Pith. sign in

REVIEW 5 major objections 5 minor 56 references

Unified Architecture and Unsupervised Speech Disentanglement for Speaker Embedding-Free Enrollment in Personalized Speech Enhancement

T0 review · 5 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read One speaker-embedding-free network can perform both conventional and personalized speech enhancement, and an unsupervised consistency loss makes it insensitive to variation in the enrollment speech.

desk verdict Useful incremental paper on unifying SE/PSE with a simple zero-enrollment trick, but the disentanglement mechanism needs a lambda=0 ablation before I'd trust the attribution. read the letter →

arxiv 2505.12288 v2 pith:OSGFRS7M submitted 2025-05-18 eess.AS cs.SD

Ziling Huang , Haixin Guan , Yanhua Long This is my paper · ORCID
classification eess.AScs.SD
keywords personalizedspeechenhancementunifiedarchitecturespeakerembedding-freenetworkunsuperviseddisentanglementenrollmentrobustnesslong-shortpairingtargetextraction
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 aims to establish that a single speaker-embedding-free architecture, previously built for personalized speech enhancement (PSE), can also handle conventional speech enhancement (SE) with no extra parameters, and that a simple unsupervised consistency constraint can make PSE robust to changes in the enrollment utterance. For the unified model USEF-PNet, the authors train the same network with either a real enrollment utterance or a zero-valued dummy enrollment, so that SE becomes the special case of PSE with no reference. For disentanglement, DSEF-PNet pairs each noisy mixture with two different utterances from the same target speaker and forces the two enhanced outputs to match, arguing that this removes emotion, wording, and other identity-irrelevant factors while preserving speaker identity. If these claims hold, one lightweight model could replace two task-specific systems and tolerate the messy enrollment speech found in real applications.

What carries the argument

The load-bearing object is the heterogeneous enrollment invariant training (HEIT) loss, $\lambda \|\hat{Y}^{\mathrm{pse}}_1 - \hat{Y}^{\mathrm{pse}}_2\|_1$, minimized alongside two SI-SDR losses in Eq. (5). It carries the disentanglement claim: because the two enrollments share only the target speaker's identity, forcing the outputs to agree is supposed to strip away emotion, speaking rate, wording, and noise while leaving identity untouched. The companion mechanism is the zero-dummy-enrollment batch construction in Eq. (1), which lets conventional SE be processed by the same network by concatenating a zero vector in place of enrollment features; this is what turns PSE into a unified task without changing the architecture. Both operate on the speaker-embedding-free SEF-PNet backbone, whose interactive speaker adaptation module feeds enrollment context directly into the time-frequency stream.

What would settle it

On the Libri2Mix 2-speaker condition, train DSEF-PNet with same-style enrollment pairs (both neutral read speech) and cross-style pairs (one neutral, one whispered or shouted) for the same speaker, matching content. If HEIT removes only identity-irrelevant variation, cross-style pairs should give the same SISDR as same-style pairs; a measurable drop would show the consistency term is discarding identity-relevant cues. The paper does not report this comparison or the value of $\lambda$ in Eq. (5), so the experiment remains open.

Watch

Extended reading notes

Core claim

The central claim is that the SEF-PNet backbone already carries enough information to serve both tasks; what changes is only how it is trained. USEF-PNet constructs each mini-batch from PSE samples (mixture plus real enrollment) and SE samples (mixture plus a zero dummy enrollment), trains with the negative scale-invariant signal-to-distortion ratio (SI-SDR) loss, and reports performance on Libri2Mix and VoiceBank-DEMAND that matches or exceeds the task-specific SEF-PNet baselines. DSEF-PNet adds heterogeneous enrollment invariant training (HEIT): the same mixture is passed through the network once with each of two same-speaker enrollments, and the L1 distance between the two complex-spectrogram outputs is added to the two SI-SDR losses, so the model is pushed to produce the same enhancement regardless of which enrollment it sees. The paper reports that this improves the Libri2Mix 2-speaker condition from 13.00 to 13.57 dB SISDR and 3.01 to 3.08 PESQ, and that combining both strategies (UDSEF-PNet) improves both the PSE and SE tasks further. The authors also explore long-short enrollment pairing (LSEP), and find that random enrollment durations during training generalize better than this pairing.

Load-bearing premise

The load-bearing premise, which enters in Eqs. (3)–(5), is that two different enrollment utterances of the same speaker share only the speaker's identity, so the HEIT consistency loss forcing their outputs to be equal removes only emotion, wording, and other irrelevant factors and never discards useful speaker information; the paper tests this only through downstream enhancement metrics and does not report the value of $\lambda$.

