Mk Systems USA Inc.

Michael Ropert

Jean Kypreos

Systems, devices and methods for encoding video data are provided. Video data is encoded to generate an encoded data structure that includes, for a given pixel block: a plurality of data layers, each associated with one of a plurality of quality levels and each including data defining: a plurality of transform coefficients encoded for the associated quality level; and a quantization step size for the given pixel block and the associated quality level. At least one of the data layers includes a header including data defining a plurality of encoder decisions made for the given pixel block, the encoder decisions common across the quality levels.

Systems and methods for scalable hybrid video encoding

Disclosed herein are systems, devices, and methods for encoding a plurality of video channels in a shared resource environment. For a plurality of frames in the video channels, complexity metrics are computed, each estimating content complexity of a corresponding frame; and encoding budget metrics are computed, each defining an allocation of bits for encoding the corresponding frame. The encoding budget metrics of a particular video channel are computed to reduce anticipated encoding distortion in the particular video channel. A resource allocation is determined that allocates a portion of total resources of the shared resource environment for encoding each of the video channels. The resource allocation is determined to reduce anticipated encoding distortion across the video channels based on the encoding budgets and the complexity metrics. Encoding configurations are selected based on the resource allocation, each for encoding a corresponding video channel.

Growing community of inventors

Rennes, France

Michael Ropert