Dynamic encoding parameters for low latency streaming
AssigneeSony Interactive Entertainment Inc.
InventorEric Chen
Life of this patent
Estimating the term…
Abstract
A method implemented in a cloud gaming system having at least one server computer is provided, including the following operations: executing a game session of a video game, wherein the execution of the game session renders gameplay video and generates game state data; using a machine learning (ML) model to determine encoding parameter settings based on the game state data; applying, by an encoder, the encoding parameter settings for processing of the gameplay video to generate compressed gameplay video; streaming the compressed gameplay video over a network to a client device.
Claims20
A method, comprising:
- executing a game session of a video game, wherein the execution of the game session renders gameplay video and generates game state data;
- providing the game state data as input to a machine learning (ML) model;
- predicting, using the machine learning (ML) model, a future in-game scene complexity based on the game state data;
- identifying, based on the game state data, a game interval associated with the future in-game scene complexity;
- determining, using the machine learning (ML) model, encoding parameter settings based on the future in-game scene complexity, wherein the encoding parameter settings correspond to the game interval;
- applying, by an encoder, the encoding parameter settings for processing of the gameplay video to generate compressed gameplay video corresponding to the game interval;
- and
- streaming the compressed gameplay video over a network to a client device.
One or more non-transitory computer-readable media storing computer-readable instructions that, when executed by one or more processors, to cause a system to perform operations comprising:
- executing a game session of a video game, wherein the execution of the game session renders gameplay video and generates game state data;
- providing the game state data as input to a machine learning (ML) model;
- predicting, using the machine learning (ML) model, a future in-game scene complexity based on the game state data;
- identifying, based on the game state data, a game interval associated with the future in-game scene complexity;
- determining, using the machine learning (ML) model, encoding parameter settings based on the future in-game scene complexity, wherein the encoding parameter settings correspond to the game interval;
- applying, by an encoder, the encoding parameter settings for processing of the gameplay video to generate compressed gameplay video corresponding to the game interval;
- and
- streaming the compressed gameplay video over a network to a client device.
A system comprising:
- a processing system;
- and
- a memory configured to store computer-executable instructions;
- and
- a processor configured to access the memory and execute the computer-executable instructions to at least:
- execute a game session of a video game, wherein the execution of the game session renders gameplay video and generates game state data;
- provide the game state data as input to a machine learning (ML) model;
- predict, using the machine learning (ML) model, a future in-game scene complexity based on the game state data;
- identify, based on the game state data, a game interval associated with the future in-game scene complexity;
- p2 determine, using the machine learning (ML) model, encoding parameter settings based on the future in-game scene complexity, wherein the encoding parameter settings correspond to the game interval;
- apply, by an encoder, the encoding parameter settings for processing of the gameplay video to generate compressed gameplay video corresponding to the game interval;
- and
- stream the compressed gameplay video over a network to a client device.
Dependencies are inferred from the claim text.