System and method for adjusting the clarity of an audio output - Patents.com

Abstract

1. A method for adjusting intelligibility of an audio output in a changing environment, comprising: receiving a content signal; applying a customized gain to the content signal; and outputting the content signal with the customized gain to at least one speaker for conversion to an acoustic signal, wherein the customized gain is applied for each frequency bin such that frequencies of lower amplitude are enhanced relative to frequencies of higher amplitude and the intelligibility of the acoustic signal is set to approximately a desired level, wherein the customized gain is determined according to at least one of a gain applied to the content signal, a bandwidth of the content signal, and a content type encoded by the content signal.

Description

[Technical field]

[0001] The present disclosure relates generally to systems and methods for adjusting the intelligibility of an audio output. Summary of the Invention [Means for solving the problem]

[0002] All examples and features mentioned below can be combined in any technically possible manner.

[0003] According to one aspect, a method for adjusting intelligibility of an audio output in a changing environment includes receiving a content signal, applying a customized gain to the content signal, and outputting the content signal with the customized gain to at least one speaker for conversion to an acoustic signal, where the customized gain is applied for each frequency bin such that lower amplitude frequencies are enhanced relative to higher amplitude frequencies and the intelligibility of the acoustic signal is set to approximately a desired level, where the customized gain is determined according to at least one of a gain applied to the content signal, a bandwidth of the content signal, and a content type encoded by the content signal.

[0004] In one example, the customized gain is based on dynamic range compression.

[0005] In one example, the customized gain is further based on equalization.

[0006] In one example, the equalization is a spectral tilt.

[0007] In one example, the gain applied to the content signal is set according to a volume control.

[0008] In one example, the audio output is generated within the vehicle cabin and the gain is set at least in part according to road noise within the vehicle cabin and according to a volume control.

[0009] In one example, the gain set according to the road noise in the vehicle cabin consists of a set of frequency dependent gain adjustments.

[0010] In one example, dynamic range compression is performed only within the bandwidth of the signal.

[0011] In one example, the content type is one of musical output or spoken word output.

[0012] In one example, the content type is detected according to a voice activity detection algorithm.

[0013] According to one aspect, a non-transitory storage medium comprising program code that, when executed by a processor, adjusts intelligibility of an audio output in a changing environment, the program code including steps of receiving a content signal, applying a customized gain to the content signal, and outputting the content signal with the customized gain to at least one speaker for conversion to an acoustic signal, the customized gain being applied for each frequency bin such that frequencies of smaller amplitude are enhanced relative to frequencies of larger amplitude and the intelligibility of the acoustic signal is set to approximately a desired level, the customized gain being determined according to at least one of a gain applied to the content signal, a bandwidth of the content signal, and a content type encoded by the content signal.

[0014] In one example, the customized gain is based on dynamic range compression.

[0015] In one example, the customized gain is further based on equalization.

[0016] In one example, the equalization is a spectral tilt.

[0017] In one example, the gain applied to the content signal is set according to a volume control.

[0018] In one example, the audio output is generated within the vehicle cabin and the gain is set at least in part according to road noise within the vehicle cabin and according to a volume control.

[0019] In one example, the gain set according to the road noise in the vehicle cabin consists of a set of frequency dependent gain adjustments.

[0020] In one example, dynamic range compression is performed only within the bandwidth of the signal.

[0021] In one example, the content type is one of musical output or spoken word output.

[0022] In one example, the content type is detected according to a voice activity detection algorithm.

[0023] The details of one or more implementations are set forth in the accompanying drawings and the description below. Other features, objects, and advantages will become apparent from the description and drawings, and from the claims.

[0024] In the drawings, like reference numbers generally refer to the same parts throughout the different views. Also, the drawings are not necessarily to scale, emphasis instead generally being placed upon illustrating the principles of various aspects. [Brief description of the drawings]

[0025] [Figure 1] FIG. 2 is a block diagram of a controller implementing an audio processing chain and a speaker according to one embodiment. [Diagram 2] FIG. 2 is a block diagram of clarity adjustment according to one embodiment. [Figure 3A]1 illustrates a plot showing input power spectral density over frequency according to one embodiment. [Figure 3B] 1 shows a plot of the difference between output spectral density and input power spectral density of dynamic range compression across frequency according to one embodiment. [Figure 4] 1 shows a plot of the difference between the output spectral density and the input power spectral density of spectral tilt over frequency according to one embodiment. [Diagram 5] 1 illustrates an interface for adjusting clarity and fine-tuning of audio output according to one embodiment. [Figure 6] FIG. 2 illustrates a plot of loudness and intelligibility mapping according to one embodiment. [Figure 7] 1 shows plots of loudness, noise compensation, and dependent intelligibility mapping according to one embodiment. [Figure 8] 1 shows a bar graph of intelligibility as a function of bandwidth for various exemplary signals. [Figure 9] 4 illustrates the difference in intelligibility as determined by the bandwidth of various input signals, according to one embodiment. [Figure 10] 13 illustrates intelligibility settings when the content type is changed between a spoken word signal and a music signal according to one embodiment. [Figure 11A] 2 illustrates a flowchart of a method for adjusting the intelligibility of an audio output in a dynamic environment, according to one embodiment. [Figure 11B] 2 illustrates a flowchart of a method for adjusting the intelligibility of an audio output in a dynamic environment, according to one embodiment. [Figure 11C] 2 illustrates a portion of a flowchart of a method for adjusting the intelligibility of an audio output in a dynamic environment, according to one embodiment. [Figure 11D] 2 illustrates a portion of a flowchart of a method for adjusting the intelligibility of an audio output in a dynamic environment, according to one embodiment. [Figure 11E] 2 illustrates a portion of a flowchart of a method for adjusting the intelligibility of an audio output in a dynamic environment, according to one embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0026] Spoken word audio output provided in a noisy context, such as a vehicle interior, is often very difficult to understand. This difficulty is often exacerbated by the fact that such noisy environments typically increase the cognitive load of the user. For example, driving requires a level of concentration, impairing the user's ability to concentrate and understand the words in a particular spoken word output. The inability to understand words in the spoken word output may be exacerbated by the low bandwidth of a particular spoken word output, such as a mobile phone call, which may further reduce intelligibility. It is therefore desirable to process audio signals to increase the intelligibility of the spoken word output in the context of a dynamic environment.

