Universal input scaling for systems with input sensors

Abstract

Systems and methods of universal input scaling perform and / or comprise receiving at least one input signal from at least one input source; ranking a signal parameter of the at least one input signal and generating a profile of the signal parameter; detecting a target signal in a full spectrum of the at least one input signal; forwarding the target signal to a signal processing unit; and processing the target signal based on the profile.

Description

UNIVERSAL INPUT SCALING FOR SYSTEMS WITH INPUT SENSORSCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of and priority to U.S. Provisional Patent Application No. 63 / 575,404, filed on April 5, 2024, titled “Universal Input Scaling Method for Systems with Input Sensors,’' the entire contents of which are each herein incorporated by reference for all purposes.STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

[0002] Not applicable.TECHNICAL FIELD

[0003] This disclosure relates to the field of input scaling for sensor-connected systems. In examples, this disclosure relates to universal input scaling for hearing aid systems.BACKGROUND

[0004] Humans have five different senses: hearing, sight, smell, taste, and touch. For each sense, different aspects of the body detect different physical changes in the environment and convert them into a sensation. In the engineering realm, sensors exist to quantify physical occurrences (e.g., sound level meters that measure sound pressure levels) or to convert one form of energy to another for further processing (e.g., microphones that detect external sounds for an amplification system.

[0005] Comparative techniques for sensing an occurrence or an event generally include using an instrument with know n sensitivity to quantify the magnitude of the occurrence / event. For example, the sensitivity of a microphone of an audio system is defined by a sound’s acoustic energy (e.g., in Pascals (Pa) or micro-Pascals (pPa)) as received by the microphone and the electrical voltage (e.g., in volts (V) or millivolts (mV)) that the microphone generates in response to external sounds. Thus, the unit of microphone sensitivity is mV / Pa. The electrical signal with its sound pressure level expressed in mV is then sent to other parts of the audiosystem for documentation or for further processing. The system that receives the signal from the microphone (e.g., ahearing aid, ahearing protection device, etc.) often provides differential processing for sounds at different levels. If the microphone sensitivity is unknown to the system, however, the level of the incoming sounds is also unknown and further processing would be impossible or inappropriate.

[0006] Accordingly, there exists a need for universal scaling systems and methods that can scale the input of any sensor without knowing the sensitivity of the input system, and that do not miss desired, important, or pertinent information.SUMMARY

[0007] The present disclosure addresses these and other needs by providing systems, devices, methods, algorithms, and / or media for universal scaling. In examples, the techniques set forth herein rank the output of sensors, generate a signal parameter (e.g., strength, magnitude, quantity7, etc.) profile, send forth the signals within a defined signal parameter range to a main device or algorithm for further processing. In some examples, detectors are also utilized to ensure desired, important, or pertinent information is not missed. The techniques set forth herein provide several advantages over the comparative techniques, including but not limited to allowing the main device or algorithm to perform parameter-dependent signal processing without knowing the sensor sensitivity. While several of the examples set forth herein are directed to audio processing as an analogue to the sense of hearing, it should be understood that the techniques described herein may be applied to develop devices with input signal scaling factors based on the rankings of the signal parameter (as opposed to absolute sensitivity as in the comparative examples) to realize processing as an analogue to other senses or for systems with different functions.

[0008] According to one aspect of the present disclosure, an input scaling method is provided. The method comprises receiving at least one input signal from at least one input source; ranking a signal parameter of the at least one input signal and generating a profile of the signal parameter; detecting a target signal in a full spectrum of the at least one input signal; forwarding the target signal to a signal processing unit; and processing the target signal based on the profile.

[0009] According to another aspect of the present disclosure, an input scaling system is provided. The system comprises at least one input source configured to receive at least one input signal from a sensor; a memory; and at least one processor configured to: rank a signal parameter of the at least one input signal and generate a profile of the signal parameter, detect a target signal from within a full spectrum of the at least one input signal, forward the target signal to a signal processing unit, and process the target signal based on the profile.BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Some examples of the disclosure are described herein with reference to the accompanying figures. The description, together with the figures, makes apparent to a person having ordinary skill in the art how some implementations of the disclosure may be practiced. The figures are for the purpose of illustrative discussion and no attempt is made to show structural details of an example in more detail than is necessary for a fundamental understanding of the teachings of the disclosures. In the drawings:

[0011] FIG. 1 illustrates an example input-output function for a comparative system.

[0012] FIG. 2 A illustrates an example input-output function for a system according to various aspects of the present disclosure.

[0013] FIG. 2B illustrates an example input-output function for a system according to various aspects of the present disclosure.

[0014] FIG. 2C illustrates an example input-output function for a system according to various aspects of the present disclosure.

[0015] FIG. 3 illustrates an example signal processor having multiple inputs according to various aspects of the present disclosure.

[0016] FIG. 4 illustrates an example scaling implementation according to various aspects of the present disclosure.

[0017] FIG. 5 illustrates an example scaling implementation according to various aspects of the present disclosure.

[0018] FIG. 6 illustrates an example scaling implementation according to various aspects of the present disclosure.

[0019] FIG. 7 illustrates an example scaling implementation according to various aspects of the present disclosure.

