Pith. sign in

REVIEW 3 major objections 5 minor 29 references

Speech Enhancement with Overlapped-Frame Information Fusion and Causal Self-Attention

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read A causal speech-enhancement system uses zero-masked 'pseudo future frames' and parallel self-attention to reach top wideband-PESQ scores on two benchmarks.

desk verdict Reasonable incremental speech-enhancement paper with an overclaimed 'future information' mechanism; the TFCA block is the stronger contribution and the benchmark numbers are competitive, but the OFIF construction needs major clarification before the paper can be trusted. read the letter →

arxiv 2501.12004 v1 pith:YARSTSRY submitted 2025-01-21 eess.AS cs.SD

Yuewei Zhang , Huanbin Zou , Jie Zhu This is my paper · ORCID
classification eess.AScs.SD
keywords speechenhancementoverlapped-frameinformationfusioncausalself-attentiontime-frequency-channelattentionDCTCRNalgorithmicdelayVoiceBankDNS-Challenge
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

Time-frequency speech enhancers are forced by the overlap-and-add step in the inverse transform to accept an algorithmic delay of one full window, yet most causal systems throw away the future information available inside that delay. The paper proposes the overlapped-frame information fusion (OFIF) scheme, which builds three 'pseudo future frames' by zero-masking different trailing fragments of the current frame, transforms them together with the real frame, and feeds the fused spectrum to the network. On top of a DCTCRN backbone with time-frequency sequence modeling, it adds a causal time-frequency-channel attention (TFCA) block that computes self-attention along time, frequency, and channel axes in parallel. The resulting OFIF-Net reports the highest WB-PESQ among the compared causal systems on both VoiceBank (3.06) and DNS-Challenge (3.30), with 2.61M parameters.

What carries the argument

The two load-bearing mechanisms are the overlapped-frame information fusion (OFIF) front-end and the time-frequency-channel attention (TFCA) block. OFIF constructs, for each frame index $t$, three pseudo frames $\tilde{x}_{t+1}, \tilde{x}_{t+2}, \tilde{x}_{t+3}$ by zero-masking successive trailing portions of the current frame (Eqs. 1–3), applies the short-time discrete cosine transform to the four-frame stack to get $\tilde{X}_t \in \mathbb{R}^{4 \times F_i}$, and feeds the fused spectral tensor into the network. This is the part that purportedly lets the model use the 'future' information already present in the overlap, without any learned synthesis window. TFCA takes an intermediate feature map $F_{in} \in \mathbb{R}^{C \times F \times T}$ and runs three parallel self-attention branches — time-wise, frequency-wise, and channel-wise — with a causal mask on the time branch and causal local pooling on the other two, then concatenates and projects the outputs. The whole pipeline (STDCT, encoder–decoder, TFSM recurrent blocks, TFCA) is causal, so the claimed gains come without increasing algorithmic delay.

What would settle it

Train OFIF-Net with an equal-sized input formed by repeating the current frame four times (or by zero-masking different, non-overlap-aligned segments), and compare WB-PESQ on VoiceBank; if the score does not drop below 3.06, the masking positions are not encoding useful overlap structure. A second check: give the same network the true future frames as input (non-causal oracle) and see whether the gap to OFIF is negligible.

Watch

Extended reading notes

Core claim

The paper's central claim is that the inherent algorithmic delay of the inverse TF transform is a free information resource that causal SE systems can exploit without adding latency. Because a reconstructed waveform frame is formed by overlapping and adding several adjacent spectral frames, future frames already contribute to the current output within the delay budget. Instead of predicting those future frames with a learned synthesis window (as earlier work does), OFIF synthesizes them by zero-masking the current frame itself: for a 32 ms window with 8 ms shift, the next three frames are approximated by masking the last 6, 4, and 2 ms blocks of the current frame. These pseudo frames carry no samples beyond the current frame, but they reshape the network's input so the model can learn to use the overlap structure already present in the spectral representation. Combined with the TFCA block, which applies causal self-attention in the time, frequency, and channel dimensions in parallel, this produces a fully causal system whose reported WB-PESQ scores surpass all compared causal baselines on both evaluation sets.

