HAMMER：面向语音增强的谐波感知并行上下文建模与无判别器感知优化
HAMMER: Harmonic-Aware Parallel Context Modeling and Discriminator-Free Perceptual Optimization for Speech Enhancement
- National Taiwan University（国立台湾大学）
- Academia Sinica（中央研究院）
- Nvidia（英伟达）
- VinUniversity（越南-新加坡大学）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
提出HAMMER语音增强方法，利用谐波感知并行注意力-Mamba和度量显式感知细化，在VoiceBank+DEMAND上以2.39M参数达到3.69 PESQ，超越或匹配判别器方法。
AI中文摘要：
最近的语音增强系统结合自注意力机制和Mamba以捕获全局交互和长距离依赖。然而，这些混合模型通常作为序列混合器运作，并未显式利用谐波周期性，而谐波周期性是噪声下保留浊音的有力线索。感知优化构成另一挑战。PESQ不可微，因此许多方法训练辅助度量判别器，这增加了复杂性并引入对抗不稳定性。我们提出\ours，一种基于谐波感知且无判别器的语音增强器，由两个组件构成。(i) 时频谐波感知注意力-Mamba（TF-HAM）块沿频谱图的两个轴并行运行自注意力和双向Mamba，然后应用语音适配的自相关前馈网络来编码局部周期结构。(ii) 度量显式感知细化（MEPR）结合可微PESQ和对数似然比损失，以在没有学习代理的情况下暴露感知度量结构。在VoiceBank+DEMAND上，\ours以仅2.39M参数达到3.69 PESQ和4.41 COVL，优于或匹配基于判别器的系统。推理时的感知对比度拉伸进一步将PESQ提升至3.79而无需重新训练。源代码将在此https URL提供。
英文摘要：
Recent speech enhancement systems combine self-attention and Mamba to capture global interactions and long-range dependencies. Yet these hybrids usually operate as sequence mixers and do not explicitly exploit harmonic periodicity, a strong cue for preserving voiced speech under noise. Perceptual optimization poses another challenge. PESQ is non-differentiable, so many methods train auxiliary metric discriminators that increase complexity and introduce adversarial instability. We propose \ours, a harmonic-aware and discriminator-free speech enhancer built around two components. (i) The Time-Frequency Harmonic-aware Attention-Mamba (TF-HAM) block runs self-attention and bidirectional Mamba in parallel along both spectrogram axes, then applies a speech-adapted autocorrelation feed-forward network to encode local periodic structure. (ii) Metric-explicit perceptual refinement (MEPR) combines differentiable PESQ and log-likelihood-ratio losses to expose perceptual metric structure without a learned surrogate. On VoiceBank+DEMAND, \ours achieves 3.69 PESQ and 4.41 COVL with only 2.39\,M parameters, outperforming or matching discriminator-based systems. Inference-time perceptual contrast stretching further raises PESQ to 3.79 without retraining. The source code will be available at https://github.com/shangfuu/HAMMER.git.
