Complex-Cycle-Consistent Diffusion Model for Monaural Speech Enhancement

Plamen Angelov; Yang Sun; Yi Li

REVIEW 4 major objections 5 minor 40 references

Complex-Cycle-Consistent Diffusion Model for Monaural Speech Enhancement

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 SEDM couples magnitude and phase diffusion with a cycle-consistent loss and reports new best scores on VCTK and DNS benchmarks.

desk verdict A promising diffusion-based speech enhancement recipe with a genuinely new magnitude-phase cycle-consistency idea, but the reverse process is underspecified around a load-bearing noise term, and the paper is not reproducible as written. read the letter →

arxiv 2412.08856 v1 pith:N67SGZI6 submitted 2024-12-12 cs.SD eess.AS

Yi Li , Yang Sun , Plamen Angelov This is my paper · ORCID

classification cs.SDeess.AS

keywords speechenhancementdiffusionmodelcycle-consistentlearningphaseestimationmagnitudeandreal-worldnoisemonauralspectrogram

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 tries to establish that a diffusion model for monaural speech enhancement works best when the magnitude and phase of the speech spectrogram are estimated by two separate networks, when the forward corruption uses real-world noise clips instead of Gaussian noise, and when a complex-cycle-consistent (CCC) loss forces the estimated magnitude and phase to map back and forth into each other. If this is right, the coupling of magnitude and phase is an important source of gains: the largest model, SEDM-L, reports PESQ 3.59 on VCTK, beating the previous best diffusion method NASE by 0.58 PESQ and 0.04 ESTOI. The result is a concrete recipe for phase-aware enhancement that other systems could adopt.

What carries the argument

The load-bearing machinery is the pair of separate diffusion paths for magnitude and phase plus the Complex-Cycle-Consistent (CCC) block. In the forward pass, real-world noise clips selected from the DEMAND database are progressively added to the clean magnitude and phase spectra through an SDE. In the noise-aware reverse pass, each step obeys $Y_{m-1} = \frac{1}{\gamma_m}(Y_m - \frac{\theta_m}{1-\gamma_m} I_m) + \sigma_m$, where $I_m$ is an estimated noise spectrum, and the two reverse branches output estimated magnitude $S_A$ and phase $S_P$. The CCC block then cycles $S_A$ through a mapping to phase and $S_P$ through a mapping to magnitude, adding cycle losses to the direct L2 losses so that each modality is trained to be consistent with the other. This cycle is the mechanism the paper claims exploits the intrinsic magnitude-phase relationship.

What would settle it

A direct test would be to run SEDM-L on VCTK test utterances while replacing the estimated noise spectrum in the reverse update with the true noise clip, and again with a zero or mismatched noise spectrum; if the PESQ difference between these runs is large, the method is silently relying on oracle noise information that the paper does not describe how to obtain at test time.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that phase and magnitude in the short-time Fourier transform are not independent targets: a cyclic mapping between them acts as a training signal that sharpens both estimates. The paper implements this in SEDM, where a forward process injects real-world noise clips into clean magnitude and phase spectra and a noise-aware reverse process estimates both clean and noise spectra at every step. Two LSTM-based cycle-consistency blocks then translate estimated magnitude to phase and estimated phase to magnitude, and the reconstruction errors are added to the direct estimation losses. With 50 diffusion blocks, SEDM-L reaches PESQ 3.59, CSIG 4.06, CBAK 4.22, COVL 4.89, and SSNR 11.8 on VCTK, and 2.93 wide-band PESQ on the DNS challenge, outperforming the compared state-of-the-art methods on most metrics. Ablations attribute the gain to the combination of the diffusion model, phase-aware estimation, and CCC, with the full model reaching PESQ 2.75 versus 2.21 without any of the three contributions.

Load-bearing premise

The reverse update formula requires the noise spectrum $I_m$ to be available or accurately estimated for the test utterance, yet the paper describes training with known noise clips and gives no loss or procedure that explains how the estimated noise is produced and fed into that equation at inference. If the test-time noise is not available in this form, the reverse process lacks a well-defined input and the reported comparisons would not transfer outside the training setup.

