Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Neural Directed Speech Enhancement with Dual Microphone Array in High Noise Scenario

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper proposes a 74.4k-parameter dual-microphone causal U-Net that steers enhancement by target angle and width, improving PESQ and downstream ASR word error.

desk verdict The width-steering idea is genuinely new, but the paper's core varied-angle claim is confounded by mismatched training data. read the letter →

arxiv 2412.18141 v2 pith:6XDLSSTX submitted 2024-12-24 eess.AS cs.SD

Wen Wen , Qiang Zhou , Yu Xi , Haoyu Li , Ziqi Gong , Kai Yu This is my paper · ORCID
classification eess.AScs.SD
keywords directedspeechenhancementdualmicrophonearraytriple-steeringspatialselectioncausalU-Netreal-timemulti-speakerseparationautomaticrecognitionbeamforming
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper claims that a dual-microphone enhancement system can be made both compact and directionally steerable by feeding a neural network three beamformer outputs: one steered at the target speaker and two steered at the edges of a tunable angular width around it. The proposed causal-directed U-Net (CDUNet) takes raw two-channel audio, the target angle, and a width value as inputs, and learns a nonlinear mask applied to the nearer microphone channel. With only 74.4k parameters it reports PESQ gains over U-Net-based and classical beamforming baselines for both fixed and varying target directions, and it lowers the word error rate of a downstream speech recognizer. If these results transfer to real devices, low-latency on-device directed enhancement could be reached with minimal hardware.

What carries the argument

The load-bearing object is the triple-steering spatial selection block: a beamformer that produces three directional outputs from two microphones, one at the target angle $\varphi_{\text{target}}$ and two at the edge angles $\varphi_{\text{target}} \pm \varphi_{\text{width}}$. These three outputs, together with the raw two-channel spectrograms, form a 10-channel input to a causal U-Net with convolutional block attention modules (CBAM) and a dual-path/LSTM sequence module; the decoder emits a mask applied to the nearer microphone channel, and training uses a combined SI-SNR plus multi-resolution STFT loss. The width input is what lets one model switch between narrow and wide spatial focus and infer the angular separation of the interferer.

What would settle it

Re-run the variable-target experiment with interfering speakers placed 3, 5, and 10 degrees from the target, or with measured room impulse responses from a real two-microphone device, and compare PESQ and word error rate: if CDUNet's advantage over the U-Net baseline disappears when the separation falls below the 15 degrees used in training, the width-steering claim is bounded to that angular regime.

Watch

Extended reading notes

Core claim

The central claim is that a single small network can learn not one spatial filter but a continuum of steerable filters indexed by angle and width. The triple-steering method forms three steering vectors at $\varphi_{\text{target}}$ and $\varphi_{\text{target}} \pm \varphi_{\text{width}}$; the beamformer outputs at those three directions, combined with the raw two-channel STFT features, tell the U-Net where the target is and how far the interfering speaker lies. The width input acts as a discriminative boundary: in the simulated data the interference is always at least 15 degrees away, and the best results occur at a width of 7 degrees, while widths at or above 15 degrees blur the boundary and degrade PESQ. In the fixed-target downstream test, CDUNet cuts the word error rate from 6.65 to 4.35 at 0 dB relative to unprocessed noisy speech.

Load-bearing premise

The results all come from simulated rectangular rooms with 30 mm microphone spacing, T60 0.2-0.5 s, and interfering speakers at least 15 degrees from the target, so the steering gains could shrink under real-device geometry, stronger reverberation, or closer interferers.

Editorial extensions

If this is right

  • A dual-microphone front-end with 74.4k parameters could run on hearing aids, earbuds, and phones, steering to a target speaker whenever the device supplies an angle.
  • Because width is an input, the same model can operate in narrow or wide focus modes without retraining, letting a device trade interferer suppression against target localization error.
  • The reported ASR gain (6.65 to 4.35 word error rate at 0 dB) means a cheap front-end can deliver cleaner speech to recognizers, not just better audio quality.
  • Variable-target training suggests that one model covers all azimuths with far fewer examples per direction than fixed-target training, lowering data collection cost.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper leaves implicit that the 15-degree minimum angular separation in the training data is a hidden condition on the claimed generality; a natural stress test is to train with separations down to a few degrees.
  • The width input suggests a self-adaptive controller: estimate the angular separation between the target and the strongest interferer and feed it as $\varphi_{\text{width}}$, so the system can track moving talkers automatically.
  • The near-microphone selection rule and 30 mm spacing tie the method to a specific array topology; binaural or phone arrays with different spacing are an open test of whether the width scale needs retuning.
  • If the learned steering generalizes across rooms, the same conditional network could serve as a fixed component in larger multi-microphone devices, with additional channels fused downstream.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper proposes CDUNet, a dual-microphone causal U-Net for target-directed speech enhancement. The core idea is a triple-steering spatial selection method: the model receives the target angle and an enhancement width, computes three beamformer outputs at the target and two edge angles, and feeds them together with raw STFT features into a compact 74.4k-parameter U-Net. Experiments report PESQ for fixed and varied target angles and WER on a downstream ASR model, with comparisons to DAS, GSC, JNF, U-Net, IPD U-Net, and BF U-Net. The authors claim superior speech quality, downstream ASR improvement, real-time operation, and suitability for on-device streaming.

