current Ith2 which is producedby a current source 70 in the first variable current reference 54. The current source 70 is coupled to a current mirror 72 which is formed from transistors Ql and Q2. The difference between the averaging current ISLOW1 and the threshold reference current Ithl is produced or mirrored at the collector of transistor Q2, i.e. the output of the current mirror 72. The design of the current mirror 72 formed from transistors Ql and Q2 is within the capability of one skilled in the art and has the characteristic of only being able to sink current at the output, i.e. the collector of transistor 2 o Referring still to Figure 2, if the averaging current ISLOW1 is less than the threshold current Ith2, then transistor Ql will not conduct and according to the operation of the current mirror 72, there will be no collector current in transistor Q2. As the averaging current ISLOW1 increases, in response to an increasing input signal 12 (i.e. louder user 2 5 environment), it will eventually exceed the value of the current Ith2 and this difference in the currents will be mirrored as the collector current of transistor Q2. The collector current of transistor Q2 is added to the first threshold current Ithl by the current summer 50 to produce the gain control current IGAIN which sets the gain of the operational amplifier 44 3 0 through the current controlled resistor 48.
From the foregoing discussion, it can be seen that there are two regions of operation for the slow averaging stage 34. In the first region of operation, the average energy of the input signal 12 is below the threshold level (i.e. the current ISLOWI is less than the reference current 5 Ith2) and the gain control current IGAIN is equal to the first threshold current Ithl- In the second region of operation, the average energy of the input signal 12 is above the threshold level. This means that the averaging current ISLOWI exceeds the current Ith2 and the difference between these two currents is mirrored in the current mirror 72 and 10 added to the current Ithl to produce the gain control current IGAIN- The gain control current IGAIN can be determined according to the following equation:
IGAIN = Ithl + (ISLOWI - Ith2) = ISLOWI ; given Ithl = Ith2 Since the threshold currents Ithl and Ith2 are designed to be identical, the gain control current IGAIN is equal to the averaging output current ISLOWI
and therefore there is a linear relationship between the control current 2 0 IGAIN and the averaging output current ISLOW1-In the preferred embodiment of the present invention,the resistance value of the current controlled resistor 48 is inversely proportional to the magnitude of the gain control current IGAIN, and above the threshold loudness level, the compression ratio is 2:1. The 2:1 25 compression ratio can be illustrated as follows. If the output signal 14 increases by a factor of 2 (or 6 dB), the gain control current IGAIN which is a function of the averaging current ISLOWI will also increase by the same factor, as shown in the above equation. Since the value of the resistor 48 is inversely proportional to the gain control current IGAlN, the input 30 signal 12 must have increased by a factor of 4 (or 12 dB). This follows from the gain equation of the inverting operational amplifier 44. In other ~, -9 ~ 3 ~ ~:
words, if the output signal 14 increases by 6 dB, then the gain will decrease by 6 dB, and therefore, the input signal 12 must have increased by 12 dB, which provides the desired 2:1 compression ratio.
Referring back to Figure 3, the operation of the fast 5 averaging stage 36 is considered. The current IRECT2 produced by current source 62 in the rectifier stage 40 is identical to current IRECT1. The current IRECT2 is applied to the fast averaging detector 36. The operation of the fast averaging detector 36 is very similar to the operation of the slow averaging detector 34 with the primary difference being the time 10 constants employed in the design of the fast averaging circuit 74. The time constants of the fast averaging circuit 74 are designed to be much shorter than those of the slow averaging circuit 64. In addition, the faster averaging detector 36 only requires one current source which generates `~
an averaging output current IFASTI ` ~
As shown in Figure 3, the averaging output current IFASTI `
is compared to the output of current source 68 which produces a current ISLOW~- The current IFAST1 is compared to current ISLOW2 through a current mirror 76 formed from two transistors Q3 and Q4. By choosing the emitter area ratios of transistor Q4 to transistor Q3 as N:1, the dynamic 2o threshold can be set to determine the amount that the fast occurring event (i.e. IFASTI) must exceed the slower moving averaging current ~ ~ ~
ISLOW2 in order to assume gain control of the amplifier stage 32. The ~ ;;dynamic thxeshold can be determined according to the following -expression:
dynamic threshold = 20 loglo (N) ;- ~
According to the above expression, if N = 2, then the fast averaging .-~.
current IFASTI must exceed the slow averaging current ISLOW2 by 6 dB.