Editorial extensions

If this is right

  • A single USEF-PNet checkpoint can replace separate SE and PSE systems: on the full Libri2Mix 2-speaker condition it matches SEF-PNet's SISDR of 13.00 dB, while on VoiceBank-DEMAND it improves SE SISDR from 19.21 to 19.69 dB.
  • PSE robustness to changes in the enrollment utterance does not cost anything at inference: DSEF-PNet uses the same weights and runtime as SEF-PNet, and improves the 2-speaker condition SISDR from 13.00 to 13.57 dB.
  • The two ideas compose: UDSEF-PNet, which trains one SEF-PNet architecture with both tricks, reports 13.34 dB SISDR on the 2-speaker condition and 19.95 dB SISDR on VoiceBank-DEMAND, better than either trick alone.
  • Training with randomly chosen enrollment durations generalizes better than always pairing a short with a long utterance or always using long enrollments, so duration variability during training is itself a robustness lever.

Reading between the lines

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

  • Beyond the paper, the zero-dummy-enrollment trick should transfer to any PSE model that concatenates an enrollment representation to the mixture, because the paper's rationale depends only on that concatenation, not on the SEF-PNet specifics.
  • Beyond the paper, HEIT's invariance is demonstrated only by final enhancement metrics, so the natural next test is a direct probe of the learned representation: if the internal enrollment features of whispered and shouted versions of the same speaker collapse to the same point, then true identity disentanglement is happening; if only the outputs collapse, the model may simply be averaging over enr
  • Beyond the paper, LSEP's negative result suggests a testable modification: instead of a hard L1 tie between short and long enrollment outputs, a soft or duration-conditioned alignment might preserve the extra speaker detail that short enrollments need.
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

5 major / 5 minor

Summary. The paper extends the authors' previous SEF-PNet speaker-encoder-free framework in two directions. USEF-PNet unifies conventional speech enhancement (SE) and personalized speech enhancement (PSE) by training the same architecture on mini-batches that contain both PSE pairs (mixture, enrollment) and SE samples with a zero-valued dummy enrollment, so that a single model can switch tasks at inference depending on whether enrollment is available. DSEF-PNet introduces Heterogeneous Enrollment Invariant Training (HEIT): for each mixture, two different enrollment utterances of the same target speaker are used in parallel, and the two enhanced outputs are forced to match with an L1 consistency loss in addition to the usual SISDR loss, with the goal of removing enrollment-content variation while retaining speaker identity. A Long-Short Enrollment Pairing (LSEP) variant is also explored. Experiments on Libri2Mix and VoiceBank-DEMAND report that USEF-PNet matches or slightly surpasses task-specific baselines, and that DSEF-PNet improves PSE metrics, most notably SISDR from 13.00 to 13.57 dB in the full Libri2Mix 2-speaker condition, while adding no extra parameters or inference overhead.

Significance. If the claims are substantiated, the paper offers an appealingly simple way to unify SE and PSE in one lightweight model, and a training-only mechanism for making PSE robust to enrollment variability—both valuable for resource-constrained deployment. The manuscript has concrete strengths: the USEF-PNet idea is clean and directly tested; the authors honestly report that the LSEP strategy does not improve over random enrollment durations; and the promise of released code and checkpoints supports reproducibility. The main weakness is that the central mechanism attribution for DSEF-PNet, namely unsupervised disentanglement of speaker identity, is not actually demonstrated. The observed gains could stem from the two-enrollment sampling acting as data augmentation rather than from the HEIT consistency loss, and no ablation or direct diagnostic separates these explanations. The significance of the DSEF-PNet contribution therefore currently rests on an under-supported inference, although the underlying idea is plausible and testable.

