Pith. sign in

REVIEW 5 major objections 4 minor 28 references

CycleFlow: Leveraging Cycle Consistency in Flow Matching for Speaker Style Adaptation

T0 review · 5 major / 4 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 CycleFlow uses cycle-consistent flow matching with a dual pitch-and-voice flow to convert timbre and pitch without paired recordings.

desk verdict The paper has a plausible engineering idea and one clear fatal flaw: the printed cycle loss doesn't implement a cycle, so the central timbre-adaptation mechanism is absent. read the letter →

arxiv 2501.01861 v1 pith:DHMMOZID submitted 2025-01-03 cs.SD eess.AS

Ziqi Liang , Xulong Zhang , Chang Liu , Xiaoyang Qu , Weifeng Zhao , Jianzong Wang This is my paper · ORCID
classification cs.SDeess.AS
keywords voiceconversioncycleconsistencyconditionalflowmatchingpitchadaptationspeakertimbresimilaritynon-paralleltrainingF0correctionDual-CFM
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

CycleFlow claims that voice conversion can transfer a source speaker's timbre and pitch to a target speaker using only non-parallel data—no two speakers uttering the same sentence—by adding a cycle-consistency regularizer to conditional flow matching and by splitting generation into a pitch-flow and a voice-flow stage. The pitch-flow stage moves the source F0 into the target's vocal range, and the voice-flow stage synthesizes the resulting mel-spectrogram in the target timbre, which together are meant to remove the hoarseness that appears when source and target pitch ranges differ widely. The paper reports that CycleFlow beats DiffVC, UUVC, CosyVoice, and Diff-HierVC on subjective naturalness (MOS) and similarity (SMOS), timbre similarity (TSIM), log-F0 Pearson correlation, and word error rate, with the largest gains in TSIM and log-F0 PCC. If these results hold, voice conversion would not need paired recordings or a perfectly decoupled content encoder, and cross-gender or wide-pitch-gap conversion would sound cleaner.

What carries the argument

The central object is the cycle-consistency regularizer $L_{x\to y\to x}$ added to a conditional flow-matching (CFM) voice-conversion model, together with a Dual-CFM decoder that splits generation into PitchCFM and VoiceCFM. The regularizer is meant to enforce transitivity: converting a source utterance to the target style and then back should reconstruct the original, so that source content and target timbre are both preserved. PitchCFM learns an optimal-transport flow from Gaussian noise to a target-style F0 contour, and VoiceCFM learns a flow from noise to the mel-spectrogram conditioned on content tokens, the target speaker embedding, and the refined F0; both regress the OT-flow vector field $v_t = x_1 - (1-\sigma)x_0$. The mechanism that carries the argument is this two-stage flow: the pitch flow adapts the F0 domain first, and the voice flow uses that adapted pitch together with the speaker embedding to synthesize speech.

What would settle it

Take a held-out source utterance, convert it to the target speaker and then back to the source with the trained CycleFlow, and compare the returned F0 contour and mel-spectrogram with the original. If the round-trip error does not decrease when $L_{x\to y\to x}$ is included in training, or if timbre similarity gains persist when that term is ablated, the cycle mechanism is not doing the work the paper attributes to it.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that cycle consistency, previously used in GAN-based non-parallel voice conversion, transfers cleanly to conditional flow matching, and that separating speech generation into two flow-matching models lets one system handle timbre and pitch adaptation at once. The training objective combines a reconstruction term, a cycle term $L_{x\to y\to x}$, and an invariance term $L_{x\to y\to y}$, together with pitch and mel vector-field matching losses. The authors claim that the cycle term is what lifts speaker timbre similarity, with intra-domain TSIM reaching 0.856 and cross-domain TSIM 0.822, and that PitchCFM is what lifts pitch adaptation, with intra-domain log-F0 PCC reaching 0.813. Their ablation shows that removing the cycle loss lowers SMOS by 0.20 and TSIM by 0.11 in cross-domain VC, while removing PitchCFM lowers log-F0 PCC by 0.10.

Load-bearing premise

The load-bearing premise is that the sum of the forward and inverse flow-matching residuals in $L_{x\to y\to x}$ truly acts as a cycle and forces the converted speech, when converted back, to reconstruct the original source; if that sum is not a real round-trip constraint, the mechanism claimed for timbre adaptation is not present in the loss.