[0020] FIG. 8 illustrates an example method according to various aspects of the present disclosure.DETAILED DESCRIPTION

[0021] In the following detailed description, reference is made to the accompanying drawings in which specific examples are shown by way of illustration. These examples are described in sufficient detail to enable those of ordinary skill in the art to practice the disclosure. It should be understood, however, that the detailed description and the specific examples, while indicating examples of embodiments of the disclosure, are given by way of illustration only and not by way of limitation. From this disclosure, various substitutions, modifications, additions rearrangements, or combinations thereof within the scope of the disclosure may be made and will become apparent to those of ordinary' skill in the art.

[0022] For example, while the following description provides examples related to hearing aids (e.g., universal input scaling for systems with audio sensors), the principles set forth herein may be applied to other human senses and other functions thereby to scale sensor input without requiring knowledge of the sensor sensitivity.

[0023] Unless otherwise indicated, the various features illustrated in the drawings may not be drawn to scale. The illustrations presented herein are not necessarily intended to be actual views of any particular method, device, or system, but are merely idealized representations that are employed to describe various embodiments of the disclosure. Accordingly, the dimensions of the various features as illustrated may be arbitrarily expanded or reduced for clarity. In addition, some of the drawings may be simplified for clarity7. Thus, the drawings may not depict all of the components of a given apparatus (e.g., device) or method. In addition, like reference numerals may be used to denote like features throughout the specification and figures.

[0024] It should be understood that any reference to an element herein using a designation such as “first,” “second,” and so forth does not limit the quantity or order of those elements, unless such limitation is explicitly stated. Rather, these designations may be usedherein as a convenient method of distinguishing between two or more elements or instances of an element. Thus, a reference to first and second elements does not mean that only two elements may be employed there or that the first element must precede the second element in some manner. Also, unless stated otherwise a set of elements may comprise one or more elements.

[0025] Unless otherwise specified or indicated by context, the terms “a,” “an,” and “the” mean “one or more.” As used herein, unless otherwise limited or defined, “or” indicates a non-exclusive list of components or operations that can be present in any variety of combinations, rather than an exclusive list of components that can be present only as alternatives to each other. For example, a list of “A, B, or C” indicates options of: A; B; C; A and B; A and C; B and C; and A, B, and C. Correspondingly, the term “or” as used herein is intended to indicate exclusive alternatives only when preceded by terms of exclusivity, such as “only one of,” or “exactly one of.” For example, a list of “only one of A. B, or C” indicates options of: A, but not B and C; B, but not A and C; and C, but not A and B. In contrast, a list preceded by “one or more” (and variations thereon) and including “or” to separate listed elements indicates options of one or more of any or all of the listed elements. For example, the phrases “one or more of A, B. or C” and “at least one of A, B. or C” indicate options of: one or more A; one or more B; one or more C: one or more A and one or more B; one or more B and one or more C; one or more A and one or more C; and one or more A, one or more B, and one or more C. Similarly, a list preceded by “a plurality of’ (and variations thereon) and including “or” to separate listed elements indicates options of one or more of each of multiple of the listed elements. For example, the phrases “a plurality of A, B. or C” and “two or more of A, B, or C” indicate options of: one or more A and one or more B; one or more B and one or more C; one or more A and one or more C; and one or more A, one or more B, and one or more C.

[0026] As used herein, the terms “include” and “including” have the same meaning as the terms “comprise” and “comprising” in that these latter terms are “open” transitional terms that do not limit claims only to the recited elements succeeding these transitional terms. The term “consisting of,” while encompassed by the term “comprising,” should be interpreted as a “closed” transitional term that limits claims only to the recited elements succeeding this transitional term. The term “consisting essentially of,” while encompassed by the term “comprising,” should be interpreted as a “partially closed” transitional term which permitsadditional elements succeeding this transitional term, but only if those additional elements do not materially affect the basic and novel characteristics of the claim.

[0027] In some examples, the systems and methods set forth herein may be implemented on or using one or more computing devices, each of which includes a processor and a memory. As used herein, a “processor” may include one or more individual electronic processors, each of which may include one or more processing cores, and / or one or more programmable hardware elements. The processor may be or include any type of electronic processing device, including but not limited to central processing units (CPUs), graphics processing units (GPUs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), microcontrollers, digital signal processors (DSPs), or other devices capable of executing software instructions. When a device is referred to as “including a processor,” one or all of the individual electronic processors may be external to the device (e.g.. to implement cloud or distributed computing). In implementations where a device has multiple processors and / or multiple processing cores, individual operations described herein may be performed by any one or more of the microprocessors or processing cores, in series or parallel, in any combination. In some implementations, one or more of the processing units or processing cores may be remote (e.g., cloud-based).

[0028] As used herein, a “memory” may be any storage medium, including a nonvolatile medium, e.g.. a magnetic media or hard disk, optical storage, or flash memory; a volatile medium, such as system memory, e.g., random access memory (RAM) such as dynamic RAM (DRAM), synchronous dynamic RAM (SDRAM), static RAM (SRAM), extended data out (EDO) DRAM, extreme data rate dynamic (XDR) RAM, double data rate (DDR) SDRAM, etc.; on-chip memory; and / or an installation medium where appropriate, such as software media, e.g., a CD-ROM. or floppy disks, on which programs may be stored and / or data communications may be buffered. The term “memory” may also include other types of memory or combinations thereof. For the avoidance of doubt, cloud storage is contemplated in the definition of memory. A memory is an example of a non-transitory computer-readable medium which stores instructions that are executable by a processor (or processors), the execution of which causes the executing device (e.g., a computer) to perform certain operations, such as those operations described herein.

