Variable-Pitch Echo Cancellation for Double-Speech Stability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing acoustic echo cancellation methods struggle to effectively handle double speech situations, leading to bias in channel estimation, divergence of filters, and perceptible residual echoes due to detection delays and sensitivity to local signals.
Innovation Solution
A method utilizing partitioned-block frequency domain NLMS (PBFD-NLMS) adaptive filtering with an optimal BLUE step size estimation, directly deriving necessary statistics from reference and microphone signals, and spectral normalization to ensure robust echo cancellation during double speech.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If adaptive filtering is performed with short-term treatment to track changes in acoustic channel, then the ability to track acoustic path changes is improved, but bias in channel estimation occurs during double speech situations
Solution Approach 1:
The patent applies dynamics by making the filter adaptation state variable - it can be either active or inactive based on double speech detection. The system dynamically adjusts the adaptation process by freezing it during double speech periods (when both near-end and far-end speech are present) and enabling it during single speech periods, thus resolving the contradiction between tracking speed and estimation accuracy.
Solution Approach 2:
The patent uses feedback through double speech detection (DSD) to monitor the acoustic environment and control the adaptive filtering process. The DSD mechanism provides feedback about the presence of double speech, which then feeds back to control whether adaptation should be active or frozen, creating a closed-loop system that resolves the contradiction adaptively.
2Measurement precision
If double speech detection systems are used to disable adaptation during double speech periods, then channel estimation bias is reduced, but detection delays cause echo feedback
Solution Approach 1:
The patent applies preliminary action by continuously monitoring for double speech conditions and being ready to freeze adaptation immediately when double speech is detected. The system prepares the adaptation mechanism in advance to respond quickly, minimizing detection delays and preventing echo feedback by preemptively freezing adaptation when double speech conditions are identified.
3Measurement precision
If adaptation is frozen during double speech periods, then channel estimation bias is reduced, but residual echo becomes perceptible when filter has not converged
Solution Approach 1:
The patent applies dynamics by making the adaptation state controllable and reversible. Rather than permanently freezing adaptation during double speech, the system dynamically adjusts the adaptation process - freezing it temporarily when double speech is detected and resuming it when single speech conditions return. This allows the filter to converge during single speech periods while preventing bias during double speech periods, resolving the contradiction between estimation accuracy and cancellation effectiveness.
4Device complexity
If NLMS technique is used for adaptive filtering, then implementation simplicity is improved, but sensitivity to local signals causes filter divergence
Solution Approach 1:
The patent uses feedback through double speech detection to control the NLMS adaptation process. When double speech is detected, the system provides feedback to freeze adaptation, preventing the filter from diverging due to sensitivity to local signals. This feedback mechanism maintains the simplicity of NLMS while adding the stability control needed to prevent divergence during challenging acoustic conditions.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The processing of a signal y(t) from a microphone (MIC) of a device further comprising at least one loudspeaker (HP) intended to be powered by a signal x(t) aims to limit an echo effect induced by the microphone (MIC) picking up a sound emitted by the loudspeaker (HP) in the surroundings of the device. Said sound emitted by the loudspeaker (HP) and any acoustic reflections thereof follow an acoustic path w from the loudspeaker (HP) to the microphone (MIC). In order to limit the echo effect, the processing comprises determining ŝ(t) a useful signal s(t) by subtracting from the signal y(t) from the microphone (MIC) a given estimate of an echo signal x(t) ∗ ŵ(t) by applying a filter ŵ(t) to the loudspeaker power supply signal x(t). The filter ŵ(t) is adaptive by variable pitch to account for a change in the acoustic path w(t) over time. Since the loudspeaker power supply signal x(t) is obtained in the form of a time sequence of frames of signal samples, the adaptive filter ŵ(t) is produced at each frame k of samples according to an update ΔW(k) of the acoustic path w for said frame k and by applying a normalisation Λ satisfying a selected criterion for minimal variance, the normalisation Λ being a function of a parameter representative of a statistical expectation of the useful signal s(t).