Abstract:In speech telecommunication, where a microphone and a loudspeaker are placed close to each other, users are always annoyed by the delayed echo. Acoustic echo cancellation system is developed to reduce the disturbance of the echo noise. In practical implementation of acoustic echo cancellation, subband technology is used to improve the system performance and reduce the computational complexity of the algorithm. The general subband construction is based on the complex-valued implementation. However, the complex-valued calculation, such as complex-valued multiplication, results in some problems. In this paper, an oversampled real-valued subband algorithm for acoustic echo cancellation is proposed. The proposed algorithm modulates the input fullband signal by general discrete Fourier transformation (GDFT) to shift the focused subband signals to corresponding frequency range. Then the signals in subbands are passed through the proto lowpass filter and decimated to produce the odd type complex-valued subband signals. The complex-valued subband signal pairs are then modulated on single side band by conjugate modulating components to realize the real-valued subband signals. However, the conjugate modulation does not occupy extra computational complexity for the modulation component feature, which can be implemented by fetching their real and imaginary or their opposing ones iteratively. The analysis on the real-valued subband algorithm demonstrates that it avoids complex computations of complex-valued signal in subbands and half computational complexity is cut off. Compared with the subband algorithm in [1] which uses single modulation component to modulate the complex-valued subband signals, the proposed subband algorithm still keeps the conjugate relation between complex-valued subband signal pairs. The simulation results of the algorithm demonstrate that the proposed algorithm has the same performance as the complex-valued subband algorithm and is superior in real-time implementation, high stability and easy realization, and it is very suitable for applications such as acoustic echo cancellation.

Higher Order Cumulants in Echo Cancellation

A low-complexity acoustic echo canceller

Deep Echo Cancellation Algorithm Based on Time-Frequency Domain Combination

Efficient Echo-Cancellation Algorithms for Full Duplex Coherent Optical Systems

Echo Canceller Implementation On Ip Telephony Gateway

Improved Echo Hiding Using Power Cepstrum and Simulated Annealing Based Synchronization Technique

Simulation and Research on Adaptive Echo Cancellation Algorithm

Echo Cancellation Algorithm Based on Two-Filters

Oversample Real-valued Subband Structure for Acoustic Echo Cancellation

Adaptive Sub-Band Acoustic Echo Cancellation Algorithms for Echo Paths Involving Nonlinear Amplifiers and Transducers

An Improved Acoustic Echo Canceller Combined with LPC-Based Speech Coder

Acoustic Echo Cancellation using Deep Complex Neural Network with Nonlinear Magnitude Compression and Phase Information

An Improved Affine Projection Algorithm and Its Application in Network Echo Cancellation

A Double-talk Robust and Fast Converging Echo Canceller

Icassp 2021 acoustic echo cancellation challenge: integrated adaptive echo cancellation with time alignment and deep learning-based residual echo plus noise suppression

A dual-filters echo canceller using adaptive threshold to handle double talk

Integrating acoustic echo cancellation in LPC-based speech coders

Acoustic Echo Cancellation by Combining Adaptive Digital Filter and Recurrent Neural Network

Echo Canceller Integrating Diverse Adaptive Filtering Algorithms

Echo Cancellation Implementation with Long Time Delay

Realization of Echo Cancellation on DSP