Pith. sign in

REVIEW 3 major objections 5 minor 43 references

How much to Dereverberate? Low-Latency Single-Channel Speech Enhancement in Distant Microphone Scenarios

T0 review · 3 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read The paper establishes that single-channel, low-latency speech enhancement is feasible when the microphone is 5–10 m from the talker in large rooms, provided the training pipeline simulates room acoustics with a volume-dependent…

desk verdict The paper's practical guidance on how much reverb to leave in the target is solid and useful, but its headline SOTA claim rests on a confounded comparison and selective metrics. read the letter →

arxiv 2505.01338 v1 pith:2GX7GZYS submitted 2025-05-02 eess.AS cs.LGcs.SD

Satvik Venkatesh , Philip Coleman , Arthur Benilov , Simon Brown , Selim Sheta , Frederic Roskam This is my paper · ORCID
classification eess.AScs.LGcs.SD
keywords speechenhancementdereverberationlowlatencydistantmicrophoneroomimpulseresponseT60DNSMOSsinglechannel
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 asks whether a single microphone can clean up speech recorded 5 to 10 meters away in large rooms like conference halls and theatres, in real time with low latency. It argues yes: the bottleneck is not the architecture but the training data. Simulating room impulse responses with a volume-dependent reverberation time ($T_{60} = 0.145\ln(V) - 0.165$, ±20%) produces more realistic training conditions, and choosing a dereverberation target that keeps the first 300 ms of decay rather than removing all reverberation gives the best perceived quality. On a distant-microphone test set the adapted model raises DNSMOS OVRL from 2.77 to 3.04, which would matter for lecture capture, drama, and stage acoustics.

What carries the argument

The argument is carried by two data-generation choices. First, volume-based $T_{60}$ sampling: a curve-fit $T_{60} = 0.145\ln(V) - 0.165$, drawn from architectural acoustics guidelines and given ±20% random variation, replaces independent random sampling of room dimensions and reverberation time. Second, a windowing rule for the dereverberation target: the target signal is clean speech convolved with the original room impulse response multiplied by a gain window that is unity up to a chosen offset and then decays to −60 dB at a rate set by the target $T_{60}^{\max}$. The paper scans offsets from 0 to 80 ms and target decays of no-decay, 150, 300, and 500 ms, and selects a 0 ms offset with a 300 ms decay as the training configuration for the final comparisons.

What would settle it

Take a set of real conference rooms and theatres, measure their volumes and reverberation times, retrain the model with those measured pairs replacing the fitted curve, and compare DNSMOS OVRL on a fixed distant-microphone test set; if the curve-trained model does not beat independent random sampling, the proposed volume–$T_{60}$ coupling is not doing the work.

Watch

Extended reading notes

Core claim

The central discovery is that realistic room impulse response simulation matters more than model choice for distant-microphone speech enhancement. Randomly sampling $T_{60}$ and room volume independently creates implausible combinations, and networks trained on those learn little new: far-microphone training without volume-based sampling performs about the same as close-microphone training. Coupling $T_{60}$ to volume through the fitted log curve improves PESQ from 2.08 to 2.17 and OVRL DNSMOS from 2.64 to 2.69. The dereverberation target also matters: at distances up to 0.5 m the network should predict the direct sound, but at 5–10 m preserving early reflections (up to 80 ms) and decaying the impulse response to a 300 ms $T_{60}$ target is best, with OVRL scores reaching 2.86–3.02 depending on room. The resulting adapted model scores 3.04 OVRL versus 2.77 for the baseline.

Load-bearing premise

The paper's gains rest on a curve linking room volume to reverberation time, fitted to three published guidelines without reported error bars; if real conference and theatre acoustics deviate from that curve, the measured improvements may not transfer.

Editorial extensions

If this is right

  • Training pipelines for large-room enhancement should sample reverberation time conditionally on room volume; independent sampling yields near-zero gain over close-microphone training.
  • The dereverberation target should be matched to microphone distance: close microphones benefit from predicting the direct sound, while distant microphones benefit from preserving 30–80 ms of early reflections.
  • Keeping 300 ms of residual reverberation gives the best overall quality at distance; extending the target to 500 ms degrades background suppression, especially in large rooms.
  • The pipeline works at both 40 ms and 20 ms latency, with the lower-latency model also improving over published baselines on the distant-microphone test set.

