Media File System Essence Data Segmentation for Indexing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing media file systems require the entire file to be available before indexing can be generated, leading to inefficiencies in storage and retrieval, especially in streaming applications where only a specific time range of the media is needed, and existing solutions like fragmented MPEG-4 files still require the entire file to be transferred before indexing is available.
Innovation Solution
Storing essence data in multiple media files, each associated with a subperiod of time, with filenames indicative of the time period, allowing for quick generation and retrieval of specific media segments without needing the entire file, and using optimized file system features for indexing and copying.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the entire essence data file is stored as a single media file, then the file structure is simple, but the index cannot be generated until the entire file is available, causing delays in retrieval
Solution Approach 1:
The patent divides the essence data into multiple media files, each corresponding to a specific time period (e.g., 0-10 seconds, 10-20 seconds). Each media file contains its own index, allowing indexes to be generated independently and in parallel rather than waiting for the entire file to be processed. This segmentation resolves the contradiction by maintaining relatively simple file structures while dramatically reducing index generation time.
2Speed
If the index is positioned at the start of the file, then access to the index is immediate, but the entire file must be created in two passes and the GOPs may be deleted before they can be used
Solution Approach 1:
By segmenting the essence data into multiple smaller media files with indexes at their respective starts, the system allows immediate index access for each segment while avoiding the need for two-pass file creation. Each media file can be independently created and used, eliminating the productivity loss associated with waiting for complete file creation and the risk of GOP deletion.
3Stability of the object's composition
If the index is positioned at the end of the file, then the entire file structure can be established first, but the entire file is required before the index is available, especially in HTTP streaming
Solution Approach 1:
The patent segments the file into multiple smaller media files, each with its index at the start. This maintains stable file structures while eliminating the time delay associated with waiting for the entire file to be transferred before the index becomes available. Each segment can be independently accessed and indexed, resolving the contradiction between structural stability and access time.
4Device complexity
If large essence data files are stored as single media files, then storage management is simplified, but retrieval of specific time ranges requires processing the entire file
Solution Approach 1:
The patent segments large essence data files into multiple smaller media files organized by time periods. This allows retrieval of specific time ranges by accessing only the relevant segments rather than processing entire files, dramatically improving retrieval speed while maintaining organized and manageable storage structures through consistent naming conventions and directory organization.
Solution Approach 2:
Each media file segment is optimized for its specific time period with local indexes and metadata, allowing rapid access to specific portions of the content. This local optimization enables efficient retrieval without requiring global file processing, resolving the contradiction between storage simplicity and retrieval efficiency.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method, computing device, computing program product, and apparatus are disclosed. The method is for storing essence data in a media file system and subsequently retrieving essence data from the media file system. The method comprises the steps of receiving essence data associated with a time period, and generating a plurality of media files, wherein each media file is associated with a subperiod of the time period of the received essence data. Each media file comprises one or more portions of essence data derived from the part of the received essence data associated with the subperiod, and an index associating times of the subperiod with locations of essence data within the media file. The method further comprises determining a filename for each media file, wherein the filename for each media file is indicative of a time of the subperiod associated with the media file, and storing the media files in the media file system with the determined filenames. Furthermore the method includes retrieving essence data from the media file system. The retrieval may comprise receiving a request for essence data, the request being indicative of a time and determining a media file, of the plurality of media files, in which the essence data is stored. The determination is done by matching the filename of the media file to the time indicated by the request, and returning the requested essence data from the determined media file.