Opera Software Ireland Limited

Performance and Privacy Ireland Limited

Vadim Arshanskiy

Erik R Swenson

Jaime Heilpern

Ron Earley

Sandeep Pombra

Nitin Bhandari

A video optimizer receives a request to optimize a source file and stream the optimized file. The request may be accompanied by a key uniquely identifying the original source file to be optimized and an optimized version of the source file to stream to the client. The key uniquely identifies the source video file as it exists on an origin server to prevent transmission of modified source video and the optimized version of the source file by optimization parameters to ensure compatibility with the user device. The video optimizer queries a cache database in response to the received request, which contains references keys and locations of optimized source files among local caches in a cluster of video optimizers. The cache database responds indicates whether a matching optimized version of the source file exists in a local cache of a given video optimizer in the cluster.

Just-in-time distributed video cache

A system and a method are disclosed for transcoding video files on a server such that they can be streamed to a client device in substantially real-time. The server receives a video file a uses a prediction algorithm to predict each video frame's size based on encoding and streaming parameters associated with the input file and an output stream and transcoding each video frame. If a transcoded frame is too large, it is compressed further, if it is too small, the video file is padded. As such, each video frame transcoded or padded until the encoded frame is the same size as a predicted size. The transcoded frames are streamed to a client device to enable the client to device to playback the video after requesting it from the server.

Real-time audio or video transcoding

Growing community of inventors

Santa Clara, CA, United States of America

Vadim Arshanskiy