Editorial extensions

If this is right

  • Non-parallel voice conversion can be trained without paired utterances because the cycle term supplies the missing correspondence signal.
  • Cross-gender and wide-vocal-range conversion becomes feasible: PitchCFM moves F0 into the target range before VoiceCFM generates the audio, which should reduce hoarseness.
  • A voice-conversion system no longer needs a perfectly decoupled content encoder to achieve high timbre similarity, since the cycle regularizer compensates for content leakage.
  • The same cycle-plus-dual-flow recipe could be applied to other conditional generation tasks that map between two style domains with no paired data.

Reading between the lines

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

  • Because $L_{x\to y\to x}$ as written is a sum of two flow residuals rather than a true composition of the forward and inverse flows, part of the reported timbre gain may actually come from the invariance loss or from shared representations; an ablation that removes only the cycle term, only the invariance term, and both jointly would isolate the mechanism.
  • A variant that composes the forward and inverse ODE flows exactly would turn the regularizer into an honest reconstruction loss and would remove the need to tune the relative weights; this is a testable reformulation of the paper's central idea.
  • The pitch encoder is trained on source F0 and target speaker embeddings, so the method's limit can be probed by evaluating on vocal-range gaps wider than those in VCTK; if PitchCFM still tracks the source contour while shifting its mean, the pitch-adaptation claim extends beyond the tested range.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 4 minor

Summary. The paper proposes CycleFlow, a voice conversion method that combines conditional flow matching with a cycle-consistency regularizer for speaker timbre adaptation and a Dual-CFM decoder (PitchCFM and VoiceCFM) for pitch correction. Experiments on LibriTTS and VCTK compare CycleFlow against DiffVC, UUVC, CosyVoice, and Diff-HierVC using MOS, SMOS, TSIM, log F0 PCC, and WER, and the authors claim significant improvements in speaker similarity and pitch adaptation.

Significance. If the proposed cycle-consistency mechanism were sound, it would offer a plausible way to train non-parallel voice conversion without paired data, and the Dual-CFM design targets a real cross-gender pitch-mismatch problem. The paper includes experiments with several strong baselines and an ablation study. However, the central equation does not implement a cycle, the main speaker-similarity metric is computed with the same speaker encoder used for conditioning, the subjective confidence intervals overlap with those of baselines, and the reported pitch metric measures source-contour preservation rather than target adaptation. The central claims are therefore not established.

major comments (5)
  1. [Section II-B, Eq. (1)] The cycle-consistency loss Lx->y->x does not implement a cycle as claimed. It is the squared norm of v_theta(yt, sx, cx) - (y1 - (1-sigma)y0) plus v_theta(xt, sy, cx) - (x1 - (1-sigma)x0). The first residual trains the model to generate a target-domain sample conditioned on the source speaker sx, and the second trains it to generate a source-domain sample conditioned on the target speaker sy; the two residuals are independent and no variable connects the output of a forward conversion to the input of an inverse conversion. A model can drive both residuals to zero while the composition of the two ODE flows does not reconstruct x1, because each flow starts from an independent noise draw. Thus the transitivity regularization described in Sec. II-B, which is the paper's primary proposed mechanism for timbre adaptation, is absent as written; if the implemented loss differs, the text does not say how.
  2. [Section III-A, III-C, Table I] The TSIM metric is computed with CAM++ [19], the same pre-trained speaker encoder used in Sec. II-A to extract the target speaker embedding that conditions the Dual-CFM decoder. Evaluation with the same embedding model used for conditioning is circular: it measures how well the output aligns with CAM++'s speaker embedding, which the model is explicitly trained to match, and it does not provide independent evidence of perceived speaker similarity. This circularity directly affects the reported TSIM improvements in Table I.
  3. [Table I, Sec. III-C] The subjective results do not support the claim of significant improvement. For intra-domain MOS, CycleFlow is 3.71 +/- 0.09 versus CosyVoice 3.65 +/- 0.08 and Diff-HierVC 3.60 +/- 0.09; for SMOS, 3.23 +/- 0.06 versus Diff-HierVC 3.15 +/- 0.11. Cross-domain values overlap similarly (MOS 3.52 +/- 0.06 vs 3.49 +/- 0.06; SMOS 3.07 +/- 0.09 vs 3.03 +/- 0.07). With 10 listeners and 25 test samples, the confidence intervals are wide and no significance testing is reported, so the headline claim that CycleFlow significantly outperforms baselines in naturalness and similarity is not statistically established.
  4. [Table I, Sec. III-B] The log F0 PCC metric is defined in the table caption as the Pearson correlation between the F0 of the source and the converted audio. This measures how well the source F0 contour is preserved, not how well the converted F0 matches the target speaker's pitch range. The paper's pitch-adaptation claim is therefore not supported by the quantitative metric; Fig. 3 is a qualitative illustration using parallel data and does not quantify target matching. An appropriate metric would compare converted F0 statistics, such as mean or distribution, to the target speaker's F0 distribution.
  5. [Table II] The ablation table is incomplete: the 'Our' row contains no absolute values, only deltas are shown for the removed-component rows, and no statistical significance is reported. The WER delta for w/o Lcycle is -0.05 percent, meaning that removing the cycle loss improves WER, which is inconsistent with the claimed advantage. Without absolute MOS, SMOS, TSIM values and significance tests, the attribution of the SMOS and TSIM drop to cycle consistency in Sec. III-E is not supported. This matters because the ablation is the only direct evidence for the paper's main mechanism.
minor comments (4)
  1. [Sec. II-B, Fig. 1] The notation y_bar0 and y1' in Fig. 1 and in Eq. (1) is not defined; the construction of c_bar_y0 in Ly->y and Lx->y->y is not specified.
  2. [Sec. III-A] The text says '8 A100-80M GPUs'; this should presumably be A100-80GB GPUs.
  3. [Abstract and Sec. II-B] There are grammatical and typographical errors, including 'there is a significant disparity in pitch between the source and target speaker style domains' in the abstract and 'training objectice' in Sec. II-B.
  4. [Reproducibility] No code, trained models, or audio samples are released, which limits verification of the implemented loss and of the subjective and objective results.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper's claims do not reduce to their inputs by construction, though Eq. (1) and the TSIM metric raise validity and reproducibility concerns.