Significance. If the reported results are reproducible, the contribution is notable: a 74.4k-parameter dual-microphone model that can steer enhancement by angle and width and also improve downstream ASR would be practically useful. The idea of using enhancement width as an input is a useful extension of spatially selective filtering. However, the current evidence is weakened by a confounded variable-angle comparison, test-set-based width selection, lack of uncertainty estimates, and an unsupported real-time claim. The paper does not provide code or machine-checked proofs, so the empirical claims rely entirely on the reported tables. The central fixed-target comparison is internally consistent, but the stronger steering claim is not yet established.

major comments (4)
  1. [§IV.C, Table III] The central claim of flexible steering across target angles is not established by the presented comparison. CDUNet is trained on the variable-target dataset, while the U-Net baselines are trained on the fixed-target dataset with the target near 90° (Section III.A). Table III then evaluates those baselines at 0°, 30°, 60°, and 90°. The collapse of U-Net, IPD U-Net, and BF U-Net at off-axis angles is therefore expected from the training/evaluation mismatch, and the table conflates architecture with training distribution. To support the steering claim, retrain the U-Net, IPD U-Net, and BF U-Net baselines on the variable-target dataset, or otherwise provide angle-conditioned baselines, and then report Table III again.
  2. [§IV.B, Table II] The enhancement width is selected on the evaluation data rather than on a held-out validation set. The text says the optimal φwidth is determined by evaluating PESQ scores with diverse input angles, and Table II is then used to pick 7° before the main comparisons in Tables I and III. If Table II uses the same test distribution, this is test-set tuning that can inflate the reported advantage. A held-out validation set should be used for width selection, and the chosen width should be reported as a hyperparameter selected before the final evaluation. In addition, because the variable-target training data always places the interference 15° from the target (Section III.A), the optimal width is tied to that fixed angular separation; the paper should evaluate at other separations before claiming the width flexibly adapts to arbitrary interference geometry.
  3. [Tables I–IV] All performance comparisons are reported as single PESQ or WER values without error bars, confidence intervals, or significance tests. Many of the claimed differences are small, for example CDUNet at 2.53 versus BF U-Net at 2.44 for one condition in Table I, and CDUNet at 3.11 versus U-Net at 3.37 in Table IV. Without multiple training seeds, utterance-level bootstrap, or a statistical test, it is unclear which differences are reliable. Please add repeated runs or significance testing, at least for the headline comparisons.
  4. [Abstract and §IV.A] The abstract and conclusion state that the model 'operates in real-time' and is 'ideal for low-latency, on-device streaming applications,' but the paper reports no runtime, real-time factor, or latency measurement. The 74.4k parameter count is a useful efficiency indicator, but it does not by itself establish real-time operation. Add an RTF or latency measurement on a target device, or explicitly qualify the real-time claim as an inference from architecture rather than a measured result.
minor comments (5)
  1. [§IV.A] There are typos: 'markable improvement' should be 'remarkable improvement', and 'CUDNet' should be 'CDUNet'.
  2. [§III.C] The sentence about JNF is confusing: 'JNF ... was conducted using a circular array comprising three microphones, leading to superior results not captured in Table I' is followed by Table I rows that do report JNF scores. Clarify whether JNF used two or three microphones in this evaluation and why its scores are low despite the stated advantage.
  3. [§IV.A, Table I] The sentence 'The first row of Table I of each SNR setup shows the PESQ scores with different interference angles' is misleading, because the first row is the unenhanced noisy-speech condition, not a model result.
  4. [§III.A, Figure 2 and Table I] There is an apparent inconsistency: Figure 2 says the fixed-target dataset has the interference 15° away from the target, but Table I evaluates interference at absolute angles from 0° to 180° with the target at 90°, which means angular separations from 0° to 90°. State explicitly how the training and evaluation angular separations differ.
  5. [§IV.A] The claim that CDUNet learns 'not only one spatial filter but 180' is not directly supported by the experiments, which test only four target angles in Table III. Clarify what '180 filters' means or soften the claim.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: the model's gains rest on external benchmark comparisons and explicit conditioning inputs; the Table III confound is an evaluation-validity issue, not circularity.

