INDEPENDENT AND CONCURRENT PROCESSING MULTIPLE AUDIO INPUT SIGNALS IN A PROSTHETIC HEARING IMPLANT

A voice detection circuit uses normal and customary signal peaks and pauses in speech to distinguish voice from mere noise. The voice detection circuit includes an audio input device for receiving an audio signal and comparing it to a threshold, and a detector for generating the VOX output signal responsive to the presence of the peaks and pauses in the audio signal which rise above the threshold. The switch is actuated upon the VOX output signal indicating that the audio signal includes selected patterns of peaks and pauses and thus comprises voice signals. Illustratively, the threshold may be varied in accordance with the input signals when the voice detection circuit indicates that the input audio signals do not comprise voice signals.

A system and method of audio processing provides enhanced speech recognition. Audio input is received at a plurality of microphones. The multi-channel audio signal from the microphones may be processed by a beamforming network to generate a single-channel enhanced audio signal, on which voice activity is detected. Audio signals from the microphones are additionally processed by an adaptable noise cancellation filter having variable filter coefficients to generate a noise-suppressed audio signal. The variable filter coefficients are updated during periods of voice inactivity. A speech recognition engine may apply a speech recognition algorithm to the noise-suppressed audio signal and generate an appropriate output. The operation of the speech recognition engine and the adaptable noise cancellation filter may advantageously be controlled based on voice activity detected in the single-channel enhanced audio signal from the beamforming network.

A sound processing method and apparatus are provided, which are capable of performing sound processing on input audio signals containing a plurality of signal components being different in desired sound processing conditions, in a manner that allows natural sound to be reproduced. An input audio signal of at least one system is separated into a plurality of separated signal components, and each signal component of at least part of the plurality of separated signal components is subjected to individual sound processing according to the signal component, and the plurality of separated signal components are outputted as at least one audio signal after each signal component of the at least part thereof is subjected to the individual sound processing. The plurality of separated signal components are synthesized into a synthesized audio signal, which is then outputted, or alternatively, the plurality of separated signal components are outputted separately as audio signals.

A bone conduction speaker and mounting system allows a person to perceive audible sounds, such as music, through his sense of touch. A bone conduction speaker or transducer is mounted against a person's body, preferably against the sternum. An audio signal is fed to the bone conduction speaker or transducer which converts the audio signal to vibrations. The bone conduction speaker or transducer transmits the vibrations to the person's rib-cage which will then resonate in synchronism with the input audio signal. The audio signal may also be simultaneously fed to regular audio speakers to allow a person to hear as well as feel the audible sounds or music.

Variations in the amplitude-frequency response characteristic of a loudspeaker are compensated by a drive circuit in which an audio input signal is summed with a plurality of delayed replicas of the input signal to provide a drive signal for the loudspeaker. The delayed replicas are derived by discrete time filtering. Noise reduction techniques applied to the least delayed of the plurality of delay signals substantially improves the S/N ratio in the drive signal.

A method of altering a social signaling characteristic of a speech signal. A statistically large number of speech samples created by different speakers in different tones of voice are evaluated to determine one or more relationships that exist between a selected social signaling characteristic and one or more measurable parameters of the speech samples. An input audio voice signal is then processed in accordance with these relationships to modify one or more of controllable parameters of input audio voice signal to produce a modified output audio voice signal in which said selected social signaling characteristic is modified. In a specific illustrative embodiment, a two-level hidden Markov model is used to identify voiced and unvoiced speech segments and selected controllable characteristics of these speech segments are modified to alter the desired social signaling characteristic.

Neuroprosthetic devices, such as cochlear and retinal implants, work by directly stimulating neurons with extracellular electrodes. This is commonly modeled using the cable equation with an applied extracellular voltage. In this paper a framework for modeling extracellular electrical stimulation is presented. To this end, a cylindrical neurite with confined extracellular space in the subthreshold regime is modeled in three-dimensional space. Through cylindrical harmonic expansion of Laplace’s equation, we derive the spatio-temporal equations governing different modes of stimulation, referred to as longitudinal and transverse modes, under types of boundary conditions. The longitudinal mode is described by the well-known cable equation, however, the transverse modes are described by a novel ordinary differential equation. For the longitudinal mode, we find that different electrotonic length constants apply under the two different boundary conditions. Equations connecting current density to voltage boundary conditions are derived that are used to calculate the trans-impedance of the neurite-plus-thin-extracellular-sheath. A detailed explanation on depolarization mechanisms and the dominant current pathway under different modes of stimulation is provided. The analytic results derived here enable the estimation of a neurite’s membrane potential under extracellular stimulation, hence bypassing the heavy computational cost of using numerical methods.