major comments (5)
  1. [Section IV-B, Eqs. (3)-(5), Table III] The DSEF-PNet training procedure changes two factors at once relative to SEF-PNet: each mixture is now paired with two distinct enrollment utterances per optimization step, and the L1 HEIT consistency term is added to the loss. The reported gains (e.g., 13.00 to 13.57 dB SISDR in the 2-speaker condition) could therefore be caused by the two-enrollment sampling acting as data augmentation or by the enlarged effective batch, without any contribution from the consistency constraint. A necessary control is a lambda=0 ablation that still uses two enrollments per step; without it, the gains cannot be attributed to HEIT. In addition, the value of lambda in Eq. (5) is never reported, so the strength of the constraint is unknown and the experiments are not reproducible.
  2. [Section IV-B and Section VI-B] The claim that HEIT 'implicitly disentangles' speaker identity from irrelevant enrollment content is not directly tested. The paper reports only downstream enhancement metrics; there is no diagnostic showing that identity-relevant cues are preserved or that emotion, content, or noise are suppressed. Since the consistency loss could equally discard useful identity information when the two enrollments differ in identity-relevant ways, the use of the term 'disentanglement' is not supported by the presented evidence. An experiment that varies emotion or content in the enrollment while holding identity fixed, or an analysis of the internal representations, would be needed to support the mechanism claim.
  3. [Tables III, IV, V and Section VI-B] The performance improvements attributed to DSEF-PNet are small (the largest SISDR gain is 0.57 dB, and several metrics are unchanged or improve by 0.03 or less), yet no error bars, number of random seeds, or significance tests are reported. The statements that DSEF-PNet 'consistently outperforms' SEF-PNet and yields 'significant' improvements are therefore not statistically substantiated, and the reader cannot determine whether the differences are within run-to-run variability.
  4. [Section VI-A2, Table II] The SEF-PNet SE baseline on VoiceBank-DEMAND is underspecified. For the Table I experiments, the text states that the ISA module is removed from SEF-PNet to create a purely SE-focused variant, but no equivalent statement appears for the Table II SEF-PNet row. If the ISA module was retained and fed a zero or dummy enrollment, the comparison with USEF-PNet is not on equal footing, and the claimed SE improvement (19.21 to 19.69 dB SISDR) may reflect the baseline configuration rather than the unified training strategy.
  5. [Section III vs. Section V-B] The backbone architecture description is internally inconsistent. Section III states that the number of Encoder Blocks was reduced from 7 to 6 compared with sDPCCN, and describes the encoder as 'an ISA adaptation module followed by 6 Encoder Blocks,' while Section V-B states that the SEF-PNet architecture 'consists of 7 Encoder Blocks, 7 corresponding Decoder Blocks.' Since every experiment in the paper builds on this backbone, the contradiction prevents exact reproduction and should be resolved.
minor comments (5)
  1. [Table II] The STOI values in Table II are reported as fractions (0.92, 0.95, 0.94) while the same metric is reported as a percentage in Tables I, III, IV, and V (e.g., 94.26 in Table IV); please unify the scale and label the units explicitly.
  2. [Section V-A and Eq. (3)] The paper should specify how the two enrollment utterances E1 and E2 are selected for DSEF-PNet: whether they are random distinct utterances from the same speaker, how the 'random' duration policy in Table V is implemented, and whether the two enrollments are always different files. The current description is insufficient for exact reproduction.
  3. [Section IV-B, Eqs. (4)-(5)] Equation (4) defines the HEIT loss on complex spectrograms, while Eq. (5) uses the SISDR loss, which is normally defined on waveforms, with the same symbol 'Y^pse' for both domains. Please clarify how the waveform is obtained from the complex spectrogram and confirm that the L1 consistency is applied in the STFT domain.
  4. [Section IV-A, Eq. (1)] The ratio M:N of PSE and SE samples in each unified mini-batch is not specified. Because this ratio controls the balance between the two tasks in USEF-PNet, it should be reported.
  5. [Abstract and Introduction] The abstract and introduction motivate robustness to variations such as emotional tone and semantic content, but the experiments only vary enrollment duration (Table V). Either add experiments that vary these factors or temper the claims and explicitly state the limitation.

Circularity Check

1 steps flagged · score 2.0 of 10

DSEF-PNet's 'disentanglement' claim restates the HEIT invariance objective, but the main benchmark results are externally evaluated and not circular.

  1. self definitional [Section IV-B, Eqs. (3)-(5), paragraph after Fig. 3]
    "The two distinct enrollment speech recordings E1_i and E2_i are with the same target speaker's identity but include heterogeneous identity-irrelevant contents. ... By aligning these two enhanced outputs, Y^pse_1 and Y^pse_2, we implicitly equip SEF-PNet with the ability to disentangle the target speaker identity. ... L_HEIT = ||Y^pse_1 - Y^pse_2||_1"

    The 'disentanglement' result is not derived from any independent measurement or theorem; it is the training objective itself. Because E1 and E2 are chosen to share only speaker identity (Eq. 3), the HEIT loss defines output invariance across them, and the paper then reads that invariance back as evidence that the model is able to focus on speaker-identity-related features. This is equivalent to assuming the conclusion: identity is what is assumed to remain after enforcing consistency. The missing lambda=0 ablation and unreported lambda further mean the performance gain in Table III cannot be attributed specifically to this consistency term rather than to two-enrollment sampling. This does not make the external benchmark comparisons circular, but the conceptual claim is definitional.