Editorial extensions

If this is right

Following the SEDM recipe, with separate magnitude and phase diffusion branches, real-world noise corruption, and a CCC loss, yields the best reported scores among the compared methods on the VCTK and DEMAND setup, with SEDM-L at PESQ 3.59.
The ablation shows each ingredient contributes, with diffusion alone lifting PESQ from 2.21 to 2.43, phase-aware estimation adding further gains, and CCC bringing the full model to 2.75.
Diffusion models trained on real-world noise clips become competitive with Gaussian-trained models as network depth grows, and remain robust on seen noise while degrading on unseen noise types.
The CCC cycle makes magnitude and phase estimates mutually consistent, so the enhanced speech no longer has to borrow the phase of the noisy mixture.

Reading between the lines

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

A step the paper leaves implicit is an explicit loss or procedure for estimating the noise spectrum $I_m$ at test time; the current reverse update appears to require that spectrum without describing how it is supervised.
The CCC principle should transfer to other complex-spectrum regression tasks, such as source separation or dereverberation, wherever magnitude and phase are estimated jointly.
The comparison against Gaussian-trained models suggests that real-world noise corruption pays off most for deep networks; one testable extension is to measure how the gap scales with the number of diffusion blocks.
If the reverse update's noise spectrum is in fact an oracle at test time, the reported gains would shrink under a blind evaluation; a public implementation with a clearly specified noise-estimation path would settle this.

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.

Desk Editor's Note

read the letter

The paper combines several existing ideas in a way that is genuinely new for speech enhancement: separate magnitude and phase diffusion networks, real-world noise clips in the forward process, and a cycle-consistent loss between magnitude and phase. The benchmark numbers are strong (e.g., PESQ 3.59 on VCTK, beating SCP-GAN by 0.07) and the ablation study is a sensible attempt to credit each contribution. If the numbers are accurate, this is a step forward for the subfield.

The problem is that the method as described cannot be reproduced. The reverse update in Eq. (4) depends on a noise term I_m, but the paper never says where I_m comes from at test time. The reverse network is said to 'estimate both clean speech and noise spectra,' yet no loss supervises the noise estimate and no procedure feeds it into Eq. (4). This is not a minor typo; it is the core generative step, so the reported tables cannot be attributed to the described algorithm. The SDE in Eq. (3) is also loosely written: dI_n is not a standard increment, and the stochastic integral is undefined. Add to that the mis-referenced equations in Algorithm 1 and the absence of code or data splits, and a reader cannot verify anything.

The empirical claims are also a bit oversold. On the DNS Challenge, SEDM-L does not beat RemixIT on WP or SI-SDR; it wins only on NP and STOI. The text acknowledges this as 'competitive,' but the abstract and conclusion claim general superiority. The gains over the strongest baselines are modest, and there are no error bars, so it is hard to know if the differences are meaningful.

That said, the core idea is clear and potentially useful. The missing details are fixable: define I_m properly, specify the training loss for the noise estimate, clean up the SDE, release code or a detailed derivation. This is exactly the kind of paper that should go to peer review, not be desk-rejected, because the conceptual contribution is solid and the flaws are mostly in presentation and specification. A careful referee could push the authors to make the method reproducible.

Who gets value from this: researchers working on diffusion-based speech enhancement, and anyone thinking about phase-magnitude coupling in generative models. I would not cite it in its current form, but I would read a revised version. Send it to review with an expectation of major revision.

Referee Report

4 major / 5 minor

Summary. The paper proposes SEDM, a diffusion-model speech enhancement method that separates magnitude and phase estimation into two networks, replaces Gaussian noise with real-world noise clips in the forward process, and uses a complex-cycle-consistent (CCC) loss to couple magnitude and phase estimates. The authors report results on IEEE, TIMIT, VCTK, and DNS challenge benchmarks, claiming state-of-the-art performance, e.g., SEDM-L reaching PESQ 3.59 on VCTK and outperforming NASE by 0.58 PESQ and 0.04 ESTOI.