Load-bearing premise

The paper's account assumes the zero-masked pseudo future frames carry useful 'future-like' structure that the network can exploit; if those gains instead come from the wider input tensor or extra model capacity, the information-utilization story fails.

Editorial extensions

If this is right

  • Each proposed component contributes measurably: on VoiceBank+DEMAND, adding OFIF to the DCTCRN+TFSM baseline raises WB-PESQ from 2.92 to 2.96 and CSIG from 4.10 to 4.20, while adding TFCA raises WB-PESQ to 2.99 and CSIG to 4.20.
  • The full OFIF-Net reaches WB-PESQ 3.06 on VoiceBank+DEMAND and 3.30 on the DNS-Challenge non-blind test set, the top reported scores among the compared causal systems, with only 2.61M parameters.
  • Since every module in the pipeline is causal, the improvements come without increasing the algorithmic delay beyond the window size already required by the inverse transform.
  • The TFCA block is inserted into both the skip connections and decoder layers of the CRN backbone, where it boosts representation capability by recalibrating features along time, frequency, and channel dimensions.

Reading between the lines

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

  • A control that replaces the three masked pseudo frames with identical copies of the current frame would separate the effect of the masking positions from the mere widening of the input tensor; the paper's 'future information' narrative predicts the masked version should win.
  • Because the pseudo frames are constructed purely from the current frame, the same zero-masking trick transfers to other frame-online tasks with inherent overlap delay, such as speech separation or neural vocoders, without changing their latency budget.
  • On DNS-Challenge, OFIF-Net's SI-SNR (18.32) trails GaGNet's (18.91); if a deployment cares more about signal-to-noise ratio than perceptual scores, the OFIF front-end may need a different loss or a post-filter.
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 proposes OFIF-Net, a causal time-frequency-domain speech enhancement system combining an overlapped-frame information fusion (OFIF) scheme with a time-frequency-channel attention (TFCA) block built on a DCTCRN backbone. The OFIF scheme constructs three pseudo future frames by zero-masking the current frame, stacks them with the current frame into a 4-channel spectral input, and claims to exploit the future speech information available within the inherent algorithmic delay of the inverse TF transformation. The TFCA block applies parallel causal self-attention along time, frequency, and channel dimensions and is inserted in the encoder, decoder, and skip connections. The system is evaluated on VoiceBank+DEMAND and DNS-Challenge, reporting WB-PESQ 3.06 and 3.30 respectively, with 2.61M parameters, and an ablation table showing small gains for OFIF and TFCA.

Significance. If the claims were fully supported, the paper would offer a parameter-efficient causal SE architecture with competitive performance and a new attention mechanism, which would be of interest to the community. The TFCA block is a reasonable and potentially useful contribution, and the reported performance levels are comparable to or better than several recent causal systems. However, the central mechanistic claim of OFIF—that pseudo future frames supply future information within the algorithmic delay—is not supported by the construction as written, and the ablations do not control for the confounded changes in input dimensionality and added network components. The significance of the paper therefore depends on whether the OFIF claim can be reformulated or substantiated; as it stands, the main novel idea is not convincingly demonstrated.

