--- id: liu26d_interspeech title: Towards Array-Invariant Speech Enhancement via Geometry-Aware Dynamic Convolution authors: - Zhenglong Liu - Wangyou Zhang - Chenda Li - Yanmin Qian year: 2026 doi: 10.21437/Interspeech.2026-548 isca_url: https://www.isca-archive.org/interspeech_2026/liu26d_interspeech.html pdf_url: https://www.isca-archive.org/interspeech_2026/liu26d_interspeech.pdf session: Multi-Channel, Beamforming and Spatial Speech Enhancement topics: - speech-enhancement - self-supervised - multilingual category: enhancement-separation labels: - robustness-noise institutions: - Shanghai Jiao Tong University - VUI Labs funding: - China NSFC - SJTU Med-X Translational Research Grant code: url: "" license: "" open_to_collaboration: false wiki_frontmatter: id: liu26d_interspeech category: enhancement-separation labels: - robustness-noise institutions: - Shanghai Jiao Tong University - VUI Labs updated: 2026-09-29 confidence: full-paper digest: v2 source: https://doi.org/10.21437/Interspeech.2026-548 pdf: https://www.isca-archive.org/interspeech_2026/liu26d_interspeech.pdf wiki_url: https://interspeech-2026-wiki.vercel.app/papers/liu26d_interspeech/ markdown_url: https://interspeech-2026-wiki.vercel.app/papers/liu26d_interspeech/markdown.md --- # Towards Array-Invariant Speech Enhancement via Geometry-Aware Dynamic Convolution *Zhenglong Liu, Wangyou Zhang, Chenda Li, Yanmin Qian* [PDF](https://www.isca-archive.org/interspeech_2026/liu26d_interspeech.pdf) · [ISCA page](https://www.isca-archive.org/interspeech_2026/liu26d_interspeech.html) · [DOI](https://doi.org/10.21437/Interspeech.2026-548) **Category:** `enhancement-separation` · **Labels:** `robustness-noise` **TL;DR** — The paper introduces Geometry-Aware Dynamic Convolution (Geo-DConv), a framework that injects explicit microphone array coordinate priors into fixed-array speech enhancement models, turning them into array-invariant systems that generalize across arbitrary microphone counts and layouts. Evaluated on the RealMAN dataset, the proposed SpatialNet-Geo-DConv and TF-GridNet-Geo-DConv achieve robust multi-channel speech enhancement without requiring device-specific retraining. ## Key contributions - Proposes Geo-DConv, a dynamic convolution module that adapts fixed-dimensional convolution kernels to arbitrary input channel counts and spatial configurations using explicit microphone coordinates. - Introduces the Topology-Aware Coordinate Transformer (TACT) to process Fourier-encoded coordinate matrices via Multi-Head Self-Attention, yielding a permutation-equivariant dynamic transformation matrix. - Bridges the gap between fixed-array and array-invariant speech enhancement, allowing powerful fixed-array models (SpatialNet, TF-GridNet) to operate across variable microphone numbers and topologies. - Demonstrates robust zero-shot cross-dataset generalization from RealMAN (trained on up to 4 microphones) to the 6-microphone CHiME-4 test set without any fine-tuning. ## Problem Multi-channel speech enhancement models typically rely on fixed microphone array geometries, causing performance degradation or complete failure when deployed on devices with different layouts, and forcing costly device-specific retraining. Existing array-agnostic methods either process channels independently (losing rich multi-channel spatial relationships) or rely on rigid geometric transformations like First-Order Ambisonics and virtual mic interpolation that fail for small arrays or lack explicit geometry modeling. Furthermore, conventional fixed-array models learn implicit spatial biases that collapse when trained on random, shifting array configurations, preventing them from benefiting from large-scale multi-channel pooling. ## Method The framework takes STFT complex spectrograms concatenated across frequency as input features $X \in \mathbb{R}^{C \times F \times T}$ along with a relative coordinate matrix $G \in \mathbb{R}^{C \times 3}$ (formatted in spherical coordinates $r, \theta, \phi$). First, coordinates are mapped using Fourier Positional Encoding ($L=6$ bands) to yield $G_{pe} \in \mathbb{R}^{C \times d_{pe}}$. These embeddings are projected to a hidden dimension ($d_{hidden}=64$) and fed into the Topology-Aware Coordinate Transformer (TACT), consisting of 2 Transformer encoder layers with 4 attention heads, generating a topology-aware feature representation. The resulting tokens produce a dynamic transformation coefficient matrix $M \in \mathbb{R}^{C \times b}$ (with basis