full rationale

The paper's derivation chain is self-contained and non-circular. The proposed model is a supervised conditional U-Net whose inputs are the raw two-channel spectra, beamformer outputs at three steering angles derived from the provided target angle and width, and the near-channel selection; the target angle and width are explicit inputs (Eq. 2), not quantities fitted to the evaluation targets. The central comparison is against external baselines (DAS, GSC, JNF, U-Net variants) under simulated RIRs, and the claimed gains are empirical benchmark results rather than consequences of a fitted parameter. Self-citations ([3], [4], [6]-[8], [10], [11]) appear only as background on downstream KWS/ASR robustness and are not load-bearing for the triple-steering method. The choice of φwidth = 7° in Table II is a hyperparameter selection performed on the evaluation set, and Table III's variable-angle comparison mixes architecture differences with training-distribution differences; these are validity and overfitting concerns, not circular reasoning, because no equation or result is defined in terms of the quantity it is said to predict.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new physical entities. The free parameters are tuning choices: the enhancement width, unreported loss weights, and the near-microphone selection threshold. The main load-bearing assumptions are the angular separation between sources, the adequacy of 30 mm dual-microphone geometry, and the representativeness of simulated room impulse responses.

free parameters (3)
  • Enhancement width phi_width = 7 degrees
    Tuned on the fixed-target evaluation data by comparing PESQ across widths 3, 5, 7, 15, 20, and 60 degrees (Table II), then used for main results. This is a hand-selected hyperparameter, not a learned weight, and it is central to the claim that width improves enhancement.
  • Loss weighting factors alpha1 and alpha2 = not reported
    Equation (5) defines the combined SI-SNR and MR-STFT loss, but the actual values chosen are never stated. The training target and therefore all results depend on them.
  • Near-microphone selection threshold = 90 degrees
    The decoder applies the mask to the nearer channel, chosen by whether the target angle is below or above 90 degrees. This binary heuristic is a design choice not justified by analysis and may create a discontinuity at broadside.
assumptions (4)
  • domain assumption Interfering speakers originate from different directions than the target and, in the variable-target dataset, are always 15 degrees away from the target.
    The problem definition (Eq. 1) and simulation assume angular separation between sources. Table I at phi_inter=90 degrees shows PESQ collapses to the noisy level, confirming the method does not work without separation.
  • domain assumption A two-microphone array with 30 mm spacing provides sufficient spatial information for the beamformer and network to distinguish directions.
    The method's spatial selectivity is assumed to hold with this compact geometry. No analysis of spatial aliasing or angular resolution limits at lower frequencies is given.
  • domain assumption Simulated RIRs with T60 0.2-0.5 s and rooms of 2.5-5 m width and 3-9 m length represent the target deployment scenario.
    All training and evaluation use synthetic room impulse responses. Real-device transfer is never tested, yet deployment claims are made.
  • standard math STFT/iSTFT and standard loss functions behave as expected.
    The pipeline relies on standard signal processing operations and established losses (SI-SNR, MR-STFT), which are not in dispute.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Neural Directed Speech Enhancement with Dual Microphone Array in High Noise Scenario." pith.science (2026). https://pith.science/paper/6XDLSSTX

@misc{pith2026241218141,
  author       = {Pith},
  title        = {Pith review of: Neural Directed Speech Enhancement with Dual Microphone Array in High Noise Scenario},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6XDLSSTX}},
  note         = {Machine review of arXiv:2412.18141}
}
read the original abstract

In multi-speaker scenarios, leveraging spatial features is essential for enhancing target speech. While with limited microphone arrays, developing a compact multi-channel speech enhancement system remains challenging, especially in extremely low signal-to-noise ratio (SNR) conditions. To tackle this issue, we propose a triple-steering spatial selection method, a flexible framework that uses three steering vectors to guide enhancement and determine the enhancement range. Specifically, we introduce a causal-directed U-Net (CDUNet) model, which takes raw multi-channel speech and the desired enhancement width as inputs. This enables dynamic adjustment of steering vectors based on the target direction and fine-tuning of the enhancement region according to the angular separation between the target and interference signals. Our model with only a dual microphone array, excels in both speech quality and downstream task performance. It operates in real-time with minimal parameters, making it ideal for low-latency, on-device streaming applications.