Reading between the lines

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

  • A natural extension is to re-fit the volume–$T_{60}$ curve to measured pairs from real conference rooms and theatres; if the fitted curve is replaced by measured data, the paper's comparisons would directly test whether the volume coupling is necessary.
  • A testable extension is to train one network with two dereverberation targets—direct sound for close microphones and a 300 ms decay for distant microphones—using distance as a conditioning input, since the paper finds distance, not room size, is the controlling variable.
  • Because the paper relies on DNSMOS and a limited set of measured impulse responses, formal listening tests would clarify whether the preserved early reflections sound natural to human listeners.
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

3 major / 5 minor

Summary. The paper addresses single-channel low-latency speech enhancement in distant microphone scenarios (talker-to-mic distances of 5–10 m in large rooms such as conference rooms and theatres). It makes three main contributions: (i) a volume-based T60 sampling strategy for synthetic RIR generation grounded in architectural acoustics data; (ii) a systematic ANOVA-based study of dereverberation target design, specifically the offset before natural decay and the maximum residual T60 (T60max); (iii) an application of the resulting training pipeline to two real-time architectures (DFN3 and HSTN), reporting DNSMOS improvements over prior state-of-the-art models. Experiments use the DNS 2022 corpus, FRA-RIR and gpuRIR for synthetic RIRs, and a 100-example test set of real distant-microphone RIRs. The main quantitative findings are that volume-based T60 sampling improves PESQ/STOI/DNSMOS over unconstrained sampling (Table 2), that preserving early reflections with offsets up to 80 ms and a T60max of 300 ms is beneficial in distant scenarios (Section 4.2), and that DFN3-d.m. reaches an OVRL DNSMOS of 3.04 versus 2.77 for the original DFN3 (Table 3).

Significance. If the results are valid, the paper fills a genuine gap in the single-channel speech enhancement literature by systematically studying distant microphone scenarios and large rooms, which are underrepresented in prior benchmarks. The volume-based T60 sampling idea is a simple but useful correction to the common practice of independently sampling room dimensions and T60, and the ANOVA-based ablation of dereverberation targets is a structured and welcome empirical contribution. The paper also ships audio examples and supplementary material, which aids reproducibility. However, the headline state-of-the-art claim is not fully supported because the key comparison in Table 3 is confounded by multiple simultaneous changes, and the authors explicitly withhold intrusive metrics for two baselines. The lack of confidence intervals and listening tests further tempers the strength of the perceptual claims. These issues are fixable but require additional experiments and analysis.

