Matrox Graphics Inc.

Mathieu Girard

Jean-Jacques Ostiguy

Nicolas Jean

Alain Champenois

Simon Garneau

Sergiu Bogdan Nicolescu

Bernard Antaki

Khalil Mawassi

A method and apparatus for enabling compression of a stream of pictures according to a target bit rate are described. A first configuration parameter for a first portion is determined based at least in part on a first relative weight of the first portion with respect to a first set of N portions, where the first set of N portions includes the first portion and N-1 portions which succeed the first portion. A second configuration parameter for a second portion of a second picture is determined based at least in part on a second relative weight of the second portion with respect to a second set of M portions of pictures, where the second set of M portions includes a subset of the N-1 portions from the first set and zero or more additional portions of pictures from the stream of pictures.

Adaptive bit rate control for image compression

Method and apparatus of compressing a stream of pictures in parallel in a compression device, where the compression device includes at least a first and a second component, are described. A first portion of a first picture of a stream of pictures is processed in the first component of the compression device, while a second portion of a second picture of the stream of pictures is simultaneously processed in the second component of the compression device according to partial compression statistics associated with the second picture, where the partial compression statistics result from the processing of the second picture in the compression device when compression of the second picture is not yet completed.

Parallel compression of image data in a compression device

A method and apparatus for enabling low latency compression of a stream of pictures are described. A first set of static regions of a current picture from the plurality of pictures is determined, where each region from the first set is static. A second set of regions of the current picture is determined, where the second set includes all regions of the current picture that are not included in the first set. Compression of the first set of regions is performed based on values of a first quantization parameter determined by a MAQ mechanism. The MAQ mechanism is operative to dynamically increase the compression quality of static regions. Compression of the second set of regions is performed based on values of a second quantization parameter determined by a rate control mechanism. The rate control mechanism is operative to compress the data stream according to a target bit rate.

Method and an apparatus for enabling ultra-low latency compression of a stream of pictures

A method and apparatus for compressing a data stream comprising a plurality of pictures are described. A first quantization parameter (QP) from a plurality of QPs is determined, for a static region in a current picture. The plurality of QPs change in accordance with a multi-step change from a start QP to a target QP and each one of the plurality of QPs is to be applied to a respective one from successive static regions in successive pictures. In response to determining, based upon statistics on static region(s) in the current picture which are associated with the first QP, that the first QP is selected, compression of the static region is caused based upon the first QP, and in response to determining that the first QP is not selected, compression of the static region is caused based upon a second QP that is greater than the first QP.

Motion-based adaptive quantization

There is described herein a method and system for dynamically controlling the bit rate of a codec during image data compression by analyzing information output by a first module that processes the group of macroblocks as part of the compression scheme, and configuring a second module to process the group of macroblocks according to this analysis.

Growing community of inventors

Dorval, Canada

Mathieu Girard