Significance. If the described algorithm were fully specified and reproducible, the reported gains would be significant for monaural speech enhancement, particularly the combination of separate magnitude/phase diffusion and a cycle-consistent coupling mechanism. The paper includes a broad set of benchmarks, an ablation study, and a claim of fair baseline reproduction for the non-diffusion competitors, which are strengths. However, the central generative step is currently underspecified, and the reported numbers cannot be assigned to the described algorithm without additional definitions.

major comments (4)

[Reverse Process, Eqs. (4)-(5)] The reverse update in Eq. (4) depends on a term I_m whose definition and test-time provenance are never given. The text states that 'the reverse network estimates both clean speech and noise spectra,' but no network output, training loss, or inference procedure is specified for I_m; during test the diffusion network is discarded and the only input is the noisy mixture. Without a precise definition of I_m (e.g., an estimated noise spectrogram with a supervised loss), Eq. (4) is not computable and the benchmark results in Tables 2 and 4 cannot be attributed to the described algorithm. This is the central load-bearing gap.
[Diffusion Process, Eqs. (2)-(3)] The forward process is written as an SDE dY_n = μ(Y_n,n)dn + σ(Y_n,n)dI_n, where dI_n is the increment of a randomly selected real-noise clip rather than a Wiener process; no stochastic integral or discrete transition q(Y_n|Y_{n-1}) is defined. Consequently, the claimed forward process is not a standard diffusion and the reverse update is not derived from it. The scheduling parameters γ_m, θ_m and the quantities \barγ_m appearing in Eq. (5) are also undefined. Please replace Eq. (3) with an explicit discrete forward transition and derive the reverse posterior, or state clearly that Eq. (4) is a heuristic update rather than the exact reverse of Eq. (3).
[Complex-Cycle-Consistent Learning, Eqs. (6)-(9) and Algorithm 1] The CCC loss definitions contain internal inconsistencies that prevent reproduction. The loss L_{SA} in Eq. (7) is never defined; the text says the magnitude loss is estimated 'with (3),' but Eq. (3) is the forward SDE. Algorithm 1 line 10 says 'Update L_{S'_A} as (8),' but Eq. (8) defines the phase reconstruction loss L_{SA→P}, and line 12 refers to a nonexistent Eq. (10). Please renumber the equations and align Algorithm 1 with the displayed loss definitions.
[Comparison to Other Diffusion Models, Table 4] For the diffusion-model comparison in Table 4, the paper does not state whether the baseline numbers were reproduced under the same training data, SNR conditions, and evaluation pipeline as the proposed SEDM, unlike the statement made for the competitors in Section 'Competitors.' Since the claimed 0.58 PESQ gain over NASE is a headline result, please specify the provenance of each baseline entry and confirm that the settings match Table 2.

minor comments (5)

[Quantitative Results] The table captions are inconsistent: Table 6 is referred to as 'Table 1' in the text and Table 7 is referred to as 'Table 2,' and the heading 'Benckmarks on TIMIT' is a typo.
[Figure 1 caption] The caption says 'left side of Figure 1' for both the diffusion and reverse networks; the reverse network should be on the right side.
[Introduction] The noise type names 'dwashing,' 'dliving,' and 'pstation' appear to be typos for the DEMAND classes 'dwashing,' 'dliving,' and 'pstation.'
[Eq. (1) and CCC notation] The symbol S_{t,f} is used for the complex STFT in Eq. (1), while S_A and S_P later denote magnitude and phase spectra; please define these symbols consistently.
[Ablation Study] The baseline 'ResNet152' is mentioned, but the input and output of this non-diffusion baseline are not described; please clarify how it is used.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the central claim is an external benchmark comparison, and the model's novelty components are training objectives rather than fitted predictions.