major comments (3)
  1. [Section II-A] The mask ranges “3/4H∼H”, “1/2H∼H”, and “1/4H∼H” use the frame shift H=8ms, but the pseudo frames are defined in R^W with W=32ms. The construction is therefore not reproducible as written: it is unclear whether the mask should range over W, over H, or over some combination. Even under the plausible correction (masks over W), the pseudo frames are masked copies of the current frame xt and contain no samples outside the time interval covered by xt. The actual future frames xt+1, xt+2, xt+3 extend beyond xt by H, 2H, and 3H samples respectively. Consequently, the claim that OFIF “utilizes future speech information within the inherent algorithmic delay” (Abstract and Section I) is not mechanistically implemented; the reported gains could arise from the altered input representation (from F×T to 4×F×T) and the additional TFCA block that the OFIF path introduces, rather than from any future information.
  2. [Table I] The ablation study reports single-run objective scores without variance or significance testing. The gains attributed to OFIF are small (WB-PESQ +0.04, CSIG +0.10, CBAK +0.02, COVL +0.08). Because the OFIF change simultaneously increases the input width by a factor of four and adds an input-stage TFCA block, the ablation does not isolate the effect of the pseudo-future-frame construction. To support the mechanistic interpretation, the authors would need a control that fixes model capacity and input width (e.g., feeding four non-masked copies of the same frame, or an equivalent wider input) and, ideally, multiple training runs to assess variance. Without such a control, the central claim that the OFIF scheme itself is beneficial is not established, even though the full OFIF-Net achieves competitive scores.
  3. [Section III-D] The performance comparisons against existing methods rely on scores quoted from prior publications rather than re-implementations under identical training conditions. The paper does not state whether the training data, SNR ranges, evaluation protocols, and metric versions are exactly the same for all systems. Given that the claimed improvements over some baselines are small (e.g., WB-PESQ 3.30 vs. 3.17 for GaGNet on DNS-Challenge), these uncontrolled differences could be comparable in magnitude to the reported gains. The paper should either provide re-implemented baselines or at least explicitly acknowledge this limitation and temper the claim that the system “outperforms” all advanced methods.
minor comments (5)
  1. [Throughout] There are numerous typographical errors, including “deep leaning” (Section I), “pesudo” (Section II-A), “strcuture” (Section II-B1), “V oiceBank” (Section III-A), and “furture” (Section IV). The paper should be carefully proofread.
  2. [Eq. (5)] The symbol K is used both for the causal mask matrix and for the key matrix K_t; this makes Eq. (5) confusing. Rename the causal mask (e.g., M) to avoid the clash.
  3. [Section II-B2] The causal adaptive pooling operations in the F-Branch and C-Branch of the TFCA block are not fully specified: the output sizes, strides, and padding details of the adaptive pooling layers are missing, which makes the implementation difficult to reproduce. Please provide precise dimensions or a pseudocode description.
  4. [Section III-D] The GitHub link at the end of Section III-D is described as containing processed audio clips, not the model code or training configuration. Making the source code available would substantially improve the reproducibility of the experimental results, especially because the paper compares against many baselines from the literature.
  5. [Eqs. (5)-(7)] The attention scoring in Eq. (5) does not apply the 1/√T scaling used in Eqs. (6) and (7). The authors should explain this asymmetry or make the scaling consistent across branches.

Circularity Check

1 steps flagged · score 2.0 of 10

OFIF's claimed future-information utilization is self-definitional: Eqs. (1)-(3) define the pseudo-future frames as masked copies of the current frame, so no sample beyond x_t enters; the empirical SE results remain independent.

  1. self definitional [Section I (claim) and Section II-A, Eqs. (1)-(3)]
    "we propose to construct 3 pseudo speech frames by zero-masking xt as: ˜xt+1 = M ask{xt} 3/4 H∼H ... The utilization of pseudo speech frames can simulate part of the information of future speech frames in the inherent algorithmic delay, thereby achieving more complete information utilization."

    By Eqs. (1)-(3), each pseudo frame is a zero-masked copy of the current frame x_t; every retained sample is a sample already present in x_t. The fused spectral feature X_t in R^{4 x F_i} is therefore a function of x_t alone and contains no sample from the actual future frames x_{t+1}, x_{t+2}, x_{t+3} or from any time beyond the current frame. Consequently, the asserted utilization of future speech information in the inherent algorithmic delay is not implemented; the OFIF contribution reduces by construction to rearranging and re-weighting current-frame samples. The +OFIF ablation gain in Table I is thus not evidence for future-information use and is confounded with the increased input dimensionality and added TFCA processing.

full rationale

The paper's central performance claims are empirical: the model is trained on VoiceBank+DEMAND and DNS-Challenge and evaluated on held-out test sets, and the ablation study compares TF-DCTCRN with and without OFIF and TFCA. No parameter is fitted to a target metric and then renamed a prediction, and no external benchmark is used as a training input. The reuse of the authors' earlier TFSM block [13] is disclosed component reuse and is not load-bearing for the OFIF/TFCA conclusions; the same-authors VSANet [22] appears only as a comparison baseline. The one circular-adjacent step is the mechanistic claim that OFIF exploits the inherent algorithmic delay: because the pseudo-future frames are defined by zero-masking the current frame (Eqs. (1)-(3)), the fused input contains no sample beyond x_t, so the future-information story is a relabeling of current-frame content. This weakens the interpretation of the +OFIF ablation, but the reported scores themselves are independent empirical results, so the overall circularity is minor.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The central claim rests on standard DSP (DCT, overlap-add) plus the unverified premise that pseudo-frame repackaging and triple-dimension causal attention improve SE. Model hyperparameters are chosen by hand without ablations. No new physical entities are introduced.

