Trick-Mode Video Stream Combining Process
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video-on-demand systems face performance limitations when generating and sending trick-mode video streams due to the high number of memory reads and writes required, which restricts the speed and concurrency of trick-mode video stream delivery.
Innovation Solution
A high-performance combining process that uses a network interface circuit to generate trick-mode video streams by combining previously stored video data with newly generated items, reducing the need for memory reads and writes, and utilizing a mapping and directing process to identify and build the necessary items for the stream.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the video server transfers selected spaced-apart portions of the normal video stream from memory into a buffer, modifies items in the buffer, and repeats until the buffer is filled, then a trick-mode video stream can be generated, but a very large amount of video data must be read from and written to memory, inherently limiting the speed of the method
Solution Approach 1:
The patent pre-calculates and stores combining directions in a data structure before actual trick-mode video stream generation is needed. These combining directions contain all the information required to efficiently assemble the trick-mode stream without repeated memory reads and writes during playback, thus resolving the contradiction between generation speed and memory operation time.
Solution Approach 2:
Instead of repeatedly reading and modifying actual video data in memory during trick-mode generation, the patent creates a simplified representation (combining directions) that copies the essential information about what needs to be combined. This allows the system to work with much smaller data structures, dramatically reducing memory read/write operations while maintaining the ability to generate correct trick-mode streams.
2Adaptability or versatility
If the video server uses the buffer-based method to generate trick-mode video streams, then video can be sent at faster or slower speeds, but the large amount of memory data processing restricts the number of concurrent streams
Solution Approach 1:
The patent replaces heavy video data copying with lightweight combining direction copying. By storing only the structural information needed to assemble trick-mode streams (combining directions) rather than actual video data, the system dramatically reduces the resource consumption per concurrent stream, enabling many more simultaneous trick-mode video streams.
Solution Approach 2:
The patent segments the trick-mode video stream generation into two phases: pre-processing where combining directions are calculated and stored, and runtime generation where these directions are executed to assemble the actual stream. This segmentation allows the system to handle multiple concurrent streams efficiently by reusing the pre-calculated directions without repeating the heavy memory operations for each stream.
3Ease of operation
If the video server reads and writes large amounts of video data to memory for trick-mode stream generation, then the video can be modified for different viewing speeds, but the processing time and memory bandwidth are consumed
Solution Approach 1:
The patent copies only the essential control information (combining directions) rather than the actual video data during pre-processing. These combining directions are stored compactly and require minimal memory bandwidth to retrieve and execute during stream generation, thus maintaining the flexibility to modify videos for different viewing speeds while dramatically reducing memory bandwidth consumption compared to processing actual video data.
Data Source
AI summary
A high-performance combining process has use in sending a trick-mode video stream, of a particular video, from a video server to a viewer network. This combining process—1) sends a series of spaced-apart requests to a computer, with each request being for combining directions which indicate how to generate a respective portion of the trick-mode video stream by combining video data that was previously stored in memory before the sending step with selected items that are generated in the computer in response to the sending step; 2) receives the requested combining directions; and 3) uses the directions to combine the selected items with the video data that was previously stored in the memory.