Microsoft Technology Licensing, LLC

Francesco Nespoli

Dushyant Sharma

Patrick A Naylor

Junho Park

Daniel Barreda

Shou-Chun Yin

DoYeong Kim

Daniel Paulino Almendro Barreda

Alvaro Martin Iturralde Zurita

William Francis Ganong, III

A method, computer program product, and computing system for processing a speech signal. A sensitive portion of the speech signal is identified. A pseudo-speech representation of the sensitive portion is generated using a voice converter system. Speech processing is performed on the speech signal and the pseudo-speech representation of the sensitive portion using a speech processing system.

System and method for secure processing of speech signals using pseudo-speech representations

A method, computer program product, and computing system for generating an obscured speech signal from an input speech signal and an obscured transcription from a transcription of the input speech signal. A speaker embedding may be extracted from the input speech signal. A speaker embedding delta may be generated based upon, at least in part, the extracted speaker embedding and a synthetic speaker embedding. A synthetic speech signal may be generated from the obscured speech signal using the synthetic speaker embedding. A residual signal may be generated based upon, at least in part, the obscured speech signal and the speaker embedding delta. A speech processing system may be trained using the obscured transcription, the synthetic speech signal, the speaker embedding delta, and the residual signal.

System and method for secure training of speech processing systems

A method, computer program product, and computing system for receiving an input speech signal. A transcription of the input speech signal may be received. A speaker embedding may be extracted from the input speech signal. Acoustic properties from the input speech signal may be extracted. An obscured transcription may be generated from the transcription, where the obscured transcription includes obscured representations of sensitive content from the transcription. An obscured speech signal may be generated based upon, at least in part, the extracted speaker embedding and the obscured transcription, where the obscured speech signal includes obscured representations of sensitive content from the input speech signal. The obscured speech signal may be augmented based upon, at least in part, the extracted acoustic properties.

System and method for secure data augmentation for speech processing systems

A method of performing distance estimation between a first recording device at a first location and a second recording device at a second location includes: estimating acoustic relative transfer function (RTF) between the first recording device and the second recording device for a sound signal, e.g., by applying an improved proportionate normalized least mean square (IPNLMS) filter; and estimating the distance between the first recording device and the second recording device based on the RTF. The at least one acoustic feature extracted from the RTF estimated between the first recording device and the second recording device includes at least one of clarity index, direct-to-reverberant ratio (DRR), and reverberation time. A distributed-gradient-boosting algorithm with regression trees is used in combination with signal-to-reverberation ratio (SRR) and the at least one acoustic feature extracted from the RTF to estimate the distance between the first recording device and the second recording device.

Growing community of inventors

London, United Kingdom

Francesco Nespoli