[0029] FIG. 1 illustrates an example input-output function used in a comparative example of a hearing aid, and uses the absolute sound pressure level (SPL) in decibels (dB). The auditory dynamic range is defined by the individual user's hearing threshold (lower limit) and uncomfortable loudness threshold (upper limit). The input-output function shows the relationship between the input SPL and the hearing aid output SPL. The gain or amplification provided by the hearing aid can be calculated as the output minus the input. In the example shown in FIG. 1, sounds with very low input levels (e.g., <20 dB SPL) are processed using expansion, which provides a smaller amount of gain for lower-level sounds; sounds with low input levels (e.g., 20-45 dB SPL) are processed using linear amplification, which provides the same amount of gain for sounds within input levels falling in this range; sounds with low to high level inputs (e.g., 45-85 dB SPL) are processed with compression, which provides less gain for higher level sounds; and sounds with high to very high level inputs (e.g., >85 dB SPL) are processed using output limiting strategies, which limit the output levels to be below the individual user’s uncomfortable loudness threshold level to prevent discomfort, or below the maximum power output of the receiver / speaker of the hearing aid to prevent distortions due to peak-clipping.

[0030] However, the comparative example requires knowledge of the sensitivity of the hearing aid in order to determine the input SPL level. Without such knowledge, the comparative example will not be able to determine the processing regime to apply in order to generate the output signal. For example, if the comparative example does not know the sensitivity of the hearing aid, it will not know whether a particular input signal represents an input SPL of <45 dB (such that linear amplification should be applied) or >45dB (such that compression amplification should be applied).

[0031] The present disclosure sets forth a universal technique for sensors that can detect the magnitude of the incoming signal and rank the magnitude according to its percentile of all sensed magnitudes. A subsequent (e.g., downstream) system may then process the signal based on the percentile of the sensed magnitudes instead of relying on the known sensitivity of the sensors. In certain applications, the magnitudes falling at different percentile ranges can be sent forward to a main signal processing unit for further processing, and the magnitude at other percentile ranges can be deleted. In such cases, to ensure that relevant (i.e., desired, important, or pertinent) signals are not inadvertently deleted from further processing, additional signalanalyzers or detectors can be implemented to monitor the presence of the relevant signals. If those signals are in fact present within the range that was not forwarded to the main signal processing unit, exceptions can be made such that those signals are forwarded for further processing to ensure they are represented in the final (output) signal. Such systems provide several advantages, including but not limited to that the systems can accept multiple and different types of sensor inputs with unknown sensitivity'. The rankings of magnitude percentile can also provide differential signal processing.

[0032] FIGS. 2A-2C illustrate example input-output functions used in an example of a hearing aid according to various aspects of the present disclosure, and use the relative SPL as a percentile rank if the sensitivity of the hearing aid is unknown. The thick arrow s in FIGS. 2A-2C illustrate the manner in which a volume control can change the output level (i.e., the amount of amplification). The dot-dash line show s that the maximum output level would not change when the volume control is adjusted, but the starting point of the output limiting region would change. The thin arrow at the left and right edge of the graph show the voltage percentile that can be varied by the sound scene detectors. In the example of FIG. 2A, sounds that generate voltages that are below a predetermined threshold percentile (in the illustrated example, the 20th percentile) of the daily environmental sounds are not brought above the user’s auditory threshold (i.e., not amplified to fall within the auditory dynamic range) because such low -level sounds are likely to be circuit noise or ambient noise and contain limited linguistic or other usable information. For sounds that generate voltages within a predetermined threshold percentile range (in the illustrated example, the 20th to 45th percentiles), the sounds receive the same amount of gain. For sounds in a predetermined threshold range above this (in the illustrated example, the 45th to 85th percentiles), the rate of gain decreases as the voltage percentile increases. Finally, for sounds above a predetermined threshold (in the illustrated example, the 85th percentile), the rate of gain reduction increases as the voltage percentiles increase.

[0033] In the example of FIG. 2B, the input-output function curve is the same as in the example of FIG. 2A, but the output dynamic range is expanded beyond the base auditory dynamic range. Thus, the example of FIG. 2B may be used to expand the range of sounds sent forward for listeners (e.g., those with normal hearing). Thus, speech that is above the threshold level of the expanded dynamic range but below the threshold of the base dynamic range ismade audible. The example of FIG. 2B may be used if speech is present at a very low level and there is a desire to render it audible for some amount of time. In the example of FIG. 2C, the output dynamic range is unchanged from the example of FIG. 2A, but the signal processing (e.g., compression) is applied to make both soft speech and sounds in the amplification range audible. The example of FIG. 2C may be implemented for listeners with hearing loss.

[0034] The examples of FIG. 2B or 2C may be usable with the example of FIG. 2A. FIG. 2A has the advantage of reducing the background noise in the environment. FIG. 2B or 2C have the advantage of bringing low-level speech into the audible range of the user, although they may temporarily increase the background noise level. Other noise reduction algorithms may be used in conjunction of the described input scaling technique to reduce the background noise before presenting the low-level speech into the audible range of the user.