full rationale

The paper's central claim is that the SEDM family outperforms existing speech enhancement methods on public datasets (IEEE, TIMIT, VCTK, DNS), with results reported in Tables 1-4. These comparisons are external and do not reduce to the model's own outputs. The three claimed contributions are: (1) using real-world noise clips from DEMAND in the forward diffusion process, (2) a noise-aware reverse process that estimates clean and noise spectra, and (3) a complex-cycle-consistent (CCC) loss that maps between magnitude and phase. None of these is a fitted parameter renamed as a prediction; the CCC terms (Eqs. 6-9) are training regularizers computed from the current estimates and the clean reference, not benchmark predictions derived from the same data. The reverse-process update in Eq. (4) contains an underspecified noise term I_m whose test-time provenance is not fully defined; however, underspecification is a reproducibility or correctness concern, not circularity, because no equation in the paper defines a target quantity in terms of itself or fits a parameter to the quantity it later claims to predict. The only self-citations (Li et al. 2021b and Li et al. 2023) appear in the introduction and architecture description as related work and architectural inspiration, and they do not carry the paper's central claim. No uniqueness theorem is imported from the authors' prior work, and no known result is renamed as a new contribution. The derivation chain is therefore self-contained with respect to circularity, and the score is 0.

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

The central result depends on the assumed validity of a real-noise diffusion process, the test-time availability of the noise clip, and the CCC loss formulation; none of these is independently verified, and two loss weights are unreported free parameters.

free parameters (4)

lambda_1 (CCC phase-to-magnitude loss weight) = not reported
Weight in Eq. (7) attenuating the cycle loss; no value or tuning procedure given.
lambda_2 (CCC magnitude-to-phase loss weight) = not reported
Weight in Eq. (9); no value or tuning procedure given.
gamma_m, theta_m (reverse-step parameters) = trainable, values not reported
Trainable scalar parameters in every reverse step (Eqs. 4-5); no initialization or schedule described.
architecture sizes (N, C, K) = small/medium/large: [30,63], [40,128], [50,128], K=3
Model capacities chosen by hand; ablation varies N only and does not control for parameter count versus baselines.

assumptions (4)

ad hoc to paper Replacing Gaussian noise with a randomly selected real-world noise clip in each diffusion embedding yields a valid forward diffusion process with a tractable reverse step (Eqs. 2-5).
The forward SDE dYn = mu*n + sigma*dIn uses a fixed clip In as a driving term without defining the resulting stochastic process or showing the reverse Eq. (4) is its posterior inverse.
ad hoc to paper At test time, the reverse network knows (or can accurately estimate) the noise clip Im that appears in the reverse update Eq. (4).
The test-stage paragraph says only that the reverse network is used to enhance noisy speech; no mechanism is given for obtaining the unknown Im, yet Eq. (4) depends on it linearly.
domain assumption The LSTM-based CCC mappings between magnitude and phase operationalize the intrinsic amplitude-phase relationship described in Shimauchi et al. 2017.
The paper cites the group-delay/instantaneous-frequency relationship as motivation but never uses it in the loss; the learned mappings are not shown to recover that relationship.
domain assumption Separately estimating STFT magnitude and phase, then recombining them, is a sufficient representation for reconstructing clean speech.
Standard assumption in phase-aware enhancement; the paper does not analyze the effect of phase wrapping or the interaction between the two estimates on final audio quality.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Complex-Cycle-Consistent Diffusion Model for Monaural Speech Enhancement." pith.science (2026). https://pith.science/paper/N67SGZI6

