Patent Application
20110033064
An embodiment of the invention relates to noise cancellation techniques that improve headset-based audio communications using a portable host device. Other embodiments are also described.
For two-way, real-time audio communications, referred to here generically as voice or video telephony, a user can wear a headset that includes a single earphone (also referred to as a headphone or a speaker) and a microphone, or a pair of stereo earphones and a microphone, that are connected to a host communications device such as a smart phone. The headset, which integrates the earphones with a microphone, may be connected to the host device through a 4-conductor electrical interface typically referred to as a headset plug and jack matching pair. The four conductors are used as follows: two of them are used for the left and right earphone signals, respectively; one of them connects a microphone signal; and the last one is a reference or power return, conventionally taken as the audio circuit reference potential. The plug that is at the end of the headset cable fits into a mating 4-conductor jack that is integrated in the housing of the host device. Connections are made within the host device from the contacts of the headset jack to various audio processing electronic components of the host device.
Packaging restrictions in host devices such as a smart phone or a cellular phone create difficult challenges for routing the signal and power lines. For example, the headset jack is often located distant from the main logic board on which the audio processing components are situated, so that the headset signal needs to be routed through a flexible circuit and one or more board-to-board connectors. The multiple connections increase the impedance of the connection, as well as the manner in which the connections are made namely through narrow or thin metal circuit board traces, can lead to the coupling of audio band noise during operation of the host device. In addition, with the shared nature of the headset's reference or ground contact (shared by the microphone and the earphones of the headset), further noise is produced at the output of the microphone preamplifier. The preamplifier provides an initial boost to the relatively small microphone signal that is received from the headset. The practical effect of such audio noise at the output of the microphone preamplifier is often that the listener at the far end of a telephone conversation hears an echo of her own voice, with a concomitant reduction in the quality of the sound.
Attempts to reduce (or, as generically referred to here, “cancel”) the noise at the output of the microphone preamplifier have been made. In one case, the concept of differentially sensing the microphone signal is used. For this purpose, a differential amplifier (in contrast with a single-ended amplifier) is used to only amplify the difference between the voltage at a sense point for the headset ground contact and the voltage at a sense point for the microphone signal contact. Using such a configuration, any audio voltage that may appear as noise between a local ground (local to the microphone preamplifier) and the ground that is near the headset jack or socket are largely rejected (that is, not significantly amplified), while the audio signal on the microphone signal contact is amplified.
Packaging constraints and compromises of the microphone and earphone signals and their common return in the host device leads to a common mode imbalance that can cause undesired common mode noise to be coupled into either a microphone signal loop or a speaker signal loop. In practice the microphone signal loop is more prone to contamination by offensive audio band noise. In addition, compromised routing of the audio signals represents a finite impedance that can act as a victim impedance for near-by sources of noise within the host device, whether of low frequency similar to the audio base bandwidth, frequencies subject to heterodyning or fold over by sampled data converters, or non-linear impedances capable of demodulating local radio frequency energy.
The differential sensing approach described above in the Background section for ameliorating microphone preamp noise falls short, when the following practical considerations are taken into account. First, there are several different types of headsets in the marketplace, each of which may have a different type of microphone circuit. Moreover, there are manufacturing variations in the microphone circuit, even for the same make and model of headset. Finally, manufacturing as well as temperature variations could also affect the electrical characteristics of a flexible circuit or board-to-board connector that is used to connect with the headset interface within the host device. Any successful attempt to cancel the microphone noise, by differentially sensing the microphone signal, will require knowledge of the precise electrical characteristics of the relevant circuitry, in each instance of the manufactured host device and headset combination. This however is not a practical solution.
An embodiment of the invention is an improved circuit for reducing microphone amplifier noise in a two-way audio communications host device. The circuit provides a more robust solution in that it is able to perform good noise reduction for different types or brands of headsets whose microphone circuits have different impedances. It can also compensate for parasitic effects in the host device that may have been caused by compromised signal or ground routing between the host headset connector and the microphone amplifier.
The microphone amplifier may be implemented as a difference amplifier having a first input and a second input; the second input is coupled to the microphone contact of an electrical interface used by a microphone-speaker combination. A variable attenuator has an input that is directly coupled to receive a signal from a sense point for a reference contact of the microphone-speaker combination electrical interface. An output of the attenuator is coupled to the first input of the difference amplifier. A controller has an output that is coupled to set the variable attenuator, in order to reduce or minimize noise. This capability is referred to here as active, real-time control of differential mode noise cancellation.