full rationale

Most of the paper's engineering claims are self-contained and externally benchmarked. USEF-PNet's unified SE/PSE result is tested on Libri2Mix and VoiceBank-DEMAND against sDPCCN and MP-SENet; using an identical architecture with zeroed enrollment for SE is a design choice, not a circular derivation. DSEF-PNet's main numeric gains (Tables III and V) are compared with the SEF-PNet baseline on held-out test sets, so they are not fitted predictions. The only definitional element is the 'unsupervised disentanglement' narrative: HEIT forces outputs from two enrollments to match, and the paper equates that invariance with 'isolating speaker identity.' That is an operational definition rather than an empirically established fact, and because lambda is unreported and no lambda=0 ablation is run, the attribution of the gains to the consistency term is a confound. Under the circularity rubric this is a minor self-definitional step, not a collapse of the derivation: the external benchmark results stand independently. No uniqueness theorem, no self-citation chain, and no renamed empirical law are load-bearing.

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

The central claims rest on empirical evaluations with a small number of hand-chosen hyperparameters; the main free parameter is λ, the HEIT loss weight, which is not reported. Key domain assumptions are that zero enrollment disables the PSE pathway and that two same-speaker enrollments share only identity.

free parameters (1)
  • λ (HEIT loss weight) = not reported
    Appears in Eq. (5) and Eq. (7) as the weight of the L1 consistency loss; no value or sensitivity analysis is given, so the reported results depend on an undisclosed choice.
assumptions (3)
  • domain assumption Zero-valued enrollment input is a valid 'no-enrollment' condition that disables the PSE pathway without harming SE capability
    Used in USEF-PNet training (Eq. 1) and inference (Section IV-A) to unify SE and PSE; the paper does not analyze how the ISA/CI modules respond to zero enrollment.
  • domain assumption Two enrollment utterances of the same speaker share only speaker identity; consistency between their outputs removes only identity-irrelevant variability
    Basis of HEIT in Section IV-B; without this, the L1 loss could discard identity-relevant cues.
  • domain assumption Libri2Mix and VoiceBank-DEMAND at 8 kHz are representative benchmarks for the PSE and SE claims
    All experiments in Section VI are on these two datasets.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Unified Architecture and Unsupervised Speech Disentanglement for Speaker Embedding-Free Enrollment in Personalized Speech Enhancement." pith.science (2026). https://pith.science/paper/OSGFRS7M

@misc{pith2026250512288,
  author       = {Pith},
  title        = {Pith review of: Unified Architecture and Unsupervised Speech Disentanglement for Speaker Embedding-Free Enrollment in Personalized Speech Enhancement},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OSGFRS7M}},
  note         = {Machine review of arXiv:2505.12288}
}
read the original abstract

Conventional speech enhancement (SE) aims to improve speech perception and intelligibility by suppressing noise without requiring enrollment speech as reference, whereas personalized SE (PSE) addresses the cocktail party problem by extracting a target speaker's speech using enrollment speech. While these two tasks tackle different yet complementary challenges in speech signal processing, they often share similar model architectures, with PSE incorporating an additional branch to process enrollment speech. This suggests developing a unified model capable of efficiently handling both SE and PSE tasks, thereby simplifying deployment while maintaining high performance. However, PSE performance is sensitive to variations in enrollment speech, like emotional tone, which limits robustness in real-world applications. To address these challenges, we propose two novel models, USEF-PNet and DSEF-PNet, both extending our previous SEF-PNet framework. USEF-PNet introduces a unified architecture for processing enrollment speech, integrating SE and PSE into a single framework to enhance performance and streamline deployment. Meanwhile, DSEF-PNet incorporates an unsupervised speech disentanglement approach by pairing a mixture speech with two different enrollment utterances and enforcing consistency in the extracted target speech. This strategy effectively isolates high-quality speaker identity information from enrollment speech, reducing interference from factors such as emotion and content, thereby improving PSE robustness. Additionally, we explore a long-short enrollment pairing (LSEP) strategy to examine the impact of enrollment speech duration during both training and evaluation. Extensive experiments on the Libri2Mix and VoiceBank DEMAND demonstrate that our proposed USEF-PNet, DSEF-PNet all achieve substantial performance improvements, with random enrollment duration performing slightly better.