This paper presents an ultra-low-power voice activity detection (VAD) system to discriminate speech from non-speech parts of audio signals. The proposed VAD system uses level-crossing sampling for voice activity detection. The useless samples in the non-speech parts of the signal are eliminated due to the activity-dependent nature of this sampling scheme. A 40 ms moving window with a 30 ms overlap is exploited as a feature extraction block, within which the output samples of the level-crossing analog-to-digital converter (LC-ADC) are counted as the feature. The only variable used to distinguish speech and non-speech segments in the audio input signal is the number of LC-ADC output samples within a time window. The proposed system achieves an average of 91.02% speech hit rate and 82.64% non-speech hit rate over 12 noise types at −5, 0, 5, and 10 dB signal-to-noise ratios (SNR) over the TIMIT database. The proposed system including LC-ADC, feature extraction, and classification circuits was designed in 0.18 µm CMOS technology. Post-layout simulation results show a power consumption of 394.6 nW with a silicon area of 0.044 mm2, which makes it suitable as an always-on device in an automatic speech recognition system.

A large number of hearing impaired people in rural sector of Pakistan cannot afford the high prices of digital hearing aids. Moreover, since the analogue devices amplify the speech and noise signals equally, they are not much flexible and have functional limitations. These problems can be solved with digital devices which are generally costly. In this research work, a microcontroller based cost-effective and portable hearing aid is designed and tested. The main components include sound detection circuit, microcontroller and a digital to analogue (DAC) convertor. Device is aimed to process human voice frequencies with an adjustable gain of up to 80 dB. It is tested with an input audio signal and it showed satisfactory results for low frequency signals up to 3.5 kHz which caters for the frequency range of most human voices. The system is now being developed for more effective noise suppression and broader frequency range.

The invention concerns audio signal processing, comprising: a first processing (205) of an audio source signal, using at least a mathematical transform applied on first sequences of samples obtained by applying first segmentation windows (505, 506, 700, 701) on the audio source signal; and a second audio processing (206) applied on second sequences of samples obtained by applying second segmentation windows (507, 508) on the signal delivered by the fist step; the two successive first windows and/or the two successive second windows overlapping, the overlaps being such that the segmentations are synchronous.

Recently, independent component analysis (ICA) is the most popular and promising statistical technique for blind audio source separation. This paper proposes the minimum beta-divergence based ICA as an adaptive robust audio source separation algorithm. This algorithm explores local structures of audio source signals in which the observed signals follow a mixture of several ICA models. The performance of this algorithm is equivalent to the standard ICA algorithms if observed signals are not corrupted by outliers and there exist only one structure of audio source signals in the entire data space, while it keeps better performance otherwise. It is able to extract all local audio source structures sequentially in presence of huge outliers. Our experimental results also support the above statements.

Image style transfer networks are used to blend images, producing images that are a mix of source images. The process is based on controlled extraction of style and content aspects of images, using pre-trained Convolutional Neural Networks (CNNs). Our interest lies in adopting these image style transfer networks for the purpose of transforming sounds. Audio signals can be presented as grey-scale images of audio spectrograms. The purpose of our work is to investigate whether audio spectrogram inputs can be used with image neural transfer networks to produce new sounds. Using musical instrument sounds as source sounds, we apply and compare three existing image neural style transfer networks for the task of sound mixing. Our evaluation shows that all three networks are successful in producing consistent, new sounds based on the two source sounds. We use classification models to demonstrate that the new audio signals are consistent and distinguishable from the source instrument sounds. We further apply t-SNE cluster visualisation to visualise the feature maps of the new sounds and original source sounds, confirming that they form different sound groups from the source sounds. Our work paves the way to using CNNs for creative and targeted production of new sounds from source sounds, with specified source qualities, including pitch and timbre.

A two‐channel single‐stage switching amplifier (SSA) based on high‐frequency link inverter is proposed. The audio input signals are modulated by phase shift between primary‐side switches and secondary‐side switches, thus independent amplifications of two channels can be obtained. Benefiting from multiplexing of primary‐side power switches for two channels, only a half bridge inverter is utilised at primary side. Clamp circuits at the secondary side of transformer eliminate spike voltages across the secondary‐side switches during the current commutation of the cycloconverters. Zero voltage switching of primary‐side switches and clamp switches are achieved by utilising magnetising inductor current. A 320 V/250W prototype is presented to verify the analysis results of the proposed two‐channel SSA.

Similar to Haptuator from Tactile Labs Inc. and Force Reactor from Alps Electric Co., DC motor can produce high fidelity vibration when it is driven by alternative current such as audio input signal. In our study, we developed a new vibration actuator using a DC motor that can generate much stronger vibrations than a normal DC motor and produce very low frequency of vibrations. We proposed that the stator of the motor could be used as the vibration mass by fixing the rotor of the actuator. To evaluate this design concept, we developed a prototype actuator that can be driven in two modes: stator mode (i.e., the new design concept) and normal mode. The experiment results revealed that stronger vibrations can be obtained on a fingertip in stator mode because the fixed part that comprises the rotor was lighter and the vibration mass using the stator was heavier. We also confirmed that the actuator can be driven at very low frequency (1 Hz) in stator mode.

This paper deals with implementing a real-time gunshot detection algorithm on the digital signal processor TMS320C6713. The developed algorithm uses 3 linear predictive coding coefficients, energy in 3 spectral bands, and frame correlation. The audio input signal is continuously processed and signal frames considered to contain a gunshot are shortly signalized by an LED indicator. Experimental results achieved 82% correctly detected gunshots and 3% false alarms. The system should increase protection of wild elephants in Africa.