Pith. sign in

REVIEW 4 major objections 4 minor 76 references

Autoregressive Guidance of Deep Spatially Selective Filters using Bayesian Tracking for Efficient Extraction of Moving Speakers

T0 review · 4 major / 4 minor · reviewed 2026-08-02 · deepseek-v4-flash

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

Pith's one-line read The paper claims that a moving speaker's direction can be tracked by feeding the spatial filter's own previous output back into a Kalman or particle filter, achieving accuracy competitive with neural trackers at negligible added compute.

desk verdict Solid extension of the authors' self-steering SSF line: autoregressive feedback into lightweight Bayesian trackers genuinely helps, but the neural baseline comparison is uncontrolled and the plug-in of enhanced speech as clean observations is the main risk. read the letter →

arxiv 2603.23723 v2 pith:UHAF3QVA submitted 2026-03-24 eess.AS cs.LGcs.SD

Jakob Kienegger , Timo Gerkmann This is my paper · ORCID
classification eess.AScs.LGcs.SD
keywords movingspeakertrackingdirectionofarrivalestimationBayesianfilteringKalmanfilterparticleautoregressiveguidancespatiallyselectivetargetextraction
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

This paper argues that a moving speaker's direction can be tracked accurately by closing the loop between a deep spatial filter and a cheap Bayesian tracker: after the filter produces its first enhanced estimate, that cleaned signal is fed back into a Kalman or particle filter as an observation, instead of the tracker relying only on the raw microphone mixture. With nothing but the speaker's initial direction as guidance, this autoregressive feedback reduces direction-of-arrival error substantially and improves the quality and intelligibility of the extracted speech. The method is compatible with arbitrary spatial filters, adds negligible computation, and generalizes to real rooms with reverberation up to 800 ms. A reader should care because it offers a way to replace expensive neural tracking networks with lightweight recursive filters in real-time hearing and communication devices.

What carries the argument

The load-bearing mechanism is temporal feedback in a frame-wise causal pipeline: at time t, the previously enhanced speech estimate replaces or conditions the tracker's likelihood in place of the raw microphone observation. A MIMO variant of the spatial filter is introduced whose output retains inter-microphone phase structure, so the enhanced signal can act as a multichannel pseudo-observation; a MISO variant instead uses the single-channel estimate as a frequency selector that emphasizes target-dominated bins in the DoA estimator. The recursive Bayesian update then couples the filter and tracker into a closed loop whose added cost is only a few million multiply-accumulate operations per se

What would settle it

Take a dataset with exact ground-truth DoA and a second independently pretrained spatial filter of different architecture. If, during a sustained speaker crossing, the AR tracker's error exceeds the concatenative tracker's, or if deliberately corrupting the fed-back signal (e.g., replacing it with the interferer's enhanced output) does not degrade tracking, then the plug-in clean-speech assumption is refuted.

Watch

Extended reading notes

Core claim

The central claim is that the enhanced output of a spatially selective filter can serve as the observation model inside a recursive Bayesian tracker, creating a self-steering loop. The paper develops two ways to do this: MISO-AR, which uses the single-channel cleaned speech to reweight frequency bins in a wrapped-Gaussian Kalman likelihood or to condition a particle-filter likelihood, and MIMO-AR, which modifies the spatial filter to output a multichannel estimate that can replace the raw mixture in the likelihood. In simulation, the bootstrap particle filter in the MISO-AR configuration reduces mean absolute DoA error from 21.65° to 6.47° and raises 10°-accuracy from 56.2% to 87.6%, while i

Load-bearing premise

The results depend on treating the spatial filter's previous enhanced output as a faithful clean-speech observation inside the tracker's likelihoods, even though the filter was trained on oracle directions and its errors are never modeled; if that output leaks the interfering speaker, the feedback loop can lock onto the wrong direction.

Editorial extensions

If this is right

  • In the strongest simulated configuration (bootstrap particle filter with MISO-AR), mean DoA error falls from 21.65° to 6.47° and 10°-accuracy rises from 56.2% to 87.6%, with PESQ improving from 1.93 to 2.04.
  • The autoregressive Kalman filter, which by itself cannot handle moving speakers (MAE above 30°), drops to 6.65° once the spatial filter is switched to its multichannel MIMO-AR form.
  • The approach costs almost nothing beyond the base filter: trackers require 0.3–2.5 MMAC/s, and the MIMO extension adds fewer than 500 parameters and about 800 kMAC/s per kHz of bandwidth.
  • AR-guided Bayesian trackers reach accuracy comparable to a reference neural tracker while using an order of magnitude less computation.
  • Spatial filters trained on pedestrian-like moving-source trajectories retain quality on unseen motion patterns, including real-world walking data.

