Multimedia Stream Segmentation for Fast Random Access
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing multimedia services experience significant delays and deterioration in Quality of Experience (QoE) due to buffering issues when users change reproduction positions or content, particularly during fast forward or random access operations, as they require substantial data buffering and result in start-up delays.
Innovation Solution
A method and apparatus that utilize a specialized multimedia stream encoded in small units, such as short Group of Pictures (GOP) and Intra frame periods, to minimize delays by enabling fast buffering and rapid content switching, using an additional multimedia stream for random access and trick mode operations, and employing an extended index (sidx) for efficient buffer management and data retrieval.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If a conventional HTTP-based multimedia streaming service is used with standard buffering, then the service can be provided with basic functionality, but start-up delay increases significantly when changing reproduction position or content
Solution Approach 1:
The multimedia stream is segmented into small units called Elementary Streaming Units (ESUs), where each ESU contains a small number of Access Units (AUs) that can be independently decoded. This segmentation allows the client to buffer and play back content in small, manageable chunks, enabling fast forward and random access operations with minimal delay. The stream is divided into multiple ESUs, each containing a limited number of AUs, allowing incremental buffering and immediate playback when changing reproduction positions.
Solution Approach 2:
The server performs preliminary segmentation of the multimedia stream into ESUs and embeds indexing information (sidx) within each ESU that indicates the number of AUs contained. This pre-organization of data structures and metadata allows the client to efficiently manage buffering without requiring large initial buffers, as the indexing information enables precise control over data retrieval and playback timing.
2Loss of time
If a specialized multimedia stream with small encoding units is used, then start-up delay is reduced and QoE is improved, but the device complexity increases due to additional stream processing requirements
Solution Approach 1:
The indexing structure (sidx) embedded in each ESU serves multiple functions: it provides information about the number of AUs in the ESU, enables buffer management, supports random access operations, and facilitates trick mode operations. This multi-functional indexing approach reduces the need for separate complex processing mechanisms while achieving fast buffering and content switching.
Solution Approach 2:
Each ESU is self-contained with embedded indexing information that allows it to be independently processed and played back. The client can buffer and play ESUs autonomously without requiring complex server-side coordination during playback operations, as each unit carries its own structural information for independent decoding and presentation.
3Loss of energy
If standard buffering thresholds are used in HTTP-based streaming, then bandwidth utilization is efficient, but QoE deteriorates during fast forward and random access operations
Solution Approach 1:
The buffering mechanism is made dynamic by allowing the client to buffer ESUs incrementally based on playback progress rather than requiring a fixed large buffer. The small ESU structure enables the buffer to be refilled in small units during fast forward operations, maintaining efficient bandwidth utilization while ensuring continuous playback capability during trick mode operations.
Data Source
Figure 1
Figure 2
Figure 3a
AI summary
A method and an apparatus for reducing a deterioration of a Quality of Experience (QoE) in a multimedia service. The method includes reproducing a first multimedia stream by the multimedia device, receiving a command for reproducing a second multimedia stream during the reproducing of the first multimedia stream, and reproducing the second multimedia stream with reference to information regarding a unit image that is independently decoded from the second multimedia stream. The information regarding the unit image includes information regarding a number of unit images per unit time included in the second multimedia stream.