Audio Coding Packet Buffering for Multiple-Description Streams

The patent segments the buffer into multiple buffer spaces, where each buffer space corresponds to a specific timestamp. Within each buffer space, multiple audio coding data packets with different description indices can be stored. This segmentation allows the system to maintain separate tracks for different descriptions at the same timestamp, preventing data loss while managing complexity through structured organization.

The patent introduces a new dimension of organization by adding the description index as an additional sorting and storage dimension. Instead of only buffering packets based on timestamp, the system now buffers packets based on both timestamp and description index. This dimensional expansion allows multiple packets with the same timestamp but different description indices to coexist in the buffer without overwriting each other.

The patent makes the audio jitter buffer universal by enabling it to handle both single description coding (SDC) and multiple description coding (MDC) bitstreams. The buffer can accommodate packets from different coding modes using the same timestamp and description index sorting mechanism. This multi-functionality allows the system to maintain processing simplicity while gaining adaptability to various coding scenarios.

The patent changes the buffering parameters from solely timestamp-based to timestamp and description index-based sorting. This parameter change enables the buffer to distinguish between different descriptions at the same timestamp, allowing support for MDC bitstreams while maintaining the simplicity of the existing buffer architecture through a straightforward sorting mechanism.

The patent segments the buffer into multiple buffer spaces organized by timestamp, with each buffer space capable of holding multiple packets with different description indices. This segmentation prevents packet loss by ensuring that packets with the same timestamp but different descriptions are stored separately, thereby maintaining audio quality while keeping buffer management organized and simple.

The patent performs preliminary sorting and organization of packets into buffer spaces before actual audio processing. By pre-sorting packets based on timestamp and description index, the system ensures that all necessary packets are available for decoding without complex real-time management during audio playback, thus maintaining both simplicity and reliability.