The difference between the fast averaging current IFAST1 and the scaled slow averaging current ISLOW2 is mirrored by a current -2 ~
mirror 78 which is formed from two transistors Q5 and Q6. The difference is reproduced or mirrored as the collector current of transistor Q6. The collector of transistor Q6 provides the third input to the current summer 50.
Similar to the operation of the current mirror 72 formed from transistors Q1 and Q2, the current mirror 78 will produce zero output current (i.e. collector current of transistor Q6) if the fast averaging current IFAST1 is less than N times the slow averaging current ISLOW2- If the fast averaging current IFAST1 exceeds the scaled slow averaging current 1 O ISLOW2~ then difference is reproduced as the collector current of transistorQ6- The current summer 50 adds the collector current of transistor Q6 to the first threshold current Ithl to produce the gain control current IGAIN
which reduces the gain of the amplifier stage 32 as explained above.
In practical systems, the current controlled resistor 48 will 15 have limitations in terms of linear operation. This non-linearity typically manifests itself as distortion at very high level input signals 12. The function of the comparator stage 42 is to compensate for this non-linearity.
As shown in Figure 3, the comparator stage 42 comprises 20 first and second comparators 80 and 82. Each comparator 80,82 can be implemented using an open loop operational amplifier and known design techniques. The first comparator 80 is configured to process positive going signals 14 with a voltage reference 84 connected to the non-inverting input of the comparator 80. The output of the comparator is 25 connected to the base of a drive transistor Q7. The emitter of the drive transistor Q7 is tied to a po-vver supply rail 90. The collector (output) of thetransistor Q7 is tied to the input of the fast averaging detector 36 which is also connected to the output of the current source 62. In a similar fashion, the second comparator 82 is coupled to a second voltage reference 86 and 30 a second drive transistor 92. The collector of the second drive transistor 92 is also tied to the input of the fast averaging detector 36. By connecting the second voltage reference 86 to the inverting input of the comparator 82, the second comparator 82 is configured to process negative going output signals 14 from the operational amplifier 44.
The function of the comparator stage 42 is to provide 5 symmetrical thresholds about the normal quiescent point of the output from the operational amplifier 44. The use of high gain comparators 80,82 connected to the fast averaging detector also effectively produces a peak detecting high compression ratio automatic gain control circuit with the accompanying short time constants (from the fast averaging circuit 74). In 10 practical terms, this means that the gain controller 13 according to the present invention produces a quick response time which lowers distortion at high signal levels while providing very fast recovery due to the short time constants in the fast averaging circuit 74. The voltage references 84,86 are selected so that they are seldom exceeded while still 15 obtaining optimum distortion performance from the amplifier as will be understood by one skilled in the art.
The invention has now been explained with reference to specific embodiments. Other embodiments will be apparent to those of ordinary skill in art in view of this description. For example, the feedback 20 circuit for the amplifier has been described as including a current controlled resistor. It would be possible to design the system in the voltage domain by using a voltage controlled resistive device in the feedback loop and generating a voltage based gain control signal. It is therefore not intended that the invention be limited except as indicated 25 bytheappendedclaims.
Claims (11)
1. A hearing aid including microphone means for receiving an audio input signal and converting the audio input signal into an electrical input signal, a variable gain amplifier coupled to the microphone means, and signal processing means coupled to transducer means for processing the electrical signal and producing an audio output signal which is coupled to the ear of a user, said variable gain amplifier comprising:
(a) an amplifier stage having an output port and an input port for receiving the electrical input signal, and including variable gain means for producing an output signal having a gain level at said output port;
(b) a rectifier stage coupled to the output port of said amplifier stage and having means for rectifying said output signal to produce a rectified signal;
(c) averaging detector means for generating an averaged signal from said rectified signal, said averaging detector means being coupled to said rectifier stage;
(d) gain controller means coupled to said averaging detector means and to said amplifier stage, said gain controller means being responsive to said averaged signal and having means for producing a gain control signal; and (e) said variable gain means having means responsive to said gain control signal for varying the gain level of said amplifier stage.
(a) an amplifier stage having an output port and an input port for receiving the electrical input signal, and including variable gain means for producing an output signal having a gain level at said output port;
(b) a rectifier stage coupled to the output port of said amplifier stage and having means for rectifying said output signal to produce a rectified signal;
(c) averaging detector means for generating an averaged signal from said rectified signal, said averaging detector means being coupled to said rectifier stage;
(d) gain controller means coupled to said averaging detector means and to said amplifier stage, said gain controller means being responsive to said averaged signal and having means for producing a gain control signal; and (e) said variable gain means having means responsive to said gain control signal for varying the gain level of said amplifier stage.