Reading between the lines

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

  • The same plug-in principle could be extended to other target cues—elevation, range, or binaural hearing-aid configurations—wherever a causal filter output is available; this is an inference, not tested here.
  • The success of the single-channel MISO-AR particle filter, without explicit spatial-cue preservation, suggests the main benefit of feedback is a cleaner, more target-dominated input to the likelihood rather than restored inter-microphone phase; a testable extension would substitute a single-channel denoiser's output to isolate that effect.
  • The real-room evaluation uses a coarse side-of-array metric rather than exact DoA ground truth; an independent motion-captured dataset could determine whether the AR gains reflect true direction accuracy or mainly reduced speaker confusion after crossings.
  • The new trajectory simulator may be useful as a training distribution for other moving-source tasks, and its public release invites cross-dataset comparisons beyond the tested room sizes and array geometry.
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 / 4 minor

Summary. The paper proposes autoregressive (AR) guidance of deep spatially selective filters (SSFs) for moving-speaker extraction, by feeding the SSF's previously enhanced speech back into lightweight Bayesian target-speaker trackers (wrapped Kalman filter and bootstrap particle filter). Two AR schemes are presented: MISO-AR, where the enhanced single-channel estimate is incorporated into the tracker's likelihood, and MIMO-AR, where a multichannel SSF estimate replaces the noisy observation in the generative model. The paper also introduces a social-force-model-based synthetic dataset for realistic speaker trajectories. Experiments on synthetic and real-room recordings show that the AR methods substantially improve tracking accuracy and enhancement quality over the concatenative baseline, with negligible added compute, and achieve accuracy competitive with neural trackers at a fraction of the computational cost.

Significance. If the central claim holds, the paper makes a useful contribution: it shows that a cheap statistical tracker can exploit the output of a deep SSF to steer it accurately, potentially replacing much more expensive neural trackers in online causal pipelines. The paper also ships a dataset-generation framework, reports confidence intervals on synthetic metrics, includes real-room evaluations, and provides compute comparisons — all strengths. The main open risk is that the AR gain depends on treating the SSF's possibly degraded output as a faithful clean-speech observation inside the tracker, and the neural-tracker comparison is not controlled for the additional feedback information available only to the Bayesian AR methods.

major comments (4)
  1. [Sec. IV-A, Eqs. (17)-(27) and Table I] The central AR mechanism relies on the plug-in substitution of enhanced speech \hat{S} for clean speech S in the tracker likelihoods, while explicitly neglecting SSF processing degradation. This is load-bearing: if \hat{S} leaks the interferer, the residual in Eq. (24) is biased and the recursive noise-covariance update in Eq. (25) is mis-specified; if the MIMO output corrupts phase structure, the steering-vector likelihood in Eq. (23) can lock onto the wrong direction. The paper acknowledges this limitation but does not bound it. Only SpatialNet is tested, and Fig. 4b shows a clear tradeoff between IPD preservation and reconstruction quality. I recommend adding an oracle-substitution experiment (feeding clean target speech or leakage-controlled mixtures into the same trackers) and evaluating at least one additional SSF or a deliberately phase-distorted variant, to demonstrate robustness
  2. [Sec. VII-B, Fig. 5] The comparison with neural trackers is not controlled for the AR feedback. The neural baselines (SELDnet, CNN/LSTM) receive only the noisy multichannel signal and the initial DoA, while the Bayesian AR methods additionally observe the SSF's enhanced output. Thus the comparison conflates tracker architecture with an extra source of information, and the claim that the Bayesian method matches or beats SELDnet at less than one-tenth the compute is only about non-AR neural trackers. A neural tracker embedded in the same AR loop might close the gap. Please either integrate the neural baselines into the AR loop (e.g., by feeding enhanced features) or explicitly restrict the claim to non-AR neural trackers, and avoid presenting the result as a pure architecture advantage.
  3. [Sec. VII-D, Fig. 8] The real-world DoA evaluation uses the regional accuracy (Re-ACC), a coarse side-of-array metric. This cannot distinguish correct target tracking from tracking an interferer who is momentarily on the expected side — precisely the failure mode associated with the leaky-feedback concern. Since the synthetic evaluation carries the main DoA claim, the real-world results should be framed primarily as supporting enhancement/generalization evidence rather than as exact DoA validation. I suggest adding a small number of manually annotated ground-truth positions or a finer event-based analysis around speaker crossings.
  4. [Sec. VI-B and Fig. 4a] The fine-tuning procedure fixes tracker hyperparameters after the first epoch and performs a final parameter sweep for the AR configuration, but the paper does not state whether the concatenative baselines received an equally thorough hyperparameter optimization. If the Concat trackers were left at less-tuned settings, part of the observed gain in Table I could be due to tuning rather than the AR mechanism. Please clarify the hyperparameter search for all compared pipelines, including the neural trackers in Fig. 5.
minor comments (4)
  1. [Throughout] Typos: "Boostrap" should be "Bootstrap" (Sec. VI-A), "competetive" should be "competitive" (Sec. VII-B), and "real-word trajectories" in the conclusion should be "real-world trajectories".
  2. [Fig. 5] The arrow labels on the y-axes ("MAE [°]←" and "ACC [%]→") are visually confusing; please use standard axis labels with a legend for the arrow direction.
  3. [Sec. VII-D] The Re-ACC definition could be stated more precisely in the text; currently it is described mainly via Fig. 7. A one-sentence mathematical definition would improve reproducibility.
  4. [Sec. V-B] The social-force model parameters are reported in the text, but a single summary table listing all model constants and their sampling distributions would aid implementation.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the AR tracking gain is an empirical result measured against ground-truth DoA, not a reduction of the prediction to its inputs.

