IMPLEMENTING NOISE REDUCTION IN RADIO SIGNALS USING SPECTRAL DUPLICATION

Abstract

A process that has: Receiving, in a noise suppression circuit of a receiver, a plurality of samples of a down-converted signal; separating the plurality of samples into a plurality of subbands; for each of corresponding pairs of the plurality of subbands, comparing a first value based on a first component of the corresponding pair with a second value based on a second component of the corresponding pair; based at least in part on the comparison, matching at least a portion of one of the first and second components of the corresponding pair with at least a portion of the other of the first and second components of the corresponding pair; Integrating the plurality of subbands into a second plurality of samples; and Outputting the second plurality of samples for demodulation.

Description

BackgroundRadio receivers are used for receiving and processing incoming radio signals. In many environments, there are noise problems caused by a variety of potentially interfering sources. Such sources may include interfering radio channels that are closely spaced from a desired radio channel. Other sources may include sources of interference locally present at the radio receiver. For example, various electrical components, either in a single integrated circuit (IC) with the radio receiver or in close proximity to the IC, may operate at frequencies that generate noise at one or more frequencies at which the radio receiver operates. The publications DE 60 2004 007 770 T2 and CN 1 02 752 248 A describe circuits and methods for reducing interference when receiving AM-modulated broadcast signals.SUMMARY OF THE INVENTIONWith respect to amplitude modulated (AM) radios, a new and increasing source of interference is found in receivers operating in hybrid electric and electric vehicles. This is because electric motors of these vehicles generate noise in the AM band, so that reception of an AM broadcast signal without undesirable noise in an automobile radio is much more difficult. Indeed, some manufacturers of electric cars have even fully charged AM radios. SUMMARY OF THE INVENTIONAccording to one aspect, a method comprises: receiving a plurality of samples of a down-converted signal in a noise suppression circuit of a receiver; separating the plurality of samples into a plurality of subbands; for each of the corresponding pairs of the plurality of subbands, comparing a first value based on a first component of the corresponding pair to a second value based on a second component of the corresponding pair; based at least in part on the comparison, matching at least a portion of one of the first and second components of the corresponding pair to at least a portion of the other of the first and second components of the corresponding pair; integrating the plurality of subbands into a second plurality of samples; and outputting the second plurality of samples for demodulation.In an example, the method further includes receiving a complex signal and forming an overlapping window with the plurality of samples. The method may also include: performing a complex fast Fourier transform to separate the plurality of samples into the plurality of subbands; and performing a complex inverse fast Fourier transform to integrate the plurality of subbands into the second plurality of samples. In one embodiment, the separating and integrating comprises a unity function.In an example: calculating the first value comprises calculating a first magnitude for the first component, the first component comprising a positive frequency component of the corresponding pair of the plurality of subbands; and calculating the second value comprises calculating a second magnitude for the second component, the second component comprising a negative frequency component of the corresponding pair of the plurality of subbands. The method also includes comparing the first magnitude with the second magnitude to obtain a comparison result. In response to the comparison result exceeding a first threshold value, one of the positive frequency component and the negative frequency component is replaced with a complex conjugate of the other of the positive and negative frequency components.In an example, the method further comprises, in response to the comparison result exceeding a second threshold, combining at least a portion of one of the positive frequency component and the negative frequency component with at least a portion of the other of the positive frequency component and the negative frequency component, wherein the second threshold is less than the first threshold.In another aspect, an apparatus includes: an analog front end circuit to receive and downconvert an AM signal to a second frequency signal; a digitizer to digitize the second frequency signal into a plurality of samples; and a noise suppression circuit coupled to the digitizer to receive the plurality of samples. In one embodiment, the noise suppression circuit includes: a window generator to generate a window having a first set of the plurality of samples; a band splitter to split the window into a plurality of pairs of symmetric frequency components; a processing circuit for each of the plurality of pairs to: compare a first magnitude of a first symmetric frequency component with a second magnitude of a second symmetric frequency component and modify one of the first symmetric frequency component and a second symmetric frequency component based on the comparison; an integrator to integrate the plurality of pairs output from the processing circuit; and a second window generator to generate a second window having a second set of samples. The apparatus may also include a demodulator