In one embodiment, the controller acts in an open loop fashion by setting the attenuator state depending upon the type of microphone-speaker combination to which the host device is to be, or is now, connected. In particular, the type of microphone circuit is determined and on that basis the attenuator is set. The determination may be detected automatically or it may be obtained via direct user input. For example, the determination may be a look up performed on a previously stored table that lists different types of microphone circuits and their respective attenuation settings that have been shown to yield improved or optimal noise cancellation. Configured in this manner, the difference amplifier will produce the boosted microphone signal with improved signal to noise ratio. The configuration process may be performed “in the field”, i.e. while the host device is used in its normal course by the end user.
In another embodiment, the controller acts in a closed loop fashion when setting the attenuation. In that case, the controller has an input coupled to an output of the difference amplifier. The controller measures the output of the difference amplifier and on that basis adjusts the attenuation until the presence of a test signal at the output of the difference amplifier is sufficiently minimized, or essentially removed. This closed loop control of the attenuator may also be done in the field, and in a manner that is generally inconspicuous to the end user.
In one embodiment, the test signal is a super-audible tone that is generated and played through a speaker contact of the microphone-speaker combination connector in the host device, while a microphone-speaker combination is connected. The output of the microphone signal difference amplifier is measured, while the microphone-speaker combination is connected and the super-audible tone is playing. The reference sense point signal that is input to the amplifier is attenuated, based on the measurement, in a manner that reduces the presence of the super-audible tone at the output of the amplifier. A final attenuation setting is selected, which may be the one for which the presence of the super-audible tone is reduced to below a given threshold or has been minimized. In that setting, the microphone amplifier is deemed calibrated, so that an uplink audio communications signal from the output of the amplifier can be transmitted, e.g. during a telephone call, with improved signal to noise ratio and reduced far end echo.
In another embodiment, the test signal is any signal applied to the speaker outputs and detected in the signal recovered from the microphone preamplifier. The test signal may therefore be constrained along fairly broad lines, examples being individual tones or combinations of tones spread above, below, and in special cases through the audio band used in the product. The significant constraint on choice of the test signal is that it not be distracting to the user. In consequence, because the application of the test signal is not necessarily continuous, its spectral characteristics can be designed to fulfill other system requirements.
The above summary does not include an exhaustive list of all aspects of the present invention. It is contemplated that the invention includes all systems and methods that can be practiced from all suitable combinations of the various aspects summarized above, as well as those disclosed in the Detailed Description below and particularly pointed out in the claims filed with the application. Such combinations have particular advantages not specifically recited in the above summary.
The embodiments of the invention are illustrated by way of example and not by way of limitation in the figures of the accompanying drawings in which like references indicate similar elements. It should be noted that references to “an” or “one” embodiment of the invention in this disclosure are not necessarily to the same embodiment, and they mean at least one.
Several embodiments of the invention with reference to the appended drawings are now explained. While numerous details are set forth, it is understood that some embodiments of the invention may be practiced without these details. In other instances, well-known circuits, structures, and techniques have not been shown in detail so as not to obscure the understanding of this description.
The host 10 may be coupled to one or more microphone-speaker combinations 11, through its headset electrical interface 12. Several different types of microphone-speaker combinations 11 that can be used are shown, including two different types of headsets (one in which a pair of earphones or headphones are in loose form, and another where a single earphone is attached to a microphone boom) and a combination microphone stand and desktop loudspeaker. Each of these microphone-speaker combinations 11 can be a separate item than the host device 10, and can be coupled to the host device 10 through a cable connector that mates with the headset electrical interface 12 in the housing of the host device 10.
Referring now to
In a typical case, all four of the contacts shown in
In some cases, there may be multiple microphones in the microphone-speaker combination 11 that share the same reference contact 15′, e.g. a headset with an integrated microphone array that can be used to implement an audio beam-forming function by the host device 10. For that scenario, the headset electrical interface 12 could have more than one microphone contact 16′, one for each of the microphones of the array.