Figures

Figures reproduced from arXiv: 2412.18141 by the authors.

Figure 1
Figure 1. Illustration of the CDUNet architecture. The beamformer output incorporates both the target direction and the width input, which captures the spatial area information crucial for enhancement. φwidth denotes the extent of the target region to be enhanced, and φtarget specifies the orientation of the target speaker. The ”Near Mic. Selection” operation selects the speech signal from the microphone that is positioned cl… view at source ↗
Figure 2
Figure 2. Illustration of the simulation setup of the first fixed-target dataset. [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. End-to-End DOA-Guided Speech Extraction in Noisy Multi-Talker Scenarios

    eess.AS 2025-07 conditional novelty 4.0 of 10

    An end-to-end DOA- and beamwidth-conditioned neural network extracts target speech from six-speaker noisy mixtures, reporting SI-SDRi of 18.3 dB and WER reductions on a simulated test set.

Reference graph

Works this paper leans on

38 extracted references · 36 canonical work pages · cited by 1 Pith paper

  1. [1]

    Two channel audio zooming system for smart- phone,

    A. Khandelwal, E. B. Goud, Y . Chand, L. Kumar, S. Prasad, N. Agar- wala, and R. Singh, “Two channel audio zooming system for smart- phone,” 2020

    work page 2020

  2. [2]

    Learning to separate voices by spatial regions,

    A. Xu and R. R. Choudhury, “Learning to separate voices by spatial regions,” in International Conference on Machine Learning , 2022, pp. 24 539–24 549

    work page 2022

  3. [3]

    Text adaptive detection for customizable keyword spotting,

    Y . Xi, T. Tan, W. Zhang, B. Yang, and K. Yu, “Text adaptive detection for customizable keyword spotting,” in Proc. IEEE ICASSP . IEEE, 2022, pp. 6652–6656

    work page 2022

  4. [4]

    TDT-KWS: Fast and accurate keyword spotting using token-and-duration transducer,

    Y . Xi, H. Li, B. Yang, H. Li, H. Xu, and K. Yu, “TDT-KWS: Fast and accurate keyword spotting using token-and-duration transducer,” 2024

    work page 2024

  5. [5]

    Music source separation with band-split rnn,

    Y . Luo and J. Yu, “Music source separation with band-split rnn,” IEEE/ACM Transactions on Audio, Speech and Language Processing , pp. 1893–1901, 2023

    work page 1901

  6. [6]

    Semi-supervised learning for code- switching ASR with large language model filter,

    Y . Xi, W. Ding, K. Yu, and J. Lai, “Semi-supervised learning for code- switching ASR with large language model filter,” 2024

    work page 2024

  7. [7]

    Contrastive learning with audio discrimination for customizable keyword spotting in continuous speech,

    Y . Xi, B. Yang, H. Li, J. Guo, and K. Yu, “Contrastive learning with audio discrimination for customizable keyword spotting in continuous speech,” 2024

    work page 2024

  8. [8]

    Text-aware speech separation for multi-talker keyword spotting,

    H. Li, B. Yang, Y . Xi, L. Yu, T. Tan, H. Li, and K. Yu, “Text-aware speech separation for multi-talker keyword spotting,” in Interspeech 2024, 2024, pp. 337–341

    2024

Show all 38 references
  1. [9]

    Distance-based sound separation,

    K. Patterson, K. Wilson, S. Wisdom, and J. R. Hershey, “Distance-based sound separation,” in Proc. ISCA Interspeech , 2022, pp. 901–905

    2022

  2. [10]

    Streaming keyword spotting boosted by cross-layer discrimination consistency,

    Y . Xi, H. Li, X. Gu, H. Li, Y . Jiang, and K. Yu, “Streaming keyword spotting boosted by cross-layer discrimination consistency,” 2024

    2024

  3. [11]

    NTC-KWS: Noise-aware CTC for robust keyword spotting,

    Y . Xi, H. Li, H. Li, J. Guo, X. Li, W. Ding, and K. Yu, “NTC-KWS: Noise-aware CTC for robust keyword spotting,” 2024