full rationale

The paper's derivation chain is not circular. The AR formulations in Sec. IV introduce clean speech S as a latent observation (Eq. 17, 20) and define likelihoods that condition on it (Eqs. 19, 22-23); during inference, enhanced estimates S-hat are used as a plug-in. This is explicitly acknowledged in Sec. IV-A: "we use the enhanced STFT segments S-hat_{1:t-1} as plug-in approximation for samples of clean speech S_{1:t-1}. While this neglects processing degradation from the SSF, we demonstrate that a high tracking accuracy can be achieved." That is an approximation/assumption, not an equivalence: the DoA estimate is not defined as a function of S-hat, and tracking performance is evaluated against ground-truth trajectories (MAE/ACC on synthetic data, Re-ACC on real data), so the reported gains are empirical rather than forced by construction. The MIMO substitution in Eq. (27) follows from the assumed generative model (Eq. 1), not from the estimator's output. The fine-tuning procedure in Sec. VI-B does involve co-adaptation between the SSF and the Bayesian trackers, and the IPD-loss weight alpha_IPD is tuned on a validation set (Fig. 4b); these are potential generalizability/overfitting concerns, but they are not definitional circularity and do not make the central claim equivalent to its inputs. Self-citations ([13]-[15]) describe prior work and extensions, but the current equations, ablations, and external benchmarks (e.g., comparison with SELDnet/CNN-LSTM in Fig. 5, real-room recordings in Fig. 8) provide independent content. The coarse Re-ACC metric in Sec. VII-D is a measurement limitation, not a circular step. Overall, no load-bearing step reduces to its own input.

Assumptions & free parameters 7 free parameters · 7 assumptions · 0 invented entities

The paper introduces no new physical entities. It relies on standard Bayesian filtering assumptions, plus two load-bearing approximations specific to this work: the plug-in replacement of clean speech by the SSF output, and the MIMO SSF's preservation of spatial cues. Several tracker hyperparameters are tuned on the validation set without final values being reported.

free parameters (7)
  • KF process/measurement noise variance ratio σ²_ξ/σ²_ν = optimal ratio chosen on validation set; absolute values not reported
    Tuned in closed-loop AR sweep (Fig. 4a); controls KF smoothing and is not derived from data.
  • PF measurement concentration κ = not reported
    Complex Watson likelihood in Eq. (16) requires a concentration parameter; no value or fitting procedure specified.
  • EMA coefficient α^(EMA) = not reported
    Noise-covariance recursion in Eq. (25) depends on hand-chosen exponential smoothing coefficient.
  • particle count N = 50 (Sec. VI-A)
    Fixed by hand; controls PF cost and accuracy.
  • adaptive resampling threshold τ^(eff) = not reported
    Alg. 1 line 9 triggers resampling when effective sample size falls below τ^(eff)N; value unspecified.
  • IPD-loss weight α^(IPD) = ≈10^-2 from Fig. 4b optimum
    Traded off against reconstruction loss in Eq. (41); selected on validation set.
  • social-force model parameters (τ=1s, A^(S)=2.1 m²/s², B^(S)=0.3 m, ε^(W)=ε^(A)=0.5 m, lookahead Δt=2 s) = as listed in Sec. V-B
    Borrowed from pedestrian dynamics or set by quasi-static energy argument; not fitted to enhancement data but shape the dataset.
assumptions (7)
  • standard math Markov assumptions and white-noise acceleration dynamics for DoA (Eq. 8)
    Recursive Bayesian filtering relies on the Markov property; the dynamics model is chosen for smoothness.
  • domain assumption Linear-phase DoA statistic Φ_t is sufficient for θ_t and wrapped-Gaussian/mode-matching approximation holds
    Borrowed from Traa et al.; limits the KF to aliasing-free bins and ignores phase wrapping.
  • domain assumption Clean speech S_t and DoA θ_t are independent, enabling factorization in (18)/(20)
    Source waveform content and direction are modeled as independent latent variables.
  • ad hoc to paper Enhanced STFT output Ŝ can be plugged in for clean speech S in likelihoods while neglecting SSF degradation
    Explicit in Sec. IV-A; central to the AR gain and not justified theoretically.
  • domain assumption Noise STFT coefficients are uncorrelated across frequency and proper complex Gaussian (23)
    Enables closed-form PF likelihood; approximate for diffuse noise and reverberation.
  • ad hoc to paper MIMO SSF estimate preserves direct-path multichannel propagation (spatial cues) sufficiently for tracking
    Trained with IPD loss; if violated, MIMO-AR tracking degrades.
  • domain assumption Far-field planar propagation; steering vector depends only on azimuth (Eqs. 2-3)
    Standard assumption for compact arrays; excludes elevation and near-field effects.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Autoregressive Guidance of Deep Spatially Selective Filters using Bayesian Tracking for Efficient Extraction of Moving Speakers." pith.science (2026). https://pith.science/paper/UHAF3QVA