[0035] Because speech is essential for hearing aid users, the system can also implement additional sound scene sensors to detect desired or important characteristics using, for example, a speech model that defines speech characteristics. When signals with speech characteristics are detected at a level that would not otherwise be forwarded for further processing (e.g., at voltage levels below the 20th percentile), the signals may be sent for further processing.

[0036] The techniques set forth herein are applicable to a wide range of implementations, one example implementation includes the development of a personalized hearing amplification application, which may be provided for smartphones running an operating system such as Android or iOS. For Android implementations in particular, the smartphones may have a wide range of specifications defined by many different smartphone manufacturers. As a result, an Android-based application will likely not have general knowledge of the SPL of the incoming sounds because different smartphones utilize microphones with different sensitivity. This lack of knowledge regarding the absolute SPL would, if comparative techniques were used, render subsequent signal processing difficult or impossible. Additionally, earbuds usable with smartphones have diverse specifications. In view of these diverse specifications and in view of the very large number of possible combinations of smartphones and earbuds, the receiver output scaling is unknown to application developers when the application is installed. Thus, implementations of comparative techniques do not provide amplification based on an individual user's hearing sensitivity profile. Instead, theyoften provide user-adjustable volume controls either for the whole input signal or different frequency regions, requiring the user to self-adjust the amplification provided.

[0037] Modem hearing aids are often implemented with speech, noise, and music detectors. Speech often contains two different types of modulations: slow modulation generated by the movement of the articulators of the vocal track, with the highest modulation rates between 3-10 Hertz (Hz); and fast modulation generated by the movements of the vocal folds, the rate of which depends on the fundamental frequencies of the speaker and the intonation of the utterances. Speech also exhibits alternating sounds with longer and shorter durations, sounds with lower and higher frequencies, and sounds with higher and lower intensities (e.g., vowels and consonants, respectively). By comparing the characteristics of incoming sounds with the known speech characteristics, the hearing aid device can infer whether speech is present in the incoming signal in different frequency channels. In some examples, the hearing aid device may be implemented with deep neural network or artificial intelligence or algorithms that would classify or identify' incoming sounds as speech, noise, music or other sounds. These implementation can also be used to detect different ty pes of desired and undesired sounds in the environment to be included or excluded in further signal processing.

[0038] According to the present disclosure, various detectors may be implemented to determine whether speech is present in the incoming signal. If speech is present, it is sent forward to the hearing amplification system for further processing. The detectors act as failsafes to ensure that the audibility7of speech is not compromised if speech fails below or outside of the percentile ranges that are predetermined to be sent for further processing. Compared to comparative techniques, the expansion algorithms used herein provide a lower amount of gain for lower-level sounds (i. e. , for sounds with SPLs lower than the expansion threshold) and the expansion algorithms may reduce the low-level sounds that do not provide a significant amount of useful information or linguistic content. Such sounds include circuit noise, microphone noise, or the pumping effect due to compression algorithms.

[0039] In the present techniques, noise reduction results from not sending signals that are below a certain percentile forward for further processing. For example, in the situation illustrated in FIG. 2A, the noise reduction effect of the present disclosure is realized because sounds generating microphone voltages lower than the 20th percentile are not sent forward forfurther signal processing. The present disclosure further improves upon comparative techniques (such as modulation-based noise reduction algorithms), at least in that the present disclosure provides techniques that reduce the gain of frequency channels with low speech contents (e.g., low estimated signal-to-noise ratios (SNRs)) and let the signals with higher speech contents (e.g., high estimated SNR) be processed.

[0040] A universal input scaling system and method is disclosed herein that can scale the input of any sensor without the need to know the sensitivity of the input system and that, at the same time, do not miss desired, important, and pertinent information. Comparative methods of defining the input to a system rely on measuring the sensor sensitivity that would convert a certain external physical quantity into another physical quantity . For example, the microphone sensitivity for an audio system is often defined as the amount of electrical voltage in millivolt generated per sound pressure level in Pa (i.e., mV / Pa).

[0041] In this disclosure, the input scaling of a system is defined by the percentile of the incoming signal and an additional analysis module is implemented to assure any desired / important / pertinent information is not missed. This ensures pertinent and useful signals are sent forward to the main processing unit for further processing. The sensor picks up the signal, ranks the strength of the signal in percentile terms, and then constructs a profile for the range of the signal strength. The generation of the profile can be accomplished by digital signal processing algorithms or learning algorithms by’ artificial intelligence (Al) before or after the sensor is implemented to the main processing unit. Depending on the purpose of the device / application, signals falling into different percentile range may be deleted without further processing or sent forward for further processing. The additional analysis module can determine whether an important signal is detected outside of the desired range of percentiles. If so. those signals will also be sent forward for further processing. Additionally, the percentile range of sounds / signals that are sent for further processing can change based on the range of sounds that are present at the time or based on the full range of sounds detected (i.e., a floating paradigm). For example, incoming sounds may be ranked based on the range of sounds in the environment day and night (i.e., 0-100th percentile). In one example, the lowest sounds are 10th percentile and the highest are 70th percentile of the overall range of sounds. The sounds that are sent for further processing can be from 15th-70th percentile. In another example, the range of sounds can be 30th-90th percentile of the full range. The sounds that are sent forfurther processing can be from 35th-90th percentile. The duration of time for the selection of range to send forward for further processing can be infinitely short or infinitely long. The advantages of being able to choose the signal range to be presented to the user are that 1) background noise can be reduced in normal operational mode (FIG.2A); 2) if the sound range presented to the user is from 35th- 90thpercentile (e.g., during the day time in a noisy environment when the overall sound levels are higher), low-level speech would be brought to the audible range of the user (FIG. 2B and 2C); 3) if the sound range presented to the user is from 15th- 70thpercentile (e.g., during the night time in a quite environment when the overall sound levels are lower), high level speech would still be presented to the user ; 4) Not presenting the full range of sounds to the user allows more flexible signal processing. For example, instead of compressing the whole input range of sounds into the audible range of the user, presenting an 80% range of the sounds would allow the user to hear the processed sounds with less compression. This is important because highly compressed sounds are often associated with lower speech understanding scores.