2. The variable gain amplifer as claimed in claim 1 wherein said averaging detector means comprises a slow averaging detector and a fast averaging detector, said slow averaging detector having output means coupled to said gain controller means and including means responsive to rectified signals in a first range for producing at said output means a slow control signal indicative of the average energy of the rectified signals in said first range, and said fast averaging detector having output means also coupled to said gain controller means and including means responsive to rectified signals in a second range for producing at said output a fast control signal indicative of the average energy of the rectified signals in said second range.
3. The variable gain amplifier as claimed in claim 2, wherein said gain controller means comprises a first current generator coupled to the output of said slow averaging detector and having means for producing a first current signal, a second current generator coupled to the output of said fast averaging detector and having means for producing a second current signal, and summing means coupled to said variable gain means for summing said first and second current signals to produce said gain control signal.
4. The variable gain amplifier as claimed in claim 3, wherein said gain controller means further includes a reference current generator coupled to said summing means and having means for producing a reference current signal, said reference current signal being used by said summing means to generate said gain control signal.
5. The variable gain amplifier as claimed in claim 4, wherein said reference current signal has a magnitude which defines the maximum gain level for said amplifier stage.
6. The variable gain amplifier as claimed in claim 1, 2, 3 or 4, wherein said variable gain means comprises feedback means and said feedback means includes a current controlled resistor having means for producing an impedance level in response to said gain control signal.
7. The variable gain amplifier as claimed in claim 1, 2, 3 or 4, further including signal limiter means for limiting output signals from said amplifier stage which exceed a predetermined range, said signal limiter means being coupled to the output of said amplifier stage and to said averaging detector means, and said signal limiter means having means for setting said predetermined range.
8. The variable gain amplifier as claimed in claim 1, 2, 3, or 4, wherein said amplifier means comprises an operational amplifier.
9. An automatic gain circuit suitable for use with an amplifer for producing a variable gain output signal from an input signal, said automatic gain circuit comprising:
(a) feedback means coupled to the amplifer, said feedback means having current controlled impedance means responsive to a control signal and having means for producing a variable impedance in response to said control signal;
(b) said feedback means being effective to produce a variable gain for the amplifier, said variable gain being determined by said current controlled impedance means;
(c) rectifier means coupled to the output of the amplifier and having means for rectifying the output signal and producing a rectified signal; and (d) averaging detector means coupled to said rectifier means and including means for determining the average energy level in said rectified signal and generating said control signal corresponding to the average energy level.
(a) feedback means coupled to the amplifer, said feedback means having current controlled impedance means responsive to a control signal and having means for producing a variable impedance in response to said control signal;
(b) said feedback means being effective to produce a variable gain for the amplifier, said variable gain being determined by said current controlled impedance means;
(c) rectifier means coupled to the output of the amplifier and having means for rectifying the output signal and producing a rectified signal; and (d) averaging detector means coupled to said rectifier means and including means for determining the average energy level in said rectified signal and generating said control signal corresponding to the average energy level.
10. A method for controlling the gain of an amplifier having means to produce an amplified output signal from an input signal, the amplifier having a feedback circuit which includes a current controlled resistor which is responsive to a gain control signal, said method comprising the steps of:
(a) determining the average energy level in the output signal;
(b) generating the gain control signal based on the average energy level determined in step (a); and (c) feeding the gain control signal to the current controlled resistor to effectively vary the gain of the amplifier by modifying the impedance of the current controlled resistor.
(a) determining the average energy level in the output signal;
(b) generating the gain control signal based on the average energy level determined in step (a); and (c) feeding the gain control signal to the current controlled resistor to effectively vary the gain of the amplifier by modifying the impedance of the current controlled resistor.