Note that in
With the microphone-speaker combination 11 connected to the host device 10, a user of the host device can hear the far end user talking during a telephone call and can speak to the far end user at the same time, via the speakers 18 and microphone circuit 20, respectively. The voice of the far end user originates in a downlink communications signal that arrives into the host 10 over a communications network. A downlink audio signal may be in digital form when it passes through a communications signal processor (not shown) with several stages that may include various digital signal processing operations, including a mixer that allows the addition of sidetone. The downlink audio signal with the sidetone is then converted into analog form using a digital to analog converter (DAC), before being applied to the headset electrical interface 12 by a speaker amplifier. At the same time, the near end user may speak into the microphone circuit 20, which picks up the voice as an uplink audio signal that passes through the headset interface 12 (in particular the microphone contacts 16, 16′). The uplink audio signal is then boosted by the microphone preamplifier and may then be converted into digital form by an analog to digital converter (ADC). This allows the generation of a digital sidetone signal (which is fed back to the speaker 18 as explained above). In addition, the uplink audio signal may be subjected to further digital signal processing before being transmitted to a remote device (e.g., the far end user's host device) over the communications network as an uplink communications signal.
Specifics of the noise cancellation circuitry in the host 10 are now described. Still referring to
Due to practical limitations, the electrical connection or direct coupling between the reference contact 15′ and the MLB ground that is at the microphone amplifier is not identically zero ohms, particularly in the audio frequency range. This may be due to various physical structures that create parasitic or stray effects, represented in
There are different types of microphone-speaker combinations 11 that can be used with the same host connector, each of which may have a different type of microphone circuit 20. For example, there are passive microphone circuits that are essentially passive acoustic transducers that produce an analog transducer signal on the microphone contact 16. There are also non-passive or active microphone circuits 20 that drive a modulated signal on the microphone contact 16. In both cases, a dc microphone bias circuit 22 may be needed in the host device 10, coupled to the microphone contact 16′ as shown, to provide a dc bias voltage for operation of the microphone circuit 20.
An attempt to cancel or reduce microphone-speaker combination noise, which appears in the uplink communications signal and may manifest itself when the far end user hears an echo of his own voice during a telephone call, calls for differentially sensing the microphone signal. As explained above in the Summary section, however, such a technique must be performed carefully else the noise reduction attempt will be ineffective. The different types of microphone circuits 20 present different impedances (both at dc and in the audio range) on the microphone contact 16′. Moreover, there are manufacturing variations in the microphone circuits 20, even for the same make and model of microphone-speaker combination. Thus, knowledge of the precise impedance characteristics of the microphone circuit 20, in addition to a good estimate of the parasitic components that cause a substantial difference between a signal at the output terminal of the microphone circuit 20 and what should be the same signal at the input terminal of the microphone amplifier in the host device 10, are needed. Such detailed knowledge however is not available to a single entity at the time of manufacture of the host 10 and the microphone-speaker combination 11, because a purchaser of the host device 10 may elect to use any one of a large variety of different types or brands of microphone-speaker combinations including some that may not be available during the time the audio processing functions of the host device 10 are being designed.
Still referring to
The hot input of the difference amplifier 28 may be AC coupled to a sense point for the microphone contact 16′, i.e. through a DC blocking capacitor 23. The capacitor 23 may be coupled as shown, where one side is at the microphone sense point, which is connected to the microphone bias circuit 22, and the other is at the hot input. The cold input of the difference amplifier 28 is coupled to a sense point for the reference contact 15′. This is also an AC coupling, i.e. though a DC blocking capacitor 25. In another embodiment, the coupling between the inputs of the difference amplifier and the microphone and reference sense points may be different, while still having constant gain through the normal and common mode bands of interest.
A variable attenuator 24 serves to attenuate a reference signal from the reference sense point, to the cold input of the difference amplifier 28. Note that in this embodiment, the dc blocking capacitor 25 is coupled between the attenuator 24 and the cold input, in other words, the attenuator 24 is in front of the capacitor 25. In another embodiment, the reverse may be true, where the capacitor 25 is in front of the attenuator 24.
The variable attenuator 24 is a voltage attenuator that can be placed into any one of several attenuation states, all of which provide a dc coupling or path to the power return plane. The attenuation states are designed to provide enough granularity and range to the attenuator for optimizing the common mode rejection (CMR) of the difference amplifier 28, for as many different types of microphone-speaker combinations 11 as expected to be practical. For example, each attenuation state may be 0.5 dB apart from its adjacent states, ranging from for example 0 dB to −30 dB. The range and granularity of the attenuation states may be determined empirically, during testing or development of the host device 10, to be that which will provide best noise reduction for all of the different, expected microphone-speaker combinations.