free parameters (6)
  • initial learning rate = 0.0002
    Chosen by hand with no sensitivity analysis; affects convergence and final metrics.
  • batch size = 16
    Standard choice; no ablation reported.
  • training epochs = 100
    Stopping point; results may vary with more epochs.
  • local pooling kernel KT = 15
    Controls causal context size in TFCA frequency and channel branches; no ablation.
  • encoder channel widths = 16,32,64,128,128
    Architectural capacity choices inherited from CRN-style models.
  • frame configuration = N=512, W=32ms, H=8ms
    Domain-standard settings that determine the overlap factor of 4 and the OFIF construction.
assumptions (4)
  • domain assumption The inverse TF transform's algorithmic delay equals the window size, making future frames within the delay available at output time.
    Stated in Section I following [9]; this is the premise that justifies exploiting future information.
  • domain assumption Zero-masking the current frame produces pseudo future frames that carry useful information for enhancement.
    Section II-A, Eqs. (1)-(3); this premise is load-bearing for the OFIF benefit but is not independently verified.
  • domain assumption Local adaptive pooling with zero-padding in the F and C branches of TFCA preserves strict causality.
    Section II-B2; an implementation error here could leak future context and inflate results.
  • domain assumption Baseline numbers in Tables II and III are directly comparable despite different training recipes and model sizes.
    The paper does not re-train baselines; it relies on values reported in prior publications.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Speech Enhancement with Overlapped-Frame Information Fusion and Causal Self-Attention." pith.science (2026). https://pith.science/paper/YARSTSRY

@misc{pith2026250112004,
  author       = {Pith},
  title        = {Pith review of: Speech Enhancement with Overlapped-Frame Information Fusion and Causal Self-Attention},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YARSTSRY}},
  note         = {Machine review of arXiv:2501.12004}
}
read the original abstract

For time-frequency (TF) domain speech enhancement (SE) methods, the overlap-and-add operation in the inverse TF transformation inevitably leads to an algorithmic delay equal to the window size. However, typical causal SE systems fail to utilize the future speech information within this inherent delay, thereby limiting SE performance. In this paper, we propose an overlapped-frame information fusion scheme. At each frame index, we construct several pseudo overlapped-frames, fuse them with the original speech frame, and then send the fused results to the SE model. Additionally, we introduce a causal time-frequency-channel attention (TFCA) block to boost the representation capability of the neural network. This block parallelly processes the intermediate feature maps through self-attention-based operations in the time, frequency, and channel dimensions. Experiments demonstrate the superiority of these improvements, and the proposed SE system outperforms the current advanced methods.

Figures

Figures reproduced from arXiv: 2501.12004 by the authors.

Figure 1
Figure 1. Time-frequency transformation & Inverse time-frequency transforma [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overall architecture of the proposed SE system combining the overlapped-frame information fusion scheme with self-attention-based DCTCRN [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Details of causal time-frequency-channel attention (TFCA) block. The “T-Branch”, “F-Branch”, and “C-Branch” respectively denote the time-wise, [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

29 extracted references · 25 canonical work pages

  1. [1]

    SEGAN: Speech Enhancement Generative Adversarial Network,

    S. Pascual, A. Bonafonte, and J. Serr `a, “SEGAN: Speech Enhancement Generative Adversarial Network,” in Proc. Interspeech 2017 , 2017, pp. 3642–3646

    work page 2017

  2. [2]

    Real Time Speech Enhancement in the Waveform Domain,

    A. D ´efossez, G. Synnaeve, and Y . Adi, “Real Time Speech Enhancement in the Waveform Domain,” in Proc. Interspeech 2020 , 2020, pp. 3291– 3295

    work page 2020

  3. [3]

    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,” in Proc. Interspeech 2018 , 2018, pp. 3229–3233

    work page 2018

  4. [4]