Figures

Figures reproduced from arXiv: 2505.12288 by the authors.

Figure 1
Figure 1. Overview of our previously proposed SEF-PNet model. All colored blocks highlight the key contributions over the original sDPCCN. [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Structure of Unified Architecture on SEF-PNet (USEF-PNet). [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. The whole training framework of DSEF-PNet with heterogeneous enrollment invariant training (HEIT). [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

56 extracted references · 52 canonical work pages

  1. [1]

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

    E. C. Cherry, “Some experiments on the recognition of speech, with one and with two ears,” The Journal of the Acoustical Society of America , vol. 25, no. 5, pp. 975–979, 1953

    work page 1953

  2. [2]

    The cocktail party problem,

    S. Haykin and Z. Chen, “The cocktail party problem,” Neural Compu- tation, vol. 17, no. 9, pp. 1875–1902, 2005

    work page 1902

  3. [3]

    SEGAN: Speech enhancement generative adversarial network,

    S. Pascual, A. Bonafonte, and J. Serra, “SEGAN: Speech enhancement generative adversarial network,” in Proceedings of the 18th Annual Conference of the International Speech Communication Association (INTERSPEECH), 2017, pp. 3642–3646

    work page 2017

  4. [4]

    MetricGAN: Generative adversarial networks based black-box metric scores optimization for speech enhancement,

    S.-W. Fu, C.-F. Liao, Y . Tsao, and S.-D. Lin, “MetricGAN: Generative adversarial networks based black-box metric scores optimization for speech enhancement,” in Proceedings of the 36th International Con- ference on Machine Learning (ICML) , 2019, pp. 2031–2041. 12

    work page 2019

  5. [5]

    MP-SENet: A speech enhancement model with parallel denoising of magnitude,

    Y .-X. Lu, Y . Ai, and Z.-H. Ling, “MP-SENet: A speech enhancement model with parallel denoising of magnitude,” in Proceedings of the 21st Annual Conference of the International Speech Communication Association (INTERSPEECH), 2020, pp. 3291–3295

    work page 2020

  6. [6]

    Real-time speech enhance- ment in the waveform domain,

    A. Defossez, G. Synnaeve, and Y . Adi, “Real-time speech enhance- ment in the waveform domain,” in Proceedings of the 21st Annual Conference of the International Speech Communication Association (INTERSPEECH), Shanghai, China, 2020, pp. 3291–3295

    work page 2020

  7. [7]

    Dual-branch attention-in-attention transformer for single-channel speech enhance- ment,

    G. Yu, A. Li, C. Zheng, Y . Guo, Y . Wang, and H. Wang, “Dual-branch attention-in-attention transformer for single-channel speech enhance- ment,” in Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , Singapore, 2022, pp. 7847– 7851

    work page 2022

  8. [8]

    DPT-FSNet: Dual-path transformer- based full-band and sub-band fusion network for speech enhancement,

    F. Dang, H. Chen, and P. Zhang, “DPT-FSNet: Dual-path transformer- based full-band and sub-band fusion network for speech enhancement,” in Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , Singapore, 2022, pp. 6857–6861

    work page 2022

Show all 56 references
  1. [9]

    PhaseN: A phase-and- harmonics-aware speech enhancement network,

    D. Yin, C. Luo, Z. Xiong, and W. Zeng, “PhaseN: A phase-and- harmonics-aware speech enhancement network,” in Proceedings of the AAAI Conference on Artificial Intelligence , vol. 34, no. 05, New York, USA, 2020, pp. 9458–9465

    2020

  2. [10]

    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 Processing Magazine, vol. 40, no. 3, pp. 8–29, May 2023

    2023

  3. [11]

    An overview of deep learning-based audio-visual speech enhancement and separation,

    D. Michelsanti et al., “An overview of deep learning-based audio-visual speech enhancement and separation,” IEEE/ACM Transactions on Audio, Speech, and Language Processing , vol. 29, pp. 1368–1396, 2021

    2021

  4. [12]

    DPCCN: Densely-connected pyramid complex convolutional network for robust speech separation and extraction,

    J. Han, Y . Long et al., “DPCCN: Densely-connected pyramid complex convolutional network for robust speech separation and extraction,” in Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2022, pp. 7292–7296

    2022

  5. [13]

    High fidelity speech enhancement with band-split rnn,

    J. Yu, H. Chen, Y . Luo, R. Gu, and C. Weng, “High fidelity speech enhancement with band-split rnn,” in Proceedings of the 24th Annual Conference of the International Speech Communication Association (INTERSPEECH), 2023, pp. 2483–2487