@misc{pith2026260323723,
  author       = {Pith},
  title        = {Pith review of: Autoregressive Guidance of Deep Spatially Selective Filters using Bayesian Tracking for Efficient Extraction of Moving Speakers},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UHAF3QVA}},
  note         = {Machine review of arXiv:2603.23723}
}
read the original abstract

Deep spatially selective filters achieve high-quality enhancement with real-time capable architectures for stationary speakers of known directions. To retain this level of performance in dynamic scenarios where only the speakers' initial directions are given, accurate, yet computationally lightweight tracking algorithms become necessary. Assuming a frame-wise causal processing style, temporal feedback allows for leveraging the enhanced speech signal to improve tracking performance. In this work, we investigate strategies to incorporate the enhanced signal into lightweight tracking algorithms and autoregressively guide deep spatial filters. Our proposed Bayesian tracking algorithms are compatible with arbitrary deep spatial filters. To increase the realism of simulated trajectories during development and evaluation, we develop a synthetic data generation framework based on the social force model. Results validate that the autoregressive incorporation significantly improves the accuracy of our Bayesian trackers, resulting in superior enhancement with none or only negligibly increased computational overhead. Real-world recordings complement these findings and demonstrate the generalizability of our methods to unseen acoustic conditions.

Figures

Figures reproduced from arXiv: 2603.23723 by the authors.

Figure 1
Figure 1. Weakly guided speaker extraction using target speaker tracking (TST) to estimate the target’s direction [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Social force motion model adapted from [31] to simulate planar two speaker ( [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Sample two-speaker ( / ) trajectories using Wrapped KF ( ) and Bootstrap PF ( ) for tracking in (top to bottom) concatentative (Fig. 1a) and our autoregressive (MISO-AR: Fig. 1b, MIMO-AR: Fig. 1c) configurations [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Closed-loop (AR) parameter optimization on the validation set using [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Computational cost (MACs) and tracking performance of our Bayesian [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 7
Figure 7. Figure 7: Two-speaker ( / ) DoA tracking under increasing reverberation (left to right) with (top to bottom) concatentative (Fig. 1a) and our autoregressive (MISO-AR: Fig. 1b, MIMO-AR: Fig. 1c) methods using Wrapped KF ( ) and Bootstrap PF ( ). Shaded areas ( ) indicate Re-ACC c…
Figure 8
Figure 8. Figure 8: Tracking (Re-ACC) and enhancement (NISQA, WER) performance of [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

76 extracted references · 1 linked inside Pith

  1. [1]

    Some experiments on the recognition of speech, with one and two ears,

    E. C. Cherry, “Some experiments on the recognition of speech, with one and two ears,”J. Acoust. Soc. Am., vol. 25, 1953

    1953

  2. [2]

    Neural target speech extraction: An overview,

    K. Zmolikova, M. Delcroix, T. Ochiai, K. Kinoshita, J. ˇCernock´y, and D. Yu, “Neural target speech extraction: An overview,”IEEE Signal Proc. Magazine, vol. 40, 2023

    2023

  3. [3]

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

    K. Tesch and T. Gerkmann, “Multi-channel speech separation using spatially selective deep non-linear filters,”IEEE/ACM TASLP, vol. 32, 2024

    2024

  4. [4]

    Spatially selective speaker separation using a DNN with a location dependent feature extraction,

    A. Bohlender, A. Spriet, W. Tirry, and N. Madhu, “Spatially selective speaker separation using a DNN with a location dependent feature extraction,”IEEE/ACM TASLP, vol. 32, 2024

    2024

  5. [5]

    Insights into deep non-linear filters for im- proved multi-channel speech enhancement,

    K. Tesch and T. Gerkmann, “Insights into deep non-linear filters for im- proved multi-channel speech enhancement,”IEEE/ACM TASLP, vol. 31, 2023

    2023

  6. [6]

    End-to-end DOA-guided speech extraction in noisy multi-talker scenarios,

    K. Jing, W. Zhang, and Y . Gao, “End-to-end DOA-guided speech extraction in noisy multi-talker scenarios,” inInterspeech, 2025

    2025

  7. [7]

    All neural low- latency directional speech extraction,

    A. Pandey, S. Lee, J. Azcarreta, D. Wong, and B. Xu, “All neural low- latency directional speech extraction,” inInterspeech, 2024