[0027] Intelligibility, as used in this disclosure, generally refers to the intelligibility of a spoken word audio output output from a speaker or set of speakers (e.g., a speaker array). Although there is necessarily a subjective perception associated with intelligibility, it can be measured by querying what percentage of words in the spoken word output were intelligible to a given listener. This measurement can be made objective by averaging responses of a set of listeners to the same query, although other objective measures of intelligibility are known and can be used.

[0028] Intelligibility input, as used in this disclosure, refers to a default or user-defined input (into an audio system, such as in a vehicle) of a desired intelligibility of the audio output. Additionally, the intelligibility input may be adjustable by the user, or may be presented to the user as an input entitled "Intelligibility," "Clarity of Spoken Word Audio," "Speech Intelligibility," etc.

[0029] The intelligibility of the audio output can be adjusted to approximate the intelligibility input (i.e., within the boundaries of a predetermined range) by applying a customized gain per frequency bin (e.g., having a variable gain for each frequency bin) based on various conditions, thereby enhancing quieter sounds (e.g., increasing the level or volume for louder sounds) and / or attenuating louder sounds (e.g., decreasing the level or volume for quieter sounds). The degree of adjustment, or processing, applied to the audio signal to approximate the intelligibility input is referred to as the intelligibility setting throughout this disclosure, with a higher intelligibility setting representing a greater adjustment made to the signal to enhance intelligibility and a lower intelligibility setting representing a lesser adjustment to the signal.

[0030] Adjusting the intelligibility setting to the maximum amount will provide the greatest improvement to intelligibility, but it often comes at the expense of a harsh, robotic, or less rich audio. Thus, the intelligibility input represents the amount of tradeoff a user is willing to make between being able to hear words at a given audio signal level and the resulting change in the tonal balance of the audio signal brought about by the intelligibility setting. For example, a user with mild to moderate hearing impairment may be willing to accept a greater degree of altered tonal balance with the benefit of greater intelligibility at lower audio levels compared to an individual with perfect hearing.

[0031] However, even for users with hearing impairments, it is typically desirable to apply an intelligibility setting that results in a desired intelligibility of the spoken word output (i.e., corresponding to the intelligibility input) so that the resulting audio output is not overcompensated. However, the amount of adjustment required to maintain the output at the desired intelligibility is context-specific and is affected by a variety of factors, such as the noise in the surrounding environment, the volume of the spoken word output, the amount and characteristics of upstream processing on the input signal, the quality of the signal, and the type of content generated within the signal, as mentioned above.

[0032] In order to provide audio output with intelligibility provided at a desired level within such dynamic environments, the intelligibility setting must be adjusted to fit the context. For example, in a vehicle cabin, as the vehicle speeds up and noise within the cabin increases, the signal-to-noise ratio within the cabin decreases, necessitating adjustment of the intelligibility setting if intelligibility of the spoken word output is to be maintained. Therefore, it is desirable to automatically determine the intelligibility setting depending on a set of predetermined factors.

[0033] 1 shows a block diagram of a processing chain of an audio system 100 implementing a context-specific intelligibility setting in an intelligibility adjustment 102. As described above, the intelligibility of the audio output d can be adjusted by the intelligibility adjustment 102 to implement an intelligibility setting that approximates the intelligibility input (i.e., within the boundaries of a predetermined range) by applying a customized gain per frequency bin (e.g., having a variable gain for each frequency bin) based on various conditions, thereby enhancing quieter sounds (e.g., increasing the level or volume for louder sounds) and / or attenuating louder sounds (e.g., decreasing the level or volume for quieter sounds). The intelligibility setting applied by the intelligibility adjustment 102 is determined by the parameters c received from the parameter mapping 104. 出力This parameter mapping implements an intelligibility setting that determines the intelligibility of the output signal d that is converted to a desired level (set by default settings or user-defined settings) by the speaker 106 or a set of speakers (e.g., organized in an array) within the context of the dynamic environment and / or input content signal, taking into account other processing performed on the content signal u (e.g., one or more gain adjustments) and the characteristics of the input content signal (e.g., bandwidth and content type). (It should be understood that to the extent that the intelligibility of the audio output d is described in this disclosure, it refers to the intelligibility of the acoustic output resulting from the speaker converting the audio output d. Furthermore, the output signal d may further include additional adjustments to prepare the signal for conversion, such as beamforming, routing, etc. In general, these processes performed on the signal after the intelligibility adjustment 102 do not affect the intelligibility of the audio signal.)