major comments (3)
  1. [Section 4.3, Table 3] The comparison between DFN3-d.m. and the original DFN3 is not an ablation of the proposed pipeline: the two models differ in training data (DNS corpus versus the new distant-mic RIR sets), in the dereverberation target ((5,500) versus (0,300)), and in the test set. Any of these changes could account for the 0.27 OVRL gap. Moreover, the authors explicitly withhold PESQ, STOI, and SI-SDR for DFN3 and FSN+ because "these values are lower and give a wrong representation," which is selective reporting and prevents an apples-to-apples comparison. To support the state-of-the-art claim, the authors should either report all metrics for all models, or run controlled ablations that isolate each component (e.g., DFN3 trained on the standard corpus but with the new RIRs, or with the new target). At minimum, the text should clearly state that the comparison is a pipeline-level comparison and not a claim about model architecture alone.
  2. [Section 2.1, Equation (1)] The volume-T60 curve T60 = 0.145 ln(V) - 0.165 is central to the improvement reported in Table 2, but the manuscript provides no information about the curve-fit itself: how many data points were used, from which of references [24,25,26], what the fit residuals are, and what error bars or confidence intervals apply. The ±20% variation around the curve is also introduced without justification. If this curve does not represent the actual acoustics of conference rooms and theatres across the volume range 10^2 to 10^5 m^3, the observed gain in Table 2 may not generalize. Please provide the underlying data, fit statistics, and a rationale for the ±20% range, or validate the distribution against measured RIRs from a larger corpus.
  3. [Sections 3.1 and 4.2] The ANOVA-based conclusions about offset and T60max are systematic and well structured, but they rest on a synthetic gpuRIR test set and use DNSMOS as the sole endpoint. No confidence intervals or paired significance values are reported for the DNSMOS differences themselves (the asterisks in Figure 3 indicate pairwise t-test results but not the magnitude uncertainty). Section 5 states that listening tests are future work, which is a significant limitation for a task whose evaluation is inherently perceptual. To strengthen the practical recommendations (e.g., choosing (0,300) over (30,300)), the authors should report confidence intervals for the main effects in Figure 3, and ideally include a small listening experiment on the real test set used in Section 4.3.
minor comments (5)
  1. [Section 3.2] Typo: "Nework Architectures" should be "Network Architectures".
  2. [Table 2 caption] The caption reads "MOS MOS MOS (SIG) (BAK) (OVL)", which is confusing because the header repeats MOS. Please revise the header to clearly separate the DNSMOS subscales.
  3. [Section 2.2, Equation (2)] The definition of N1 as "the direct sound" is ambiguous; specify whether N1 is the sample index of the direct-sound arrival or the end of the direct-sound region.
  4. [Section 4.2] The text reports that in the small-room distant-microphone condition, "(0, 300) and (30, 300) obtain the highest OVRL score of 2.86", while later it says "(30, 300) obtains the highest OVRL score of 3.02" for the large room. These numbers are presented without a direct link to Figure 3 panels, which may confuse readers; please clarify which condition each value refers to and ensure consistency with Figure 3.
  5. [References] Reference [24] (Harris handbook) lacks volume and page details; please complete the bibliographic information or use a consistent citation style.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper's claims are empirical and evaluated on external test data with external metrics.

full rationale

The paper does not derive its conclusions from fitted parameters that are then presented as predictions. Equation 1 (T60 = 0.145 ln(V) - 0.165) is explicitly curve-fitted to external acoustics guidelines [24, 25, 26], and the later Table 2 comparison evaluates the sampling strategy on held-out real RIRs from [31], [32], and in-house recordings; the improvement is not an algebraic consequence of the fitted curve. The dereverberation target (0, 300 ms) is selected after an ANOVA-style comparison on a gpuRIR test set, which is model selection rather than a prediction forced by construction. The final claim (DFN3-d.m. OVRL 3.04 vs DFN3 2.77) is measured with DNSMOS, an external non-intrusive metric trained on human ratings [35], on a test set of 100 examples. The paper's self-citation of HSTN [33] provides an architecture used as a secondary model, but that citation is not load-bearing evidence for the central dereverberation or distance-scenario claims. The manuscript itself flags that intrusive metrics for the original DFN3 are withheld and that listening tests are deferred to future work; these are correctness and evidence-strength concerns, not circularity. No step reduces to its own input by definition, and no uniqueness theorem or prior result by the same authors is invoked to forbid alternatives. Therefore the circularity score is 0.

Assumptions & free parameters 1 free parameters · 3 assumptions · 0 invented entities

The paper introduces no new physical entities. It relies on one fitted empirical relation (Equation 1), a synthetic-RIR realism assumption, and a neural-metric validity assumption. The central empirical findings are otherwise measured against external test sets and external metrics.

free parameters (1)
  • T60-volume curve coefficients a and b = a=0.145, b=0.165
    Curve-fitted to literature guidelines for T60 versus volume in conference rooms from references [24,25,26]; no fit statistics or error bars are reported. The plus-or-minus 20 percent spread around the curve is also chosen by hand (Section 2.1, Equation 1).
assumptions (3)
  • domain assumption Synthetic room impulse responses from gpuRIR and FRA-RIR accurately represent real distant microphone, large-room acoustics
    Training uses synthetic RIRs (Section 3.1), and the controlled study in Section 4.2 tests on synthetic gpuRIR RIRs to precisely control room size and distance, assuming these match real conditions when drawing conclusions about offsets and decay targets.
  • domain assumption DNSMOS is a valid proxy for human perceived quality
    All conclusions about dereverberation target and early reflection preservation (Section 4.2) and the state-of-the-art comparison (Table 3) are based on DNSMOS, a non-intrusive neural network predictor, not human listening tests. The authors note listening tests as future work.
  • ad hoc to paper T60 = 0.145 ln(V) - 0.165 is a representative summary of conference-room and theatre acoustics across volumes from about 10^2 to 10^5 cubic meters
    Introduced in Equation 1 as a curve-fit to three literature sources [24,25,26] with no fit quality reported; it is central to the volume-based T60 sampling recommendation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of How much to Dereverberate? Low-Latency Single-Channel Speech Enhancement in Distant Microphone Scenarios." pith.science (2026). https://pith.science/paper/2GX7GZYS

