Weighted Noise Injection for Audio Frame Loss Correction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current frame loss correction techniques in digital signal processing, particularly in audio signal decoding, often result in audible phase discontinuities and unpleasant hearing imperfections due to inadequate handling of signal energy and periodicity, leading to artifacts like metallic sounds and energy inconsistencies.

Innovation Solution

A method that generates a replacement signal structure using spectral components from valid samples, injecting blocks with variable parameters to ensure energy homogenization and reduce periodicity effects, thereby smoothing transitions and maintaining consistent signal energy levels.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If frame loss correction is performed by repeating the decoded spectrum in the last received frame, then the technique is inexpensive in terms of resources and does not require additional delay, but it produces audible phase discontinuities and significant audio artifacts

Engineering Contradiction:
Improveprocessing complexityVSAvoidaudio artifacts
Core Design Contradiction:
Device complexityVSObject-generated harmful factors

Solution Approach 1:

The patent applies dynamics by making the frame loss correction adaptive rather than static. The system dynamically adjusts the correction parameters (pitch period, spectral components, energy levels) based on the actual signal characteristics in the lost frame region, allowing the correction to evolve with the signal rather than applying a fixed repetition pattern

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes key parameters including pitch period estimation, spectral components, and energy levels to match the surrounding valid frames. By adjusting these parameters dynamically based on signal analysis, the system maintains spectral and temporal consistency while avoiding the artifacts produced by simple frame repetition

Inventive Principle:
Principle #35Parameter changes

2Loss of time

If the overlap time between frames is reduced to achieve low delay, then the delay is minimized, but the phase discontinuities become more audible and artifacts increase

Engineering Contradiction:
Improvedecoder delayVSAvoidphase discontinuities
Core Design Contradiction:
Loss of timeVSObject-generated harmful factors

Solution Approach 1:

The patent performs preliminary analysis of the signal characteristics (pitch period, spectral components, energy) from valid frames before attempting correction. This preliminary characterization allows the system to pre-calculate appropriate correction parameters that will maintain consistency even with reduced overlap, preventing phase discontinuities before they occur

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system dynamically adjusts the correction strength and parameters based on the actual overlap duration. When overlap is reduced, the correction algorithm adapts by using more aggressive parameter matching and energy adjustment to compensate for the shorter transition zone, maintaining quality despite lower delay

Inventive Principle:
Principle #15Dynamics

3Reliability

If synthesized signal is generated from pitch period structure to replace lost frames, then frame loss correction is achieved, but the physical properties do not correspond to the original signal causing unpleasant hearing imperfections

Engineering Contradiction:
Improveframe loss correctionVSAvoidhearing imperfections
Core Design Contradiction:
ReliabilityVSObject-generated harmful factors

Solution Approach 1:

The patent carefully adjusts parameters including pitch period, spectral envelope, and energy levels to match the characteristics of the original signal. By deriving these parameters from actual signal analysis rather than using fixed synthesized patterns, the correction maintains physical properties consistent with the original audio

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The system copies spectral and temporal characteristics from valid frames to reconstruct the lost frame content. Rather than generating entirely synthetic signals, it copies the essential signal properties (spectral components, pitch structure, energy distribution) from surrounding valid frames, ensuring fidelity to the original signal's physical properties

Inventive Principle:
Principle #26Copying

4Reliability

If energy levels of correctly received signal and reconstructed signal are significantly different, then frame loss correction is achieved, but it causes auditory sensation of jumping with noise level changes

Engineering Contradiction:
Improvesignal reconstructionVSAvoidenergy inconsistencies
Core Design Contradiction:
ReliabilityVSObject-generated harmful factors

Solution Approach 1:

The patent dynamically adjusts energy parameters including overall signal level, spectral energy distribution, and transient energy to match between valid and reconstructed frames. This energy normalization prevents audible jumps by ensuring smooth transitions in noise levels and signal intensity

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The system uses feedback from energy analysis of valid frames to adjust the reconstruction parameters. By measuring the energy characteristics of surrounding valid frames and using this information to guide the reconstruction energy levels, the system maintains energy consistency and eliminates auditory jumping sensations

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP2987165B1Frame loss correction by weighted noise injection
Publication Date: 2018.10.10 ORANGE SA
  • EP2987165B1 patent drawingFigure 1A~1B
  • EP2987165B1 patent drawingFigure 1C
  • EP2987165B1 patent drawingFigure 2

AI summary

The invention relates to a method for processing a digital signal, implemented during decoding of the signal, in order to replace a succession of samples lost during decoding, the method comprising steps of: generating a structure of a signal for replacing the lost succession, this structure comprising spectral components determined from valid samples received during decoding before the succession of lost samples; generating a residue between a digital signal available to the decoder, comprising received valid samples, and a signal generated from the spectral components; and extracting blocks from the residue, method in which window weighted blocks are injected into the structure using an overlap-add approach, the injected blocks partially overlapping in time.