Abstract: An audio transmitter including an audio input interface receiving first audio signals indicative of input audio from an audio source, a wireless transmitter wirelessly streaming second audio signals indicative of the input audio to a first audio reproducing device, and an audio output interface connectable to a second audio reproducing device, the audio transmitter arranged to supply third audio signals indicative of the input audio to the audio output interface. The audio transmitter facilitates connection to the audio transmitter to control wireless streaming of the second audio signals. A characteristic of sound of the wirelessly streamed second audio signals is controllable independently of a characteristic of sound of the third audio signals. An initial connection is established between the audio transmitter and a computing device in response to detection that the location of the computing device relative to the audio transmitter is less than a defined threshold.

Abstract: During text-to-speech processing, a sequence-to-sequence neural network model may process text data and determine corresponding spectrogram data. A normalizing flow component may then process this spectrogram data to predict corresponding phase data. An inverse Fourier transform may then be performed on the spectrogram and phase data to create an audio waveform that includes speech corresponding to the text.

Abstract: During text-to-speech processing, a speech model creates synthesized speech that corresponds to input data. The speech model may include an encoder for encoding the input data into a context vector and a decoder for decoding the context vector into spectrogram data. The speech model may further include a voice decoder that receives vocal characteristic data representing a desired vocal characteristic of synthesized speech. The voice decoder may process the vocal characteristic data to determine configuration data, such as weights, for use by the speech decoder.

Abstract: A speech model is trained using multi-task learning. A first task may correspond to how well predicted audio matches training audio; a second task may correspond to a metric of perceived audio quality. The speech model may include, during training, layers related to the second task that are discarded at runtime.

Abstract: During text-to-speech processing, a speech model creates output audio data, including speech, that corresponds to input text data that includes a representation of the speech. A spectrogram estimator estimates a frequency spectrogram of the speech; the corresponding frequency-spectrogram data is used to condition the speech model. A plurality of acoustic features corresponding to different segments of the input text data, such as phonemes, syllable-level features, and/or word-level features, may be separately encoded into context vectors; the spectrogram estimator uses these separate context vectors to create the frequency spectrogram.

Abstract: A speech model includes a sub-model corresponding to a vocal attribute. The speech model generates an output waveform using a sample model, which receives text data, and a conditioning model, which receives text metadata and produces a prosody output for use by the sample model. If, during training or runtime, a different vocal attribute is desired or needed, the sub-model is re-trained or switched to a different sub-model corresponding to the different vocal attribute.

Abstract: During text-to-speech processing, a speech model creates output audio data, including speech, that corresponds to input text data that includes a representation of the speech. A spectrogram estimator estimates a frequency spectrogram of the speech; the corresponding frequency-spectrogram data is used to condition the speech model. A plurality of acoustic features corresponding to different segments of the input text data, such as phonemes, syllable-level features, and/or word-level features, may be separately encoded into context vectors; the spectrogram estimator uses these separate context vectors to create the frequency spectrogram.

Abstract: An audio transmitter including an audio input interface receiving first audio signals indicative of input audio from an audio source, a wireless transmitter wirelessly streaming second audio signals indicative of the input audio to a first audio reproducing device, and an audio output interface connectable to a second audio reproducing device, the audio transmitter arranged to supply third audio signals indicative of the input audio to the audio output interface. The audio transmitter facilitates connection to the audio transmitter to control wireless streaming of the second audio signals. A characteristic of sound of the wirelessly streamed second audio signals is controllable independently of a characteristic of sound of the third audio signals. An initial connection is established between the audio transmitter and a computing device in response to detection that the location of the computing device relative to the audio transmitter is less than a defined threshold.

Abstract: A speech model includes a sub-model corresponding to a vocal attribute. The speech model generates an output waveform using a sample model, which receives text data, and a conditioning model, which receives text metadata and produces a prosody output for use by the sample model. If, during training or runtime, a different vocal attribute is desired or needed, the sub-model is re-trained or switched to a different sub-model corresponding to the different vocal attribute.

Abstract: A system is disclosed for configuring a hearing device. The system comprises an audiogram processing component arranged to process image data indicative of an image of an audiogram associated with a person so as to produce audiogram data indicative of the audiogram, and a hearing device configuration component arranged to produce hearing device configuration data based on the audiogram data. The configuration data is indicative of configuration settings for a hearing device that will cause the hearing device to assist the hearing of the person.

Abstract: A speech model is trained using multi-task learning. A first task may correspond to how well predicted audio matches training audio; a second task may correspond to a metric of perceived audio quality. The speech model may include, during training, layers related to the second task that are discarded at runtime.

