Echo Path Power Estimation With FIR Filtering Under Clock Drift
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing echo cancellation and suppression techniques in audio devices face challenges in accurately estimating echo power, particularly due to clock-drift, nonlinearities, and changes in the echo path, which affect the quality of audio calls by introducing interference.
Innovation Solution
A method using a Finite Impulse Response (FIR) filter dynamically adapted in the time domain to estimate the echo power, allowing for effective echo suppression by modeling the impulse response of the echo path and applying frequency-dependent suppression.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional echo cancellation techniques are used, then echo removal is attempted, but measurement precision of echo power estimation deteriorates due to clock-drift, nonlinearities, and echo path changes
Solution Approach 1:
The patent replaces traditional time-domain adaptive filtering approaches with a frequency-domain power spectral density-based method. Instead of directly estimating echo power in the time domain where clock-drift and nonlinearities cause errors, the invention transforms the problem to frequency domain using power spectral density estimates, which are more robust to these impairments. This substitution of the estimation domain fundamentally improves measurement precision while maintaining reliability.
Solution Approach 2:
The patent introduces power spectral density as an intermediary representation between the raw audio signals and the final echo power estimate. By computing power spectral density of the reference signal and applying it to the filtered received signal, the method creates a stable intermediate representation that is less sensitive to clock-drift and nonlinearities, thereby improving echo power estimation accuracy.
2Reliability
If echo suppression is applied to improve call quality, then near-end speech clarity is improved, but device complexity increases due to need for accurate echo power estimation
Solution Approach 1:
The patent creates a model of the echo path behavior through power spectral density estimation rather than attempting to perfectly reconstruct the actual echo signal. This copied representation captures the essential characteristics needed for suppression without requiring complex real-time echo path tracking, thereby reducing device complexity while maintaining call quality improvement.
Solution Approach 2:
The patent changes the parameters used for echo estimation from time-domain filter coefficients to frequency-domain power spectral density values. This parameter transformation simplifies the estimation process by working with energy distributions across frequencies rather than temporal signal variations, reducing computational complexity while improving robustness to environmental changes.
3Adaptability or versatility
If dynamic echo path modeling is implemented to handle varying conditions, then adaptability improves, but measurement precision deteriorates due to errors in modeling
Solution Approach 1:
The patent performs preliminary power spectral density estimation and filtering before computing the final echo power estimate. By pre-processing the signals to obtain stable spectral characteristics and applying appropriate filtering, the method prepares accurate intermediate values that remain valid even as echo paths vary, thereby maintaining measurement precision across different adaptive conditions.
Solution Approach 2:
The patent maintains continuous updating of power spectral density estimates and echo power calculations as conditions change during the audio call. This continuous adaptation allows the system to track echo path variations while maintaining precision through ongoing spectral analysis rather than relying on periodic or discrete updates that would be less accurate under dynamic conditions.
Data Source
AI summary
Method, user device and computer program product for suppressing echo. An audio signal is output from a speaker. A microphone receives an audio signal, wherein the received audio signal includes an echo resulting from said outputted audio signal. A Finite Impulse Response filter estimate ĥ(t) is dynamically adapted in the time domain based on the outputted audio signal and the received audio signal to model an echo path h(t) of the echo in the received audio signal. At least one power response is determined from the filter estimate ĥ(t) and used to estimate the echo power of the echo in the received audio signal. The estimated echo power is used to apply echo suppression to the received audio signal, thereby suppressing the echo in the received audio signal.