@misc{pith2026250501338,
  author       = {Pith},
  title        = {Pith review of: How much to Dereverberate? Low-Latency Single-Channel Speech Enhancement in Distant Microphone Scenarios},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2GX7GZYS}},
  note         = {Machine review of arXiv:2505.01338}
}
read the original abstract

Dereverberation is an important sub-task of Speech Enhancement (SE) to improve the signal's intelligibility and quality. However, it remains challenging because the reverberation is highly correlated with the signal. Furthermore, the single-channel SE literature has predominantly focused on rooms with short reverb times (typically under 1 second), smaller rooms (under volumes of 1000 cubic meters) and relatively short distances (up to 2 meters). In this paper, we explore real-time low-latency single-channel SE under distant microphone scenarios, such as 5 to 10 meters, and focus on conference rooms and theatres, with larger room dimensions and reverberation times. Such a setup is useful for applications such as lecture demonstrations, drama, and to enhance stage acoustics. First, we show that single-channel SE in such challenging scenarios is feasible. Second, we investigate the relationship between room volume and reverberation time, and demonstrate its importance when randomly simulating room impulse responses. Lastly, we show that for dereverberation with short decay times, preserving early reflections before decaying the transfer function of the room improves overall signal quality.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 36 canonical work pages

  1. [1]

    INTRODUCTION Speech enhancement (SE) or speech restoration tries to improve the intelligibility and quality of speech contaminated by additive noise [1, 2], reverberation [3, 4], clipping, and low sampling rates [5]. The topic has been addressed by various machine learning challenges such as Deep Noise Suppression (DNS) [6, 7], REVERB [8], Com- putation H...

    work page 2021

  2. [2]

    PROPOSED METHOD 2.1. V olume-based T60 Sampling Most studies for dereverberation use synthetic RIRs generated through the image source method (ISM), or a hybrid model that combines ISM and diffuse reverberation [21, 4, 22]. These RIR sim- ulators are fed random values within specified ranges. For example, room dimensions in the range of (3, 3, 2.5) m and ...

    work page Pith review arXiv 2025

  3. [3]

    Dataset We used the clean speech, noise, and RIR dataset provided by the DNS challenge 2022 [6]

    EXPERIMENTAL SETUP 3.1. Dataset We used the clean speech, noise, and RIR dataset provided by the DNS challenge 2022 [6]. The clean speech is a corpora of vari- ous datasets such as VCTK [28], PTDB [29], and read speech from Wall street journal [6]. To manage compute storage constraints, we trained only on the English dataset, similar to [27]. To generate ...

    work page 2022

  4. [4]

    RESULTS To evaluate the SE models, we adopt Perceptual Evaluation of Speech Quality (PESQ), Short-Time Objective Intelligibility (STOI), and Scale-Invariant Signal-to-Distortion Ratio (SI-SDR). We also present the Deep Noise Suppression Mean Opinion Score (DNS- MOS), which is a neural network that predicts the perceptual eval- uation score for three facto...

    work page

  5. [5]

    We demonstrated that SE under such challenging scenarios is feasible and obtained state-of-the-art per- formance for this task

    CONCLUSION In this paper, we investigated real-time low-latency SE under dis- tant microphone scenarios. We demonstrated that SE under such challenging scenarios is feasible and obtained state-of-the-art per- formance for this task. When simulating RIRs, it was helpful to consider the volume of the simulated room. This helps us generate more realistic RIR...