    DCCRN: Deep Complex Convolution Recurrent Network for Phase-Aware Speech Enhancement,

    Y . Hu, Y . Liu, S. Lv, M. Xing, S. Zhang, Y . Fu, J. Wu, B. Zhang, and L. Xie, “DCCRN: Deep Complex Convolution Recurrent Network for Phase-Aware Speech Enhancement,” in Proc. Interspeech 2020 , 2020, pp. 2472–2476

    work page 2020

  5. [5]

    Fullsubnet: A full-band and sub- band fusion model for real-time single-channel speech enhancement,

    X. Hao, X. Su, R. Horaud, and X. Li, “Fullsubnet: A full-band and sub- band fusion model for real-time single-channel speech enhancement,” in ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2021, pp. 6633–6637

    work page 2021

  6. [6]

    Two heads are better than one: A two-stage complex spectral mapping approach for monaural speech enhancement,

    A. Li, W. Liu, C. Zheng, C. Fan, and X. Li, “Two heads are better than one: A two-stage complex spectral mapping approach for monaural speech enhancement,” IEEE/ACM Transactions on Audio, Speech, and Language Processing, vol. 29, pp. 1829–1843, 2021

    2021

  7. [7]

    CMGAN: Conformer-based Metric GAN for Speech Enhancement,

    R. Cao, S. Abdulatif, and B. Yang, “CMGAN: Conformer-based Metric GAN for Speech Enhancement,” in Proc. Interspeech 2022 , 2022, pp. 936–940

    work page 2022

  8. [8]

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

    J. Chen, Z. Wang, D. Tuo, Z. Wu, S. Kang, and H. Meng, “Fullsubnet+: Channel attention fullsubnet with complex spectrograms for speech enhancement,” in ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2022, pp. 7857– 7861

    work page 2022

Show all 29 references
  1. [9]

    Improving frame-online neural speech enhancement with overlapped-frame prediction,

    Z.-Q. Wang and S. Watanabe, “Improving frame-online neural speech enhancement with overlapped-frame prediction,” IEEE Signal Process- ing Letters , vol. 29, pp. 1422–1426, 2022

    2022

  2. [10]

    Real-time monaural speech enhancement with short-time discrete cosine transform,

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

    2021 arXiv

  3. [11]

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

    S. Zhao, B. Ma, K. N. Watcharasupat, and W.-S. Gan, “Frcrn: Boosting feature representation using frequency recurrence for monaural speech enhancement,” in ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2022, pp. 9281– 9285

    2022

  4. [12]

    Iccrn: Inplace cepstral convolutional recurrent neural network for monaural speech enhancement,

    J. Liu and X. Zhang, “Iccrn: Inplace cepstral convolutional recurrent neural network for monaural speech enhancement,” in ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2023, pp. 1–5

    2023

  5. [13]

    A two-stage framework in cross-spectrum domain for real-time speech enhancement,

    Y . Zhang, H. Zou, and J. Zhu, “A two-stage framework in cross-spectrum domain for real-time speech enhancement,” in ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2024, pp. 12 587–12 591

    2024

  6. [14]

    Discrete cosine transform,

    N. Ahmed, T. Natarajan, and K. Rao, “Discrete cosine transform,” IEEE Transactions on Computers , vol. C-23, no. 1, pp. 90–93, 1974

    1974

  7. [15]

    Dual-path rnn: Efficient long sequence modeling for time-domain single-channel speech separation,

    Y . Luo, Z. Chen, and T. Yoshioka, “Dual-path rnn: Efficient long sequence modeling for time-domain single-channel speech separation,” in ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2020, pp. 46–50

    2020

  8. [16]

    Investi- gating rnn-based speech enhancement methods for noise-robust text-to- speech

    C. Valentini-Botinhao, X. Wang, S. Takaki, and J. Yamagishi, “Investi- gating rnn-based speech enhancement methods for noise-robust text-to- speech.” in SSW, 2016, pp. 146–152

    2016

  9. [17]

    The INTERSPEECH 2020 Deep Noise Suppression Challenge: Datasets, Subjective Testing Framework, and Challenge Re- sults,

    C. K. Reddy, V . Gopal, R. Cutler, E. Beyrami, R. Cheng, H. Dubey, S. Matusevych, R. Aichner, A. Aazami, S. Braun, P. Rana, S. Srinivasan, and J. Gehrke, “The INTERSPEECH 2020 Deep Noise Suppression Challenge: Datasets, Subjective Testing Framework, and Challenge Re- sults,” i...