[0042] A sample implementation can be a hearing amplification system that has one or multiple microphones or wired / wireless inputs. FIG. 3 illustrates one example implementation where multiple inputs 302 are provided to a main processing device 304. The voltage output generated by microphones in response to external sounds or the voltages of the wired / wireless inputs (with unknown sensitivity) are logged to generate a voltage percentile profile or profiles in different frequency regions. The user input can be a program implemented by the designer of the hearing amplification system or the individual user of the app gives instructions on signals failing within a certain percentile range (e.g., 20th - 100th percentile) to the amplification system for further processing. Additional sound scene detectors in the analysis module monitor the presence of speech signals in the incoming sounds using a speech model with known characteristics of speech. If speech signals are detected at voltages <20th percentile, those signals are also sent forward to the amplification system for further processing.

[0043] In FIG. 3, the inputs 302 include a plurality of microphone inputs, a wireless input, and a streaming (e g., from the Internet) input. However, this is merely by way of example and not limitation. In implementations, any number of inputs 302 may be present, each configured to receive at least one input signal from a sensor (e.g., a sensor that generatesan electronic signal in response to a detected physical phenomenon such as a microphone, a physical measurement parameter, etc.). Where multiple inputs 302 are present, the main processing device 304 may be configured to receive (and operate on) the multiple inputs 302 simultaneously. The physical measurement parameter, where implemented, may be a time of day, a day of the week, a season, an overall frequency spectrum, an identification of an event in one or more frequency regions, a directionality of an input signal, combinations thereof, and the like.

[0044] The main processing device 304 includes a memory 306 and a processor 308. The processor 308 is configured to perform a series of operations, for example by loading or executing instructions stored in the memory 306. These operations include, but are not limited to, ranking a signal parameter of the signal input(s) and generating a profile of the signal parameter; detecting a target signal from within a full spectrum of the signal(s). forwarding the target signal to a signal processing unit (which may be the same as or separate from the processor 308); and processing the target signal based on the profile. The parameter may be a signal strength, a signal magnitude, a signal quantity, or combinations thereof. In examples, as noted above, the signal is processed based on the profile and not based on a value of the signal parameter itself. The main processing device 304 may be configured to process the target signal using a differential processing operation. In some examples, the main processing device 304 may be configured to invoke one or more Al models, for example by providing the full spectrum of the input signal(s) to the Al model and receiving the target signal from the Al model. In such examples, the Al model may be local to the main processing device 304, may be remote (e.g., cloud-based), or both.

[0045] In another sample implementation, a hearing protection system (e.g., for hunters) is provided to send forward sounds from Oth - 90th percentiles for further processing so that the users can have environmental awareness; and to not present sounds with voltages for >90th percentile. Hearing protection systems are generally inserted in the ear canals of their users (e.g.. earplugs) or worn over the ears of the users (e g., headphones), or both. They provide 30-40 dB of passive attenuation to environmental sounds when the electronic components inside the earplugs or the headphones are shut off. As sounds such as gunshot sounds are loud and can damage hearing, the hearing protection system using the input scaling method set forth herein can shut off the electronic circuits when the microphone receivessounds at the >90th percentile so that the sound is heard after the earplug / headphone's passive attenuation.

[0046] Although the examples are described using percentile rankings, the disclosed input scaling technique generally uses rankings of quantity / magnitude of physical attributes / characteristics / events / occurrences, instead of the absolute / standard measurements of the quantity / magnitude of the attribute / characteristics / events / occurrences. Thus, ranking methodologies other than percentiles may be used. Such ranking can be based on the rankings of a quantity (e.g., voltages) or based on the number of occurrences (e.g., frequency).

[0047] FIGS. 4-6 illustrates various implementations of the present disclosure, including sensors, an analysis module (which may include an option for user input), and a main processing unit. The sensors illustrated in FIGS. 4-6 may be examples of the input sources 302 of FIG. 3, and the analysis module and main processing units illustrated in FIGS. 4-6 may be examples of the main processing device 304 of FIG. 3. Whether the input scaling set forth herein is implemented in an analog or digital system, an analysis module may be provided. If the sensor output is analog, the analysis module can be implemented immediately after the microphone / wired / wireless inputs (as in FIG. 4) or after the signal is converted to the digital format via analog-to-digital conversation (as in FIG. 5). The analysis module can also be implemented immediately after the microphone / wired / wireless inputs and then followed by the analog-to-digital conversion, which can happen at any point in the main device signal processing pathway (e.g., as in FIG. 6). Similarly, if the sensor produces digital output and analog signal processing is desired, the analysis nodule can be implemented before or after digital-to-analog conversion.

