Neural Cascade Architecture for Multi-Channel Acoustic Echo Suppression

Zhang, Hao; Wang, De Liang

doi:10.1109/taslp.2022.3192104

Cited by 18 publications

(6 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, it has been suggested that ancient rock art is often created at places that can generate maximal echo intensities [ 42 ], and there are a number of places that are famous for generating echoes, such as the echo wall at the Temple of Heaven in Beijing and the whispering gallery of St Paul’s Cathedral in London. During modern teleconferencing, echoes are frequently encountered: An echo is generated when the voice from one side of the conversation is picked up by the microphone on another side and transmitted back [ 32 ]. In real teleconference recordings, echoes almost always have >100 ms latency [ 43 ].…”

Section: Discussionmentioning

confidence: 99%

“…In daily reverberant environments, however, the power of sound reflection decays exponentially as the delay of the reflection increases [ 31 ]. Echoic environments are prevalent during online conferencing [ 32 ]. Previous studies have shown that the influence of reverberation on speech envelope can be compensated through basic neural adaptation mechanisms [ 9 , 33 ].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Original speech and its echo are segregated and separately processed in the human brain

Gao,

Chen,

Fang

et al. 2024

PLoS Biol

View full text Add to dashboard Cite

Speech recognition crucially relies on slow temporal modulations (<16 Hz) in speech. Recent studies, however, have demonstrated that the long-delay echoes, which are common during online conferencing, can eliminate crucial temporal modulations in speech but do not affect speech intelligibility. Here, we investigated the underlying neural mechanisms. MEG experiments demonstrated that cortical activity can effectively track the temporal modulations eliminated by an echo, which cannot be fully explained by basic neural adaptation mechanisms. Furthermore, cortical responses to echoic speech can be better explained by a model that segregates speech from its echo than by a model that encodes echoic speech as a whole. The speech segregation effect was observed even when attention was diverted but would disappear when segregation cues, i.e., speech fine structure, were removed. These results strongly suggested that, through mechanisms such as stream segregation, the auditory system can build an echo-insensitive representation of speech envelope, which can support reliable speech recognition.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Original speech and its echo are segregated and separately processed in the human brain

Gao,

Chen,

Fang

et al. 2024

PLoS Biol

View full text Add to dashboard Cite

show abstract

“…With recent advances in deep neural networks, deep learningbased methods [5,6,7] have been utilized for AEC, and their ability to model nonlinear relations leads to promising results, even in challenging noisy or double-talk scenarios. Such methods usually treat AEC as a source separation problem and directly estimate the nearend signal based on the microphone and far-end reference signal.…”

Section: Introductionmentioning

confidence: 99%

NeuralKalman: A Learnable Kalman Filter for Acoustic Echo Cancellation

Zhang¹,

Meng²,

Zhang³

et al. 2023

Preprint

View full text Add to dashboard Cite

The Kalman filter is widely used for addressing acoustic echo cancellation (AEC) problems due to their robustness to double-talk and fast convergence. However, the inability to model nonlinearity and the need to tune control parameters cast limitations on such adaptive filtering algorithms. In this paper, we integrate the frequency domain Kalman filter (FDKF) and deep neural networks (DNNs) into a hybrid method, called NeuralKalman, to leverage the advantages of deep learning and adaptive filtering algorithms. Specifically, we employ a DNN to estimate nonlinearly distorted far-end signals, a transition factor, and the nonlinear transition function in the state equation of the FDKF algorithm. Experimental results show that the proposed NeuralKalman improves the performance of FDKF significantly and outperforms strong baseline methods.

show abstract

“…Besides, Peng et al [14] described a threestage AEC and suppression framework for the ICASSP 2021 AEC Challenge, where the partitioned block frequency domain least mean square (PBFDLMS) with a time alignment was firstly implemented to cancel the linear echo components, and two deep learning networks were then proposed to suppress the residual echo and the non-speech residual noise simultaneously. In addition, Zhang et al [15] proposed a neural cascade architecture, including a CRN module and an LSTM module, which is used for joint acoustic echo and noise suppression to address both single-channel and multi-channel AEC problems. More recently, Cheng et al [16] proposed a deep complex multi-frame filtering network for stereophonic AEC, where two deep learning-based modules were separately used for suppression of the linear and residual echo components.…”

Section: Introductionmentioning

confidence: 99%

Deep Learning-Based Acoustic Echo Cancellation for Surround Sound Systems

Zheng

et al. 2023

Applied Sciences

View full text Add to dashboard Cite

Surround sound systems that play back multi-channel audio signals through multiple loudspeakers can improve augmented reality, which has been widely used in many multimedia communication systems. It is common that a hand-free speech communication system suffers from the acoustic echo problem, and the echo needs to be canceled or suppressed completely. This paper proposes a deep learning-based acoustic echo cancellation (AEC) method to recover the desired near-end speech from the microphone signals in surround sound systems. The ambisonics technique was adopted to record the surround sound for reproduction. To achieve a better generalization capability against different loudspeaker layouts, the compressed complex spectra of the first-order ambisonic signals (B-format) were sent to the neural network as the input features directly instead of using the ambisonic decoded signals (D-format). Experimental results on both simulated and real acoustic environments showed the effectiveness of the proposed algorithm in surround AEC, and outperformed other competing methods in terms of the speech quality and the amount of echo reduction.

show abstract

Neural Cascade Architecture for Multi-Channel Acoustic Echo Suppression

Cited by 18 publications

References 47 publications

Original speech and its echo are segregated and separately processed in the human brain

Original speech and its echo are segregated and separately processed in the human brain

NeuralKalman: A Learnable Kalman Filter for Acoustic Echo Cancellation

Deep Learning-Based Acoustic Echo Cancellation for Surround Sound Systems

Contact Info

Product

Resources

About