    2024

  8. [8]

    ReZero: Region-customizable sound extraction,

    R. Gu and Y . Luo, “ReZero: Region-customizable sound extraction,” IEEE/ACM TASLP, 2024

    2024

Show all 76 references
  1. [9]

    Multichannel-to-multichannel target sound extraction using direction and timestamp clues,

    D. Choi and J.-W. Choi, “Multichannel-to-multichannel target sound extraction using direction and timestamp clues,” inIEEE ICASSP, 2025

    2025

  2. [10]

    Location-aware target speaker extraction for hearing aids,

    D.-J. A. Padilla, N. L. Westhausen, S. Vivekananthan, and B. T. Meyer, “Location-aware target speaker extraction for hearing aids,” in Interspeech, 2025

    2025

  3. [11]

    Continuous speech separation: Dataset and analysis,

    Z. Chen, T. Yoshioka, L. Lu, T. Zhou, Z. Meng, Y . Luo, J. Wu, X. Xiao, and J. Li, “Continuous speech separation: Dataset and analysis,” inIEEE ICASSP, 2020

    2020

  4. [12]

    The fifth ’CHiME’ speech separation and recognition challenge: Dataset, task and base- lines,

    J. Barker, S. Watanabe, E. Vincent, and J. Trmal, “The fifth ’CHiME’ speech separation and recognition challenge: Dataset, task and base- lines,” inInterspeech, 2018

    2018

  5. [13]

    Steering deep non-linear spatially selective filters for weakly guided extraction of moving speakers in dynamic scenarios,

    J. Kienegger and T. Gerkmann, “Steering deep non-linear spatially selective filters for weakly guided extraction of moving speakers in dynamic scenarios,” inInterspeech, 2025

    2025

  6. [14]

    Self-steering deep non-linear spatially selective filters for efficient extraction of moving speakers under weak guidance,

    J. Kienegger, A. Mannanova, H. Fang, and T. Gerkmann, “Self-steering deep non-linear spatially selective filters for efficient extraction of moving speakers under weak guidance,” inIEEE WASPAA, 2025

    2025

  7. [15]

    Adaptive rotary steering with joint autoregression for robust extraction of closely moving speakers in dynamic scenarios,

    J. Kienegger and T. Gerkmann, “Adaptive rotary steering with joint autoregression for robust extraction of closely moving speakers in dynamic scenarios,” inIEEE ICASSP, 2026

    2026

  8. [16]

    Robust sound source tracking using SRP-PHAT and 3D convolutional neural networks,

    D. Diaz-Guerra, A. Miguel, and J. R. Beltran, “Robust sound source tracking using SRP-PHAT and 3D convolutional neural networks,” IEEE/ACM TASLP, vol. 29, 2021

    2021

  9. [17]

    Exploiting temporal context in CNN based multisource DoA estimation,

    A. Bohlender, A. Spriet, W. Tirry, and N. Madhu, “Exploiting temporal context in CNN based multisource DoA estimation,”IEEE/ACM TASLP, vol. 29, 2021

    2021

  10. [18]

    SRP-DNN: Learning direct-path phase difference for multiple moving sound source localization,

    B. Yang, H. Liu, and X. Li, “SRP-DNN: Learning direct-path phase difference for multiple moving sound source localization,” inIEEE ICASSP, 2022

    2022

  11. [19]

    FN-SSL: Full-band and narrow-band fusion for sound source localization,

    Y . Wang, B. Yang, and X. Li, “FN-SSL: Full-band and narrow-band fusion for sound source localization,” inInterspeech, 2023

    2023

  12. [20]

    TF-Mamba: A time-frequency network for sound source localization,

    Y . Xiao and R. K. Das, “TF-Mamba: A time-frequency network for sound source localization,” inInterspeech, 2025

    2025

  13. [21]

    A convolutional recurrent neural network for real-time speech enhancement,

    K. Tan and D. Wang, “A convolutional recurrent neural network for real-time speech enhancement,” inInterspeech, 2018

    2018

  14. [22]

    Real time speech enhancement in the waveform domain,

    A. D ´efossez, G. Synnaeve, and Y . Adi, “Real time speech enhancement in the waveform domain,” inInterspeech, 2020

    2020

  15. [23]

    Towards efficient models for real-time deep noise suppression,

    S. Braun, H. Gamper, C. K. Reddy, and I. Tashev, “Towards efficient models for real-time deep noise suppression,” inIEEE ICASSP, 2021

    2021

  16. [24]

    Conv-TasNet: Surpassing ideal time–frequency magnitude masking for speech separation,

    Y . Luo and N. Mesgarani, “Conv-TasNet: Surpassing ideal time–frequency magnitude masking for speech separation,”IEEE/ACM TASLP, vol. 27, 2019

    2019

  17. [25]