coupled to the noise suppression circuit to demodulate the second set of samples and output a demodulated AM signal.In one example, the band splitter includes a fast Fourier transform engine and the integrator includes an inverse fast Fourier transform engine. The processing circuit may include a plurality of individual circuits, each of the plurality of individual circuits including: a first value generator for generating the first magnitude of the first balanced frequency component, the first balanced frequency component including a complex signal; a second value generator for generating the second magnitude of the second balanced frequency component, the second balanced frequency component including a complex signal; and a comparator for comparing the first magnitude with the second magnitude and outputting a comparison result based thereon.In an example, the apparatus further comprises: a controller to receive the comparison result and output one or more control signals based on a comparison of the comparison result with at least one threshold; and a mixing circuit to modify the one of the first balanced frequency component and the second balanced frequency component based on the one or more control signals. The controller may replace the second frequency component with the first frequency component when the comparison result exceeds the at least one threshold. The at least one threshold may be a dynamic threshold based on a dynamically calculated noise carpet. The controller may cause the mixing circuit to output the first balanced frequency component and the second balanced frequency component unchanged when the comparison result is less than the at least one threshold. The window generator may generate the window having at least some overlapping samples of a previous window, and the second window generator may generate the second window overlapping at least a portion of the previous window. The noise suppression circuit may remove noise in a first sideband of the AM signal caused by one or more electrical components of a vehicle, the device being a radio receiver of the vehicle.In yet another aspect, a system includes: an antenna for receiving an AM signal; a tuner coupled to the antenna to receive and process the AM signal to output a down-converted modulated signal; a noise suppression circuit coupled to the tuner to: generate a window having a first set of a plurality of samples of the down-converted modulated signal; divide the window into a plurality of pairs of balanced frequency components; for one or more of the plurality of pairs: compare a first magnitude of a first balanced frequency component with a second magnitude of a second balanced frequency component; and modify one of the first balanced frequency component and the second balanced frequency component based at least in part on the comparison; and integrating the plurality of pairs into a second window having a second set of samples.The system may also include a demodulator coupled to the noise suppression circuit for demodulating the second set of samples to output a signal stream and an output device for playing back the signal stream. The noise suppression circuit may modify those of the first balanced frequency component and the second balanced frequency component by replacing the second frequency component with a representation of the first frequency component when a comparison result of the first magnitude with the second magnitude exceeds a threshold value. The system may be a vehicle infotainment system, wherein the noise suppression circuit is configured to compensate for a noise source present in an electric vehicle or a hybrid electric vehicle with the vehicle infotainment system.List of FiguresBrief Description of the DrawingsFIG. 1 is a block diagram of a receiver according to an embodiment. FIG. 2 is a block diagram of a noise suppression circuit according to an embodiment. FIG. 3 is a block diagram of a noise suppression circuit according to another embodiment. FIG. 4 is an illustration of the various operations performed according to an embodiment. FIG. 5 is a block diagram of a processing circuit according to an embodiment. FIG. 6 is a block diagram of a processing circuit according to another embodiment. FIG. 7 is a flow diagram of a method according to an embodiment. FIG. 8 is a flow diagram of a method according to another embodiment. FIG. 9 is a block diagram of a system according to an embodiment.Detailed DescriptionIn various embodiments, a radio receiver may be equipped with a noise suppression circuit to remove unwanted noise in a particular channel of interest. Embodiments are described herein in the context of AM signals having upper and lower side bands that have identical complex conjugates of each other for a perfect signal. At a high level, embodiments may identify unwanted noise in one of two sideband that contain the same content (although inverted with respect to each other) in a perfect situation. When noise is identified in one of these sidebands, one of several signal processing techniques may be performed to remove this noise to allow demodulation and processing of a noise compensated signal.More specifically, embodiments herein may process a series of time-based samples that form a window, convert these time-based samples to frequency-based samples, identify (in some cases) noise, perform corresponding processing, and reintegrate the frequency domain information back into time domain samples for demodulation and further processing.Referring now to FIG. 1, a block diagram of a receiver is shown according to an embodiment. As shown in FIG. 1, the receiver 100 is a receiver for processing incoming AM signals. In one embodiment, the receiver 100 may be