    work page

  6. [6]

    ICASSP 2022 deep noise suppression challenge,

    Harishchandra Dubey, Vishak Gopal, Ross Cutler, et al., “ICASSP 2022 deep noise suppression challenge,” in IEEE ICASSP, 2022, pp. 9271–9275

    work page 2022

  7. [7]

    Music source separation in the waveform domain,

    Alexandre D ´efossez, Nicolas Usunier, L´eon Bottou, and Francis Bach, “Music source separation in the waveform domain,” arXiv preprint arXiv:1911.13254, 2019

    arXiv 1911

  8. [8]

    V oicefixer: A uni- fied framework for high-fidelity speech restoration,

    Haohe Liu, Xubo Liu, Qiuqiang Kong, et al., “V oicefixer: A uni- fied framework for high-fidelity speech restoration,” arXiv preprint arXiv:2204.05841, 2022

    arXiv 2022

Show all 43 references
  1. [9]

    Real-time single-channel dereverbera- tion and separation with time-domain audio separation network.,

    Yi Luo and Nima Mesgarani, “Real-time single-channel dereverbera- tion and separation with time-domain audio separation network.,” in Interspeech, 2018, pp. 342–346

    2018

  2. [10]

    Whamr!: Noisy and reverberant single-channel speech separation,

    Matthew Maciejewski, Gordon Wichern, Emmett McQuinn, and Jonathan Le Roux, “Whamr!: Noisy and reverberant single-channel speech separation,” in Proc. IEEE ICASSP, 2020, pp. 696–700

    2020

  3. [11]

    Audio super- resolution using neural nets,

    V olodymyr Kuleshov, S Zayd Enam, and Stefano Ermon, “Audio super- resolution using neural nets,” in ICLR (Workshop Track), 2017

    2017

  4. [12]

    Real time speech enhancement in the waveform domain,

    Alexandre Defossez, Gabriel Synnaeve, and Yossi Adi, “Real time speech enhancement in the waveform domain,” arXiv preprint arXiv:2006.12847, 2020

    2006 arXiv

  5. [13]

    ICASSP 2023 deep speech enhancement challenge,

    Harishchandra Dubey, Ashkan Aazami, Vishak Gopal, et al., “ICASSP 2023 deep speech enhancement challenge,” arXiv preprint arXiv:2303.11510, 2023

    2023 arXiv

  6. [14]

    A summary of the REVERB challenge: state-of-the-art and remaining challenges in reverberant speech processing research,

    Keisuke Kinoshita, Marc Delcroix, Sharon Gannot, et al., “A summary of the REVERB challenge: state-of-the-art and remaining challenges in reverberant speech processing research,” EURASIP Journal on Ad- vances in Signal Processing, vol. 2016, pp. 1–19, 2016

    2016

  7. [15]

    The CHiME-7 DASR challenge: Distant meeting transcription with mul- tiple devices in diverse scenarios,

    Samuele Cornell, Matthew Wiesner, Shinji Watanabe, et al., “The CHiME-7 DASR challenge: Distant meeting transcription with mul- tiple devices in diverse scenarios,” arXiv preprint arXiv:2306.13734 , 2023

    2023 arXiv

  8. [16]

    TPARN: Triple-path attentive recur- rent network for time-domain multichannel speech enhancement,

    Ashutosh Pandey, Buye Xu, Anurag Kumar, Jacob Donley, Paul Calamia, and DeLiang Wang, “TPARN: Triple-path attentive recur- rent network for time-domain multichannel speech enhancement,” in Proc. IEEE ICASSP, 2022, pp. 6497–6501

    2022

  9. [17]

    Deft-an: Dense frequency-time attentive network for multichannel speech enhancement,

    Dongheon Lee and Jung-Woo Choi, “Deft-an: Dense frequency-time attentive network for multichannel speech enhancement,” IEEE Signal Processing Letters, vol. 30, pp. 155–159, 2023

    2023

  10. [18]

    Here, the decay rate of the window is adaptive and depends on the T60 of the original impulse response

    proposed reverb-time shortening, where they naturally decay the room impulse response to a targetT60, instead of a constant win- dow function. Here, the decay rate of the window is adaptive and depends on the T60 of the original impulse response. In this case, q ={3/(T max 60 ...