full rationale

CycleFlow's derivation is not circular in the sense of this pass. The cycle-consistency objective in Eq. (1) is the paper's own proposed definition and is tested against external baselines (DiffVC, UUVC, CosyVoice, Diff-HierVC) rather than being derived from those baselines or from a fitted parameter. The TSIM metric uses the same CAM++ speaker encoder [19] that supplies the conditioning embedding (Sec. II-A and Table I), but this is a metric-alignment and possible overfitting concern, not an equation-level equivalence: a model conditioned on a CAM++ embedding can still produce speech whose CAM++ embedding does not match the target, so the claim is not forced by construction. The same applies to log F0 PCC, which correlates converted F0 with source F0; this is a measurement-validity issue, not circularity. The self-citations ([3], [5], [17]) appear only as related work and are not load-bearing. The serious problems in this paper are technical correctness (as written, Lx→y→x sums two independent CFM residuals and does not instantiate a compositional cycle) and lack of released code or audio; these are not circularity and do not raise the circularity score.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central derivation depends on the OT-CFM framework, on pretrained disentanglement models, and on the correctness of the cycle loss. The main free parameters are the loss weights and the OT interpolation constant, none of which are reported. No new physical entities are introduced.

free parameters (2)
  • sigma = not specified
    Smoothing constant in the OT-CFM interpolation in Eq. (4) and Eq. (8); it must be set by hand but the paper does not report its value.
  • lambda_1, lambda_2, lambda_3 = not specified
    Loss weights in Eq. (2); the paper does not report their values or tuning procedure.
assumptions (4)
  • standard math Standard optimal-transport conditional flow matching theory is valid.
    The method builds on OT-CFM from Lipman et al. and Tong et al. without re-deriving it.
  • domain assumption The pre-trained speech tokenizer, RMVPE, CAM++, and HiFTNet provide well-disentangled content, pitch, and timbre representations and a high-quality vocoder.
    The entire pipeline assumes these external models generalize to the training data and that the CAM++ speaker embedding is a sufficient timbre condition.
  • ad hoc to paper The cycle loss Lx→y→x in Eq. (1) enforces forward/inverse transitivity.
    This is the central methodological premise; as written the loss is not a compositional cycle, so the premise is likely false.
  • domain assumption Non-parallel data can train a domain mapping via cycle consistency.
    Assumed from prior GAN-based voice conversion work; not proven in this paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of CycleFlow: Leveraging Cycle Consistency in Flow Matching for Speaker Style Adaptation." pith.science (2026). https://pith.science/paper/DHMMOZID

@misc{pith2026250101861,
  author       = {Pith},
  title        = {Pith review of: CycleFlow: Leveraging Cycle Consistency in Flow Matching for Speaker Style Adaptation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DHMMOZID}},
  note         = {Machine review of arXiv:2501.01861}
}
read the original abstract

