Patent Application
20230155562
A loudspeaker, or speaker, is an electroacoustic transducer, which converts an electrical audio signal into acoustic energy that can produce a corresponding sound. A sound source, such as a music recording or music track, is amplified via an audio power amplifier, and a resulting electrical audio signal is provided to the loudspeaker. A common example of a loudspeaker includes a dynamic speaker to produce sound from the electrical audio signal. The dynamic speaker includes a voice coil suspended near a permanent magnet. The voice coil is also coupled to a movable diaphragm. An alternating current electrical audio signal is provided to the voice coil. Passing an alternating current through the voice coil creates a magnetic field via induction. When the electrical audio signal causes the polarity of the magnetic field in the voice coil to be the same as the permanent magnet, the magnetic forces repel one another and the voice coil can push on the diaphragm. When the electrical audio signal causes the polarity of the magnetic field in the voice coil to be the opposite of the permanent magnet, the magnetic forces attract one another and the voice coil can pull on the diaphragm. Movement of the diaphragm causes changes in the air pressure, which provides the acoustic energy.
In contrast to a dynamic speaker, a vibration speaker is a loudspeaker that does not include a diaphragm. Rather, the voice coil is coupled to a movable plate included in an inducer, or exciter. The movable plate of the exciter is disposed against a surface to transfer vibrations into the surface to produce acoustic energy. Examples of a surface can include a table, a desk, a bookshelf, a wall, and a window. The exciter can be disposed on the table, desk, bookshelf, wall, and window, such as placed or laid down on the table, desk, and bookshelf and attached to or pressed against the wall or window. As current of the electrical audio signal alternates in the voice coil, the voice coil causes the movable plate to vibrate. The movable plate vibrates against the surface, which can cause sound corresponding with the electrical audio signal.
Resonance sound amplification devices incorporate vibration speakers to produce acoustic energy, such as sound, that corresponds with an electric audio signal. Examples of resonance sound amplification devices can include portable speakers, smart speakers, invisible speakers, audio output devices, and computing devices such as laptops, docks, and mobile devices. Resonance sound amplification devices are apt in circumstances of small form factors that have limited space for dynamic loudspeakers. Resonance sound amplification devices include an exciter that is disposed on a surface, and the exciter can cause the surface to vibrate and produce the sound. Often, the quality of the sound produced with the resonance sound amplification device is based on the surface on which the resonance sound amplification device is dispose. For instance, an exciter disposed on a desk may produce a sound of a first quality and the exciter disposed on a table or bookshelf may produce a sound of a different quality. In the case of a resonance sound amplification device implemented in a portable computing device or portable speaker, for example, the resonance sound amplification device may be disposed on many different surfaces, and the quality of the sound can vary depending on the surface on which the resonance sound amplification device is disposed.
In one example, the electronic device 100 includes a resonance sound amplification device to generate an audio output via a resonance. For instance, the resonance sound amplification device includes an exciter disposed on a surface to vibrate the surface and produce a sound. The exciter can be included in a vibration speaker and can receive an electrical audio signal having an equalization setting, such as a first equalization setting. The exciter vibrates against the surface, which can cause the audio output corresponding with the electrical audio signal. For example, the exciter includes a movable plate or member to cause a vibration. In one example, the surface is external to the resonance sound amplification device and not included as part of the resonance sound amplification device. The audio output can include an audible range of frequencies and an inaudible range of frequencies.
A resonance sample of the audio output is detected and received. For example, the audio output can be generated over a period of time, and the resonance sample of the audio output can include a subset of time of the period of time. In one example, the resonance sample is detected via a transducer disposed on the surface. The transducer may include an accelerometer. The transducer converts the vibrations on the surface into an electrical input signal, and the electrical input signal, or a portion of the electrical input signal, is applied as the resonance sample. In another example, the resonance sensor is detected by another sensor such as a microphone or other sensor that can detect vibrations such as a camera.
The received resonance sample of the audio output is categorized in an example at 108. For example, the received resonance sample of the audio output is categorized to quantifiably determine the nature and quality aspects of the audio output based on the received resonance sample. Aspects of the audio output based on the received resonance sample can include the type of corresponding audio recording, such as whether the audio recording is music speech or whether the audio recording is a music track and can include the type of music track such as whether the music track is of classic rock or classical music. Further, the received resonance sample can represent a recording of a performance in a live concert hall or a studio recording. Further, aspects of the received resonance sample include the resonance and characteristics of the surface, such as whether the surface is relatively large or relatively small, or the type of material of the surface. In an additional example, the aspects of the received resonance can account for the equalization settings applied to the audio output, such as the aspects of the received resonance can include the first equalization setting. In one example, the received resonance sample is transformed into a determined profile.