    An investigation of incorporating Mamba for speech enhancement,

    R. Chao, W.-H. Cheng, M. L. Quatra, S. M. Siniscalchi, C.-H. H. Yang, S.-W. Fu, and Y . Tsao, “An investigation of incorporating Mamba for speech enhancement,” inIEEE Spoken Language Tech. Workshop, 2024

    2024

  18. [26]

    Iterative autoregression: A novel trick to improve your low-latency speech enhancement model,

    P. Andreev, N. Babaev, A. Saginbaev, I. Shchekotov, and A. Alanov, “Iterative autoregression: A novel trick to improve your low-latency speech enhancement model,” inInterspeech, 2023

    2023

  19. [27]

    PARIS: Pseudo-autoregressive siamese training for online speech separation,

    Z. Pan, G. Wichern, F. G. Germain, K. Saijo, and J. Le Roux, “PARIS: Pseudo-autoregressive siamese training for online speech separation,” in Interspeech, 2024

    2024

  20. [28]

    ARiSE: Auto-regressive multi- channel speech enhancement,

    P. Shen, X. Zhang, and Z.-Q. Wang, “ARiSE: Auto-regressive multi- channel speech enhancement,” inInterspeech, 2025

    2025

  21. [29]

    A wrapped Kalman filter for azimuthal speaker tracking,

    J. Traa and P. Smaragdis, “A wrapped Kalman filter for azimuthal speaker tracking,”IEEE Signal Proc. Letters, vol. 20, 2013. JOURNAL OF LATEX CLASS FILES, VOL. XX, NO. X, MONTH XXXX 10

    2013

  22. [30]

    Particle filtering algorithms for tracking an acoustic source in a reverberant environment,

    D. Ward, E. Lehmann, and R. Williamson, “Particle filtering algorithms for tracking an acoustic source in a reverberant environment,”IEEE Trans. on Speech and Audio Proc., vol. 11, 2003

    2003

  23. [31]

    Social force model for pedestrian dynamics,

    D. Helbing and P. Molnar, “Social force model for pedestrian dynamics,” Physical review E, vol. 51, 1995

    1995

  24. [32]

    Benesty, G

    J. Benesty, G. Huang, J. Chen, and N. Pan,Microphone Arrays. Springer, 2024

    2024

  25. [33]

    S ¨arkk¨a,Bayesian Filtering and Smoothing

    S. S ¨arkk¨a,Bayesian Filtering and Smoothing. Cambridge University Press, 2013

    2013

  26. [34]

    A new approach to linear filtering and prediction problems,

    R. E. Kalman, “A new approach to linear filtering and prediction problems,”Journal of Basic Engineering, vol. 82, 1960

    1960

  27. [35]

    Survey of maneuvering target tracking. Part I. Dynamic models,

    X. Rong Li and V . Jilkov, “Survey of maneuvering target tracking. Part I. Dynamic models,”IEEE Trans. on Aerospace and Electronic Systems, vol. 39, 2003

    2003

  28. [36]

    Particle filtering approaches for multiple acoustic source detection and 2-D direction of arrival estimation using a single acoustic vector sensor,

    X. Zhong and A. B. Premkumar, “Particle filtering approaches for multiple acoustic source detection and 2-D direction of arrival estimation using a single acoustic vector sensor,”IEEE Trans. on Signal Proc., vol. 60, 2012

    2012

  29. [37]

    Particle filter algorithm for DoA tracking using co-prime array,

    F. Dong, L. Xu, and X. Li, “Particle filter algorithm for DoA tracking using co-prime array,”IEEE Comm. Letters, vol. 24, 2020

    2020

  30. [38]

    Multichannel source separation and tracking with phase dif- ferences by random sample consensus,

    J. Traa, “Multichannel source separation and tracking with phase dif- ferences by random sample consensus,” Master’s thesis, University of Illinois at Urbana-Champaign, 2013

    2013

  31. [39]

    A low complexity weighted least squares narrowband DOA estimator for arbitrary array geometries,

    O. Thiergart, W. Huang, and E. A. Habets, “A low complexity weighted least squares narrowband DOA estimator for arbitrary array geometries,” inIEEE ICASSP, 2016

    2016

  32. [40]

    K. V . Mardia and P. E. Jupp,Directional statistics. John Wiley & Sons, 2000

    2000

  33. [41]

    Novel approach to nonlinear/non-Gaussian Bayesian state estimation,

    N. Gordon, D. Salmond, and A. Smith, “Novel approach to nonlinear/non-Gaussian Bayesian state estimation,”IEE proc. F, vol. 140, 1993

    1993

  34. [42]

    A tutorial on particle filters for online nonlinear/non-Gaussian Bayesian tracking,

    M. Arulampalam, S. Maskell, N. Gordon, and T. Clapp, “A tutorial on particle filters for online nonlinear/non-Gaussian Bayesian tracking,” IEEE Trans. on Signal Proc., vol. 50, 2002

    2002

  35. [43]