implemented on a single semiconductor die as incorporated into a particular integrated circuit. Understand that only one AM receiver is shown for purposes of illustration and discussion, while a particular receiver IC may have signal processing paths for multiple bands including AM, FM, satellite, and so forth.As shown, the receiver 100 is coupled to an antenna 105 configured to receive incoming RF signals. For purposes of discussion, it will be understood that the RF signals comprise an AM radio channel of interest. The received RF signals are provided to an analog front end (AFE) circuit 110, which may perform various processing of the incoming RF signal, including amplification, e.g., in a low noise amplifier (LNA), down-conversion via a complex mixer to a lower frequency, such as baseband or intermediate frequency, filtering and additional gain control, and so forth. The downconverted signal, in turn, is applied to a digitizer, namely an analog-to-digital converter (ADC) 120, which digitizes the information and provides a digital current to a noise suppression circuit 130.In embodiments, the noise suppression circuit 130 may detect unwanted interference in the received signal stream. As described herein, the noise suppression circuit 130 may be configured to identify the presence of such undesirable noise by identifying differences between the signal content in positive and negative sidebands of a channel of interest that contain the same message information. That is, an AM signal is composed of two sidebands which are identical at the initial transmission except for an inversion of the imaginary component. A interferer is represented as an imbalance between the upper and lower sideband.To identify noise, the noise suppression circuit 130 may be configured to filter short periods of the incoming complex signal (as converted down) into multiple frequency bands symmetrically about a carrier frequency. Thereafter, the noise suppression circuit 130 may analyze pairs of symmetric bands in magnitude and phase. In one embodiment, the band that least likely contains interference may be selected and duplicated into its complement, and the short portion of the incoming complex signal is recovered and concatenated to generate a new output signal. The new output signal may then be demodulated to produce a noise reduced audio signal.In particular, in embodiments, the noise suppression circuit 130 may use a band splitting technique to split each of the positive and negative sideband into a plurality of subbands. The signal processing is then carried out on the corresponding subbands. Such processing in one embodiment comprises comparisons, matching at least one of the sidebands and optionally other processing and then reintegrating the processed subbands into a processed modulated data stream.In the embodiment of FIG. 1, the noise suppression circuit 130 is implemented as a hardware circuit. However, in other embodiments, the noise suppression as described herein may be performed in a general purpose processor such as a digital signal processor, microcontroller, or so forth. To this end, such a hardware processor may execute instructions to perform the noise cancellation. In various implementations, these instructions may be stored on one or more non-transitory storage media, such as firmware memory, flash memory, or so forth.As further illustrated in FIG. 1, the processed modulated data stream may be provided to a demodulator 140, which may demodulate the signal stream and provide demodulated content to an audio processor 150, which may perform additional optional audio processing, resulting in audio output. Understand that many variations and alternatives are possible, while this is shown at a high level in the embodiment of FIG. 1.Referring now to FIG. 2, a block diagram of a noise suppression circuit according to an embodiment is shown. As shown in FIG. 2, the noise suppression circuit 200 may correspond to a high-level view of the noise suppression circuit 130 of FIG. 1. As can be seen, incoming complex input signals from a tuner portion of a receiver are received in a window overlap generator 210. In embodiments, the window overlap generator 210 may generate overlapping windows of received samples. In a particular embodiment described below, a Hann window generator may be used to form overlapping windows, e.g., of 848 samples. As can be seen, an overlapping window is provided to a band-separating filter bank 220 that can split these samples of the overlapping window into a plurality of subbands, e.g., N subbands. In one example, N may be equal to 848.In turn, respective pairs of subbands, positive and negative frequency components, are supplied to a plurality of processing circuits 230o-230n / 2. Although various implementations are possible, in one embodiment, each processing circuit 230 may be implemented as a compare, select, and duplicate circuit for comparing the positive and negative frequency components, select one of the positive and negative frequency components based on the comparison, and duplicate (optionally complex-conjugated) the selected component to the other side, thereby replacing the original unselected frequency component with a representation of the selected frequency component. The resulting processed subbands output from each of the processing circuits 230 are fed to a band reintegrator 240 which reintegrates the individual processed subbands in a unity process into a set of samples which are provided to a window overlap generator 250. The window overlap generator 250 converts the processed sample block into a stream of samples of complex signals which are output to a demodulator. Understand that many variations and alternatives are possible, while