Audio Spectral Bit Allocation for Bitrate-Quality Tradeoffs

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing audio codecs face challenges in achieving an optimal tradeoff between audio quality, complexity, and bitrate, particularly at high bitrates where truncating high-frequency coefficients can severely degrade performance, and at low bitrates where suboptimal global-gain settings lead to inefficient bit usage.

Innovation Solution

An audio decoder and encoder that jointly encode and decode two or more most significant bits per spectral value using arithmetic encoding, while selectively omitting the encoding and decoding of least significant bits based on available bit budget, to maintain good audio quality and reduce bit demand.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If global-gain is increased to reduce the number of bits consumed by the arithmetic encoder, then bitrate is reduced, but the distortion introduced by quantization increases resulting in worse audio quality

Engineering Contradiction:
Improvenumber of bits consumedVSAvoidaudio quality
Core Design Contradiction:
Quantity of substanceVSManufacturing precision

Solution Approach 1:

The patent segments the spectral coefficients into different groups based on their significance and frequency position. Instead of applying a uniform quantization approach to all coefficients, the encoder processes significant coefficients (those above a threshold) differently from insignificant ones, and further divides significant coefficients into high-frequency and low-frequency groups. This segmentation allows selective application of quantization and truncation strategies to optimize the bitrate-quality tradeoff.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies different quantization and encoding strategies to different parts of the spectral data based on their local characteristics. Significant coefficients retain full precision with arithmetic encoding, while insignificant coefficients are truncated to zero. This local differentiation ensures that bits are allocated to the most perceptually important coefficients, maintaining audio quality where it matters most while reducing overall bitrate.

Inventive Principle:
Principle #3Local quality

2Quantity of substance

If high-frequency coefficients are truncated to stay within bit budget, then bitrate is reduced, but signal-to-noise ratio degrades severely at high bitrates

Engineering Contradiction:
ImprovebitrateVSAvoidsignal-to-noise ratio
Core Design Contradiction:
Quantity of substanceVSReliability

Solution Approach 1:

The patent implements a dynamic truncation strategy where the decision to truncate high-frequency coefficients is not fixed but adapts based on the available bit budget and the characteristics of the current audio frame. The encoder calculates the number of significant coefficients and compares it against the bit budget, dynamically adjusting the truncation point. This dynamic approach allows the system to maintain high signal-to-noise ratio when bitrate permits while still achieving compression when needed.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes the parameter of coefficient precision dynamically based on available resources. Instead of using a fixed truncation threshold, the system adjusts the number of bits allocated to different frequency regions based on the bit budget. When bitrate is充足, more coefficients are encoded with full precision; when bitrate is constrained, truncation is applied selectively. This parameter adaptation resolves the contradiction between maintaining high signal-to-noise ratio and achieving efficient compression.

Inventive Principle:
Principle #35Parameter changes

3Manufacturing precision

If rate-loop iterative approach is used to find optimal global-gain, then audio quality is improved, but computational complexity increases

Engineering Contradiction:
Improveaudio qualityVSAvoidcomputational complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent performs preliminary classification of spectral coefficients into significant and insignificant groups before the main encoding process. By pre-identifying which coefficients exceed the quantization threshold, the encoder can make immediate decisions about which coefficients to encode with full precision and which to truncate. This preliminary action eliminates the need for iterative rate-loop adjustments, as the optimal quantization strategy is determined in advance based on the coefficient distribution.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent employs a self-adaptive encoding strategy where the encoder automatically determines the optimal number of coefficients to encode based on the bit budget and coefficient significance without requiring external iterative control. The system serves itself by internally calculating the tradeoff between bitrate and quality and making encoding decisions autonomously. This self-service mechanism replaces the complex iterative rate-loop with a single-pass efficient algorithm.

Inventive Principle:
Principle #25Self-service

Data Source

PatentUS11380339B2Audio encoders, audio decoders, methods and computer programs adapting an encoding and decoding of least significant bits
Publication Date: 2022.07.05 FRAUNHOFER GESELLSCHAFT ZUR FORDERUNG DER ANGEWANDTEN FORSCHUNG EV
  • US11380339B2 patent drawing
  • US11380339B2 patent drawing
  • US11380339B2 patent drawing

AI summary

An audio decoder for providing a decoded audio information on the basis of an encoded audio information is configured to obtain decoded spectral values on the basis of an encoded information representing the spectral values. The audio decoder is configured to jointly decode two or more most significant bits per spectral value on the basis of respective symbol codes for a set of spectral values using an arithmetic decoding, wherein a respective symbol code represents two or more most significant bits per spectral value for one or more spectral values. The audio decoder is configured to decode one or more least significant bits associated with one or more of the spectral values in dependence on how much least significant bit information is available, such that one or more least significant bits associated with one or more of the spectral values are decoded.