dimension $b=8$), which combines through linear combination with basis convolution kernels $K \in \mathbb{R}^{b \times O \times K_f \times K_t}$ (output channel size $O=16$) to generate input-specific dynamic weights $W_{dyn} \in \mathbb{R}^{C \times O \times K_f \times K_t}$. Point-wise Fourier PE and permutation-equivariant multi-head self-attention ensure that spatial input permutations result in an identically permuted dynamic kernel, mathematically guaranteeing output invariance under channel reordering. The dynamically weighted features pass through Layer Normalization (LN) and a PReLU activation function, converting standard fixed-array backbones like SpatialNet and TF-GridNet into array-invariant processors that map variable-dimensional inputs into fixed-dimensional representations. ## Experimental setup Evaluated on the RealMAN dataset containing 83.7 hours of real-recorded multi-channel speech across 32 scenes and 144.5 hours of background noise across 31 scenes, alongside zero-shot cross-dataset evaluation on the 6-microphone CHiME-4 test set. Baselines include single-channel BSRNN, array-agnostic FaSNet-TAC and USES2-comp, and fixed-array SpatialNet and TF-GridNet. Signals are resampled to 8 kHz with a 256-point STFT window and 128-sample frame shift, evaluated using SDR, SI-SDR, PESQ, STOI, P808, SIG, BAK, and OVRL metrics. ## Results SpatialNet-Geo-DConv achieves an SI-SDR of 4.22 dB and an OVRL of 2.68 under the geometry-invariant setting, closely matching or outperforming array-agnostic baselines while maintaining a tiny footprint (1.3M parameters, 6.08 G/s MACs). TF-GridNet-Geo-DConv reaches a PESQ of 2.59 and an OVRL of 2.77, establishing state-of-the-art performance among array-invariant models at the cost of higher compute (8.3M parameters, 73.12 G/s MACs). In zero-shot cross-dataset transfer to CHiME-4 (6 mics), SpatialNet-Geo-DConv and TF-GridNet-Geo-DConv improve the unprocessed DNSMOS OVRL score from 1.42 to 2.64 and 2.73, respectively, outperforming USES2-comp (2.55). | System / Condition | SI-SDR (dB) | PESQ | OVRL | |---|---|---|---| | Unprocessed (RealMAN) | -9.47 | 1.54 | 1.49 | | FaSNet-TAC (Array-Agnostic) | -0.81 | 1.74 | 2.05 | | USES2-comp (Array-Agnostic) | 4.17 | 2.52 | 2.56 | | SpatialNet-Geo-DConv (Ours) | 4.22 | 2.48 | 2.68 | | TF-GridNet-Geo-DConv (Ours) | 3.90 | 2.59 | 2.77 | ## Limitations The current evaluation is restricted to maximum 4-microphone training configurations, and performance on ultra-dense arrays or highly irregular non-coplanar geometries beyond those in RealMAN and CHiME-4 requires further study. The framework assumes accurate relative coordinate information is available for the active microphones. Computationally, incorporating TACT adds a modest transformer overhead over standard convolutions, though it remains efficient for lightweight architectures like SpatialNet. ## Why read this Speech researchers and audio engineers seeking to merge disparate multi-channel datasets or deploy robust spatial audio models across arbitrary hardware form-factors should read this paper to learn how to inject explicit geometry priors into dynamic convolutional adapters. ## Code None released (as of this page's `updated` date). If you are an author with a repo, please claim this entry — see CONTRIBUTING.md. ## Applications Multi-microphone smart speakers, hearing aids, wearable audio devices, and teleconferencing hardware requiring robust speech enhancement across diverse, unconstrained array geometries. ## Institutions / 機構 Shanghai Jiao Tong University, VUI Labs **Funding / 經費:** China NSFC, SJTU Med-X Translational Research Grant ## Related - [Improving Multichannel Speech Enhancement through Accurate Room-Acoustic Simulations](gotz26_interspeech.md) — same problem · relatedness 2.2/3 - [Spatial-Magnifier: Spatial upsampling for multichannel speech enhancement](lee26k_interspeech.md) — same problem · relatedness 2.2/3 - [Joint Learning of Covariance Estimation and White Noise Gain for Robust MVDR Beamforming](deng26d_interspeech.md) — same problem · relatedness 2.1/3 - [Bridging the Distribution Gap in Real-World Far-Field Speech Enhancement via Lightweight Latent Representation Alignment](liu26s_interspeech.md) — same problem · relatedness 2.0/3 - [SPOT-TSE: Spatial Point-Guided Target Speech Extraction](ryu26c_interspeech.md) — same problem · relatedness 2.0/3 All 950k paper pairs scored by TypeSafe Jev (`scripts/related/`); relatedness 0 = unrelated … 3 = directly comparable.