Voice Conversion (VC) aims to convert the style of a source speaker, such as timbre and pitch, to the style of any target speaker while preserving the linguistic content. However, the ground truth of the converted speech does not exist in a non-parallel VC scenario, which induces the train-inference mismatch problem. Moreover, existing methods still have an inaccurate pitch and low speaker adaptation quality, there is a significant disparity in pitch between the source and target speaker style domains. As a result, the models tend to generate speech with hoarseness, posing challenges in achieving high-quality voice conversion. In this study, we propose CycleFlow, a novel VC approach that leverages cycle consistency in conditional flow matching (CFM) for speaker timbre adaptation training on non-parallel data. Furthermore, we design a Dual-CFM based on VoiceCFM and PitchCFM to generate speech and improve speaker pitch adaptation quality. Experiments show that our method can significantly improve speaker similarity, generating natural and higher-quality speech.

Figures

Figures reproduced from arXiv: 2501.01861 by the authors.

Figure 1
Figure 1. The conversion cycle in CFM-based VC. two domains with unparallel training samples is a more effec￾tive approach. Some generative adversarial network (GAN)- based methods [14]–[17] employ cycle consistency to learn this mapping without any parallel data. Cycle consistency enforces transitivity between forward and inverse conversion by regularizing pairs of samples, particularly in non-parallel VC where no correspond… view at source ↗
Figure 2
Figure 2. Overall pipeline of proposed CycleFlow. The CycleFlow includes a forward conversion from the source speech [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Pitch contours without normalization of voice conver [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Objective evaluation results for VC. F: Female; M: [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

28 extracted references · 24 canonical work pages

  1. [19]

    CAM++: A fast and efficient network for speaker verification using context-aware masking,

    Hui Wang, Siqi Zheng, Yafeng Chen, Luyao Cheng, and Qian Chen, “CAM++: A fast and efficient network for speaker verification using context-aware masking,” in 24th Annual Conference of the International Speech Communication Association, Interspeech 2023, Dublin, Ireland, August 20-24, 2023 , 2023, pp. 5301–5305

    work page 2023

  2. [1]

    A unified one-shot prosody and speaker conversion system with self-supervised discrete speech units,

    Li-Wei Chen, Shinji Watanabe, and Alexander Rudnicky, “A unified one-shot prosody and speaker conversion system with self-supervised discrete speech units,” in IEEE International Conference on Acoustics, Speech and Signal Processing ICASSP 2023, Rhodes Island, Greece, June 4-10, 2023 , 2023, pp. 1–5

    work page 2023

  3. [2]

    Speech resynthesis from discrete disentangled self-supervised representations,

    Adam Polyak, Yossi Adi, Jade Copet, Eugene Kharitonov, Kushal Lakhotia, Wei-Ning Hsu, Abdelrahman Mohamed, and Emmanuel Dupoux, “Speech resynthesis from discrete disentangled self-supervised representations,” in Interspeech, 2021, pp. 3615–3619

    work page 2021

  4. [3]

    EAD-VC: enhancing speech auto-disentanglement for voice conversion with IFUB estimator and joint text-guided consistent learning,

    Ziqi Liang, Jianzong Wang, Xulong Zhang, Yong Zhang, Ning Cheng, and Jing Xiao, “EAD-VC: enhancing speech auto-disentanglement for voice conversion with IFUB estimator and joint text-guided consistent learning,” in International Joint Conference on Neural Networks, IJCNN 2024, Yokohama, Japan, June 30 - July 5, 2024 , 2024, pp. 1–7

    work page 2024

  5. [4]

    Neural analysis and synthesis: Reconstructing speech from self-supervised representations,

    Hyeong-Seok Choi, Juheon Lee, Wansoo Kim, Jie Lee, Hoon Heo, and Kyogu Lee, “Neural analysis and synthesis: Reconstructing speech from self-supervised representations,” in Advances in Neural Information Processing Systems 34: Annual Conference on Neural Information Processing Systems 2021, NeurIPS 2021, December 6-14, 2021, virtual , 2021, pp. 16251–16265

    work page 2021

  6. [5]

