mlsanityData to 19 Oct 2026

Don't Listen to Me: A Lightweight, Low-Latency Model for Own-Voice Cancellation in Far-Field Speech Enhancement

arXiv 2606.23332 ↗First posted 22 Jun 2026eess.AScited by 0

On the map: Speech and audio language models → Speech enhancement and noise suppression

Abstract (arXiv)

We introduce own-voice cancellation (OVC): removing a target (enrolled) speaker from a noisy multi-speaker mixture while preserving any remaining speech. Framed as the complement of target speaker extraction, OVC addresses latency-induced own-voice artifacts that arise when a far-field device streams enhanced audio back to the user, as the round-trip time easily exceeds the perceptual threshold for own-voice distortion. We condition a time-domain model with only 2 ms algorithmic latency on a short enrollment utterance and benchmark TD-SpeakerBeam alongside a lighter Mamba-MinGRU masker built from Mamba blocks with MinGRU temporal mixing. Replacing the ConvTasNet-based auxiliary network with a linear RNN encoder improves both signal-to-distortion ratio and predicted MOS while reducing compute. Results establish OVC as a practical, low-latency enhancement objective for far-field denoising.

Key facts. No citing papers in the map yet (data through 19 Oct 2026). Placed in Speech enhancement and noise suppression, one of 11 directions in Speech and audio language models. Builds on 22 of its 33 references that are in the map. First posted on arXiv 22 Jun 2026.

Builds on

22 of 33 references.

Closest ancestors · 1

Same direction, newest first

  1. 2023ICASSP 2023 Acoustic Echo Cancellation Challenge22 Sep

Other references · 21

  1. 2016Layer Normalizationcited by 5,866
  2. 2023Mamba: Linear-Time Sequence Modeling with Selective State Spacescited by 4,879
  3. 2017Sigmoid-Weighted Linear Units for Neural Network Function Approximation in Reinforcement Learningcited by 960
  4. 2018SDR - half-baked or well done?cited by 626
  5. 2023Resurrecting Recurrent Neural Networks for Long Sequencescited by 415
  6. 2022Diagonal State Spaces are as Effective as Structured State Spacescited by 353
  7. 2019Dual-path RNN: efficient long sequence modeling for time-domain single-channel speech separationcited by 275
  8. 2019WHAM!: Extending Speech Separation to Noisy Environmentscited by 246
  9. 2020LibriMix: An Open-Source Dataset for Generalizable Speech Separationcited by 219
  10. 2017TasNet: time-domain audio separation network for real-time, single-channel speech separationcited by 157
  11. 2024Hydra: Bidirectional State Space Models Through Generalized Matrix Mixerscited by 47
  12. 2020Improving speaker discrimination of target speech extraction with time-domain SpeakerBeamcited by 45
9 more
  1. 2024Were RNNs All We Needed?cited by 43
  2. 2025Straight to Zero: Why Linearly Decaying the Learning Rate to Zero Works Best for LLMscited by 37
  3. 2025Distillation and Pruning for Scalable Self-Supervised Representation-Based Speech Quality Assessmentcited by 25
  4. 2024TF-Locoformer: Transformer with Local Modeling by Convolution for Speech Separation and Enhancementcited by 20
  5. 2024Separate and Reconstruct: Asymmetric Encoder-Decoder for Speech Separationcited by 17
  6. 2024SpeakerBeam-SS: Real-time Target Speaker Extraction with Lightweight Conv-TasNet and State Space Modelingcited by 8
  7. 2022Listen only to me! How well can target speech extraction handle false alarms?cited by 6
  8. 2024Target Speech Extraction with Pre-trained Self-supervised Learning Modelscited by 6
  9. 2024SepMamba: State-space models for speaker separation using Mambacited by 6

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2021Personalized PercepNet: Real-time, Low-complexity Target Voice Separation and Enhancement
  2. 2026One Model, Many Latencies: Universal Speech Enhancement for Diverse Real-Time Applications
  3. 2021On TasNet for Low-Latency Single-Speaker Speech Enhancement
  4. 2025EDNet: A Versatile Speech Enhancement Framework with Gating Mamba Mechanism and Phase Shift-Invariant Training
  5. 2022Personalized Acoustic Echo Cancellation for Full-duplex Communications ↗
  6. 2022A Universally-Deployable ASR Frontend for Joint Acoustic Echo Cancellation, Speech Enhancement, and Voice Separation
  7. 2023CheapNET: Improving Light-weight speech enhancement network by projected loss function
  8. 2021A Conformer-based ASR Frontend for Joint Acoustic Echo Cancellation, Speech Enhancement and Speech Separation