In the embodiment of
In one embodiment, the controller 26 automatically detects the type of microphone-speaker combination 11 that is coupled to the host connector and then accesses a previously stored look up table to determine the appropriate attenuation setting for the given type of microphone-speaker combination. This may be done by using a circuit (not shown) that measures the impedance seen from the host device 10 out through the microphone contact 16′, for example relative to the reference contact 15′. Different types of microphones can be expected to have different impedances; the entries of the look up table could be empirically determined and filled in advance, to include the different types of microphone by referencing their respective impedances. Other ways of automatically detecting the microphone-speaker combination type are possible, e.g. by reading a stored digital or analog code value through the speaker contact 14′ or the microphone contact 16′.
In another embodiment, the controller 26 can be operated “manually”, with direct user input. In that case, the controller 26 can obtain the desired attenuation setting, based on receiving user input regarding microphone-speaker combination type (e.g., the user could indicate his selection from a stored list of microphone-speaker combination types that are being displayed to him on a display screen of the host device 10).
The controller 26 may be implemented as a programmed processor (e.g., an applications processor in a smart phone that is executing software or firmware) designed to manage the overall process of configuring a microphone signal difference amplifier, for improved noise reduction.
Referring now to
In one embodiment, the controller 32 may be designed to have access to a previously stored indication of what is an acceptably low level of microphone-speaker combination noise at the output of the difference amplifier 28. In other words, values representing the lowest acceptable level of microphone-speaker combination noise, also referred to as a noise threshold, may be stored in memory or other storage within the portable device 10. This allows the controller 32 to adjust the attenuator 24 while monitoring the output of the difference amplifier 28, until the expected noise threshold is detected.
Alternatively, the controller 32 may be designed to adjust the attenuator 24 until it detects a minimum at the output of the difference amplifier 28, where the lowest point of the minimum represents the lowest possible noise level. In one embodiment, a super-audible tone generator 30 is included, having an output coupled to the speaker contact 14′. In that case, the controller 32 may be designed to signal the generator 32 to generate a super-audible tone that is played through the speaker contact 14′. This may be viewed as a calibration or test signal. The test signal may be played for a relatively short period of time, e.g. a few seconds, while the attenuation state of the variable attenuator 24 is automatically swept over an attenuation range that is sufficiently broad as to produce the expected minimum at the monitored output of the difference amplifier 28. The attenuation state that yields the minimum is accepted as the final setting that provides improved or optimized CMR for the current microphone-speaker combination that is being used with the host device 10. Note that by virtue of being super-audible, the test signal even though driving the connected speaker 18 cannot be heard by the end user of the host device 10, and is close enough to the audible spectrum to be useful in the noise cancellation control process.
Turning now to
In one embodiment, the attenuator 24 is implemented using a voltage divider network that has at least one series resistor Ras and at least one shunt resistor Rah. In the embodiment of
If the difference amplifier 28 also has variable gain, then the above described control process may be performed either before or after having set the gain.
In operation 86, the host device 10 configures the difference amplifier 28 (of a microphone amplifier block). This occurs by setting a variable attenuator at the reference sense point input of the difference amplifier, in accordance with any one of the techniques described above. These may include: open loop manual, which is based on received direct input from the near end user regarding the type of speaker-microphone combination (e.g., headset type) that is to be used with the host; open loop automatic, based on automatic measurement of microphone-speaker combination impedance or automatic detection of a microphone-speaker identification code; and closed loop, based on monitoring the output of the difference amplifier while sweeping the variable attenuator. The output of the difference amplifier provides the improved, uplink audio communications signal for the telephone call.
In operation 88, the telephone call is performed with the benefit of noise cancellation being obtained from the difference amplifier 28 as configured in operation 86. Thus, the far end user of the call should be able to better hear the near end user (in the uplink signal originating at the output of the difference amplifier), with higher signal to noise ratio and/or diminished echo of his own voice.
It should be noted that the selection in operation 86 could occur either before the call is established in operation 84, or it could occur during the call (e.g., as soon as the conversation begins—during operation 88).
While certain embodiments have been described and shown in the accompanying drawings, it is to be understood that such embodiments are merely illustrative of and not restrictive on the broad invention, and that the invention is not limited to the specific constructions and arrangements shown and described, since various other modifications may occur to those of ordinary skill in the art. For example, although the host device is described in several instances as being a portable device, the noise reduction circuitry could also be useful in certain non-portable host devices such as desktop personal computers that also have similar limitations regarding interior signal routing and a shared reference contact in the headset electrical interface. Also, the concept need not be limited to the described combination of one microphone and one or two speakers. The technique disclosed can be used without loss of generality or performance to m microphones and s speakers, requiring, in general between 2(m+s) to m+s+1 separate connections through the headset electrical interface. Finally, although the microphone amplifier block in