[0034] The desired level can be determined by a default setting or by a received user input c 入力 The intelligibility setting can be adjusted based on user-defined fine-tuning inputs to tailor the intelligibility setting to the hearing needs of a particular user. 入力 Thus, in this example, the parameter for adjusting the intelligibility setting is the fine-tuned output t 出力 It may further include.

[0035] In one example, the signal processing chain of the audio system 100 may be implemented by a controller 108 comprising one or more processors along with any associated hardware. In one example, some or all of the steps performed by the audio processing chain may be stored in one or more non-transitory storage media and executed by one or more processors. In certain examples, portions of the audio system processing chain, such as volume control or noise compensation (discussed below), may be executed in a processor or hardware (circuitry) separate from the controller 108. The controller 108 and the speaker 106 or set of speakers may be located in a vehicle and configured to generate audio output within the vehicle cabin. However, in alternative examples, the controller 108 and the speaker or set of speakers 106 may be located elsewhere, such as in a sound bar or any device that generates audio output to a dynamic environment.

[0036] An example of an intelligibility adjustment 102 is shown in FIG. 2, which includes dynamic range compression 202 and spectral tilt 204. In this example, the intelligibility setting can be adjusted by changing the parameters of the dynamic range compression, e.g., implemented thresholds, associated gain reduction, etc., in a manner designed to improve the intelligibility of the spoken word output. In general, a higher compression of the dynamic range, which reduces the difference in sound level between louder and quieter content, is associated with a higher intelligibility of the speech. The gain applied within a frequency bin can be determined, at least in part, by the characteristics of the input signal (typically peaks or power) and can be achieved by downward compression, upward compression, or both. Downward compression reduces the gain when the signal exceeds a predefined threshold, and upward compression increases the gain when the signal falls below a predefined threshold. Digitally, dynamic range compression is performed on a frequency bin basis by performing compression for a given frame according to the energy within the frequency bin. Dynamic range compression is commonly known, and will not be described in greater detail in this disclosure.

[0037] In this example, the intelligibility setting can be varied between a setting that applies no dynamic range compression, and therefore does not modify the input signal, and a full intelligibility setting that applies the highest predetermined dynamic range adjustment, resulting in the highest degree of intelligibility of the spoken word output. 入力 A change in one direction can be associated with an increase in intelligibility, and a change in the other direction can be associated with a decrease in intelligibility. In other words, in this example, adjusting to a higher intelligibility setting would adjust the parameters of the applied dynamic range compression to increase the intelligibility of the speech.

[0038] A higher intelligibility setting may mean that more compression is applied to certain frequencies, but it does not always follow that higher compression is applied to all frequencies. Rather, a higher intelligibility setting results in dynamic range compression parameters that shape the output signal to provide higher speech intelligibility. Such dynamic range compression parameters can be determined theoretically (e.g., by inverting the average hearing threshold curve), empirically (e.g., by querying users about the hearing improvement achieved with different compression parameters), or by a combination of theoretical and empirical approaches.

[0039] 3A and 3B show an example dynamic range compression for a sample input signal. More specifically, FIG. 3A shows the input power spectral density (PSD) of the sample input signal versus frequency. FIG. 3B shows the difference between the input and output PSD of the dynamic range compression 202 to illustrate the gain adjustment performed by the dynamic range compression 202. As shown, frequency bins with higher amplitudes receive little or no gain adjustment. Meanwhile, frequency bins with relatively low amplitudes are increased through upward compression such that quiet sounds are expanded relative to loud sounds.

[0040] Furthermore, while dynamic range compression that improves intelligibility for an average user can be determined theoretically or empirically, the resulting dynamic range compression may not improve intelligibility for all users: certain users may experience hearing impairments in atypical ways, and dynamic compression tuned for the average user will not address how a given user perceives spoken word content.

[0041] Thus, the intelligibility of the audio output can be further tailored to a particular user by applying equalization, such as spectral tilt, to the signal. In the example of FIG. 出力 A spectral tilt 202 adjusted according to increases the gain of high frequencies and decreases the gain of low frequencies, or when reversed increases the gain of low frequencies and decreases the gain of high frequencies. An example of this is shown in FIG. 4, which shows the difference between the output PSD and the input PSD. At one end of the spectral tilt setting, the tilt frequency f t The gain at frequencies below the slope frequency f t Above that, the gain is increased (represented by curve 402). At the other end of the spectral tilt setting, the tilt frequency f t The gain at frequencies below the slope frequency f t Above that, the gain is increased (represented by curve 404). Spectral tilt addresses the needs of different listeners who may have hearing impairments in the low or high ranges by allowing the user to emphasize the signal at one end or the other.