  11. [19]

    Speech enhancement with multichannel wiener filter techniques in multimicrophone binaural hearing aids,

    Tim Van den Bogaert, Simon Doclo, Jan Wouters, and Marc Moo- nen, “Speech enhancement with multichannel wiener filter techniques in multimicrophone binaural hearing aids,” The Journal of the Acous- tical Society of America , vol. 125, no. 1, pp. 360–371, 2009

    2009

  12. [20]

    An analysis of environment, microphone and data simulation mismatches in robust speech recognition,

    Emmanuel Vincent, Shinji Watanabe, Aditya Arie Nugraha, Jon Barker, and Ricard Marxer, “An analysis of environment, microphone and data simulation mismatches in robust speech recognition,” Com- puter Speech & Language, vol. 46, pp. 535–557, 2017

    2017

  13. [21]

    ICASSP 2021 deep noise suppression challenge,

    Chandan KA Reddy, Harishchandra Dubey, Vishak Gopal, et al., “ICASSP 2021 deep noise suppression challenge,” in Proc. IEEE ICASSP, 2021, pp. 6623–6627

    2021

  14. [22]

    Fullsubnet+: Channel atten- tion fullsubnet with complex spectrograms for speech enhancement,

    Jun Chen, Zilin Wang, Deyi Tuo, et al., “Fullsubnet+: Channel atten- tion fullsubnet with complex spectrograms for speech enhancement,” in Proc. IEEE ICASSP, 2022, pp. 7857–7861

    2022

  15. [23]

    Effects of early and late reflections on intelligibility of reverberated speech by cochlear implant listeners,

    Yi Hu and Kostas Kokkinakis, “Effects of early and late reflections on intelligibility of reverberated speech by cochlear implant listeners,” The Journal of the Acoustical Society of America , vol. 135, no. 1, pp. EL22–EL28, 2014

    2014

  16. [24]

    Speech dereverberation with a reverberation time shortening target,

    Rui Zhou, Wenye Zhu, and Xiaofei Li, “Speech dereverberation with a reverberation time shortening target,” in ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2023, pp. 1–5

    2023

  17. [25]

    Towards efficient models for real-time deep noise suppression,

    Sebastian Braun, Hannes Gamper, Chandan KA Reddy, and Ivan Ta- shev, “Towards efficient models for real-time deep noise suppression,” in Proc. IEEE ICASSP, 2021, pp. 656–660

    2021

  18. [26]

    Deepfil- ternet: Perceptually motivated real-time speech enhancement,

    Hendrik Schr ¨oter, Tobias Rosenkranz, Andreas Maier, et al., “Deepfil- ternet: Perceptually motivated real-time speech enhancement,” arXiv preprint arXiv:2305.08227, 2023

    2023 arXiv

  19. [27]

    Pyroomacous- tics: A python package for audio room simulation and array processing algorithms,

    Robin Scheibler, Eric Bezzam, and Ivan Dokmani ´c, “Pyroomacous- tics: A python package for audio room simulation and array processing algorithms,” in Proc. IEEE ICASSP, 2018, pp. 351–355

    2018

  20. [28]

    gpurir: A python library for room impulse response simulation with gpu acceler- ation,

    David Diaz-Guerra, Antonio Miguel, and Jose R Beltran, “gpurir: A python library for room impulse response simulation with gpu acceler- ation,” Multimedia Tools and Applications , vol. 80, no. 4, pp. 5653– 5671, 2021

    2021

  21. [29]

    Multi-scale temporal frequency convolutional network with axial at- tention for speech enhancement,

    Guochang Zhang, Libiao Yu, Chunliang Wang, and Jianqiang Wei, “Multi-scale temporal frequency convolutional network with axial at- tention for speech enhancement,” in Proc. IEEE ICASSP , 2022, pp. 9122–9126

    2022

  22. [30]

    1960, McGraw-Hill New York, 1957

    Cyril M Harris and Cyril Manton Harris, Handbook of noise control , vol. 1960, McGraw-Hill New York, 1957

    1960

  23. [31]

