Method and device for indicating quantizer parameters in a video coding system

Patent No. US8175148 (titled "Method and device for indicating quantizer parameters in a video coding system") on Jul 26, 2007. The application was issued on May 8, 2012.

What is this patent about?

’148 is related to the field of digital video compression, specifically the efficient signaling of quantization parameters (QP) within an encoded bit-stream. In standard video codecs, the QP determines the precision of transform coefficients, directly balancing image quality against the required data rate. Because video content varies, encoders often adjust the QP at the picture or slice level to maintain a target bitrate; however, repeatedly transmitting these absolute values consumes significant bandwidth, particularly in low-bitrate environments like mobile telecommunications.

The underlying idea behind ’148 is to replace the repetitive transmission of absolute quantization values with a hierarchical signaling structure based on a sequence level quantization parameter (SQP). Instead of treating every slice or frame as an independent data point, the invention establishes a global default reference for the entire video sequence. By signaling only the statistical difference (ΔQP) between this sequence-level baseline and the actual QP used for a specific segment, the encoder significantly reduces the overhead required for control information.

The claims of ’148 focus on a method and apparatus for defining a default level of quantization at the sequence level and providing an indication of this default to the decoding process. The independent claims specifically cover the application of this default to transform coefficients derived from motion-compensated prediction error values. This architecture ensures that the decoder can reconstruct the specific quantization level for any given block by combining the globally defined sequence parameter with localized updates provided in the bit-stream.

In practice, the invention works by inserting the SQP into a sequence header at the start of a video stream. For each subsequent slice, the encoder evaluates whether the default is sufficient or if a deviation is necessary. If the QP remains constant, the system can signal the use of the default with as little as a single bit, rather than the five or six bits typically required for an absolute value. When changes are needed, the encoder transmits a small signed integer representing the delta, which is typically smaller and more compressible than the full parameter.

This approach differs from prior solutions, such as H.26L, which generally signaled absolute QP values in every picture or slice header. By leveraging the fact that quantization levels often remain stable across a sequence, the invention achieves a reduction in transmission bit-rate without sacrificing image quality. This efficiency is particularly critical for mobile multimedia terminals where radio link bandwidth is limited, allowing more of the available data budget to be allocated to actual visual information rather than redundant control metadata.

How does this patent fit in bigger picture?

Technical Landscape

In the early 2000s when ’148 was filed, digital video coding systems were typically implemented using block-based motion compensation and transform coding where quantization parameters were defined at the picture or slice level. At a time when systems commonly relied on transmitting absolute quantization values for each discrete segment of a video stream, the overhead associated with signaling these parameters consumed a significant portion of the available bandwidth. When hardware and software constraints made high-bitrate transmission non-trivial for mobile or bandwidth-limited applications, the repetitive signaling of full-scale quantization values created an engineering bottleneck that limited overall compression efficiency.

Prosecution Position

The disclosed invention represents a technical advancement by shifting the signaling architecture from absolute per-segment quantization values to a hierarchical delta-based scheme. By integrating a sequence-level default quantization parameter that serves as a global reference, the system enables the transmission of statistically smaller difference values for individual frames or slices rather than full-bit-depth parameters. This architectural shift overcomes the constraint of high signaling overhead, achieving a reduction in transmission bit-rate while maintaining the capability to adjust quantization levels dynamically at the frame or segment level through additive reconstruction at the decoder.

Claims

The patent contains 23 claims, with claims 1, 12, and 23 serving as the independent claims. These independent claims focus on a method and corresponding encoder systems for processing digital video sequences by applying motion compensated prediction and transform coding, specifically emphasizing the definition and communication of a default quantization level for the encoding process. The dependent claims serve to elaborate on the transmission of the bit-stream, the dynamic updating of quantization levels, the use of differential values to indicate deviations from the default quantization level at the frame or segment level, and the implementation of the encoder within specific hardware environments like multimedia terminals and radio telecommunications devices.

Key Claim Terms New

Definitions of key terms used in the patent claims.

Term (Source)Support for SpecificationInterpretation
Default level of quantization
(Claim 1, Claim 12, Claim 23)
The present invention improves prior art solutions for indicating QP related information by introducing a sequence level QP. This allows the encoder application to decide a video sequence dependent reference QP to be used in coding of the picture/slice QPs. Instead of coding the absolute values of picture/slice QPs, it is enough to code the difference between the reference sequence QP and the actually used picture/slice QP.A sequence-level reference quantization parameter (QP) used as a baseline for the entire video sequence or application, rather than an absolute value for each individual picture or slice.
Encoded video bit-stream
(Claim 1, Claim 12, Claim 23)
An indication of the default level of quantization is provided in the encoded bit-stream representative of the digital video sequence. The encoded bit-stream, including the indication of the default level of quantization to be used throughout encoding of the digital video sequence, is transmitted to a video decoding device.The compressed digital output containing the video data and the indication of the default quantization level for transmission to a decoder.
Motion compensated prediction
(Claim 1, Claim 12, Claim 23)
The method comprises encoding a frame of the digital video sequence by applying motion compensated prediction to blocks of pixels, thereby producing corresponding blocks of prediction error values. A transform coding technique is applied to the blocks of prediction error values to produce sets of transform coefficient values representative of the blocks of prediction error values.A technique used to encode frames by predicting pixel blocks based on other frames to reduce temporal redundancy, resulting in prediction error values.
Prediction error values
(Claim 1, Claim 12, Claim 23)
The method comprises encoding a frame of the digital video sequence by applying motion compensated prediction to blocks of pixels, thereby producing corresponding blocks of prediction error values. A transform coding technique is applied to the blocks of prediction error values to produce sets of transform coefficient values representative of the blocks of prediction error values.The residual difference between the original pixel blocks and the predicted blocks generated through motion compensation.
Transform coefficient values
(Claim 1, Claim 12, Claim 23)
A transform coding technique is applied to the blocks of prediction error values to produce sets of transform coefficient values representative of the blocks of prediction error values and a level of quantization is applied to the sets of transform coefficient values to yield sets of quantized transform coefficient values.Data values resulting from applying a mathematical transform (such as DCT) to prediction error values, which are subsequently quantized.

Litigation Cases New

US Latest litigation cases involving this patent.

Case NumberFiling DateTitle
0:24-cv-04269Nov 25, 2024Element Television Company, Llc V. Nokia Corporation
1:23-cv-01236Oct 31, 2023Nokia Technologies Oy V. Amazon.Com, Inc.

Patent Family

Patent Family

File Wrapper

The dossier documents provide a comprehensive record of the patent's prosecution history - including filings, correspondence, and decisions made by patent offices - and are crucial for understanding the patent's legal journey and any challenges it may have faced during examination.

  • Get instant alerts for new documents

US8175148

SEP
Application Number
US11881367A
Filing Date
Jul 26, 2007
Status
Expired
Publication Date
May 8, 2012
External Links
Slate, USPTO , Google Patents