[0048] Generally, the implementations shown in FIGS. 4-6 include one or more sensors(e.g., a microphone with or without pre-amplifier and other wired or wireless inputs), an analysis module, and a main signal processing unit hosted in the main device (e.g., an amplification system or a hearing protection system) that implements subsequent signal processing algorithms. The analysis module includes an overall level detector or a set of frequency -specific level detectors to create one or more voltage profiles for the incoming signals, a set of sound scene detectors to implement models of different essential functions (e.g., speech, noise, music, location detection), and a user interface to permit the user to define the processing parameters.

[0049] FIG. 7 shows an implementation of the present disclosure in the visual modality. Instead of level detectors, the system has one or an array of image sensors. The user interface can be, for example, a safety monitoring app in a factory in which a camera including the image sensor automatically makes recordings of the space. The signal detector could be any camera (e.g., a commercially available device) and for which the user does not know the photosensitivity. The app can make recordings at a light range of, for example, 25th-100th percentile. If the recording light sensitivity is adjusted in one manner, the image in high light settings would white out as a result of pixel saturation. If the sensitivity is adjusted to be lower, the image for low light settings would be darkened and not detectable as a result of low pixel output signals. Using the disclosed input scaling technique, the app can adjust the recording sensitivity of the system to record images below 25th percentile of lighting when motion is detected and return to the 25th- 100th percentile setting after the motion ceases. Alternatively, it can be set to record 0-70th percentile of lighting at night and record 30th-100th percentile of lighting during the day. Although a variety of paradigms can be used to set the recording sensitivity, the system of FIG. 7 permits the user to monitor the space with clear imagery without knowing the photo-sensitivity of the camera. In examples, the visual modality implementation of FIG. 7 may implement the input-output curve from any of FIGS. 2A-2C.

[0050] The systems and methods set forth herein may rank input signals, a set of detectors for detecting desired / important / relevant characteristics (which may be omitted in some implementations), a set of instructions, and a main device system that utilizes the ranked inputs to further process the signals.

[0051] The ranked input signals can be audio, haptic, sight, taste, or smell, or can also be other physical phenomena / occurrences / events that are measured for algorithms / devices to perform different functions. The input signals can be ranked based on their physical properties (e.g., voltages generated by a microphone) or the frequency of occurrences ( e.g., number of times certain voltages are generated by a microphone). Multiple input signals can be ranked at the same time (e.g., voltages generated by two microphones, sounds coming from multiple directions, voltages generated by a microphone at different frequency regions). The rankings of the signal strength / magnitude / quantity can be before or after transformation of the signal strength / magnitude / quantity using standard or custom formula (e.g., to the dB scale, absolute temperature scale in Kelvin, log scale, base 10 log scale, natural log scale, etc.).

[0052] The set of detectors are implemented to detect the characteristics of the incoming signals that are desired / important / relevant to the design purposes of the main devices. They can sen e as fail-safe measures as well as auxiliary information to enhance the performance of the main device / algorithm. The type of the detectors implemented depends on the design purpose of the main device / algorithm.

[0053] The set of instructions provided through the user interface can be predetermined or modified in an ongoing manner. Artificial intelligence can be employed in the pre-determination and the ongoing modification processes.

[0054] The following describes systems in the audio and auditory domain as sample embodiments of the disclosed applications. It should be understood that similar principles can be applied to develop devices with input signal scaling factors based on the rankings of the signal strenglh / magnitude / quantity instead of absolute sensitivity of the sensors for different human senses, e.g., touch, or for systems with different functions, e.g., magnetism.

[0055] In comparative audio / electrical / electronic systems, input transducers are utilized to convert one form of energy in a certain physical quantity to a different form of energy in a different physical quantity7. For example, the sensitivity of a microphone of an amplification system is defined by a sound's acoustic energy in SPL (e.g., in Pa) that the microphone receives and the electrical voltage in millivolt or volt that the microphone generates in response to the sound. Thus, the unit of microphone sensitivity is mV / Pa. The electrical signal with its sound pressure level expressed in mV is then sent to other parts of the amplification system for further processing. The comparative systems exhibit several disadvantages; for example, the amplification system must know the microphone sensitivity in order to properly process the audio signals to be suitable for the user of the amplification system, and the initial signal processing is purely based on the voltages generated by the microphone without the consideration of what is important for the user.

[0056] In the universal input scaling technique set forth herein, the output of the sensors for sounds or other human senses, or sensors for other applications are ranked in percentile or other counting methods. In one example implementation of the disclosed input scaling method in an amplification system, the system includes an analysis module to conduct at least two types of surveillance on the output(s) of one or multiple microphone(s): level detection, to measure and log the voltages of the input sound signals that the microphone generates and tocompile a voltage percentile profile; and sound scene detection, to monitor the characteristics of the incoming audio signals at all percentiles for speech and speech from different directions;

[0057] The input scaling system can make decisions based on either the individual results of the level detector and the sound scene detector, or in a combination of the two or other factors. In this example of the amplification system, instead of classifying sounds at different sound pressure levels, the sounds that generate voltages in a certain percentile range (e.g., 20th - 100th percentile) are sent for further processing and then presented above the hearing thresholds at different frequencies or within the auditory dynamic ranges of the user of the system. In one particular example for such applications, voltages <20th percentile are excluded because they are most likely to be low-level microphone noise, low-level environmental sounds that do not convey important information.