this is shown at a high level in the embodiment of FIG. 2.While various ways of performing the above-described band-splitting and processing operations are possible, such as wavelet decomposition or overlapping band pass filters, in a particular embodiment, a fast Fourier transform (FFT) may be used for the band-splitting operation and a corresponding inverse FFT (IFFT) may be used for the reintegration process.Referring now to FIG. 3, a block diagram of a noise suppression circuit according to another embodiment is shown. As shown in FIG. 3, the noise suppression circuit 300 is another implementation of the noise suppression circuit 130 of FIG. 1. An overlapping window is provided to an N-point complex FFT engine 320, which may split these samples in the time domain of the overlapping window into a plurality of subbands in the frequency domain, e.g., N subbands.In turn, respective pairs of subbands, namely, respective positive and negative frequency components, are supplied to a plurality of processing circuits 3301-330n / 2. Based on identifying a particular pair of the corresponding pairs that has a higher noise content and a particular policy, a noise suppression process may be performed on at least one of the frequency components. As will be described in more detail, each processing circuit 330 may perform one or more computations and analysis to identify the higher noise frequency component. In response to this identification, processing circuit 330 may then take actions with respect to at least one of the frequency components to modify, adjust, or otherwise compensate for this identified noise.The resulting processed subbands output by each of the processing circuits 330 are provided to an N-point complex IFFT engine 340 that reintegrates, in a unity process, the individual processed subbands from the frequency domain into a set of time domain samples provided to an N-point Hanning window generator 350. The window generator 350 converts the processed sample block into a stream of samples of complex signals which are output to a demodulator. As with the input window, it should be noted that the output window generated in the window generator 350 may be a Hann window overlapping by 50%. Understand that many variations and alternatives are possible, while this is shown at a high level in the embodiment of Figure 4. For example, in other embodiments, depending on the desired complexity and frequency characteristics, a different type of window function and shape may be chosen, provided that an input / output pair of window functions leaves the sample stream unchanged. For example, the input window function could be a 50% overlap triangular window and the output window could be a 50% overlap rectangular window.As discussed above, in a particular embodiment, a Hann window function may be used to generate overlapping windows to be processed as described herein. Referring now to FIG. 4, an illustration of the various operations performed to generate overlapping windows for processing as described herein and then reintegrate these overlapping windows into a set of output samples is shown. As shown in FIG. 4 at point 410, a plurality of input samples are provided. These input samples are multiplied by a Hann window of length N, shown at point 420, namely a half period of a sine wave (extending from 0 to n). The resulting set of window samples 430 is provided for band division, processing, and reintegration. Thereafter, the processed samples after reintegration (at point 440) are multiplied by the same Hann window at point 450, resulting in a second set of window samples 460. Then, as shown, the window samples 460 are combined with the previously output samples at Pos. 470 so as to output a set of output samples 480. Previous samples 470 may be an N / 2 overlap of an output sample stream. Note that the use of a square Hann window overlapped by half results in a unit window pair, resulting in good transient and frequency characteristics.Referring now to FIG. 5, a block diagram of a processing circuit according to an embodiment is shown. As shown in FIG. 5, processing circuit 500 may be a single one of a plurality of such processing circuits, each configured to operate on a pair of corresponding subbands, e.g., a given positive and negative frequency component. As can be seen, incoming complex signals for the positive and negative frequency components are received in respective value generators 510p,n. In one embodiment, each value generator 510 may generate a magnitude value corresponding to a magnitude of the complex frequency components. In one embodiment, each value generator 510 may generate a magnitude as: I2+Q2. In another embodiment, the magnitude may be a square root of this sum. Of course, other magnitude calculations are also possible in other embodiments.In any case, the magnitudes generated in the value generators 510 p,nare supplied to a comparator 520. In one embodiment, the comparator 520 may be configured to output a comparison value corresponding to a difference between the magnitudes. If the magnitudes are different, this is an indication of interference signal or other unwanted noise. In one embodiment, the comparator 520 may indicate the difference value and indicate which of the magnitudes is greater. This comparison output is provided to a controller 530.In embodiments, the controller 530 is configured to control a mixing circuit 540. As shown, the mixer circuit 540 receives the incoming positive and negative frequency components. Based on the control signals of the controller 530 (which in turn generates the control signals based on the comparison result), the mixing circuit 540 may output potentially adjusted positive and negative frequency