Abstract: A monitoring system (1) comprises sensors (102) adapted to be worn by a user, and, a processor (101, 302) linked with the sensor. The processor receives sensor data and processes this data to determine user posture data including data indicative of vertical distance between level of the user's heart and ankle (?h, Vd 1, Vd2, Vd3). Based on the posture data together with a value for degree of user chronic venous insufficiency and/or blood density, generate an estimate of user static venous pressure while the user is static, without calf muscle pump activity. The processor (101, 302) also processes the sensor data to determine if there is calf muscle pump activity, and generates an estimate of user active venous pressure according to the static venous pressure estimate, rate of calf muscle activity, and a value for degree of user chronic venous insufficiency. The processor (101, 302) may generate the venous pressure estimate in real time, and may control an NMES device accordingly.

Abstract: A handrail includes a support assembly for attachment to a concrete slab having an upper surface with attachment formations anchored at the upper surface. The assembly includes a handrail support component and a connector component. The connector component is attachable to the support component and an attachment formation. The support and connector components can be tightened to each other to tighten the connector component to the attachment formation for tightening the support component onto the slab. The invention extends to a handrail system including the slab, and to a method of installing handrail components on the slab.

Abstract: A monitoring system comprises sensors adapted to be worn by a user, and, a processor linked with the sensor. The processor receives sensor data and processes this data to determine user posture data including data indicative of vertical distance between level of the user's heart and ankle. Based on the posture data together with a value for degree of user chronic venous insufficiency and/or blood density, generate an estimate of user static venous pressure while the user is static, without calf muscle pump activity. The processor also processes the sensor data to determine if there is calf muscle pump activity, and generates an estimate of user active venous pressure according to the static venous pressure estimate, rate of calf muscle activity, and a value for degree of user chronic venous insufficiency. The processor may generate the venous pressure estimate in real time, and may control an NMES device accordingly.

Abstract: A pneumocardial function monitor includes a carrier configured to be mounted about at least a part of a trunk of a body of a subject. A sensing arrangement is mounted on the carrier, the sensing arrangement including at least one element for monitoring changes in volume of the part of the subject's body. A signal processing module is in communication with the sensing arrangement for processing signals output from the sensing arrangement, the signal processing module having at least one output for outputting a signal related to respiratory function and/or cardiac function.

Abstract: A monitoring system has biomechanical sensors, physiological sensors and a controller which receive sensory inputs from the sensors to provide output signals for the output device, and it detects from the sensory inputs risk of a syncopal event The bio-mechanical sensors include sensors arranged to allow the processor to detect a user postures and posture transitions. The processor operates a finite state machine, in which there is a state corresponding to each of a plurality of user physical postures and to each of a plurality of transitions between said postures, and the processor determines a relevant state depending on the sensory inputs. A device output may be muscle stimulation to prevent syncope, and there are stimulation permissions associated with the finite state machine states.

Abstract: A blood volume monitor includes a carrier mountable on a subject's body part. A measuring arrangement is mounted on the carrier. The measuring arrangement provides data relating to a change in volume of the body part, and outputs signals representative of the change. A control unit is in communication with the measuring arrangement to process data output by the measuring arrangement to determine, using the volume data, the volume of the body part and determines blood volume in the body part. The control unit uses the response signals received from the measuring arrangement and the determined volume of the part of the body to determine the blood volume.

Abstract: A pneumocardial function monitor includes a carrier configured to be mounted about at least a part of a trunk of a body of a subject. A sensing arrangement is mounted on the carrier, the sensing arrangement including at least one element for monitoring changes in volume of the part of the subject's body. A signal processing module is in communication with the sensing arrangement for processing signals output from the sensing arrangement, the signal processing module having at least one output for outputting a signal related to respiratory function and/or cardiac function.

Abstract: A monitoring system has biomechanical sensors, physiological sensors and a controller which receive sensory inputs from the sensors to provide output signals for the output device, and it detects from the sensory inputs risk of a syncopal event The bio-mechanical sensors include sensors arranged to allow the processor to detect a user postures and posture transitions. The processor operates a finite state machine, in which there is a state corresponding to each of a plurality of user physical postures and to each of a plurality of transitions between said postures, and the processor determines a relevant state depending on the sensory inputs. A device output may be muscle stimulation to prevent syncope, and there are stimulation permissions associated with the finite state machine states.

Abstract: A handrail includes a support assembly for attachment to a concrete slab having an upper surface with attachment formations anchored at the upper surface. The assembly includes a handrail support component and a connector component. The connector component is attachable to the support component and an attachment formation. The support and connector components can be tightened to each other to tighten the connector component to the attachment formation for tightening the support component onto the slab. The invention extends to a handrail system including the slab, and to a method of installing handrail components on the slab.