@misc{pith2026241208856,
  author       = {Pith},
  title        = {Pith review of: Complex-Cycle-Consistent Diffusion Model for Monaural Speech Enhancement},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/N67SGZI6}},
  note         = {Machine review of arXiv:2412.08856}
}

read the original abstract

In this paper, we present a novel diffusion model-based monaural speech enhancement method. Our approach incorporates the separate estimation of speech spectra's magnitude and phase in two diffusion networks. Throughout the diffusion process, noise clips from real-world noise interferences are added gradually to the clean speech spectra and a noise-aware reverse process is proposed to learn how to generate both clean speech spectra and noise spectra. Furthermore, to fully leverage the intrinsic relationship between magnitude and phase, we introduce a complex-cycle-consistent (CCC) mechanism that uses the estimated magnitude to map the phase, and vice versa. We implement this algorithm within a phase-aware speech enhancement diffusion model (SEDM). We conduct extensive experiments on public datasets to demonstrate the effectiveness of our method, highlighting the significant benefits of exploiting the intrinsic relationship between phase and magnitude information to enhance speech. The comparison to conventional diffusion models demonstrates the superiority of SEDM.

Figures

Figures reproduced from arXiv: 2412.08856 by the authors.

**Figure 2.** The proposed diffusion block and reverse block. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** The proposed complex-cycle-consistent learning [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Figures from the paper (4 more)

**Figure 4.** Figure 4: The pipeline of (diffusion model: ✓, phase-aware: ✗, and CCC: ✓). The clean speech spectra and the corresponding reconstruction are only converted into magnitude and phase components before CCC module. The models are trained and tested on the IEEE dataset as in Sectio…

**Figure 5.** Figure 5: IEEE dataset visualization with diffusion t-SNE for different numbers of embeddings N. [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

**Figure 6.** Figure 6: Speech enhancement performance over VCTK [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

**Figure 7.** Figure 7: The spectra of different signals. The experiment [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 33 canonical work pages

[1]

, " * write output.state after.block = add.period write newline

ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...
[2]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...
[3]

Cao, R.; Abdulatif, S.; and Yang, B. 2022. CMGAN: Conformer-based metric GAN for speech enhancement . IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

work page 2022
[4]

T.; and Shah, M

Croitoru, F.-A.; Hondru, V.; Ionescu, R. T.; and Shah, M. 2023. Diffusion models in vision: a survey . IEEE Transactions on Pattern Analysis and Machine Intelligence, 45: 10850--10869

work page 2023
[5]

R.; Zhang, C.; and Li, Y

Deng, F.; Jiang, T.; Wang, X. R.; Zhang, C.; and Li, Y. 2020. NAAGN: noise-aware attention-gated network for speech enhancement . Interspeech

work page 2020
[6]

S.; Lamel, L

Garofolo, J. S.; Lamel, L. F.; Fisher, W. M.; Fiscus, J. G.; Pallett, D. S.; and Dahlgren, N. L. 1993. TIMIT acoustic phonetic continuous speech corpus CD-ROM . Linguistic Data Consortium

work page 1993
[7]

Han, S.; and Lee, J. 2022. NU-Wave 2: a general neural audio upsampling model for various sampling rates . Interspeech

work page 2022
[8]

D.; Horaud, R.; and Li, X

Hao, X.; Su, X. D.; Horaud, R.; and Li, X. F. 2021. FullSubNet: a full-band and sub-band fusion model for real-time single-channel speech enhancement . IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

work page 2021

Show all 40 references

[9]

He, K.; Zhang, X.; Ren, S.; and Sun, J. 2015. Delving deep into rectifiers: surpassing human-level performance on imagenet classification . IEEE International Conference on Computer Vision

2015
[10]

He, K.; Zhang, X.; Ren, S.; and Sun, J. 2016. Deep residual learning for image recognition . IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

2016
[11]

Hu, Y.; Chen, C.; Li, R.; Zhu, Q.; and Chng, E. S. 2020. Denoising diffusion probabilistic models . Processdings of Neural Information Processing Systems (NeurIPS)

2020
[12]

Hu, Y.; Chen, C.; Li, R.; Zhu, Q.; and Chng, E. S. 2023. Noise-aware speech enhancement using diffusion probabilistic model . arXiv preprint arXiv:2307.08029

2023 arXiv
[13]

