公开(公告)号：US20230229892A1

公开(公告)日：2023-07-20

申请号：US17927929

申请日：2021-05-31

Applicant: DOLBY INTERNATIONAL AB

Inventor： Arijit BISWAS ,  Simon PLAIN

IPC: G06N3/0455 ,  G10L19/26 ,  G10L25/30 ,  G10L25/69 ,  G06N3/082

CPC classification number: G06N3/0455 ,  G10L19/26 ,  G10L25/30 ,  G10L25/69 ,  G06N3/082

Abstract: Described herein is a method of determining parameters for a generative neural network for processing an audio signal, wherein the generative neural network includes an encoder stage mapping to a coded feature space and a decoder stage, each stage including a plurality of convolutional layers with one or more weight coefficients, the method comprising a plurality of cycles with sequential processes of: pruning the weight coefficients of either or both stages based on pruning control information, the pruning control information determining the number of weight coefficients that are pruned for respective convolutional layers; training the pruned generative neural network based on a set of training data; determining a loss for the trained and pruned generative neural network based on a loss function; and determining updated pruning control information based on the determined loss and a target loss. Further described are corresponding apparatus, programs, and computer-readable storage media.

公开(公告)号：US20230267938A1

公开(公告)日：2023-08-24

申请号：US18004197

申请日：2021-07-07

Applicant: DOLBY INTERNATIONAL AB

Inventor： Harald MUNDT ,  Stefan BRUHN ,  Heiko PURNHAGEN ,  Simon PLAIN ,  Michael SCHUG

IPC: G10L19/005 ,  G10L19/008 ,  G10L19/02

CPC classification number: G10L19/005 ,  G10L19/008 ,  G10L19/0204

Abstract: Described are methods of processing an audio signal for packet loss concealment. The audio signal comprises a sequence of frames, each frame containing representations of a plurality of audio channels and reconstruction parameters for upmixing the plurality of audio channels to a predetermined channel format. One method includes: receiving the audio signal; and generating a reconstructed audio signal in the predefined channel format based on the received audio signal. Generating the reconstructed audio signal comprises: determining whether at least one frame of the audio signal has been lost; and if a number of consecutively lost frames exceeds a first threshold, fading the reconstructed audio signal to a predefined spatial configuration. Also described is a method of encoding an audio signal. Yet further described are apparatus for carrying out the methods, as well as corresponding programs and computer-readable storage media.