    Acoustics and architecture in italian catholic churches,

    Etore Cirillo and Francesco Martellotta, “Acoustics and architecture in italian catholic churches,” inInternational Symposium on Room Acous- tics (ISRA), 2007

    2007

  24. [32]

    Influence of proportion towards speech intelligibility in mosque’s praying hall,

    Ahmad Ridzwan Othman and Mohamed Rizal Mohamed, “Influence of proportion towards speech intelligibility in mosque’s praying hall,” Procedia-Social and Behavioral Sciences, vol. 35, pp. 321–329, 2012

    2012

  25. [33]

    Deepfilternet2: Towards real-time speech enhancement on embedded devices for full-band audio,

    Hendrik Schr ¨oter, A Maier, Alberto N Escalante-B, and Tobias Rosenkranz, “Deepfilternet2: Towards real-time speech enhancement on embedded devices for full-band audio,” in IEEE IWAENC, 2022, pp. 1–5

    2022

  26. [34]

    CSTR VCTK Corpus: English multi-speaker corpus for CSTR voice cloning toolkit (version 0.92),

    Christophe; MacDonald Kirsten Yamagishi, Junichi; Veaux, “CSTR VCTK Corpus: English multi-speaker corpus for CSTR voice cloning toolkit (version 0.92),” University of Edinburgh. The Centre for Speech Technology Research (CSTR), 2019

    2019

  27. [35]

    A pitch tracking corpus with evaluation on multipitch tracking sce- nario.,

    Gregor Pirker, Michael Wohlmayr, Stefan Petrik, and Franz Pernkopf, “A pitch tracking corpus with evaluation on multipitch tracking sce- nario.,” in Interspeech, 2011, pp. 1509–1512

    2011

  28. [36]

    Fra-rir: Fast random approximation of the image-source method,

    Yi Luo and Jianwei Yu, “Fra-rir: Fast random approximation of the image-source method,” arXiv preprint arXiv:2208.04101, 2022

    2022 arXiv

  29. [37]

    S3a room impulse re- sponses,

    P Coleman, L Remaggi, and PJB Jackson, “S3a room impulse re- sponses,” 2020

    2020

  30. [38]

    Openair: An interactive au- ralization web resource and database,

    Damian T Murphy and Simon Shelley, “Openair: An interactive au- ralization web resource and database,” in Audio Engineering Society Convention 129. Audio Engineering Society, 2010

    2010

  31. [39]

    Real-time low-latency music source separation using hybrid spectrogram-tasnet,

    Satvik Venkatesh, Arthur Benilov, Philip Coleman, and Frederic Roskam, “Real-time low-latency music source separation using hybrid spectrogram-tasnet,” in Proc. IEEE ICASSP, 2024, pp. 611–615

    2024

  32. [40]

    FRCRN: Boosting feature representation using frequency recurrence for monaural speech enhancement,

    Shengkui Zhao, Bin Ma, Karn N Watcharasupat, and Woon-Seng Gan, “FRCRN: Boosting feature representation using frequency recurrence for monaural speech enhancement,” in Proc. IEEE ICASSP, 2022, pp. 9281–9285

    2022

  33. [41]

    Dnsmos: A non- intrusive perceptual objective speech quality metric to evaluate noise suppressors,

    Chandan KA Reddy, Vishak Gopal, and Ross Cutler, “Dnsmos: A non- intrusive perceptual objective speech quality metric to evaluate noise suppressors,” in Proc. IEEE ICASSP, 2021, pp. 6493–6497

    2021

  34. [42]

    Data augmentation and loss nor- malization for deep noise suppression,

    Sebastian Braun and Ivan Tashev, “Data augmentation and loss nor- malization for deep noise suppression,” in International Conference on Speech and Computer. Springer, 2020, pp. 79–86

    2020

  35. [43]

    Icassp 2023 speech signal im- provement challenge,

    Ross Cutler, Ando Saabas, Babak Naderi, Nicolae-C ˘at˘alin Ristea, Se- bastian Braun, and Solomiya Branets, “Icassp 2023 speech signal im- provement challenge,” IEEE Open Journal of Signal Processing, 2024. 5

    2023

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.