    PMVC: data augmentation-based prosody modeling for expressive voice conversion,

    Yimin Deng, Huaizhen Tang, Xulong Zhang, Jianzong Wang, Ning Cheng, and Jing Xiao, “PMVC: data augmentation-based prosody modeling for expressive voice conversion,” in Proceedings of the 31st ACM International Conference on Multimedia, MM 2023, Ottawa, ON, Canada, 29 October 2023- 3 November 2023 , 2023, pp. 184–192

    work page 2023

  7. [6]

    wav2vec 2.0: A framework for self-supervised learning of speech representations,

    Alexei Baevski, Yuhao Zhou, Abdelrahman Mohamed, and Michael Auli, “wav2vec 2.0: A framework for self-supervised learning of speech representations,” in Advances in Neural Information Processing Systems 33: Annual Conference on Neural Information Processing Systems 2020, NeurIPS 2020, December 6-12, 2020, virtual , 2020

    work page 2020

  8. [7]

    Hubert: Self- supervised speech representation learning by masked prediction of hidden units,

    Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakho- tia, Ruslan Salakhutdinov, and Abdelrahman Mohamed, “Hubert: Self- supervised speech representation learning by masked prediction of hidden units,” IEEE ACM Trans. Audio Speech Lang. Process. , vol. 29, pp. 3451–3460, 2021

    work page 2021

Show all 28 references
  1. [8]

    Contentvec: An im- proved self-supervised speech representation by disentangling speakers,

    Kaizhi Qian, Yang Zhang, Heting Gao, Junrui Ni, Cheng-I Lai, David D. Cox, Mark Hasegawa-Johnson, and Shiyu Chang, “Contentvec: An im- proved self-supervised speech representation by disentangling speakers,” in International Conference on Machine Learning, ICML 2022, 17-23 Jul...

    2022

  2. [9]

    Cosyvoice: A scalable multilingual zero-shot text-to-speech synthesizer based on supervised semantic tokens,

    Zhihao Du, Qian Chen, Shiliang Zhang, Kai Hu, Heng Lu, Yexin Yang, Hangrui Hu, Siqi Zheng, Yue Gu, Ziyang Ma, Zhifu Gao, and Zhijie Yan, “Cosyvoice: A scalable multilingual zero-shot text-to-speech synthesizer based on supervised semantic tokens,” CoRR, vol. abs/2407.05407, 2024

    2024 arXiv

  3. [10]

    Soundstorm: Efficient parallel audio generation,

    Zal ´an Borsos, Matthew Sharifi, Damien Vincent, Eugene Kharitonov, Neil Zeghidour, and Marco Tagliasacchi, “Soundstorm: Efficient parallel audio generation,” CoRR, vol. abs/2305.09636, 2023

    2023 arXiv

  4. [11]

    Sef-vc: Speaker embedding free zero-shot voice conversion with cross attention,

    Junjie Li, Yiwei Guo, Xie Chen, and Kai Yu, “Sef-vc: Speaker embedding free zero-shot voice conversion with cross attention,” in ICASSP, 2024, pp. 12296–12300

    2024

  5. [12]

    LM-VC: zero-shot voice conversion via speech generation based on language models,

    Zhichao Wang, Yuanzhe Chen, Lei Xie, Qiao Tian, and Yuping Wang, “LM-VC: zero-shot voice conversion via speech generation based on language models,” IEEE Signal Process. Lett. , vol. 30, pp. 1157–1161, 2023

    2023

  6. [13]

    VQ-CL: learning disentangled speech representations with contrastive learning and vector quantization,

    Huaizhen Tang, Xulong Zhang, Jianzong Wang, Ning Cheng, and Jing Xiao, “VQ-CL: learning disentangled speech representations with contrastive learning and vector quantization,” in IEEE International Conference on Acoustics, Speech and Signal Processing ICASSP 2023, Rhodes Islan...

    2023

  7. [14]

    High-quality nonparallel voice conversion based on cycle- consistent adversarial network,

    Fuming Fang, Junichi Yamagishi, Isao Echizen, and Jaime Lorenzo- Trueba, “High-quality nonparallel voice conversion based on cycle- consistent adversarial network,” in 2018 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2018, Calgary, AB, Cana...

    2018

  8. [15]

    Cyclegan-vc2: Improved cyclegan-based non-parallel voice conversion,

    Takuhiro Kaneko, Hirokazu Kameoka, Kou Tanaka, and Nobukatsu Hojo, “Cyclegan-vc2: Improved cyclegan-based non-parallel voice conversion,” in IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2019, Brighton, United Kingdom, May 12-17, 2019 , 2019,...

    2019

  9. [16]

    Maskcyclegan-vc: Learning non-parallel voice conversion with filling in frames,

    Takuhiro Kaneko, Hirokazu Kameoka, Kou Tanaka, and Nobukatsu Hojo, “Maskcyclegan-vc: Learning non-parallel voice conversion with filling in frames,” in IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2021, Toronto, ON, Canada, June 6-11, 2021 ,...