[0042] While spectral tilt has been shown to provide good results, other equalizations can be used to improve the intelligibility of the audio output. For example, additional adjustments can increase high frequencies without reducing low frequencies, or increase low frequencies without reducing high frequencies, although other audio equalizations are contemplated.

[0043] As shown in FIG. 5, clarity and spectral tilt can be presented to the user as settings that are each adjustable through an interface, such as a knob, slider, touch screen (e.g., in conjunction with a dedicated or multi-purpose program application), a remote control, or any other suitable interface. Both settings can be adjusted over a range of values. As an example, the lowest clarity setting is 0 and the highest clarity setting is 100, with values ​​selectable therebetween, although it will be understood that the end points and intermediate values ​​are arbitrary and any suitable value can be selected. Additional adjustments to the clarity setting can similarly be made according to user input, and the user can be presented with inputs entitled "tweak," "tuning," "treble / bass," etc.

[0044] The signal processing chain is configured to provide at least one gain adjustment before applying the intelligibility setting. In the example of Figure 1, the signal processing chain performs two gain adjustment operations on the input content signal u: (1) a volume control gain adjustment at volume control 110, and (2) a noise compensation gain adjustment at noise compensation 112.

[0045] The gain adjustment of the volume control 110 represents the volume control v set by the user. The volume control may be input according to an actuator or touch interface located on the dashboard, for example, or may be accessible via a remote control or via other known methods.

[0046] Noise compensation 112 applies gain to the content signal u to compensate for ambient noise, such as road noise, according to a noise compensation algorithm. A simpler example of a noise compensation algorithm applies a uniform gain across frequencies as road noise increases in the vehicle cabin (e.g., in response to vehicle speed, gear position, etc.). In more complex noise compensation algorithms, such as the Bose AudioPilot algorithm, the noise compensation may apply independent gains to each frequency band as road noise increases.

[0047] The noise compensation 112 can be applied according to a road noise signal n indicative of road noise in the vehicle cabin. In one example, the road noise signal n can be an output from one or more microphones positioned in the vehicle cabin to detect road noise. Alternatively or additionally, the road noise signal n can be a signal indicative of vehicle speed, gear position, or other metric output from a vehicle computer and indicative of road noise in the vehicle cabin.

[0048] Since the total gain of the output signal d is related to the signal-to-noise ratio and thus the intelligibility of the output acoustic signal, the parameter mapping 104 needs to take into account the applied gain. More generally, the intelligibility setting may be reduced as the gain increases to account for the increased signal-to-noise ratio. In the example of FIG. 1, the parameter mapping 104 receives the applied gain from the volume control 110 as a context parameter p1 and, through manipulation of the volume input, may decrease the intelligibility setting as the gain increases. This may be achieved through a predefined gain-intelligibility mapping, as shown in FIG. 6. The points of the gain-intelligibility mapping may be predefined through a set of user-defined points, for example as default settings designed to maintain the same perceived intelligibility as the volume increases through a set of user-defined points. (To avoid setting each point individually, the user may define two or more points, and the remaining points may be determined by an appropriate method of interpolation or extrapolation. In either case, manipulation of the gain via the volume knob adjusts the intelligibility setting according to the mapping.

[0049] The parameter mapping 104 may further take into account the gain adjustment implemented by the noise compensation 112, which is received as a context parameter p2. Because the gain adjustment performed by the noise compensation 112 already takes into account the increased noise in the vehicle cabin, and because the gain adjustment is often applied frequency-dependently, a separate intelligibility mapping that takes into account the nature of the gain adjustment, the increased noise in the vehicle cabin, and the gain adjustment already performed by the volume control 110 must be applied to properly configure the intelligibility setting of the intelligibility adjustment 102. Figure 7 shows an intelligibility mapping that takes into account both the gain adjustment from the volume control 110 and the noise compensation 112. The axis labeled "Volume Gain" is the gain-intelligibility mapping as represented in Figure 6. On the axis labeled "Noise Compensation Gain" there is an additional intelligibility mapping that is adjusted as the gain from the noise compensation increases. In this example, the intelligibility applied according to the gain parameters p1 and p2 thus follows the surface created by the gain-intelligibility mappings of the volume and noise compensation.

[0050] In one example, the shape of the surface of the plot in FIG. 7 is determined by equal “intelligibility” contours. In other words, the contours of the surface in FIG. 7 are determined such that for a given volume, the intelligibility of the audio output is approximately maintained across different noise compensation gains. For example, at a volume setting of “20”, when the vehicle is stationary, the noise compensation 112 applies no gain to the signal. As the vehicle starts to move and road noise increases, the noise compensation 112 starts to apply gain in at least one frequency band. Assuming the user does not adjust the volume knob, the intelligibility is adjusted along the contour 702 according to the gain applied by the noise compensation 112. When the user adjusts the volume, the intelligibility setting is determined at the new contour, and for example, if the user changes the volume to “80”, the intelligibility setting is determined along the contour 704 according to the gain applied by the noise compensation 112. In general, the noise compensation 112 gain is lower at higher volumes and higher at lower volumes, so smaller adjustments to the intelligibility setting are needed to maintain a desired amount of intelligibility as the volume increases.