    2020

  10. [18]

    P. 862.2: Wideband extension to recommendation p. 862 for the assessment of wideband telephone networks and speech codecs,

    I. Rec, “P. 862.2: Wideband extension to recommendation p. 862 for the assessment of wideband telephone networks and speech codecs,” International Telecommunication Union, CH–Geneva , 2005

    2005

  11. [19]

    Evaluation of objective quality measures for speech enhancement,

    Y . Hu and P. C. Loizou, “Evaluation of objective quality measures for speech enhancement,” IEEE Transactions on Audio, Speech, and Language Processing, vol. 16, no. 1, pp. 229–238, 2008

    2008

  12. [20]

    Comp- net: Complementary network for single-channel speech enhancement,

    C. Fan, H. Zhang, A. Li, W. Xiang, C. Zheng, Z. Lv, and X. Wu, “Comp- net: Complementary network for single-channel speech enhancement,” Neural Networks , vol. 168, pp. 508–517, 2023

    2023

  13. [21]

    Glance and gaze: A collaborative learning framework for single-channel speech enhancement,

    A. Li, C. Zheng, L. Zhang, and X. Li, “Glance and gaze: A collaborative learning framework for single-channel speech enhancement,” Applied Acoustics, vol. 187, p. 108499, 2022

    2022

  14. [22]

    Vsanet: Real-time speech enhancement based on voice activity detection and causal spatial attention,

    Y . Zhang, H. Zou, and J. Zhu, “Vsanet: Real-time speech enhancement based on voice activity detection and causal spatial attention,” in 2023 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU), 2023, pp. 1–8

    2023

  15. [23]

    Sicrn: Advancing speech enhancement through state space model and inplace convolution techniques,

    C. Zhao, S. He, and X. Zhang, “Sicrn: Advancing speech enhancement through state space model and inplace convolution techniques,” in ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2024, pp. 10 506–10 510

    2024

  16. [24]

    PoCoNet: Better Speech Enhancement with Frequency- Positional Embeddings, Semi-Supervised Conversational Data, and Bi- ased Loss,

    U. Isik, R. Giri, N. Phansalkar, J.-M. Valin, K. Helwani, and A. Kr- ishnaswamy, “PoCoNet: Better Speech Enhancement with Frequency- Positional Embeddings, Semi-Supervised Conversational Data, and Bi- ased Loss,” in Proc. Interspeech 2020 , 2020, pp. 2487–2491

    2020

  17. [25]

    Inter-subnet: Speech enhancement with subband interaction,

    J. Chen, W. Rao, Z. Wang, J. Lin, Z. Wu, Y . Wang, S. Shang, and H. Meng, “Inter-subnet: Speech enhancement with subband interaction,” in ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2023, pp. 1–5

    2023

  18. [26]

    Speech Enhancement with Fullband-Subband Cross-Attention Network,

    J. Chen, W. Rao, Z. Wang, Z. Wu, Y . Wang, T. Yu, S. Shang, and H. Meng, “Speech Enhancement with Fullband-Subband Cross-Attention Network,” in Proc. Interspeech 2022 , 2022, pp. 976–980

    2022

  19. [27]

    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,” in 2001 IEEE International Con- ference on Acoustics, Speech, and Signal Processing. Proceedings (Cat....

    2001

  20. [28]

    An algorithm for intelligibility prediction of time–frequency weighted noisy speech,

    C. H. Taal, R. C. Hendriks, R. Heusdens, and J. Jensen, “An algorithm for intelligibility prediction of time–frequency weighted noisy speech,” IEEE Transactions on Audio, Speech, and Language Processing , vol. 19, no. 7, pp. 2125–2136, 2011

    2011

  21. [29]

    Sdr – half-baked or well done?

    J. L. Roux, S. Wisdom, H. Erdogan, and J. R. Hershey, “Sdr – half-baked or well done?” in ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2019, pp. 626– 630

    2019

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.