11. The method according to claim 10, further including the step of rectifying the output signal before determining the average energy level in step (a).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CA 2090531 CA2090531A1 (en) | 1993-02-26 | 1993-02-26 | Dual time constant audio compression system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CA 2090531 CA2090531A1 (en) | 1993-02-26 | 1993-02-26 | Dual time constant audio compression system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CA2090531A1 true CA2090531A1 (en) | 1994-08-27 |
Family
ID=4151216
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CA 2090531 Abandoned CA2090531A1 (en) | 1993-02-26 | 1993-02-26 | Dual time constant audio compression system |
Country Status (1)
| Country | Link |
|---|---|
| CA (1) | CA2090531A1 (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1351550A1 (en) * | 2002-09-10 | 2003-10-08 | Phonak Ag | Method for adapting a signal amplification in a hearing aid and a hearing aid |
| EP1802168A1 (en) * | 2005-12-21 | 2007-06-27 | Oticon A/S | System for controlling transfer function of a hearing aid |
| EP2747454A1 (en) * | 2012-12-20 | 2014-06-25 | Starkey Laboratories, Inc. | Separate inner and outer hair cell loss compensation |
| CN114327040A (en) * | 2021-11-25 | 2022-04-12 | 歌尔股份有限公司 | Vibration signal generation method, device, electronic device and storage medium |
-
1993
- 1993-02-26 CA CA 2090531 patent/CA2090531A1/en not_active Abandoned
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1351550A1 (en) * | 2002-09-10 | 2003-10-08 | Phonak Ag | Method for adapting a signal amplification in a hearing aid and a hearing aid |
| US7123732B2 (en) | 2002-09-10 | 2006-10-17 | Phonak Ag | Process to adapt the signal amplification in a hearing device as well as a hearing device |
| EP1802168A1 (en) * | 2005-12-21 | 2007-06-27 | Oticon A/S | System for controlling transfer function of a hearing aid |
| AU2006252157B2 (en) * | 2005-12-21 | 2011-02-24 | Oticon A/S | System for controlling a transfer function of a hearing aid |
| US8014550B2 (en) | 2005-12-21 | 2011-09-06 | Oticon A/S | System for controlling a transfer function of a hearing aid |
| CN1988737B (en) * | 2005-12-21 | 2012-05-23 | 奥迪康有限公司 | System for controlling a transfer function of a hearing aid |
| EP2747454A1 (en) * | 2012-12-20 | 2014-06-25 | Starkey Laboratories, Inc. | Separate inner and outer hair cell loss compensation |
| US8873782B2 (en) | 2012-12-20 | 2014-10-28 | Starkey Laboratories, Inc. | Separate inner and outer hair cell loss compensation |
| US9408001B2 (en) | 2012-12-20 | 2016-08-02 | Starkey Laboratories, Inc. | Separate inner and outer hair cell loss compensation |
| CN114327040A (en) * | 2021-11-25 | 2022-04-12 | 歌尔股份有限公司 | Vibration signal generation method, device, electronic device and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US4466119A (en) | Audio loudness control system | |
| US5862238A (en) | Hearing aid having input and output gain compression circuits | |
| CA2268918C (en) | Compression systems for hearing aids | |
| JP2529129B2 (en) | Signal compression apparatus and method | |
| US5553151A (en) | Electroacoustic speech intelligibility enhancement method and apparatus | |
| US6549630B1 (en) | Signal expander with discrimination between close and distant acoustic source | |
| US5832097A (en) | Multi-channel synchronous companding system | |
| US5822442A (en) | Gain compression amplfier providing a linear compression function | |
| US5278912A (en) | Multiband programmable compression system | |
| US8045720B2 (en) | Method for dynamic determination of time constants, method for level detection, method for compressing an electric audio signal and hearing aid, wherein the method for compression is used | |
| US6628795B1 (en) | Dynamic automatic gain control in a hearing aid | |
| WO1998018294A9 (en) | Compression systems for hearing aids | |
| US4475230A (en) | Hearing aid | |
| JP4018207B2 (en) | Method for automatically limiting distortion of audio equipment and circuit arrangement for implementing this method | |
| GB2120903A (en) | Headphone level protection circuit | |
| JPH0216852A (en) | Transmitting level control circuit for telephone set | |
| CA2294713A1 (en) | Hearing aid having input agc and output agc | |
| CA2090531A1 (en) | Dual time constant audio compression system | |
| US7539320B2 (en) | Hearing aid with automatic excessive output sound control | |
| CA2190787C (en) | A gain compression amplifier providing a linear compression function | |
| US3991272A (en) | Audio AGC amplifier | |
| JPH02113698A (en) | Hearing aid | |
| EP0101469A1 (en) | A method and a device for the damping of transients | |
| CA1199080A (en) | Audio loudness control system | |
| US5220287A (en) | Voice processing apparatus |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| FZDE | Dead |