[0051] It should further be understood that in addition to volume control and noise compensation gain adjustments, other gain adjustments can be made, for example audio equalization can be implemented solely on the quality of the audio output (e.g., to adjust bass, midrange, or treble according to user preferences) and taken into account by parameter mapping 104 through appropriate adjustments to the intelligibility setting.

[0052] Although the context parameters p1 and p2 are shown as being received from the volume control 110 and the noise compensation 112, it should be understood that in various examples, the parameters representing the gains applied to the content signal u may be detected from the signal rather than being received directly as parameters. For example, the gains applied to the signal may be detected by comparing the signals input to and output from the amplifiers of the volume control 110 or the amplifiers of the noise compensation 112. Alternatively, the volume input v or the noise signal n may be input to the parameter mapping 104 to determine which gains are applied by the volume control v or the noise compensation n according to these inputs. In this example, the volume input v or the road noise signal n may be considered as context parameters because they are used by the parameter mapping 104 as proxies for the gain application parameter mapping 104 to repeat the same calculations or adjustments performed by the volume control 110 and the noise compensation 112 to determine the gains applied by each.

[0053] In addition to gain adjustments, the characteristics of the content signal u determine the need for intelligibility adjustments to maintain the desired amount of intelligibility. For example, in general, intelligibility decreases with bandwidth. Cell phone calls tend to have low bandwidths, while various voice-over-IP services have been shown to have higher but varying bandwidths. Streaming services such as podcasts, or audio files such as audiobooks, tend to have very high bandwidths. Figure 8 shows the intelligibility of various input signals. As shown, cell phone calls tend to have the lowest bandwidth and therefore the lowest intelligibility. HD cell phone calls have higher bandwidths and higher intelligibility. These two are followed by three voice-over-IP services of various bandwidths: VoIP service #1, VoIP service #1, VoIP service #1. In this example, VoIP service #1 has the lowest bandwidth and therefore the lowest intelligibility of the VoIP services. Meanwhile, VoIP service #3 has the highest bandwidth and therefore the highest intelligibility. Generally speaking, very high quality signals have a bandwidth of 16 kHz or greater, with quality and intelligibility decreasing as the bandwidth decreases from that value.

[0054] As shown in FIG. 1, the bandwidth detection 114 receives and determines the bandwidth of the content signal u according to any suitable bandwidth detection method, and outputs a context parameter p3 representing the bandwidth of the content signal u to the parameter mapping 104. To maintain the intelligibility at a desired amount, i.e. according to the intelligibility input, the parameter mapping 104 makes a larger intelligibility adjustment (i.e. the intelligibility setting is set to a higher amount) for the low bandwidth signal and a smaller intelligibility adjustment (i.e. the intelligibility setting is set to a lower amount) for the high bandwidth signal. This is represented by the "desired intelligibility" line in FIG. 8. Each signal receives a different intelligibility adjustment (represented by the respective arrows) that depends on the bandwidth of the signal. FIG. 9 represents an example of the adjustment of the intelligibility and fine-tuning settings for signals of different bandwidths (in addition to the intelligibility and fine-tuning settings already applied according to other factors taking into account the intelligibility input). Although fine-tuning as described herein is typically used as a way to adjust the implementation of the intelligibility setting, the applicant recognizes that increasing the treble and decreasing the bass of the low bandwidth signal generally improves intelligibility. 9, the intelligibility setting (e.g., a parameter for dynamic range compression) represented by line 902 is higher for lower bandwidth signals and is linearly decreased for very high quality signals, such as fullband podcasts, until no intelligibility adjustment is made beyond the intelligibility setting already determined taking into account the intelligibility input and according to other factors. Similarly, the amount of spectral tilt represented by line 904 represents a linear adjustment such that treble is increased and bass is decreased for lower quality signals, but there is no additional spectral tilt for fullband podcasts.

[0055] Regardless of the quality of the signal, no amount of gain is required to be applied outside the audio bandwidth of the signal. It is generally desirable to avoid applying gain increases outside the bandwidth of the signal, since there is typically undesirable noise outside the band. Indeed, thresholding can be applied to frequencies outside the bandwidth, where no gain is applied or the signal is removed to reduce any unwanted sounds, such as hiss, that exist outside the audio bandwidth. In addition, such thresholding can be applied within the bandwidth of the signal to remove or avoid increasing the gain of any signal below a predetermined threshold selected for signals that are likely to be noise.

[0056] Because the intelligibility adjustment, as described in this disclosure, is designed for use in increasing the intelligibility of spoken word audio output, applying the intelligibility adjustment to music or other types of audio is generally undesirable as it may alter the output in a way that may degrade the quality of the music. Thus, the parameter mapping 104 may receive as a context parameter p4 an output from the content type detection 116, which detects whether the content signal u is a spoken word audio output or another type of audio output. In one example, the content type detection 116 may detect the presence of spoken word audio output using a suitable voice activity algorithm. In an alternative example, a method for detecting the presence of music may be used, such as detecting a spectral distribution that is characteristic of a musical output.