The received resonance sample of the audio output is categorized with a neural network to determine the inference profile in an example at 108. For example, the received resonance sample can be converted to a spectrograph, such as an audio spectrograph in the time domain, included in a determined profile. In one example, the determined profile is compared to a library of stock profiles to determine a match or a close approximation. The spectrograph compared to other audio spectrographs having known aspects. The other audio spectrographs can be included in a library of spectrographs. In one example, a deep neural network is used in the comparison. A deep neural network is an example of a neural network having multiple layers between an input and output. An example of a deep neural network suitable for use to categorize the received resonance samples includes a convolutional neural network, or CNN. The CNN can output the inference profile.
In one example, the inference profile can be associated with or correspond with an inference profile equalization setting. The equalization setting can alter features of the audio data associated with the electrical audio signal including frequency response, time response, and phase response. In one example, the equalization setting can be affected via the application of analog filters or digital filters. For instance, filters can be applied to the audio data to adjust bass and treble tones in an example equalizer or equalizer component.
The equalization setting for the audio output, such as the equalization setting that is applied to generate the audio output, can be adjusted based on the inference profile in an example at 110. For instance, an equalizer applied to audio data used to generate the electrical audio signal can be adjusted from the first equalization setting to the designated equalization setting corresponding with the inference profile. In another example, the equalizer applied to audio data used to generate the electrical audio signal can be adjusted from the first equalization setting to a setting approximating the designated equalization setting corresponding with the inference profile. In one example, the electronic device can be coupleable to a resonance sound amplification device to adjust the equalization setting at 110.
In one example, the method can be repeated for the audio output. In this example, the received resonance sample is periodically received, or a periodically received resonance sample. For example, the method can be repeated every selected amount of time, for a threshold change in the electrical audio signal, or for a change in track for a periodically received resonance sample.
In one example, the audio output used for the received resonance sample is a user-selected audio output played with the resonance sound amplification device. Examples of the user-selected audio output can include a music track, a conversation, and audio from an audio/video that is played with the resonance sound amplification device. In another example, the audio output is a device-selected audio output played with the resonance sound amplification device. The device-selected audio output can be generated via the processor used to categorize the received resonance sample. For instance, the device-selected audio output may be played to tune the resonance sound amplification device to the surface and include a known or a predetermined frequency and phase. The device-selected audio output can include a plurality of predetermined frequencies and phases output concurrently or over time. In still another example, the audio output can include a combination of a user-selected audio output and a device-selected audio output. In one example, the user-selected audio output may be determined to include a preponderance of low frequencies, and the device-selected audio output may be generated to include a higher frequency to tune the resonance sound amplification device in anticipation of, for example, subsequent music tracks. The device-selected audio output can be selected from an audible range of frequencies and an inaudible range of frequencies.
The audio generator 202 can convert the sound source into audio data for processing. In the example, the audio generator 202 includes an equalizer 208. The equalizer 208 can apply an equalizer setting, such as the first equalizer setting to the audio data during processing to generate an electrical audio signal. The electrical audio signal is provided to the exciter 204. The exciter 204 generates a vibration corresponding with the electrical audio signal.
The vibration can be applied to a surface 210 if the exciter is disposed against the surface 210. The surface 210 is external to and separate from the electronic device 200. In the example, the exciter 204 is disposed on the surface 210, and the exciter 204 can cause the surface to vibrate and generate an audio output, such as a sound, corresponding with the electrical audio signal. The input transducer 206 receives the audio output and generates a transducer signal based on the received audio output. The transducer 206, such as an accelerometer, in one example, can be disposed on the surface 210. For instance, the components of the electronic device 200 can be enclosed in a case 220, and the exciter 204 and an accelerometer may extend from the case 220 as feet of the electronic device 200. In some examples, the electronic device 200 can include a plurality of exciters to provide independent audio channels or to focus or direct vibrations. The electronic device 200 can also include a plurality of transducers 206 including a plurality of different types of transducers such as an accelerometer and a microphone.
The electronic device 200 also includes a controller 212 to receive the transducer signal from the transducer 206. The controller 212, in one example, includes a processor 214 and memory 216. The memory 216 stores computer executable instructions. The processor 214 executes the instructions to generate a resonance sample of the audio output from the transducer signal. The processor 214 also executes the instructions to categorize the resonance sample of an audio output with a neural network, such as a CNN, to obtain an inference profile. The processor 214 is operably coupled to the audio generator 202, and the processor 214 executes the instructions to adjust the selected equalization setting for the audio output based on the inference profile. The audio generator 202 can apply the adjusted equalization settings to the sound source to produce the audio electrical signal.
