AI 中文总结
本文针对强混响环境下神经定向滤波（NDF）的空间线索感知局限，提出基于GAN的神经去混响与定向滤波（NDDF）方法，结合指向性图案估计技术，在VDM信号重建任务中性能优于级联基线。
AI 中文摘要
近期，神经定向滤波（Neural Directional Filtering，NDF）可实现具有期望指向性图案的虚拟定向麦克风（Virtual Directional Microphone，VDM）信号重建，通过保留空间线索精准呈现多声源场景。在强混响环境中，空间线索难以被感知区分，限制了基于NDF的空间声音采集。本文针对该局限提出三项贡献：其一，提出神经去混响与定向滤波（Neural Dereverberation and Directional Filtering，NDDF）方法，用于重建去混响的VDM信号；其二，NDDF采用经判别式训练的模型及基于生成对抗网络（Generative Adversarial Network，GAN）的模型实现，与级联去混响和定向滤波基线方法对比，实验结果显示NDDF始终优于级联基线，且在处理高阶VDM目标时，基于GAN的NDDF性能优于判别式变体；其三，提出一种仅依赖输入输出信号的指向性图案估计方法，该方法适用于基于信号映射的空间滤波，可直接合成输出信号，无需显式滤波或掩蔽操作。
英文摘要
Recently, neural directional filtering (NDF) enables reconstruction of a virtual directional microphone (VDM) with a desired directivity pattern, accurately rendering multi-source scenes by preserving spatial cues. In strongly reverberant environments, spatial cues become perceptually difficult to distinguish, limiting NDF-based spatial sound capture. This paper addresses this limitation with three contributions: First, we propose a neural dereverberation and directional filtering (NDDF) approach to reconstruct dereverberated VDM signals. Second, NDDF is implemented with discriminatively trained and generative adversarial network (GAN)-based models, compared with cascaded dereverberation and directional-filtering baselines. Experimental results indicate that the NDDF consistently surpasses the cascaded baselines. Additionally, the GAN-based NDDF outperforms the discriminative variant when addressing a high-order VDM target. Third, we introduce a method for directivity pattern estimation that relies solely on the input and output signals. This method is suitable for signal-mapping-based spatial filtering, which synthesizes the output signal directly without explicit filtering or masking.
CommentsAccepted to the IEEE International Workshop on Acoustic Signal Enhancement (IWAENC) 2026
