Nuvoton Technology Corporation

Powen Ru

Jin Hee Kim

Nathan Wong

Dung B Nguyen

Tawei Yang

Alex Paiz

Pratul Saini

Andrew Zamansky

A method is disclosed to estimate the delay between an original signal and the corresponding captured signal. The signals are transformed and buffered to two sets of spectral descriptors for a similarity measure. The method advantageously offers robust delay estimation for inconsistent delays and adverse spectral distortions.

Delay estimation using frequency spectral descriptors

A frequency dependent dynamic range control method is employed in a signal emitting system characterized by lower impedance in the high frequency region. An efficient technique is implemented to dynamically estimate current spectrum and conditioning parameters to lower the current in the high frequency region. The method advantageously avoids current overloading without using a series resistor as a current limiter.

Frequency dependent dynamic range control

A recurrent neural network is employed in a loudspeaker system to compensate the distortion of the system based upon a source signal (content) and the sensing output of a sensing circuit (context). A frequency domain transform is selected to provide mapping between the source signal and a recorded signal; and enable reconstruction of desirable playback. Various sensing-related features and source-related features are derived to serve as the auxiliary information. A desirable content is therefore generated based upon the original content and the context.

Deep learning speaker compensation

A beamforming system comprises an input module, a phase-difference module, a delay distribution module, and a delay estimation module configured to make a final delay estimation based on the delay distribution. The final delay estimation is applied to align one of two selected channels and combine the two channels to obtain a signal of interest. The phase-difference module calculates phase differences for a range of frequency bins. The delay estimation module considers all possible delays derived from the phase differences, including multiples of &#xb1;2&#x3c0; to overcome the ambiguity in high frequency bins when the spacing between two acoustic sensors is longer than the half of the wavelengths.

Beamforming system based on delay distribution model using high frequency phase difference

A voice activity detection (VAD) system includes an input processing module configured to receive an acoustic signal, convert the acoustic signal into an analog signal, and subsequently, a digital signal; an energy-based detection module configured to receive one of the analog/digital signals and determine a sound activity decision; an area-function-based detection module configured to derive an area-related function from the digital signal and use a machine learning method to output an area-based decision according to the area related function; and a VAD decision module configured to make a final VAD decision based on the sound activity decision from the energy-based detection module and the area-based decision from the area-function-based detection module.

Growing community of inventors

San Jose, CA, United States of America

Powen Ru