Grouping of image frames in video coding

What is this patent about?

’134 is related to the field of compressed video decoding, specifically addressing the management of reference picture buffers during the playback of streaming multimedia. In modern video compression, decoders rely on a sequence of numbered frames to maintain synchronization and perform temporal prediction. However, when frames are intentionally omitted—such as during bit rate adjustments, scene transitions, or the insertion of secondary content like commercials—the resulting gaps in frame numbering often trigger erroneous error-correction routines or buffer synchronization failures in standard decoders.

The underlying idea behind ’134 is to provide a signaling mechanism that distinguishes between accidental packet loss and intentional frame removal, allowing the decoder to maintain a synchronized state without performing unnecessary recovery actions. By embedding an explicit indication of an intentional discontinuity within the bitstream, the inventor realized that a decoder could be instructed to proactively manage its buffer state. This insight ensures that the reference picture indexing remains consistent between the encoder and decoder, even when the physical transmission of certain frames is bypassed to save bandwidth or handle network congestion.

The claims of ’134 focus on a method and apparatus for decoding a video sequence by identifying an indication of an intentional numbering discontinuity and reconfiguring a buffer memory in response. Specifically, the independent claims cover the process of decoding this indication from the bitstream and subsequently configuring the buffer to provide a specific number of frames that correspond to the gap in numbering. These provided frames are then utilized within the decoding process to ensure that subsequent inter-predicted frames can correctly reference the expected indices in the buffer.

In practice, when the decoder encounters the intentional discontinuity signal, it populates its buffer with filler frames or memory indicators designated as invalid. This prevents the decoder from interpreting the missing sequence numbers as a protocol error or a network failure. By filling these gaps with placeholder data, the sliding window or adaptive memory control remains aligned with the encoder’s state. This mechanism is particularly useful in scalable coding environments where enhancement layers or specific sub-sequences are dropped to match the available transmission bandwidth or the processing limits of the terminal device.

This approach differs from prior solutions that relied on rigid, consecutive frame numbering where any gap was treated as a transmission failure requiring retransmission or complex error concealment. By utilizing sub-sequence identifiers and intentional discontinuity flags, the invention allows for flexible traffic shaping at the server or network level without breaking the decoder's temporal prediction chain. This ensures a fluid playback experience even during drastic scene transitions or bit rate fluctuations, as the decoder is always aware of the structural integrity of the received frame sequence.