[0058] Such a rationale may be applied to hearing aid processing such that sounds below a certain sound pressure levels are either not provided any gain or processed with progressively lower gain using expansion algorithms because they are mostly circuit noises or sounds with limited linguistic contents or useful information. To prevent the user of the amplification system from missing important information, a sound scene detector can be employed to detect the characteristics of incoming sounds at all voltage percentiles. In case speech-like sounds are detected at voltages below the 20th percentile, those sounds will be sent forward and amplified to be above the user's hearing thresholds or within the auditory dynamic range of the user.

[0059] The techniques set forth herein allow the amplification system to take the input from microphones without knowing their sensitivity, or take different microphones with different sensitivity. The techniques also reduce low-level background noises, while ensuring that desired / important information is not missed.

[0060] Similar principles can be applied to systems with other functions. For example, in a hearing protection system / device (e.g., for hunters), it is important for hunters to hear low- level sounds generated by game and, at the same time, to be protected from loud gunshot sounds. In one example for such an application, the desirable sounds can be defined as those in the Oth - 90th percentile range which are sent forth to the hearing protection system / device for further processing. The processing provided by the hearing protection system for hunters can amplify sounds between Oth - 20th percentile to increase the low-level sound awareness,and can refrain from presenting the level of sounds above 90th percentile to protect the hearing of the hunter. For sounds above 90th percentile, the hearing protector devices that are worn in the hunter's ears or that are covering the hunter's ears become passive attenuators (i. e. , no audio signals flow through the electrical pathway of the hearing protection devices). As many hearing protection devices can only provide a maximum of 30-40 dB of attenuation when they are inserted in the ear canal or covering the ear / pinna, the hunter can still be able to hear the gunshot sounds at 30-40 dB lower sound pressure level than the original gunshot sound.

[0061] Similar principles can be applied to systems with other types of sensors. For example, in a safety monitoring system, the camera’s photo-sensitivity is unknown to a system but it is important to capture clear images (i.e., without white out or black out). In one example for such an application, the described input scaling technique can adjust the recording sensitivity of the system to a light range of 30th-100thpercentile and monitor movements below 30thpercentile of light. When movements are detected, the system automatically adjusts the light range it records (e.g., 0th-70thpercentile range) to make sure the images are recorded clearly. Such change in the recording light range can be sustained as long as the movement in low-light setting lasts or until the movement ceases. In one example for such application, the system stays at this new recording sensitivity range and monitors movement at higher light range (i.e., >70thpercentile). Using the disclosed input scaling technique, the app can adjust the recording light sensitivity to capture light images from 30th-100thpercentile if movements at higher light range occur (e.g.. a highly light reflective surface). Although a variety of paradigms can be used to set the recording sensitivity, the system of FIG. 7 permits the user to monitor the space with clear imagery without knowing the photo-sensitivity of the camera.

[0062] These three examples show different implementations of the input scaling method, which can provide input specific information and the subsequent main device / algorithm can further process the signal. The level detector(s) can detect the voltage generated by the microphone for the whole signal spectrum or in different frequency channels. The percentile of sounds sent for ard to the system for further processing will depend on the purpose of the system.

[0063] FIG. 8 illustrates an example of an input scaling method 800 in accordance with various aspects of the present disclosure. The method 800 may be performed by the systems illustrated in FIGS. 3-7 and described above. The method 800 includes an operation 802 ofreceiving at least one input signal from at least one input source. The at least one input source may include a sensor configured to generate an electrical signal in response to a physical occurrence (e.g., a sound). For example, the input source may be ahearing aid. Where operation 802 includes receiving multiple input signals from multiple input sources, the input signals may be received simultaneously. Operation 802 may include receiving the input signal(s) as a wireless transmission from the input source(s). In such examples, the input source(s) may be local (e.g., within a Wi-Fi, Bluetooth, near-field communication (NFC) range, etc.), or may be remote (e.g., connected via the Internet). In other examples, the input signal(s) may be received via a wired connection.

[0064] Operation 804 includes ranking a signal parameter of at least one input signal and generating a profile of the signal parameter. The signal parameter may be a signal strength, a signal magnitude, a signal quantity, or combinations thereof. In some example, operation 804 may include ranking components of the signal parameter or ranking a transformation of the components of the signal parameter. The ranking may include ordering each component (or its transformation) according to the ranking, for example by sorting it into its corresponding percentile rank with regard to the signal parameter.

[0065] Operation 806 includes detecting a target signal in a full spectrum of the at least one input signal. In examples, operation 806 includes identifying a purpose of the signal processing unit (e.g.. that the signal processing unit is a hearing aid), selecting a signal range from within the full spectrum based on the purpose, and setting components of the full spectrum that are within the signal range as the target signal. Operation 806 may be performed by an Al model, as noted above.

[0066] At operation 808, the target signal is forwarded to a signal processing unit and at operation 810, the target signal is processed based on the profile (i.e., the profile generated at operation 804). Operation 808 may include analyzing additional components of the full spectrum of the input signal that are outside of the signal range and, if it is determined that the additional components include a candidate target signal, forwarding the candidate target signal to the signal processing unit. This additional analysis may be performed to identify’ relevant signals and ensure that they are not inadvertently deleted and / or to ensure that they are included in the processed signal output at operation 810.