    DOA-estimation based on a complex Watson kernel method,

    L. Drude, F. Jacob, and R. Haeb-Umbach, “DOA-estimation based on a complex Watson kernel method,” inEUSIPCO, 2015

    2015

  36. [44]

    Target speaker localization based on the complex Watson mixture model and time-frequency selection neural network,

    Z. Wang, J. Li, and Y . Yan, “Target speaker localization based on the complex Watson mixture model and time-frequency selection neural network,”Applied Sciences, vol. 8, 2018

    2018

  37. [45]

    On optimal multichannel mean-squared error estimators for speech enhancement,

    R. C. Hendriks, R. Heusdens, U. Kjems, and J. Jensen, “On optimal multichannel mean-squared error estimators for speech enhancement,” IEEE Signal Proc. Letters, vol. 16, 2009

    2009

  38. [46]

    P. J. Schreier and L. L. Scharf,Statistical signal processing of complex- valued data: The theory of improper and noncircular signals. Cam- bridge University Press, 2010

    2010

  39. [47]

    Particle filter with integrated voice activity detection for acoustic source tracking,

    E. A. Lehmann and A. M. Johansson, “Particle filter with integrated voice activity detection for acoustic source tracking,”EURASIP J. on Adv. in Signal Proc., vol. 2007, 2006

    2007

  40. [48]

    GCC-Speaker: Target speaker localization with optimal speaker-dependent weighting in multi-speaker scenarios,

    G. Li, W. Xue, W. Liu, J. Yi, and J. Tao, “GCC-Speaker: Target speaker localization with optimal speaker-dependent weighting in multi-speaker scenarios,” inIEEE ICASSP, 2023

    2023

  41. [49]

    LocSelect: Target speaker localization with an auditory selective hearing mechanism,

    Y . Chen, X. Qian, Z. Pan, K. Chen, and H. Li, “LocSelect: Target speaker localization with an auditory selective hearing mechanism,” inIEEE ICASSP, 2024

    2024

  42. [50]

    Robust frame-level speaker localization in reverberant and noisy envi- ronments by exploiting phase difference losses,

    S. S. Battula, H. Taherian, A. Pandey, D. Wong, B. Xu, and D. Wang, “Robust frame-level speaker localization in reverberant and noisy envi- ronments by exploiting phase difference losses,” inIEEE ICASSP, 2025

    2025

  43. [51]

    Librispeech: An ASR corpus based on public domain audio books,

    V . Panayotov, G. Chen, D. Povey, and S. Khudanpur, “Librispeech: An ASR corpus based on public domain audio books,” inIEEE ICASSP, 2015

    2015

  44. [52]

    LibriMix: An open-source dataset for generalizable speech separation,

    J. Cosentino, M. Pariente, S. Cornell, A. Deleforge, and E. Vincent, “LibriMix: An open-source dataset for generalizable speech separation,”

  45. [53]

    gpuRIR: A Python library for room impulse response simulation with GPU acceleration,

    D. Diaz-Guerra, A. Miguel, and J. R. Beltr ´an, “gpuRIR: A Python library for room impulse response simulation with GPU acceleration,” Multimedia Tools and Applications, vol. 80, 2018

    2018

  46. [54]

    Image method for efficiently simulating small- room acoustics,

    J. Allen and D. Berkley, “Image method for efficiently simulating small- room acoustics,”J. Acoust. Soc. Am., vol. 65, 1979

    1979

  47. [55]

    Generating sensor signals in isotropic noise fields,

    E. Habets and S. Gannot, “Generating sensor signals in isotropic noise fields,”J. Acoust. Soc. Am., vol. 122, 2007

    2007

  48. [56]

    Mask-based neural beamforming for moving speakers with self-attention-based tracking,

    T. Ochiai, M. Delcroix, T. Nakatani, and S. Araki, “Mask-based neural beamforming for moving speakers with self-attention-based tracking,” IEEE/ACM TASLP, vol. 31, 2023

    2023

  49. [57]

    Array geometry-robust attention-based neural beamformer for moving speakers,

    M. Tammen, T. Ochiai, M. Delcroix, T. Nakatani, S. Araki, and S. Doclo, “Array geometry-robust attention-based neural beamformer for moving speakers,” inInterspeech, 2024

    2024

  50. [58]

    Characterization of moving sound sources direction-of-arrival estimation using different deep learning architectures,

    J. Rusrus, S. Shirmohammadi, and M. Bouchard, “Characterization of moving sound sources direction-of-arrival estimation using different deep learning architectures,”IEEE TIM, vol. 72, 2023

    2023

  51. [59]

    Goldstein, J

    H. Goldstein, J. Safko, , and C. Poole,Classical mechanics. Addison- Wesley, 2002

    2002

  52. [60]

    Outdoor walking speeds of apparently healthy adults: A systematic review and meta-analysis,