[0057] Thus, the parameter mapping 104 can reduce the intelligibility setting when music is detected. An example of this is shown in the plot depicted in Figure 10. When speech is detected, the intelligibility setting is reduced by the intelligibility input c inputThe intelligibility setting is set according to the desired intelligibility of the music. When music is detected, the intelligibility setting is reduced to 0 or some small value. In some examples, the value at which intelligibility is applied to the music audio output may be a user-defined parameter that applies an absolute intelligibility setting value, or it may be a reduction in the intelligibility setting (i.e., a reduction to an intelligibility setting determined by other parameters and intelligibility inputs). For example, a user may decide whether to apply a 100%, 50%, etc. reduction in the intelligibility setting when music is present.

[0058] It should also be appreciated that the intelligibility setting can be applied to safety warning signals such as blind spot warnings, seat belt warnings, proximity warnings, etc. In one example, these types of warning signals can be detected by the controller 108 via any suitable means or by a signal sent from a computer or system that generates the warning signal, and the controller 108 can apply the intelligibility setting to the warning signal. The intelligibility setting applied to the warning signal can be the same as that applied to the spoken word output, or can be applied according to an amount determined by the user (either as an absolute setting or a reduction to the intelligibility setting as otherwise applied to the spoken word output).

[0059] 11A shows a flowchart of a method 1100 for adjusting the intelligibility of an audio output in a dynamic environment. The method steps 1100 may be performed by a controller (such as controller 108) that includes one or more processors and any associated hardware. The method steps 1100 may also be stored on one or more non-transitory storage media and executed by the controller.

[0060] At step 1102, a content signal is received. The content signal may include one more gain adjustment applied by the amplifier, for example, according to a volume control and / or according to a noise compensation algorithm such as Bose AudioPilot. It should be appreciated that in the latter case, the gain adjustment may include multiple gain adjustments applied across different frequency bands.

[0061] In step 1104, a customized gain is applied to the content signal on a frequency bin basis such that the intelligibility of the resulting acoustic signal (e.g., as output in step 1106) is set to approximately a desired level according to at least one of the gain applied to the content signal, the bandwidth of the content signal, and the content type encoded by the content signal. The customized gain is selected to enhance quieter sounds (e.g., increase the level or volume for louder sounds) and / or reduce louder sounds (e.g., decrease the level or volume for quieter sounds). The desired level is determined by a default setting or by received user input.

[0062] In one example, the customized gain (i.e., intelligibility setting) may be implemented, at least in part, by dynamic range compression and spectral tilt. In this example, the intelligibility setting may be adjusted by changing parameters of the dynamic range compression (e.g., implemented thresholds, associated gain reduction, etc.) in a manner designed to improve the intelligibility of the spoken word output according to the desired intelligibility.

[0063] Although dynamic range compression that improves intelligibility for an average user can be determined theoretically or empirically, the resulting dynamic range compression does not improve intelligibility for all users. Certain users may experience hearing impairments in an anomalous way, and dynamic compression tailored to an average user does not address the way a given user perceives spoken word content. Thus, the intelligibility of the audio output can be further tailored to a particular user by applying equalization, such as spectral tilt, to the signal. Spectral tilt addresses the needs of different listeners who may have hearing impairments in low or high ranges by allowing the user to emphasize the signal at one end or the other. Although spectral tilt has been shown to provide good results, other equalizations can be used to improve the intelligibility of the audio output. For example, additional adjustments can increase high frequencies without reducing low frequencies, or increase low frequencies without reducing high frequencies, although other audio equalizations are contemplated.

[0064] As discussed above, the customized gain is applied according to one or more gain adjustments, e.g., as applied according to a volume control. Thus, as shown in Figure 11B, step 1104 may further include sub-step 1104B, in which the customized gain is adjusted according to a gain applied to the content signal according to a volume control. In one example, the gain adjustment applied to the content signal according to a volume control may be determined from a received parameter (e.g., a volume control input from a user) or by comparing a signal input to an amplifier and a signal output from the amplifier, which adjusts the gain according to a desired volume.

[0065] The customized gain can be adjusted through a predefined gain-intelligibility mapping (e.g., as described above in connection with FIG. 6). The points of the gain-intelligibility mapping can be predefined, e.g., as default settings designed to maintain the same perceived intelligibility as the volume increases through a user-defined set of points, or via a user-defined set of points. (To avoid setting each point individually, the user can define two or more points, and the remaining points can be determined by a suitable method of interpolation or extrapolation.) In either case, manipulation of the gain via the volume knob adjusts the intelligibility setting according to the mapping.

[0066] The customized gain further takes into account the gain adjustment performed according to the noise compensation algorithm. Thus, as shown in Fig. 11C, step 1104 may further include sub-step 1104C, in which the customized gain is adjusted according to the gain applied to the content signal according to the noise compensation algorithm. Similar to the volume gain adjustment, the gain adjustment applied to the content signal according to the noise compensation may be determined from received parameters (e.g., output from the noise compensation algorithm) or by comparing the signal input to the amplifier that adjusts the gain according to the noise compensation algorithm with the signal output from the amplifier.