    2024

  4. [12]

    A consol- idated perspective on multimicrophone speech enhancement and source separation,

    S. Gannot, E. Vincent, S. Markovich-Golan, and A. Ozerov, “A consol- idated perspective on multimicrophone speech enhancement and source separation,” IEEE/ACM Transactions on Audio, Speech, and Language Processing, pp. 692–730, 2017

    2017

  5. [13]

    Complex neural spatial filter: Enhancing multi-channel target speech separation in complex domain,

    R. Gu, S.-X. Zhang, Y . Zou, and D. Yu, “Complex neural spatial filter: Enhancing multi-channel target speech separation in complex domain,” IEEE Signal Processing Letters , pp. 1370–1374, 2021

    2021

  6. [14]

    Multi- channel overlapped speech recognition with location guided speech extraction network,

    Z. Chen, X. Xiao, T. Yoshioka, H. Erdogan, J. Li, and Y . Gong, “Multi- channel overlapped speech recognition with location guided speech extraction network,” in IEEE Spoken Language Technology Workshop (SLT), 2018

    2018

  7. [15]

    Multi-channel deep clustering: Discriminative spectral and spatial embeddings for speaker- independent speech separation,

    Z.-Q. Wang, J. Le Roux, and J. R. Hershey, “Multi-channel deep clustering: Discriminative spectral and spatial embeddings for speaker- independent speech separation,” in Proc. IEEE ICASSP , 2018, pp. 1–5

    2018

  8. [16]

    Combining spectral and spatial features for deep learning based blind speaker separation,

    Z.-Q. Wang and D. Wang, “Combining spectral and spatial features for deep learning based blind speaker separation,” IEEE/ACM Transactions on Audio, Speech, and Language Processing , pp. 457–468, 2019

    2019

  9. [17]

    Dbnet: Doa-driven beamforming network for end-to-end reverberant sound source separation,

    A. Aroudi and S. Braun, “Dbnet: Doa-driven beamforming network for end-to-end reverberant sound source separation,” in Proc. IEEE ICASSP, 2021, pp. 211–215

    2021

  10. [18]

    All-pole modeling of speech based on the minimum variance distortionless response spectrum,

    M. Murthi and B. Rao, “All-pole modeling of speech based on the minimum variance distortionless response spectrum,” IEEE Transactions on Speech and Audio Processing , pp. 221–239, 2000

    2000

  11. [19]

    Multi-microphone neural speech separation for far-field multi-talker speech recognition,

    T. Yoshioka, H. Erdogan, Z. Chen, and F. Alleva, “Multi-microphone neural speech separation for far-field multi-talker speech recognition,” in Proc. IEEE ICASSP , 2018, pp. 5739–5743

    2018

  12. [20]

    On time-frequency mask estimation for mvdr beamforming with application in robust speech recognition,

    X. Xiao, S. Zhao, D. L. Jones, E. S. Chng, and H. Li, “On time-frequency mask estimation for mvdr beamforming with application in robust speech recognition,” in Proc. IEEE ICASSP , 2017, pp. 3246–3250

    2017

  13. [21]

    Digital speech transmission: Enhancement, coding and error concealment,

    P. Vary and R. Martin, “Digital speech transmission: Enhancement, coding and error concealment,” 2006. [Online]. Available: https: //api.semanticscholar.org/CorpusID:62240326

    2006

  14. [22]

    Generalized sidelobe canceler beamforming combined with eigenspace-wiener postfilter for medical ultrasound imaging,

    J. Yang, X. Chen, H. Cai, and Y . Wang, “Generalized sidelobe canceler beamforming combined with eigenspace-wiener postfilter for medical ultrasound imaging,” Technology and Health Care , pp. 501–512, 2022

    2022

  15. [23]

    Emgse: Acous- tic/emg fusion for multimodal speech enhancement,

    K.-C. Wang, K.-C. Liu, H.-M. Wang, and Y . Tsao, “Emgse: Acous- tic/emg fusion for multimodal speech enhancement,” in Proc. IEEE ICASSP, 2022, pp. 1116–1120

    2022

  16. [24]

    Vsegan: Visual speech enhancement generative adversarial network,

    X. Xu, Y . Wang, D. Xu, Y . Peng, C. Zhang, J. Jia, and B. Chen, “Vsegan: Visual speech enhancement generative adversarial network,” in Proc. IEEE ICASSP, 2022, pp. 7308–7311

    2022

  17. [25]