In one example, the processor 214 may include a plurality of main processing cores to run an operating system and perform general-purpose tasks on an integrated circuit. The processor 214 may also include built-in logic or a programmable functional unit, also on the same integrated circuit with a heterogeneous instruction-set architecture. In additional to multiple general-purpose, main processing cores and the application processing unit, controller 212 can include other devices or circuits such as graphics processing units, neural network processing units, and an audio decoder/encoder, which may include heterogeneous or homogenous instruction set architectures with the main processing cores. For example, the controller 212 may be used to perform other tasks such as in the case of a computing device including the resonance sound amplification device. Additionally, features of the audio generator 202 may be included in or performed with the controller 212.
Memory 216 is an example of computer storage media. Computer storage media includes RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile discs (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, USB flash drive, flash memory card, or other flash storage devices, or other storage medium that can be used to store the desired information and that can be accessed by the processor 214. Accordingly, a propagating signal by itself does not qualify as storage media. Any such computer storage media may be part of the electronic device 200 and implemented as memory 216.
In some examples, the electronic device 200 can include a dynamic speaker 218 coupled to the audio generator 202 to generate an audio output. The audio generator 202 can selectively implement the dynamic speaker 218, such as to produce an audio output instead of the exciter 204, in addition to the exciter 204, or to not produce an audio output. For example, the dynamic speaker 218 can be applied to supplement the sound from the exciter 204 or instead of the exciter 204 in circumstances in which a vibration speaker may not be able to produce a sound that surpasses a quality threshold.
In one example, the electronic device 200 can be implemented using a laptop computer, a desktop computer, a tablet computer, a smart phone, etc. In such cases, the electronic device 200 can include additional components such as display, touchscreen, and keyboard. In other examples, the electronic device 200 can be implemented as an audio component including elements to store and execute computer readable instructions such as a portable speaker with a transducer 206 and a controller 212, such as an onboard processing circuit. In some examples, the electronic device 200 may include connections with the audio generator 202 that are coupleable to other audio output devices such as dynamic speakers or other vibration speakers.
A resonance sample of the audio output is received at 604. The transducer 206 is applied to receive the audio output and produce a transducer signal to the controller 212. The transducer signal can be applied as the resonance sample or the transducer signal can be sampled to produce the resonance sample. A transducer 206 can be used to detect the vibrations directly from the surface or vibrations in a medium such as air. In some examples, multiple transducers can be applied to generate multiple transducer signals.
The resonance sample is categorized to generate an inference profile at 606. In one example, the resonance sample can be categorized via a CNN using a co-processor device such as a neural network processor for processor 214. The generated inference profile corresponds with an equalization setting. The equalization setting of the inference profile is compared with the equalization setting of the audio generator 202.
If the equalization setting of the inference profile is a match with the equalization setting of the audio generator at 608, in which a match is if the equalization setting of the audio generator is the same or within an accepted threshold of the equalization setting of the inference profile audio generator, no changes are made to the equalization setting of the audio generator. Another resonance sample of the audio output can be received at 604, such as in a subsequent music track, after an amount of time has elapsed, if a characteristic of the audio output have been changed via a threshold amount, such as if bass and treble frequency characteristics of the audio output have been changed enough as predetermined in the electronic device 200.
If the equalization setting of the inference profile is not a match with the equalization setting of the audio generator at 608, in which a not a match is if the equalization setting of the audio generator is outside of an accepted threshold of the equalization setting of the inference profile audio generator, the changes to the equalization setting of the audio generator to match the equalization setting of the inference profile are determined at 610. The determined changes to the equalization setting of the audio generator to match the equalization setting of the inference profile are applied to the audio generator 202 such as to the equalizer 208 at 612. Another resonance sample of the audio output can be received at 604, such as in a subsequent music track, after an amount of time has elapsed, if a characteristic of the audio output has been changed by a threshold amount.
In some circumstances, errors may be generated in the method 600. For example, the categorization at 600 may at times not produce an inference profile, such as if the exciter 204 has been moved and is not on a surface, or the determined changes to the equalization setting at 610 are outside the scope of the equalizer 208. In such circumstances, method 600 may revert to features including initialization at 602, sampling again at 604, and disabling the exciter and enabling dynamic speaker 218.
Although specific examples have been illustrated and described herein, a variety of alternate and/or equivalent implementations may be substituted for the specific examples shown and described without departing from the scope of the present disclosure. This application is intended to cover any adaptations or variations of the specific examples discussed herein. Therefore, it is intended that this disclosure be limited only by the claims and the equivalents thereof.
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/US20/29142 | 4/21/2020 | WO |