    2023

  6. [14]

    X-SEPFORMER: End-to-end speaker extraction network with explicit optimization on speaker con- fusion,

    K. Liu, Z. Du, X. Wan, and H. Zhou, “X-SEPFORMER: End-to-end speaker extraction network with explicit optimization on speaker con- fusion,” in Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2023, pp. 1–5

    2023

  7. [15]

    Deep- FilterNet: Perceptually motivated real-time speech enhancement,

    H. Schr ¨oter, T. Rosenkranz, A. N. Escalante-B., and A. Maier, “Deep- FilterNet: Perceptually motivated real-time speech enhancement,” in Proceedings of the 24th Annual Conference of the International Speech Communication Association (INTERSPEECH) , 2023

    2023

  8. [16]

    SEF-Net: Speaker embedding free target speaker extraction network,

    B. Zeng, H. Suo, Y . Wan, and M. Li, “SEF-Net: Speaker embedding free target speaker extraction network,” in Proceedings of the 24th Annual Conference of the International Speech Communication Association (INTERSPEECH), 2023, pp. 3452–3456

    2023

  9. [17]

    Target speaker extraction by directly exploiting contextual information in the time-frequency domain,

    X. Yang, C. Bao, J. Zhou, and X. Chen, “Target speaker extraction by directly exploiting contextual information in the time-frequency domain,” in Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2024, pp. 10 476–10 480

    2024

  10. [18]

    Personalized speech enhancement without a separate speaker embedding model,

    T. P ¨arnamaa and A. Saabas, “Personalized speech enhancement without a separate speaker embedding model,” in Proceedings of the 25th Annual Conference of the International Speech Communication Association (INTERSPEECH), 2024, pp. 4863–4867

    2024

  11. [19]

    Urgent Challenge: Universality, robustness, and generalizability for speech enhancement,

    W. Zhang, R. Scheibler, K. Saijo, S. Cornell, C. Li, Z. Ni, J. Pirklbauer, M. Sach, S. Watanabe, T. Fingscheidt, and Y . Qian, “Urgent Challenge: Universality, robustness, and generalizability for speech enhancement,” in Proceedings of the 25th Annual Conference of the Interna...

    2024

  12. [20]

    AnyEnhance: A unified generative model with prompt-guidance and self-critic for voice enhancement,

    J. Zhang, J. Yang, Z. Fang, Y . Wang, Z. Zhang, Z. Wang, F. Fan, and Z. Wu, “AnyEnhance: A unified generative model with prompt-guidance and self-critic for voice enhancement,” arXiv preprint arXiv:2501.15417, 2025

    2025

  13. [21]

    Uniaudio: Towards universal audio generation with large language models,

    D. Yang, J. Tian, X. Tan, R. Huang, S. Liu, H. Guo, X. Chang, J. Shi, J. Bian, Z. Zhao et al. , “Uniaudio: Towards universal audio generation with large language models,” in Proceedings of the Forty- first International Conference on Machine Learning , 2024

    2024

  14. [22]

    Metis: A foundation speech generation model with masked generative pre- training,

    Y . Wang, J. Zheng, J. Zhang, X. Zhang, H. Liao, and Z. Wu, “Metis: A foundation speech generation model with masked generative pre- training,” arXiv preprint arXiv:2502.03128 , 2025

    2025 arXiv

  15. [23]

    Target confusion in end-to-end speaker extraction: Analysis and approaches,

    Z. Zhao, D. Yang, R. Gu, H. Zhang, and Y . Zou, “Target confusion in end-to-end speaker extraction: Analysis and approaches,” in Proceedings of the 23rd Annual Conference of the International Speech Communi- cation Association (INTERSPEECH) , 2022, pp. 5333–5337

    2022

  16. [24]

    Self-supervised disentangled representation learning for robust target speech extraction,

    Z. Mu, X. Yang et al. , “Self-supervised disentangled representation learning for robust target speech extraction,” in Proceedings of the AAAI Conference on Artificial Intelligence , 2024, pp. 18 815–18 823

    2024

  17. [25]

    Centroid estimation with transformer-based speaker embedder for robust target speaker extraction,

    W. Heo, J. Maeng et al. , “Centroid estimation with transformer-based speaker embedder for robust target speaker extraction,” inProceedings of the 25th Annual Conference of the International Speech Communication Association (INTERSPEECH), 2024, pp. 4333–4337

    2024

  18. [26]

    SEF-PNet: Speaker encoder- free personalized speech enhancement with local and global contexts aggregation,

    Z. Huang, H. Guan, H. Wei, and Y . Long, “SEF-PNet: Speaker encoder- free personalized speech enhancement with local and global contexts aggregation,” in Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2025, pp. 1–5

    2025

  19. [27]

    ECAPA-TDNN: Emphasized channel attention, propagation and aggregation in tdnn based speaker verification,

    B. Desplanques, J. Thienpondt, and K. Demuynck, “ECAPA-TDNN: Emphasized channel attention, propagation and aggregation in tdnn based speaker verification,” in Proceedings of the 21st Annual Confer- ence of the International Speech Communication Association (INTER- SPEECH), 202...