Hu, Y.; and Loizou, P. C. 2008. Evaluation of objective quality measures for speech enhancement . IEEE Transactions on Audio, Speech and Language Processing, 16(1): 229--238

2008
[14]

IEEE Audio and Electroacoustics Group . 1969. IEEE recommended practice for speech quality measurements . IEEE Transactions on Audio, Speech and Language Processing, AE-17(3): 225--246

1969
[15]

Li, Q.; Gao, F.; Guan, H.; and Ma, K. 2021 a . Real-time monaural speech enhancement with short-time discrete cosine transform . arXiv preprint arXiv: 2102.04629

2021 arXiv
[16]

Li, Y.; Sun, Y.; Horoshenkov, K.; and Naqvi, S. M. 2021 b . Domain adaptation and autoencoder based unsupervised speech enhancement . IEEE Transactions on Artificial Intelligence, 3(1): 43 -- 52

2021
[17]

Li, Y.; Sun, Y.; Wang, W.; and Naqvi, S. M. 2023. U-shaped Transformer with frequency-band aware attention for speech enhancement . IEEE/ACM Transactions on Audio, Speech and Language Processing, 31: 1511–1521

2023
[18]

X.; Ma, H

Liu, Z. X.; Ma, H. T.; and Chen, F. 2017. A new data-driven band-weighting function for predicting the intelligibility of noise-suppressed speech . Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)

2017
[19]

Lu, Y.-J.; Tsao, Y.; and Watanabe, S. 2021. A study on speech enhancement based on diffusion probabilistic model . Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)

2021
[20]

Lu, Y.-J.; Wang, Z.-Q.; Watanabe, S.; Richard, A.; Yu, C.; and Tsao, Y. 2022. Conditional diffusion probabilistic model for speech enhancement . IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

2022
[21]

Macartney, C.; and Weyde, T. 2018. Improved speech enhancement with the wave-u-net . arXiv preprint arXiv:1811.11307

2018 arXiv
[22]

Rahman, A.; J. M. J. Valanarasu, I. H.; and Patel, V. M. 2023. Ambiguous medical image segmentation using diffusion models . IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

2023
[23]

Reddy, C.; Dubey, H.; Koishida, K.; Nair, A.; Gopal, V.; Cutler, R.; Braun, S.; Gamper, H.; Aichner, R.; and Srinivasan, S. 2021. Interspeech 2021 deep noise suppression challenge . Interspeech

2021
[24]

Rogers, L. C. G.; and Williams, D. 2000. Diffusions, Markov processes and martingales, . Cambridge University Press

2000
[25]

Shimauchlt, S.; Kudo, S.; Koizumli, Y.; and Furuva, K. 2017. On relationships between amplitude and phase of short-time fourier transform . IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

2017
[26]

Sohl-Dickstein, J.; Weiss, E.; Maheswaranathan, N.; and Ganguli, S. 2015. Deep unsupervised learning using nonequilibrium thermodynamics. International Conference on Machine Learning (ICML)

2015
[27]

Thiemann, J.; Ito, N.; and Vincent, E. 2013. The diverse environments multi-channel acoustic noise database: a database of multichannel environmental noise recordings . The Journal of the Acoustical Society of America, 133(5): 3591 -- 3591

2013
[28]

K.; Xu, B.; Smaragdis, P.; and Kumar, A

Tzinis, E.; Adi, Y.; Ithapu, V. K.; Xu, B.; Smaragdis, P.; and Kumar, A. 2022. RemixIT: continual self-training of speech enhancement models via bootstrapped remixing . IEEE Journal of Selected Topics in Signal Processing, 16(6): 1329--1341

2022
[29]

Veaux, C.; Yamagishi, J.; and King, S. 2013. The voice bank corpus: design, collection and data analysis of a large regional accent speech database . IEEE Conference on Asian Spoken Language Research and Evaluation (O-COCOSDA/CASLRE)

2013
[30]