[0067] Because the gain adjustments performed by noise compensation already take into account the increased noise in the vehicle cabin, and because the gain adjustments are often applied in a frequency-dependent manner, a separate intelligibility mapping (e.g., as described in connection with FIG. 7) that takes into account the nature of the gain adjustments, the increased noise in the vehicle cabin, and the gain adjustments already performed by the volume gain adjustments must be applied to adjust the intelligibility setting appropriately. As shown in FIG. 7, the additional mapping can be determined by equal intelligibility contours along which, for a given volume setting, the customized gain is adjusted such that the intelligibility remains approximately the same as road noise increases, and the noise compensation algorithm adjusts the gain accordingly.

[0068] It should further be understood that in addition to volume control and noise compensation gain adjustments, other gain adjustments can be made, for example, audio equalization can be implemented solely on the quality of the audio output (e.g., to adjust bass, midrange, or treble according to user preferences) and can be taken into account when adjusting the customized gain to reach a desired intelligibility.

[0069] The customized gain may further take into account the bandwidth of the content signal. Thus, as shown in FIG. 11D, step 1104 may include a subsubject 1104D in which the customized gain is adjusted according to the bandwidth of the content signal. The bandwidth of the signal may be determined according to any suitable method.

[0070] To adjust the customized gain so that the output acoustic signal approximates the desired intelligibility, a larger intelligibility adjustment (i.e., the intelligibility setting is set to a higher amount) can be made for low bandwidth signals, and a smaller intelligibility adjustment (i.e., the intelligibility setting is set to a lower amount) can be made for high bandwidth signals. As described in connection with FIG. 9, the intelligibility setting represented by line 902 is adjusted to be higher for lower bandwidth signals and linearly decreased until no additional intelligibility adjustment is made for very high quality signals, such as full-band podcasts. Similarly, the amount of spectral tilt represented by line 904 represents a linear adjustment such that treble is increased and bass is decreased for low quality signals, but no additional spectral tilt is made for full-band podcasts.

[0071] Regardless of the quality of the signal, no amount of gain is required to be applied outside the audio bandwidth of the signal. It is generally desirable to avoid applying gain increases outside the bandwidth of the signal, since there is typically undesirable noise outside the band. Indeed, thresholding can be applied to frequencies outside the bandwidth, where no gain is applied or the signal is removed to reduce any unwanted sounds, such as hiss, that exist outside the audio bandwidth. In addition, such thresholding can be applied within the bandwidth of the signal to remove or avoid increasing the gain of any signal below a predetermined threshold selected for signals that are likely to be noise.

[0072] Furthermore, the customized gain may be adjusted according to the type of content signal (e.g., spoken word audio signal or music). Thus, as shown in FIG. 11E, step 1104 may include sub-step 1104E of adjusting the customized gain according to the type of content signal. The content type may be determined according to any suitable method, for example, through a voice activity detection algorithm or through identifying spectral content typically associated with music.

[0073] When speech is detected, the intelligibility setting is set according to the desired intelligibility of the intelligibility input (related to the desired intelligibility of the spoken word output). When music is detected, the intelligibility setting is reduced to 0 or some small value. In some examples, the value at which customized gain is applied to the music audio output is an absolute intelligibility setting value or a user-defined parameter that applies a reduction in the intelligibility setting determined by other parameters. For example, a user can decide whether to apply a 100%, 50%, etc. reduction in the intelligibility setting when music is present.

[0074] In step 1106, the content signal with the customized gain is output to at least one speaker for conversion to an acoustic signal. This step may further include applying beamforming, routing to various speakers, etc. It should be understood that these processes applied to the signal after the customized gain has been applied typically do not substantially affect the intelligibility of the acoustic signal as heard by the user.

[0075] The functionality or portions thereof, and various modifications thereof (hereinafter "functionality") described herein may be implemented, at least in part, via a computer program product (e.g., a computer program tangibly embodied in an information carrier, such as one or more non-transitory machine-readable media or storage devices, for execution by or to control the operation of one or more data processing apparatus, e.g., a programmable processor, computer, multiple computers, and / or programmable logic components).

[0076] The computer program may be written in any form of programming language, including compiled or interpreted languages, and it may be arranged in any form, including as a stand-alone program or as a module, component, subroutine or other unit suitable for use in a computing environment. The computer program may be arranged to be executed on one computer or on multiple computers at one site, or may be distributed across multiple sites and interconnected by a network.

[0077] The operations associated with implementing all or a portion of the functionality may be performed by one or more programmable processors executing one or more computer programs to perform the functions of the calibration process. All or a portion of the functionality may be implemented as special purpose logic circuitry, such as an FPGA and / or an ASIC (Application Specific Integrated Circuit).