components. In particular, based on the comparison value, the controller 530 may cause the mixing circuit 540 to output the received positive and negative frequency components, either unchanged or adjusted in any form. Depending on the embodiment, the adaptation can be a replacement of one of the frequency components with the other frequency component. Or the adaptation can take place in the form of a combination of the two frequency components, which are described in more detail below. Understand that many variations and alternatives are possible, while this is shown at a high level in the embodiment of Figure 5.In a particular embodiment, a processing circuit may be configured to perform a comparison, selection, and duplication function, wherein based on the comparison performed between the frequency components, a higher order of the frequency components than the likely noise-stronger of the frequency components is identified and discarded, and the other frequency component is duplicated into the discarded frequency component (with an inverse of the imaginary axis).Referring now to FIG. 6, a block diagram of a processing circuit according to another embodiment is shown. As shown in FIG. 6, processing circuit 600 may be a single one of a plurality of such processing circuits, each configured to operate on a pair of corresponding subbands. As can be seen, incoming complex signals for the positive and negative frequency components are received in respective magnitude computers 610p,n. In one embodiment, each magnitude calculator 610 may generate a magnitude value according to: I2+Q2. The magnitudes generated in magnitude computers 610 p,nare provided to a comparator 620, which is configured to output a comparison value that is a difference between the magnitudes (and is based on, e.g., the sign that the magnitude is greater). This comparison output is provided to a controller 630.In embodiments, the controller 630 is configured to control a pair of selection circuits 6401 and 6402, which may be implemented as multiplexers. Based on the control signals of the controller 630 (which in turn generates the control signals based on the comparison result), the selection circuits 640 are controlled to pass the lower magnitude frequency component and replace the higher magnitude frequency component with the lower magnitude frequency component complex conjugate.Thus, in processing circuit 600, the magnitude of the two frequency components is compared and the smaller is duplicated into the opposite band with an inverse of the imaginary axis. If there is a interferer at this frequency, then the larger magnitude (in most cases) represents the corrupted signal and the smaller is likely to represent the undamaged signal. Understand that many variations and alternatives are possible, while this is shown at a high level in the embodiment of Figure 6. For example, other algorithms for comparing and selecting may be used. As such an example, time-delayed data from previous and future sample blocks may be used to refine the selection, or adjacent frequency bins may be used to modify the selection process. For example, if the interferer is at the same frequency as a desired signal and out of phase, then the lower power will be on the corrupted sideband, which would be the wrong choice for a simple comparison, selection and duplication function as in Figure 6. Because the phase of the interferer is very unlikely to be exactly false for multiple sample block periods, the selection may be made from historical information as additional criteria in some embodiments.Referring now to FIG. 7, a flowchart of a method according to an embodiment is shown. As shown in FIG. 7, method 700 is an example implementation of a method for noise cancelling in AM signal processing. As such, the method 700 may be performed by hardware circuitry, firmware, software, and / or combinations thereof. In a particular embodiment, the method 700 may be performed in a hardware noise suppression circuit as described herein.As seen, the method 700 begins with receiving a complex signal from a tuner (block 710). This complex signal can be received as a signal stream of samples of incoming I and Q values. Subsequently, at block 720, an overlapping block or window of samples of the complex signal may be generated. As described herein, the overlapping block may have a width of 848 samples, and the overlapping block contains a portion of all 848 samples, although the system proceeds only 1024 samples at a time by overlapping 1024 of the samples with the previous block and storing 1024 samples to be used to overlap the next block.Thereafter, at block 730, this overlapping window is filtered into multiple sets of filtered pairs. More specifically, 2046 pairs of filter sets in the frequency domain, 1023 positive and 1023 negative frequency components, can be generated, with DC (tone 0) and FS / 2 (tone 1024) treated differently. Note that these pairs are symmetrical to DC, such that each filtered pair has a positive frequency component and a negative frequency component. And again, each frequency component itself is formed of a complex value, namely I and Q values.Next, one of the filtered pairs of each set is selected (block 740). More specifically, based on one or more integrity criteria and the samples themselves, the one of these filtered pairs may be selected with less noise content. In general, the criteria may be based on a comparison of the balanced components to determine which of the two balanced frequency components is likely to have a higher noise content. The selected one of the filtered pairs may thus be that of the two balanced frequency components with less noise content.Control next proceeds to block 