[0067] While FIG. 8 illustrates the operations 802-810 being performed in a particular order, the present disclosure is not limited to the ordering implied by FIG. 8. In some implementations, certain operations may be performed in a different order than the illustrated order, or certain sets of operations may be performed in parallel. For example, operation 802 may be performed continuously to monitor the input signal, and operations 804-810 may be performed intermittently (e.g., at predetermined intervals) such that operation 802 is performed in parallel with the remaining operations. Additionally or alternatively, in some implementations operation 806 and operation 808 may be performed in parallel, such that the signal (or a range thereof) is sent forward for processing while the target signal is detected in the full signal spectrum. In such implementations, the method 800 may further include an operation of adjusting the range of the signal sent forth for processing (e.g., adjusting the output of modified operation 808 using the output of modified operation 806) prior to operation 810.

[0068] Moreover, the target signal range processed in operation 810 may include multiple ranges, which in some implementations may overlap. For example, instead of the above-described modification in which the signal range is adjusted, operation 808 may include forwarded two or more signal processing ranges (e.g., 0-70th percentile and 30th-100th percentile). This may be especially useful when processing image sensor data, for example to overlay the visual images for two different range such that both bright portions of the spectrum and dark portions of the spectrum can be show n or displayed.

[0069] Other examples and uses of the disclosed technology will be apparent to those having ordinary skill in the art upon consideration of the specification and practice of the invention disclosed herein. The specification and examples given should be considered exemplary only, and it is contemplated that the appended claims will cover any other such embodiments or modifications as fall within the true scope of the invention.

[0070] The Abstract accompany ing this specification is provided to enable the UnitedStates Patent and Trademark Office and the public generally to determine quickly from a cursory inspection the nature and gist of the technical disclosure and in no way intended for defining, determining, or limiting the present invention or any of its embodiments.

Claims

CLAIMSWhat is claimed is:

1. An input scaling method comprising: receiving at least one input signal from at least one input source; ranking a signal parameter of the at least one input signal and generating a profile of the signal parameter; detecting a target signal in a full spectrum of the at least one input signal; forwarding the target signal to a signal processing unit; and processing the target signal based on the profile.

2. The method of claim 1, wherein the operation of receiving at least one input signal from at least one input source includes simultaneously receiving a plurality of input signals from a plurality' of input sources.

3. The method of claim 1. wherein the operation of ranking the signal parameter includes ranking components of the signal parameter or ranking a transformation of the components of the signal parameter; and the operation of generating the profile of the signal parameter includes ordering the components or the transformation of the components according to the ranking.

4. The method of claim 1, wherein the operation of detecting the target signal includes: identifying a purpose of the signal processing unit; selecting a signal range within the full spectrum of the input signal based on the purpose; and setting, as the target signal, first components of the full spectrum of the input signal that are within the signal range.

5. The method of claim 4. wherein the operation of forwarding the target range to the signal processing unit further includes:analyzing second components of the full spectrum of the input signal that are outside the signal range; and in response to a determination that one of the second components includes a candidate target signal, forwarding the candidate target signal to the signal processing unit.

6. The method of claim 1, wherein the operation of detecting the target signal is performed by an algorithm or an artificial intelligence model.

7. The method of claim 1, wherein the signal parameter includes at least one of a signal strength, a signal magnitude, a signal quantity, or a frequency of occurrence.

8. The method of claim 1. wherein receiving the at least one input signal includes receiving a wireless transmission from the at least one input source, wherein the at least one input source is remotely located.

9. The method of claim 1, wherein the at least one input source is sensor configured to generate an electrical signal in response to a physical occurrence.

10. The method of claim 9, wherein the physical occurrence is a sound, and the at least one input source is a hearing aid.

11. An input scaling system comprising: at least one input source configured to receive at least one input signal from a sensor; a memory device; and at least one processor configured to: rank a signal parameter of the at least one input signal and generate a profile of the signal parameter, detect a target signal from within a full spectrum of the at least one input signal, forward the target signal to a signal processing unit, and process the target signal based on the profile.

12. The system of claim 11, wherein the at least one processor is configured to process the target signal based on the profile or the rank of the signal parameter and not based on a value of the signal parameter itself.

13. The system of claim 11, wherein the at least one input signal is a plurality of input signals, and the at least one input source is configured to simultaneously receive the plurality of input signals.

14. The system of claim 11, further comprising a user interface configured to receive a user input, wherein the at least one processor is configured to rank the signal parameter and generate the profile based on the user input.

15. The system of claim 11, wherein the sensor includes a first sensing device configured to generate the at least one input signal and a second sensing device configured to generate a physical measurement parameter.

16. The system of claim 15, wherein the physical measurement parameter includes at least one of a time, a day, a month, a season, an overall frequency spectrum, an identification of an event in a frequency region, or a direction of a source of a physical phenomenon.

17. The system of claim 11. wherein the at least one processor is configured to detect the target signal by providing the full spectrum to an algorithm or an artificial intelligence model and receiving the target signal from the algorithm or the artificial intelligence model.

18. The system of claim 11. wherein the at least one processor is configured to process the target signal using a differential processing operation.

19. The system of claim 11, wherein the signal parameter includes at least one of a signal strength, a signal magnitude, a signal quantity, or a frequency of occurrence.

20. The system of claim 11 , wherein the at least one processor is configured to rank the signal parameter using a floating sub-range of a full spectral range of the at least one input signal.

Images