    2020

  20. [28]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, J. Sun et al. , “Deep residual learning for image recognition,” in Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 2016, pp. 770–778

    2016

  21. [29]

    Improving speaker dis- crimination of target speech extraction with time-domain speakerbeam,

    M. Delcroix, T. Ochiai, K. Zmolikova et al. , “Improving speaker dis- crimination of target speech extraction with time-domain speakerbeam,” in Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2020, pp. 691–695

    2020

  22. [30]

    SpEx: Multi-scale time domain speaker extraction network,

    C. Xu, W. Rao, E. S. Chng et al. , “SpEx: Multi-scale time domain speaker extraction network,” IEEE/ACM Transactions on Audio, Speech, and Language Processing , vol. 28, pp. 1370–1384, 2020

    2020

  23. [31]

    SpEx+: A complete time domain speaker extraction network,

    M. Ge, C. Xu, L. Wang, E. S. Chng, J. Dang, and H. Li, “SpEx+: A complete time domain speaker extraction network,” in Proceedings of the 21st Annual Conference of the International Speech Communication Association (INTERSPEECH), 2020, pp. 1406–1410

    2020

  24. [32]

    MC-SpEx: Towards effective speaker extraction with multi- scale interfusion and conditional speaker modulation,

    J. Chen, W. Rao, Z. Wang, J. Lin, Y . Ju, S. He, Y . Wang, and Z. Wu, “MC-SpEx: Towards effective speaker extraction with multi- scale interfusion and conditional speaker modulation,” in Proceedings of the 24th Annual Conference of the International Speech Communication Associ...

    2023

  25. [33]

    Contrastive self-supervised speaker embedding with sequential disentanglement,

    Y . Tu, M.-W. Mak, and J.-T. Chien, “Contrastive self-supervised speaker embedding with sequential disentanglement,” IEEE/ACM Transactions on Audio, Speech, and Language Processing , pp. 2704–2715, 2024

    2024

  26. [34]

    Noise- disentanglement metric learning for robust speaker verification,

    Y . Sun, H. Zhang, L. Wang, K. A. Lee, M. Liu, and J. Dang, “Noise- disentanglement metric learning for robust speaker verification,” in Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) . IEEE, 2023, pp. 1–5

    2023

  27. [35]

    Disentangled speaker embedding for robust speaker verification,

    Y . Lu and M.-W. Mak, “Disentangled speaker embedding for robust speaker verification,” in Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) . IEEE, 2022, pp. 7662–7666

    2022

  28. [36]

    Disentangling voice and content with self-supervision for speaker recognition,

    T. Liu, K. A. Lee, Q. Wang, and H. Li, “Disentangling voice and content with self-supervision for speaker recognition,” Advances in Neural Information Processing Systems , vol. 36, pp. 50 221–50 236, 2023

    2023

  29. [37]

    Time-domain target-speaker speech separation with waveform-based speaker embedding,

    J. Zhao, S. Gao, and T. Shinozaki, “Time-domain target-speaker speech separation with waveform-based speaker embedding,” in Proceedings of the 19th Annual Conference of the International Speech Communication Association (INTERSPEECH). ISCA, 2020, pp. 1436–1440

    2020

  30. [38]

    Univer- sal speech enhancement with score-based diffusion,

    J. Serr `a, S. Pascual, J. Pons, R. O. Araz, and D. Scaini, “Univer- sal speech enhancement with score-based diffusion,” arXiv preprint arXiv:2206.03065, 2022

    2022 arXiv

  31. [39]

    Conditional diffusion probabilistic model for speech enhancement,

    Y .-J. Lu, Z.-Q. Wang, S. Watanabe, A. Richard, C. Yu, and Y . Tsao, “Conditional diffusion probabilistic model for speech enhancement,” in Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) . IEEE, 2022, pp. 7402–7406

    2022

  32. [40]

    V oiceFixer: A unified framework for high-fidelity speech restoration,

    H. Liu, X. Liu, Q. Kong, Q. Tian, Y . Zhao, D. Wang, C. Huang, and Y . Wang, “V oiceFixer: A unified framework for high-fidelity speech restoration,” in Proceedings of the 23rd Annual Conference of the Inter- national Speech Communication Association (INTERSPEECH) , 2022, pp. ...