750, where the selected one of the filtered pairs may be duplicated and inverted to replace the other one of the filtered pairs identified as having a higher noise content. Thereafter, the processed filtered pairs are passed to a reintegrator where, at block 760, the multiple updated sets may be integrated into an overlapping window of the complex signal. At block 770, this overlapping window may then be reconverted as the sampling current of the complex signal back to the time domain, which is forwarded to the further signal processing circuit, e.g. for demodulation. Note that the division and reintegration of the signal stream may be performed via a unity process so that no underlying content is lost within the signal stream. While a compare, select, and duplicate function is performed in the high level illustration of FIG. 7, other signal processings may also be performed in other embodiments to remove unwanted noise components from at least one of the two sideband.Referring now to FIG. 8, a flowchart of a method according to another embodiment is shown. As shown in FIG. 8, method 800 is an example implementation of a method for noise cancelling in AM signal processing. As such, the method 800 may be performed by hardware circuitry, firmware, software, and / or combinations thereof. In a particular embodiment, the method 800 may be performed in a hardware noise suppression circuit as described herein.At block 810, positive and negative frequency components of a filtered pair may be received within the processing circuit. Subsequently, at block 820, magnitudes for the various frequency components may be calculated. More specifically, a first magnitude for the positive frequency component and a second magnitude for the negative frequency component may be calculated. Although the embodiments are not limited in this respect, in one embodiment, the magnitude may be determined by executing a sum of squares of the individual complex (I and Q) components.Still referring to FIG. 8, at block 830, a comparison may be made between these magnitudes to obtain a comparison result. In one embodiment, the comparison may be implemented as a difference calculation so as to identify a magnitude of the difference, along with an identification of which of the two components has a larger magnitude (e.g., as determined based on the sign of the comparison result).Still referring to FIG. 8, it is determined whether the comparison result exceeds a first threshold (diamond 840). This threshold may correspond to a relatively small deviation, e.g. in the order of about 5% to 10% deviation with respect to the power in the entire signal. If it is determined that the comparison result does not exceed this threshold, control passes to block 845 where the positive and negative frequency components (unmodified) of the filtered pair may be output. Note that these frequency components are output to a reintegrator where the reintegration process can be performed using these frequency components of this filtered pair along with the possibly adjusted frequency components of the additional frequency pairs.Still referring to FIG. 8, if the comparison result is determined to exceed the first threshold, control passes to diamond 850 to determine if the comparison result exceeds a second, higher threshold. In one embodiment, this second threshold may be set to a higher level to indicate a greater difference between the magnitudes, e.g., on the order of between about 10% and 50% deviation. Note that in some embodiments, this second threshold and the first threshold may be dynamically updated based on a calculated noise floor. If it is determined that the comparison result does not exceed this second threshold, control passes to block 855 where an adjustment operation may be performed. More specifically, in this matching operation, at block 855, the positive and negative frequency components may be combined according to a particular policy. As an example, a kind of combination ratio may be used to combine the frequency components. For example, the larger magnitude may have an undesirable amount of noise. Thus, a certain percentage of the lower magnitude frequency component may be combined with a certain percentage of the higher magnitude frequency component. As a particular example, a smaller percentage of the lower magnitude frequency component may be combined with a larger percentage of the higher magnitude frequency component. Of course, other examples are possible and it is also possible to modify the component of the lower magnitude. Thereafter, these updated positive and negative frequency components of a filtered pair may be output as discussed above (block 870).If instead it is determined at diamond 850 that the comparison result actually exceeds the second threshold, control passes to block 860. At block 860, a duplication process may be performed to replace the signal information of the frequency component with a higher noise content with the signal information of the frequency component with a lower noise content. As illustrated at block 860, this duplication process may result in the complex conjugate of the selected one of the filtered pair being duplicated (and thus replaced) into the other one of the filtered pair. Understand that other examples of determining when and what type of adjustment is to be performed may occur than shown with this particular type of selection criteria.Referring now to FIG. 9, a block diagram of a system according to an embodiment is shown. As shown in FIG. 9, the system 900 is at least a portion of a vehicle infotainment system. As shown, system 900 includes various components that are adapted on a printed circuit board, as well as additional components that may be