Wang, Y.; Narayanan, A.; and Wang, D. 2014. On training targets for supervised speech separation . IEEE/ACM Transactions on Audio, Speech and Language Processing, 22(12): 1849--1858

2014
[31]

Welker, S.; Richter, J.; and Gerkmann, T. 2022 a . Speech enhancement with score-based generative models in the complex STFT domain . Interspeech

2022
[32]

Welker, S.; Richter, J.; and Gerkmann, T. 2022 b . Speech enhancement with score-based generative models in the complex STFT domain . Interspeech

2022
[33]

S.; Wang, Y.; and Wang, D

Williamson, D. S.; Wang, Y.; and Wang, D. 2016. Complex ratio masking for monaural speech separation. IEEE/ACM Transactions on Audio, Speech and Language Processing, 24(3): 483 -- 492

2016
[34]

Wu, Q.; Han, M.; Jiang, T.; Fan, H.; Zeng, B.; and Liu, S. 2023. Realistic noise synthesis with diffusion models . arXiv preprint arXiv: 2305.14022

2023 arXiv
[35]

Yang, Y.; Pandey, A.; and Wang, D. 2023. Time-domain speech enhancement for robust automatic speech recognition . Interspeech

2023
[36]

W.; Zhou, J.; Wang, H

Yu, W. W.; Zhou, J.; Wang, H. B.; and Tao, L. 2021. SETransformer: speech enhancement transformer . Cognitive Computation

2021
[37]

Zadorozhnyy, V.; and Q. Ye, K. K. 2022. SCP-GAN: self-correcting discriminator optimization for training consistency preserving metric GAN on speech enhancement tasks . arXiv preprint arXiv: 2210.14474

2022 arXiv
[38]

H.; and Ma, B

Zhao, S.; Nguyen, T. H.; and Ma, B. 2021. Monaural speech enhancement with complex convolutional block attention module and joint time frequency losses . IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

2021
[39]

Zhao, Y.; and Wang, D. L. 2020. Noisy-reverberant Speech Enhancement Using DenseUNet with Time-frequency Attention . Interspeech

2020
[40]

Zhu, Q.-S.; Zhang, J.; Zhang, Z.-Q.; and Dai, L.-R. 2023. A joint speech enhancement and self-supervised representation learning framework for noise-robust speech recognition . IEEE/ACM Transactions on Audio, Speech and Language Processing, 31: 1927--1939

2023

Pith tools

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

[1] [1]

, " * write output.state after.block = add.period write newline

ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

[2] [2]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

[3] [3]

Cao, R.; Abdulatif, S.; and Yang, B. 2022. CMGAN: Conformer-based metric GAN for speech enhancement . IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

work page 2022

[4] [4]

T.; and Shah, M

Croitoru, F.-A.; Hondru, V.; Ionescu, R. T.; and Shah, M. 2023. Diffusion models in vision: a survey . IEEE Transactions on Pattern Analysis and Machine Intelligence, 45: 10850--10869

work page 2023

[5] [5]

R.; Zhang, C.; and Li, Y

Deng, F.; Jiang, T.; Wang, X. R.; Zhang, C.; and Li, Y. 2020. NAAGN: noise-aware attention-gated network for speech enhancement . Interspeech

work page 2020

[6] [6]

S.; Lamel, L

Garofolo, J. S.; Lamel, L. F.; Fisher, W. M.; Fiscus, J. G.; Pallett, D. S.; and Dahlgren, N. L. 1993. TIMIT acoustic phonetic continuous speech corpus CD-ROM . Linguistic Data Consortium

work page 1993

[7] [7]

Han, S.; and Lee, J. 2022. NU-Wave 2: a general neural audio upsampling model for various sampling rates . Interspeech

work page 2022

[8] [8]

D.; Horaud, R.; and Li, X

Hao, X.; Su, X. D.; Horaud, R.; and Li, X. F. 2021. FullSubNet: a full-band and sub-band fusion model for real-time single-channel speech enhancement . IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

work page 2021