A Hybrid Approach for Low-Complexity Joint Acoustic Echo and Noise Reduction

Shrishti Saha Shetu,Naveen Kumar Desiraju,Jose Miguel Martinez Aponte,Emanuël A. P. Habets,Edwin Mabande

2024-08-28

Abstract:Deep learning-based methods that jointly perform the task of acoustic echo and noise reduction (AENR) often require high memory and computational resources, making them unsuitable for real-time deployment on low-resource platforms such as embedded devices. We propose a low-complexity hybrid approach for joint AENR by employing a single model to suppress both residual echo and noise components. Specifically, we integrate the state-of-the-art (SOTA) ULCNet model, which was originally proposed to achieve ultra-low complexity noise suppression, in a hybrid system and train it for joint AENR. We show that the proposed approach achieves better echo reduction and comparable noise reduction performance with much lower computational complexity and memory requirements than all considered SOTA methods, at the cost of slight degradation in speech quality.

Audio and Speech Processing,Sound

What problem does this paper attempt to address?

The paper attempts to address the problem of achieving real-time Acoustic Echo and Noise Reduction (AENR) on low-resource platforms. Specifically, while existing deep learning-based methods perform well in terms of performance, they typically require a large amount of memory and computational resources, making them unsuitable for deployment on low-resource platforms such as embedded devices. Therefore, this paper proposes a low-complexity hybrid approach that uses a single model to simultaneously suppress residual echo and noise components, aiming to achieve better echo suppression and comparable noise suppression performance while significantly reducing computational complexity and memory requirements.

The main contributions of the paper include:
1. **Proposing a low-complexity hybrid system**: Combining the advanced ULCNet model with adaptive filters (such as Kalman filters) to achieve efficient AENR.
2. **Model modification**: Appropriately modifying the ULCNet model to handle three input signals (far-end signal, error signal, and echo estimate).
3. **Experimental validation**: Validating the performance of the proposed method in various scenarios through multiple experiments and comparing it with existing state-of-the-art methods, demonstrating its advantages in terms of computational complexity and memory requirements.

Overall, the paper aims to develop a high-quality and low-complexity AENR solution for real-time application on embedded devices.