Abstract: Techniques are described for generating parallel data for real-time speech form conversion. In an embodiment, based at least in part on input speech data of an original form, a speech machine learning (ML) model generates parallel speech data. The parallel speech data includes the input speech data of the original form and temporally aligned output speech data of a target form different than the original form. Each frame of the input speech data temporally corresponds to the corresponding output speech frame of the target speech form and contains a same portion of the particular content. The techniques further include training a teacher machine learning model that is offline and is substantially larger than a student machine learning model for converting speech form. Transferring “knowledge” from the trained Teacher model for training the Production Student Model that performs the speech form conversion on an end-user computing device.

Abstract: Techniques are described for generating parallel data for real-time speech form conversion. In an embodiment, based at least in part on input speech data of an original form, a speech machine learning (ML) model generates parallel speech data. The parallel speech data includes the input speech data of the original form and temporally aligned output speech data of a target form different than the original form. Each frame of the input speech data temporally corresponds to the corresponding output speech frame of the target speech form and contains a same portion of the particular content. The techniques further include training a teacher machine learning model that is offline and is substantially larger than a student machine learning model for converting speech form. Transferring “knowledge” from the trained Teacher model for training the Production Student Model that performs the speech form conversion on an end-user computing device.

Abstract: Techniques are described for generating parallel data for real-time speech form conversion. In an embodiment, based at least in part on input speech data of an original form, a speech machine learning (ML) model generates parallel speech data. The parallel speech data includes the input speech data of the original form and temporally aligned output speech data of a target form different than the original form. Each frame of the input speech data temporally corresponds to the corresponding output speech frame of the target speech form and contains a same portion of the particular content. The techniques further include training a teacher machine learning model that is offline and is substantially larger than a student machine learning model for converting speech form. Transferring “knowledge” from the trained Teacher model for training the Production Student Model that performs the speech form conversion on an end-user computing device.

Abstract: Techniques are described for generating parallel data for real-time speech form conversion. In an embodiment, based at least in part on input speech data of an original form, a speech machine learning (ML) model generates parallel speech data. The parallel speech data includes the input speech data of the original form and temporally aligned output speech data of a target form different than the original form. Each frame of the input speech data temporally corresponds to the corresponding output speech frame of the target speech form and contains a same portion of the particular content. The techniques further include training a teacher machine learning model that is offline and is substantially larger than a student machine learning model for converting speech form. Transferring “knowledge” from the trained Teacher model for training the Production Student Model that performs the speech form conversion on an end-user computing device.

Abstract: Example speech enhancement and noise suppression systems and methods are described. In one implementation, a method receives an audio file comprising a combination of voice data and noise data, and divides the audio file into multiple frames. The method performs a discrete Fourier transform on each frame of a first subset of the multiple frames to provide a plurality of frequency-domain outputs, which are input to a neural network. A ratio mask is obtained as an output from the neural network and clean voice coefficients are computed using the ratio mask. The method outputs an audio file having enhanced speech based on the computed clean voice coefficients.