located in other parts of a vehicle.In the embodiment illustrated in FIG. 9, incoming radio frequency (RF) signals are received from multiple antennas 910a, 910b (generically antennas 910). Understand that while two antennas are shown for discussion purposes, in various implementations, a vehicle may be adapted with only a single antenna or more than two antennas, as needed for a particular level of radio solution as well as desired operating bands. For purposes of discussion herein, it is to be assumed that antennas 910 are configured to receive digital radio communication in accordance with one or more digital radio standards such as DAB, HD radio, digital radio mondiale (DRM), or so forth, analog radio transmitters (e.g., AM and FM), and potentially other broadcast signals, including video signals, e.g., in accordance with a digital media broadcast (DMB) standard.Received RF signals from antennas 910 are in turn provided to an integrated circuit 920, which in one embodiment may comprise a single semiconductor die. As shown, integrated circuit 920 may include multiple signal processing paths. For purposes of illustration only two signal processing paths are illustrated, however, it should be appreciated that a particular receiver and tuner IC may have more than two such paths to receive, down convert, and process radio signals of many different bands and modulation schemes. As shown, the RF signals received by antennas 910 are provided to respective tuners 922a, 922b (generic tuner 922). Tuners 922 may be multiband tuners to receive and process RF signals of different bands. In general, tuners 922 may include RF front end circuits such as LNA and other gain control circuits, mixers, filters, digitizers, etc., that are for receiving and processing the RF signal and generating a resulting digitized signal having a down-converted frequency. For example, tuners 922 may be configured to output signals at baseband, at the zero intermediate frequency (ZIF), or at low intermediate frequency or another down-converted level. In embodiments, tuners 922 may output such signals in digitized form.As further illustrated in Figure 9, the resulting digitized signals are provided to demodulators 932a,b. Demodulation circuits 932 a,bmay receive incoming signal information from one of the tuners 922 a, 922 b. The demodulator circuits 932, in turn, demodulate the incoming signals received in modulated form. Generally, demodulator circuits 932 may include various circuits, including asynchronous sample rate converters, decoder circuits, and so forth. Demodulator circuits 932 output demodulated signals output from the IC 920.Note that with respect to the first signal processing path with tuner 922 aand demodulator 932 a, a noise suppression circuit 925 is located therebetween. The noise suppression circuit 925 may implement an embodiment herein to perform the noise suppression of down-converted AM signals. To this end, the noise suppression circuit 925 may eliminate a noise source that may be generated by an electric motor or other electrical components in a vehicle, such as an electric or hybrid vehicle. As such, noise suppression circuit 925 may output noise compensated signals for demodulation in demodulator 932 a.Still referring to FIG. 9, demodulated signals output from the demodulators 932 are provided to a system on chip (SoC) 950, which is a main processor of the infotainment system 900. As shown in FIG. 9, SoC 950 includes a processing machine 955. Although a single processing machine is illustrated for purposes of illustration, it should be appreciated that multiple processing machines may be provided in various implementations. As examples, the processing engine 955 may be implemented as one or more general purpose processor cores, one or more DSPs, and / or one or more other programmable logic circuits.SoC 950 is additionally shown to include a radio application 960, which in one embodiment may be a high level radio application of the system and which may be executed on the processing engine 955. The radio application 960 may serve as an interface to receive user input (e.g., a request for a particular radio station) and provide instructions to additional components to achieve the desired functionality.Note that although the radio application 960 is shown as a separate component within the SoC 950, it should be understood that it may be implemented as software and / or firmware executing on the processing engine 955 or other programmable circuitry within the SoC 950. Thus, regardless of whether implemented as software or firmware, the instructions are stored on a non-transitory storage medium. This storage medium may be implemented within the SoC 950 itself, such as internal non-volatile memory or external memory such as external flash memory of the system 900. Further understand that while a particular description of the components is illustrated in FIG. 9, other implementations are possible.As further illustrated in FIG. 9, the processing engine 955 outputs audio signals that may be output to an optional audio processor 980. The audio processor 980 may perform additional audio processing such as post processing, balance control, fading, etc. Audio processor 980, in turn, outputs audio signals to one or more speakers 990. Understand that many variants and alternatives are possible, while this is shown at high level in the embodiment of Figure 9.While the present invention has been described with reference to a limited number of embodiments, those skilled in the art will appreciate numerous modifications and variations thereof. It is intended that the appended claims cover all modifications and variations that fall within the true spirit and scope of this present invention.