    Metricgan-u: Unsupervised speech enhancement/ dereverberation based only on noisy/ reverberated speech,

    S.-W. Fu, C. Yu, K.-H. Hung, M. Ravanelli, and Y . Tsao, “Metricgan-u: Unsupervised speech enhancement/ dereverberation based only on noisy/ reverberated speech,” in Proc. IEEE ICASSP , 2022, pp. 7412–7416

    2022

  18. [26]

    Hgcn: Harmonic gated compensation network for speech enhancement,

    T. Wang, W. Zhu, Y . Gao, J. Feng, and S. Zhang, “Hgcn: Harmonic gated compensation network for speech enhancement,” in ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2022, pp. 371–375

    2022

  19. [27]

    Tf-gridnet: Integrating full- and sub-band modeling for speech separation,

    Z.-Q. Wang, S. Cornell, S. Choi, Y . Lee, B.-Y . Kim, and S. Watan- abe, “Tf-gridnet: Integrating full- and sub-band modeling for speech separation,” IEEE/ACM Transactions on Audio, Speech, and Language Processing, pp. 3221–3236, 2023

    2023

  20. [28]

    Fullsubnet+: Channel attention fullsubnet with complex spectrograms for speech enhancement,

    J. Chen, Z. Wang, D. Tuo, Z. Wu, S. Kang, and H. Meng, “Fullsubnet+: Channel attention fullsubnet with complex spectrograms for speech enhancement,” in Proc. IEEE ICASSP , 2022, pp. 7857–7861

    2022

  21. [29]

    Guided speech enhancement network,

    Y . Yang, S.-F. Shih, H. Erdogan, J. Menjay Lin, C. Lee, Y . Li, G. Sung, and M. Grundmann, “Guided speech enhancement network,” in Proc. IEEE ICASSP, 2023, pp. 1–5

    2023

  22. [30]

    Binaural angular separation network,

    Y . Yang, G. Sung, S.-F. Shih, H. Erdogan, C. Lee, and M. Grundmann, “Binaural angular separation network,” in Proc. IEEE ICASSP , 2024, pp. 1201–1205

    2024

  23. [31]

    Dsenet: Directional signal extraction network for hearing improvement on edge devices,

    A. Kovalyov, K. Patel, and I. Panahi, “Dsenet: Directional signal extraction network for hearing improvement on edge devices,” IEEE Access, vol. 11, pp. 4350–4358, 2023

    2023

  24. [32]

    Spatially selective deep non-linear filters for speaker extraction,

    K. Tesch and T. Gerkmann, “Spatially selective deep non-linear filters for speaker extraction,” in Proc. IEEE ICASSP , 2023, pp. 1–5

    2023

  25. [33]

    Multi-channel speech separation using spatially selective deep non-linear filters,

    ——, “Multi-channel speech separation using spatially selective deep non-linear filters,” IEEE/ACM Transactions on Audio, Speech, and Language Processing, p. 542–553, 2024

    2024

  26. [34]

    U-net: Convolutional networks for biomedical image segmentation,

    O. Ronneberger, P. Fischer, and T. Brox, “U-net: Convolutional networks for biomedical image segmentation,” in Medical Image Computing and Computer-Assisted Intervention (MICCAI) , 2015, pp. 234–241

    2015

  27. [35]

    Cbam: Convolutional block attention module,

    S. Woo, J. Park, J.-Y . Lee, and I. S. Kweon, “Cbam: Convolutional block attention module,” in Proceedings of the European Conference on Computer Vision (ECCV) , 2018, pp. 3–19

    2018

  28. [36]

    Optimal scale-invariant signal-to-noise ratio and curriculum learning for monaural multi-speaker speech separation in noisy environment,

    C. Ma, D. Li, and X. Jia, “Optimal scale-invariant signal-to-noise ratio and curriculum learning for monaural multi-speaker speech separation in noisy environment,” in 2020 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) , 2...

    2020

  29. [37]

    Time-domain speech enhancement assisted by multi-resolution frequency encoder and decoder,

    H. Shi, M. Mimura, L. Wang, J. Dang, and T. Kawahara, “Time-domain speech enhancement assisted by multi-resolution frequency encoder and decoder,” in ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2023, pp. 1–5

    2023

  30. [38]

    Librispeech: an asr corpus based on public domain audio books,

    V . Panayotov et al., “Librispeech: an asr corpus based on public domain audio books,” in Proc. IEEE ICASSP , 2015, pp. 5206–5210

    2015

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.