System and method for efficiently processing digital video

A system for processing video information, the system comprising:

• a first memory configured to store video frames from a stream of digitally encoded video;
• and
• a video processing pipeline, comprising:
• a frame reader configured to read a series of video frames from the first memory into First-In First-Out (FIFO) queues including reading luminance information into a first FIFO queue and reading chrominance information into a second FIFO queue;
• a simultaneous resizing and chroma upsampling stage configured to simultaneously resize and chroma upsample video frames from a native video resolution to a target resolution for a destination video window, the simultaneous resizing and chroma upsampling stage further configured to resize the luminance information of the first FIFO queue and to resize and chroma upsample the chrominance information of the second FIFO queue without accessing an external memory being external to the video processing pipeline;
• a color space conversion stage configured to perform color space conversion on each pixel in the resized and chroma upsampled video frames without accessing the external memory;
• and
• an output queue configured to output a series of pixel data bits representing an output series of video frames of the digitally encoded video in the destination video window without accessing the external memory.

A digital video system for rendering a video signal, the digital video system comprising:

• a screen buffer configured to store a digital representation of a computer display having at least one application window, one of the application windows comprising a full-motion video window;
• a video decoder configured to decode a digital video stream into video frames stored in a video buffer;
• a video processing pipeline configured to, without accessing an external memory external to the video processing pipeline:
• read the video frames from the video buffer including reading luminance information of the video frames into a first queue of the video processing pipeline and reading chrominance information of the video frames into a second queue of the video processing pipeline;
• simultaneously resize and chroma upsample the video frames from a native video resolution to a target resolution for the full-motion video window including resizing the luminance information and resizing and chroma upsampling the chrominance information;
• and
• output a series of pixel bits for display in the full-motion video window;
• and
• a video display driver configured to combine the digital representation of the computer display with the series of pixel data bits in an output video signal.

A method of rendering a display on a terminal device, the method being performed by a computing device, the method comprising:

• maintaining a screen buffer in the terminal device, the screen buffer comprising a digital representation of a computer display having at least one application window, one of the application windows comprising a full-motion video window;
• receiving a digital video stream in the terminal device;
• decoding the digital video stream into video frames stored in a video buffer;
• reading the video frames into a video processing pipeline including reading luminance information of the video frames into a first queue of the video processing pipeline and reading chrominance information of the video frames into a second queue of the video processing pipeline;
• simultaneously resizing and chroma upsampling the video frames from a native video resolution to a target resolution for the full-motion video window including resizing the luminance information and resizing and chroma upsampling the chrominance information;
• outputting a series of pixel data bits from the video processing pipeline for display in the full-motion video window;
• and
• combining the digital representation of the computer display with the series of pixel data bits in an output video signal;
• wherein the simultaneously resizing and chroma upsampling and outputting are performed without accessing a memory external to the video processing pipeline.