Hybrid Transform-Time Domain Codebook for CELP Audio Encoding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

ACELP codebooks face limitations in quality gain at higher bit rates compared to transform coding and vector quantization, particularly in controlling bit allocation across frequency components of sound signals, leading to suboptimal performance at bit rates above 16 kbits/s.

Innovation Solution

A modified CELP model incorporating an additional transform-domain codebook stage that encodes transform-domain coefficients, allowing for better bit allocation and quality improvement by combining time-domain and transform-domain codebook searches based on signal characteristics and bit rate, with a selector determining the order of codebooks for optimal performance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If ACELP codebook size is increased to improve quality at higher bit rates, then quality gain is limited, but device complexity and bit rate increase without proportional quality improvement

Engineering Contradiction:
Improvequality gainVSAvoidcodebook size
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The codebook is divided into two separate codebooks: a time-domain CELP codebook and a transform-domain codebook. This segmentation allows each codebook to specialize in different domains, with the transform-domain codebook handling frequency component allocation more efficiently, thereby achieving better quality gain without proportionally increasing overall codebook size and complexity

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a transform-domain dimension alongside the traditional time-domain approach. By adding this new dimension (transform domain), the system can allocate bits across frequency components more effectively, achieving superior quality at higher bit rates without simply increasing the size of the original time-domain codebook

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Measurement precision

If transform coding and vector quantization are used instead of ACELP, then quality gain at higher bit rates is improved, but device complexity increases

Engineering Contradiction:
Improvequality gain at higher bit ratesVSAvoidcoding structure
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent merges the advantages of both ACELP and transform coding by combining a time-domain CELP codebook with a transform-domain codebook in a unified hybrid structure. This merging allows the system to achieve the quality gains of transform coding while retaining the efficiency benefits of ACELP, without fully adopting the more complex transform coding framework

Inventive Principle:
Principle #5Merging (Combining)

3Productivity

If ACELP model is used for encoding, then encoding efficiency is good, but control over bit allocation across frequency components is insufficient

Engineering Contradiction:
Improveencoding efficiencyVSAvoidbit allocation control
Core Design Contradiction:
ProductivityVSAdaptability or versatility

Solution Approach 1:

The patent introduces dynamic adaptability by making the codebook structure flexible and selectable based on signal characteristics and bit rate. The system can dynamically choose between time-domain and transform-domain codebooks or combine them, enabling adaptive bit allocation control across frequency components while maintaining encoding efficiency

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP2707687B1Transform-domain codebook in a CELP coder and decoder
Publication Date: 2018.03.28 VOICEAGE CORPORATION
  • EP2707687B1 patent drawingFigure 1
  • EP2707687B1 patent drawingFigure 2
  • EP2707687B1 patent drawingFigure 3

AI summary

A codebook arrangement for use in coding an input sound signal comprises first and second codebook stages. The first codebook stage includes one of a time-domain CELP codebook and a transform-domain codebook. The second codebook stage follows the first codebook stage and includes the other of the time-domain CELP codebook and the transform-domain codebook. A third codebook stage comprising an adaptive codebook may be provided before the first codebook stage. A selector may be provided to select an order of the time-domain CELP codebook and the transform-domain codebook in the first and second codebook stages, respectively, as a function of characteristics of the input sound signal. The selector may also be responsive to both the characteristics of the input sound signal and a bit rate of the codec using the codebook arrangement to bypass the second codebook stage. The codebook arrangement can be used in a coder of an input sound signal.