    2022

  33. [41]

    Toward uni- versal speech enhancement for diverse input conditions,

    W. Zhang, K. Saijo, Z. Wang, S. Watanabe, and Y . Qian, “Toward uni- versal speech enhancement for diverse input conditions,” in 2023 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU) . IEEE, 2023, pp. 1–6

    2023

  34. [42]

    Improving design of input condition invariant speech enhancement,

    W. Zhang, J. Jung, and Y . Qian, “Improving design of input condition invariant speech enhancement,” in Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) . IEEE, 2024, pp. 10 696–10 700

    2024

  35. [43]

    Multi-stage speaker extraction with utterance and frame-level reference signals,

    M. Ge, C. Xu, L. Wang, E. S. Chng, J. Dang, and H. Li, “Multi-stage speaker extraction with utterance and frame-level reference signals,” in 13 Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2021, pp. 6109–6113

    2021

  36. [44]

    Short-duration speaker verification (sdsv) challenge 2021: the challenge evaluation plan,

    H. Zeinali, K. Lee, J. Alam, and L. Burget, “Short-duration speaker verification (sdsv) challenge 2021: the challenge evaluation plan,” arXiv preprint arXiv:1912.06311, 2021

    2021 arXiv

  37. [45]

    Discriminative neural embed- ding learning for short-duration text-independent speaker verification,

    S. Wang, Z. Huang, Y . Qian, and K. Yu, “Discriminative neural embed- ding learning for short-duration text-independent speaker verification,” IEEE/ACM Transactions on Audio, Speech, and Language Processing , vol. 27, no. 11, pp. 1686–1696, 2019

    2019

  38. [46]

    ERes2NetV2: Boosting short-duration speaker verification performance with computational efficiency,

    Y . Chen, S. Zheng, H. Wang, L. Cheng, Q. Chen, S. Zhang, and J. Li, “ERes2NetV2: Boosting short-duration speaker verification performance with computational efficiency,” in Proceedings of the 25th Annual Conference of the International Speech Communication Association (INTERS...

    2024

  39. [47]

    Augmenting short enrollment speech via synthesis for target speaker ex- traction,

    Z. Huang, J. Lin, M. Ge, Y . Jiang, X. Wang, L. Wang, and J. Dang, “Augmenting short enrollment speech via synthesis for target speaker ex- traction,” in Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) . IEEE, 2025, pp. 1–5

    2025

  40. [48]

    Target speaker extraction with ultra-short reference speech by ve-ve framework,

    L. Yang, W. Liu, L. Tan, J. Yang, and H.-G. Moon, “Target speaker extraction with ultra-short reference speech by ve-ve framework,” in Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2023, pp. 1–5

    2023

  41. [49]

    Sdr–half-baked or well done?

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

    2019

  42. [50]

    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,” arXiv preprint arXiv:2005.11262 , 2020

    2005 arXiv

  43. [51]

    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

  44. [52]

    On the importance of power compression and phase estimation in monaural speech dereverberation,

    A. Li, C. Zheng, R. Peng, and X. Li, “On the importance of power compression and phase estimation in monaural speech dereverberation,” JASA Express Letters, p. 014802, 2021

    2021

  45. [53]

    Adam: A method for stochastic optimization,

    P. Diederik, “Adam: A method for stochastic optimization,” 2015

    2015

  46. [54]

    Why gradient clipping accelerates training: A theoretical justification for adaptivity,

    J. Zhang, T. He, S. Sra, A. Jadbabaie et al. , “Why gradient clipping accelerates training: A theoretical justification for adaptivity,” 2020

    2020

  47. [55]

    Pesq (perceptual evaluation of speech quality) wrapper for python users,

    M. Wang, C. Boeddeker, R. Dantas, A. Seelan et al., “Pesq (perceptual evaluation of speech quality) wrapper for python users,” Zenodo, 2022

    2022

  48. [56]

    A short- time objective intelligibility measure for time-frequency weighted noisy speech,

    C. H. Taal, R. C. Hendriks, J. Heusdens, and R. Jensen, “A short- time objective intelligibility measure for time-frequency weighted noisy speech,” in Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2010, pp. 4214–4217

    2010

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.