[0078] Processors suitable for the execution of a computer program include, by way of example, both general and special purpose microprocessors, and any one or more processors of any kind of digital computer. Generally, a processor will receive instructions and data from a read-only memory or a random access memory, or both. Elements of a computer include a processor for executing instructions and one or more memory devices for storing instructions and data.

[0079] While several embodiments of the invention have been described and illustrated herein, those skilled in the art will readily envision various other means and / or structures for performing the functions and / or obtaining the results and / or advantages of one or more of the advantages described herein, and each of such variations and / or modifications are deemed to be within the scope of the embodiments of the invention described herein. More generally, those skilled in the art will readily appreciate that all of the parameters, dimensions, materials, and configurations described herein are exemplary, and that the actual parameters, dimensions, materials, and / or configurations will depend on the specific application or applications for which the teachings of the invention are used. Those skilled in the art will recognize, or be able to ascertain using no more than routine experimentation, many equivalents to the specific embodiments of the invention described herein. Thus, it should be understood that the foregoing embodiments have been presented by way of example only, and that, within the scope of the appended claims and their equivalents, embodiments of the invention may be practiced otherwise than as specifically described and claimed. The inventive embodiments of the present disclosure relate to each individual feature, system, article, material, and / or method described herein. Furthermore, any combination of two or more such features, systems, articles, materials and / or methods, if such features, systems, articles, materials and / or methods are not mutually inconsistent, is included within the inventive scope of the present disclosure. [Explanation of symbols]

[0080] 100 Audio System 102 Clarity adjustment 104 Parameter Mapping 106 Speaker 108 Controller 110 Volume Control 112 Noise Compensation 114 Bandwidth Detection 116 Content Type Detection 202 Dynamic Range Compression 204 Spectral Tilt 402 curve 404 Curve 702 Contour Lines 704 Contour Lines 902 line 904 line 1100 methods 1102 Steps 1104 Steps 1104B Substep 1104C Substep 1104D Sub-subject 1104E Substep 1106 Steps

Claims

1. 1. A method for adjusting intelligibility of an audio output in a changing environment, comprising: Receiving a content signal; applying a customized gain to the content signal; outputting the content signal with the customized gain to at least one speaker for conversion to an acoustic signal, wherein the customized gain is applied for each frequency bin such that frequencies of lower amplitude are enhanced relative to frequencies of higher amplitude and the intelligibility of the acoustic signal is set to approximately a desired level, the customized gain being determined according to at least one of a gain applied to the content signal, a bandwidth of the content signal, and a content type encoded by the content signal.

2. The method of claim 1 , wherein the customized gain is based on dynamic range compression.

3. The method of claim 2 , wherein the customized gain is further based on equalization.

4. The method of claim 3 , wherein the equalization is a spectral tilt.

5. The method of claim 1 , wherein the gain applied to the content signal is set according to a volume control.

6. The method of claim 1 , wherein the audio output is generated within a vehicle cabin and the gain is set at least in part according to road noise within the vehicle cabin and according to a volume control.

7. The method of claim 1 , wherein the gains that are set according to the road noise in the vehicle cabin consist of a set of frequency dependent gain adjustments.

8. The method of claim 2 , wherein the dynamic range compression is performed only within the bandwidth of the signal.

9. The method of claim 1 , wherein the content type is one of musical output or spoken word output.

10. The method of claim 1 , wherein the content type is detected according to a voice activity detection algorithm.

11. 1. A non-transitory storage medium comprising program code that, when executed by a processor, adjusts intelligibility of an audio output in a changing environment, the program code comprising: receiving a content signal; applying a customized gain to the content signal; and outputting the content signal with the customized gain to at least one speaker for conversion to an acoustic signal, wherein the customized gain is applied per frequency bin such that frequencies of lower amplitude are enhanced relative to frequencies of higher amplitude and an intelligibility of the acoustic signal is set to approximately a desired level, and the customized gain is determined according to at least one of a gain applied to the content signal, a bandwidth of the content signal, and a content type encoded by the content signal.

12. 12. The non-transitory storage medium comprising the program code of claim 11, wherein the customized gain is based on dynamic range compression.

13. The non-transitory storage medium comprising the program code of claim 12 , wherein the customized gain is further based on equalization.

14. 14. The non-transitory storage medium comprising the program code of claim 13, wherein the equalization is a spectral tilt.

15. The non-transitory storage medium comprising program code as recited in claim 11 , wherein the gain applied to the content signal is set according to a volume control.

16. 12. The non-transitory storage medium comprising program code of claim 11, wherein the audio output is generated within a vehicle cabin and the gain is set at least in part according to road noise within the vehicle cabin and according to a volume control.

17. 12. The non-transitory storage medium comprising program code as recited in claim 11, wherein the gain set according to the road noise in the vehicle cabin consists of a set of frequency dependent gain adjustments.

18. 13. A non-transitory storage medium comprising program code as claimed in claim 12, wherein the dynamic range compression is performed only within the bandwidth of the signal.

19. 12. The non-transitory storage medium comprising program code of claim 11, wherein the content type is one of a musical output or a spoken word output.

20. The non-transitory storage medium comprising program code according to claim 11 , wherein the content type is detected according to a voice activity detection algorithm.

Images