Audio Decoder Transition Between Transform and Predictive Coding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing audio coding technologies face complexity and inefficiency in transitioning between transform coding and predictive coding modes, particularly in conversational applications, due to the need for adaptive dictionary updates and resampling, which increases processing complexity and can result in artifacts.

Innovation Solution

A method for decoding digital audio signals that involves inverse transform decoding of previous frames and predictive decoding of current frames without using an adaptive dictionary from the previous frame, with reinitialization of predictive decoding states to predetermined default values, and an addition-recovery step combining signal segments from both decoding methods to simplify transitions and reduce complexity.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If adaptive dictionary updates and resampling are performed during mode transitions, then coding accuracy is improved, but processing complexity increases

Engineering Contradiction:
Improvecoding accuracyVSAvoidprocessing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies preliminary action by pre-calculating and storing transition frames in advance during transform coding modes. When switching to predictive coding mode, these pre-computed transition frames are directly utilized, eliminating the need for real-time adaptive dictionary updates and resampling operations. This approach maintains coding accuracy while significantly reducing processing complexity during mode transitions.

Inventive Principle:
Principle #10Preliminary action

2Stability of the object's composition

If adaptive dictionary updates are performed during mode transitions, then transition smoothness is improved, but processing time increases

Engineering Contradiction:
Improvetransition smoothnessVSAvoidprocessing time
Core Design Contradiction:
Stability of the object's compositionVSLoss of time

Solution Approach 1:

The patent computes transition frames in advance during transform coding and stores them for later use during predictive coding mode. This preliminary computation ensures smooth transitions between coding modes without requiring time-consuming adaptive dictionary updates at the moment of mode switching, thus maintaining transition smoothness while reducing processing time.

Inventive Principle:
Principle #10Preliminary action

3Measurement precision

If resampling is performed during mode transitions, then frequency accuracy is improved, but computational load increases

Engineering Contradiction:
Improvefrequency accuracyVSAvoidcomputational load
Core Design Contradiction:
Measurement precisionVSPower

Solution Approach 1:

The patent performs resampling operations in advance during transform coding to generate transition frames with accurate frequency characteristics. These pre-resampled transition frames are then reused during predictive coding mode transitions, maintaining frequency accuracy while avoiding the computational burden of real-time resampling operations.

Inventive Principle:
Principle #10Preliminary action

4Manufacturing precision

If complex transition procedures are implemented, then transition quality is improved, but algorithmic delay increases

Engineering Contradiction:
Improvetransition qualityVSAvoidalgorithmic delay
Core Design Contradiction:
Manufacturing precisionVSLoss of time

Solution Approach 1:

The patent generates and stores transition frames in advance during transform coding modes. During predictive coding mode transitions, these pre-prepared transition frames are directly applied, ensuring high transition quality without requiring complex real-time processing procedures that would increase algorithmic delay.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3069340B1Transition from a transform coding/decoding to a predictive coding/decoding
Publication Date: 2017.09.20 ORANGE SA
  • EP3069340B1 patent drawingFigure 1
  • EP3069340B1 patent drawingFigure 2~4
  • EP3069340B1 patent drawingFigure 3

AI summary

The invention pertains to a method of decoding a digital audio signal, comprising the steps of decoding (E602) according to an inverse transform decoding of a previous frame of samples of the digital signal, which frame is received and coded according to a transform coding, decoding (E608) according to a predictive decoding of a current frame of samples of the digital signal, which frame is received and coded according to a predictive coding. The predictive decoding of the current frame is a transition predictive decoding which does not use any adaptive dictionary arising from the previous frame and the method furthermore comprises a step of reinitialization (E606) of at least one state of the predictive decoding to a predetermined default value, an add-overlap step (E609) which combines a signal segment synthesized by predictive decoding of the current frame and a signal segment synthesized by inverse transform decoding, corresponding to a stored segment of the decoding of the previous frame. The invention pertains correlatively to a method of coding comprising a reinitialization of at least one state of the predictive coding to a predetermined default value. It pertains to a coder and decoder implementing the respective methods.