    E. M. Murtagh, J. L. Mair, E. Aguiar, C. Tudor-Locke, and M. H. Murphy, “Outdoor walking speeds of apparently healthy adults: A systematic review and meta-analysis,”Sports Medicine, vol. 51, 2021

    2021

  53. [61]

    Specification of the social force pedestrian model by evolutionary adjustment to video tracking data,

    A. Johansson, D. Helbing, and P. K. Shukla, “Specification of the social force pedestrian model by evolutionary adjustment to video tracking data,”Advances in complex systems, vol. 10, 2007

    2007

  54. [62]

    SpatialNet: Extensively learning spatial information for multichannel joint speech separation, denoising and dereverberation,

    C. Quan and X. Li, “SpatialNet: Extensively learning spatial information for multichannel joint speech separation, denoising and dereverberation,” IEEE/ACM TASLP, vol. 32, 2024

    2024

  55. [63]

    Mamba: Linear-time sequence modeling with selective state spaces,

    A. Gu and T. Dao, “Mamba: Linear-time sequence modeling with selective state spaces,” inFirst conference on language modeling, 2024

    2024

  56. [64]

    Multichannel long-term streaming neural speech enhancement for static and moving speakers,

    C. Quan and X. Li, “Multichannel long-term streaming neural speech enhancement for static and moving speakers,”IEEE Signal Proc. Letters, vol. 31, 2024

    2024

  57. [65]

    Leveraging sound source trajectories for universal sound separation,

    D. Wu, X. Wu, and T. Qu, “Leveraging sound source trajectories for universal sound separation,”IEEE/ACM TASLP, 2025

    2025

  58. [66]

    Recurrent deep stacking networks for supervised speech separation,

    Z.-Q. Wang and D. Wang, “Recurrent deep stacking networks for supervised speech separation,” inIEEE ICASSP, 2017

    2017

  59. [67]

    Perceptual evaluation of speech quality (PESQ)-a new method for speech quality assessment of telephone networks and codecs,

    A. Rix, J. Beerends, M. Hollier, and A. Hekstra, “Perceptual evaluation of speech quality (PESQ)-a new method for speech quality assessment of telephone networks and codecs,” inIEEE ICASSP, 2001

    2001

  60. [68]

    An algorithm for predicting the intelligibility of speech masked by modulated noise maskers,

    J. Jensen and C. H. Taal, “An algorithm for predicting the intelligibility of speech masked by modulated noise maskers,”IEEE/ACM TASLP, vol. 24, 2016

    2016

  61. [69]

    Sound event localization and detection of overlapping sources using convolutional recurrent neural networks,

    S. Adavanne, A. Politis, J. Nikunen, and T. Virtanen, “Sound event localization and detection of overlapping sources using convolutional recurrent neural networks,”IEEE J. Sel. Topics Signal Proc., vol. 13, 2019

    2019

  62. [70]

    6DoF SELD: Sound event localization and detection using microphones and motion tracking sensors on self-motioning human,

    M. Yasuda, S. Saito, A. Nakayama, and N. Harada, “6DoF SELD: Sound event localization and detection using microphones and motion tracking sensors on self-motioning human,” inIEEE ICASSP, 2024

    2024

  63. [71]

    The LOCATA challenge: Acoustic source localization and tracking,

    C. Evers, H. W. L ¨ollmann, H. Mellmann, A. Schmidt, H. Barfuss, P. A. Naylor, and W. Kellermann, “The LOCATA challenge: Acoustic source localization and tracking,”IEEE/ACM TASLP, vol. 28, 2020

    2020

  64. [72]

    SDR – Half- baked or well done?

    J. L. Roux, S. Wisdom, H. Erdogan, and J. R. Hershey, “SDR – Half- baked or well done?” inIEEE ICASSP, 2019

    2019

  65. [73]

    Fairbanks,Voice and Articulation Drillbook

    G. Fairbanks,Voice and Articulation Drillbook. Harper, 1960

    1960

  66. [74]

    NISQA: A deep CNN- self-attention model for multidimensional speech quality prediction with crowdsourced datasets,

    G. Mittag, B. Naderi, A. Chehadi, and S. M ¨oller, “NISQA: A deep CNN- self-attention model for multidimensional speech quality prediction with crowdsourced datasets,” inInterspeech, 2021

    2021

  67. [75]

    NeMo: A toolkit for building AI applica- tions using neural modules,

    O. Kuchaiev, J. Li, H. Nguyen, O. Hrinchuk, R. Leary, B. Ginsburg, S. Kriman, S. Beliaev, V . Lavrukhin, J. Cook, P. Castonguay, M. Popova, J. Huang, and J. M. Cohen, “NeMo: A toolkit for building AI applica- tions using neural modules,” 2019. Jakob Kienegger(Student Member, I...

    2019

  68. [2020]

    Available: https://arxiv.org/abs/2005.11262

    [Online]. Available: https://arxiv.org/abs/2005.11262

    2005 arXiv

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.