Claims

A method comprising: receiving, in a noise suppression circuit of a receiver, a plurality of samples of a down-converted signal; separating the plurality of samples into a plurality of subbands; for each of corresponding pairs of the plurality of subbands, comparing a first value based on a first component of the corresponding pair to a second value based on a second component of the corresponding pair; based at least in part on the comparison, matching at least a portion of one of the first and second components of the corresponding pair to at least a portion of the other of the first and second components of the corresponding pair; integrating the plurality of subbands into a second plurality of samples; and outputting the second plurality of samples for demodulation.The method of claim 1, further comprising receiving a complex signal and forming an overlapping window with the plurality of samples.The method of claim 1, further comprising: performing a complex fast Fourier transform to separate the plurality of samples into the plurality of subbands; and performing a complex inverse fast Fourier transform to integrate the plurality of subbands into the second plurality of samples.The method of claim 1, wherein the separating and integrating comprises a unity function.The method of claim 1, wherein: calculating the first value comprises calculating a first magnitude for the first component, the first component comprising a positive frequency component of the corresponding pair of the plurality of subbands; and calculating the second value comprises calculating a second magnitude for the second component, the second component comprising a negative frequency component of the corresponding pair of the plurality of subbands.The method of claim 5, further comprising comparing the first magnitude with the second magnitude to obtain a comparison result.The method of claim 6, further comprising, in response to the comparison result exceeding a first threshold, replacing one of the positive frequency component and the negative frequency component with a complex conjugate of the other of the positive and negative frequency components.The method of claim 6, further comprising, in response to the comparison result exceeding a second threshold, combining at least a portion of one of the positive frequency component and the negative frequency component with at least a portion of the other of the positive frequency component and the negative frequency component, wherein the second threshold is less than the first threshold.An apparatus comprising: an analog front end circuit for receiving and downconverting an amplitude modulation (AM) signal to a second frequency signal; a digitizer to digitize the second frequency signal into a plurality of samples; a noise suppression circuit coupled to the digitizer for receiving the plurality of samples, the noise suppression circuit comprising: a window generator to generate a window having a first set of the plurality of samples; a band splitter to split the window into a plurality of pairs of balanced frequency components; a processing circuit for each of the plurality of pairs for: comparing a first magnitude of a first balanced frequency component with a second magnitude of a second balanced frequency component and modifying one of the first balanced frequency component and a second balanced frequency component based on the comparison; an integrator for integrating the plurality of pairs output by the processing circuit; and a second window generator for generating a second window having a second set of samples; and a demodulator coupled to the noise cancellation circuit for demodulating the second set of samples to output a demodulated AM signal.The apparatus of claim 9, wherein the band splitter comprises a fast Fourier transform engine and the integrator comprises an inverse fast Fourier transform engine.The apparatus of claim 9, wherein the processing circuit comprises a plurality of individual circuits, each of the plurality of individual circuits comprising: a first value generator for generating the first magnitude of the first balanced frequency component, the first balanced frequency component comprising a complex signal; a second value generator for generating the second magnitude of the second balanced frequency component, the second balanced frequency component comprising a complex signal; and a comparator for comparing the first magnitude with the second magnitude and outputting a comparison result based thereon.The apparatus of claim 11, further comprising: a controller to receive the comparison result and output one or more control signals based on a comparison of the comparison result with at least one threshold; and a mixing circuit to modify the one of the first balanced frequency component and the second balanced frequency component based on the one or more control signals.The apparatus of claim 12, wherein the controller is configured to replace the second frequency component with the first frequency component when the comparison result exceeds the at least one threshold.The apparatus of claim 12, wherein the at least one threshold comprises a dynamic threshold based on a dynamically calculated noise carpet.The apparatus of claim 12, wherein the controller is configured to cause the mixing circuit to output the first balanced frequency component and the second balanced frequency component unchanged when the comparison result is less than the at least one threshold.The apparatus of claim 11, wherein the window generator is configured to generate the window having at least some overlapping samples of a previous window, and the second window generator is configured to generate the second window overlapping at least a portion of the previous window.The apparatus of claim 9, wherein the noise suppression circuit is configured to remove the noise present in a first sideband of the AM signal caused by one or more electrical components of a vehicle, the apparatus comprising a radio receiver of the vehicle.A system comprising: an antenna for receiving an amplitude modulation (AM) signal; a tuner coupled to the antenna for receiving and processing the AM signal to output a down-converted modulated signal; a noise suppression circuit coupled to the tuner for: generating a window having a first set of a plurality of samples of the down-converted modulated signal; dividing the window into a plurality of pairs of balanced frequency components; for one or more of the plurality of pairs, for: comparing a first magnitude of a first balanced frequency component to a second magnitude of a second balanced frequency component; and modifying one of the first balanced frequency component and the second balanced frequency component based at least in part on the comparison; and integrating the plurality of pairs into a second window having a second set of samples; a demodulator coupled to the noise cancelling circuit for demodulating the second set of samples to output a signal stream; and an output device for playing back the signal stream.The system of claim 18, wherein the noise suppression circuit is configured to modify one of the first balanced frequency component and the second balanced frequency component by replacing the second frequency component with a representation of the first frequency component when a comparison result of the first magnitude with the second magnitude exceeds a threshold.The system of claim 18, wherein the system comprises a vehicle infotainment system, wherein the noise suppression circuit is configured to compensate for a noise source present in an electric vehicle or a hybrid electric vehicle with the vehicle infotainment system.

Images