    2021

  10. [17]

    Cyclegean: Cycle generative enhanced adversarial network for voice conversion,

    Xulong Zhang, Jianzong Wang, Ning Cheng, Edward Xiao, and Jing Xiao, “Cyclegean: Cycle generative enhanced adversarial network for voice conversion,” in IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2021, Cartagena, Colombia, December 13-17, 2021 , 2021, p...

    2021

  11. [18]

    RMVPE: A robust model for vocal pitch estimation in polyphonic music,

    Haojie Wei, Xueke Cao, Tangpeng Dan, and Yueguo Chen, “RMVPE: A robust model for vocal pitch estimation in polyphonic music,” in 24th Annual Conference of the International Speech Communication Association, Interspeech 2023, Dublin, Ireland, August 20-24, 2023 , 2023, pp. 5421–5425

    2023

  12. [20]

    Diffusion-based voice conversion with fast maximum likelihood sampling scheme,

    Vadim Popov, Ivan V ovk, Vladimir Gogoryan, Tasnima Sadekova, Mikhail Sergeevich Kudinov, and Jiansheng Wei, “Diffusion-based voice conversion with fast maximum likelihood sampling scheme,” in The Tenth International Conference on Learning Representations, ICLR 2022, Virtual E...

    2022

  13. [21]

    Diff-hiervc: Diffusion-based hierarchical voice conversion with robust pitch genera- tion and masked prior for zero-shot speaker adaptation,

    Ha-Yeong Choi, Sang-Hoon Lee, and Seong-Whan Lee, “Diff-hiervc: Diffusion-based hierarchical voice conversion with robust pitch genera- tion and masked prior for zero-shot speaker adaptation,” in 24th Annual Conference of the International Speech Communication Association, Int...

    2023

  14. [22]

    Naturalspeech 2: Latent diffusion models are natural and zero-shot speech and singing synthesizers,

    Kai Shen, Zeqian Ju, Xu Tan, Eric Liu, Yichong Leng, Lei He, Tao Qin, Sheng Zhao, and Jiang Bian, “Naturalspeech 2: Latent diffusion models are natural and zero-shot speech and singing synthesizers,” in The Twelfth International Conference on Learning Representations, ICLR 202...

    2024

  15. [23]

    Flow matching for generative modeling,

    Yaron Lipman, Ricky T. Q. Chen, Heli Ben-Hamu, Maximilian Nickel, and Matthew Le, “Flow matching for generative modeling,” in The Eleventh International Conference on Learning Representations, ICLR 2023, Kigali, Rwanda, May 1-5, 2023 , 2023

    2023

  16. [24]

    Improving and generalizing flow-based generative models with mini- batch optimal transport,

    Alexander Tong, Kilian Fatras, Nikolay Malkin, Guillaume Huguet, Yanlei Zhang, Jarrid Rector-Brooks, Guy Wolf, and Yoshua Bengio, “Improving and generalizing flow-based generative models with mini- batch optimal transport,” Trans. Mach. Learn. Res. , vol. 2024, 2024

    2024

  17. [25]

    Conformer: Convolution-augmented transformer for speech recognition,

    Anmol Gulati, James Qin, Chung-Cheng Chiu, Niki Parmar, Yu Zhang, Jiahui Yu, Wei Han, Shibo Wang, Zhengdong Zhang, Yonghui Wu, and Ruoming Pang, “Conformer: Convolution-augmented transformer for speech recognition,” in 21st Annual Conference of the International Speech Communi...

    2020

  18. [26]

    Libritts: A corpus derived from librispeech for text-to-speech,

    Heiga Zen, Viet Dang, Rob Clark, Yu Zhang, Ron J. Weiss, Ye Jia, Zhifeng Chen, and Yonghui Wu, “Libritts: A corpus derived from librispeech for text-to-speech,” in 20th Annual Conference of the International Speech Communication Association, Interspeech 2019, Graz, Austria, Se...

    2019

  19. [27]

    Super- seded - cstr vctk corpus: English multi-speaker corpus for cstr voice cloning toolkit,

    Christophe Veaux, Junichi Yamagishi, and Kirsten MacDonald, “Super- seded - cstr vctk corpus: English multi-speaker corpus for cstr voice cloning toolkit,” 2016

    2016

  20. [28]

    Hift- net: A fast high-quality neural vocoder with harmonic-plus-noise filter and inverse short time fourier transform,

    Yinghao Aaron Li, Cong Han, Xilin Jiang, and Nima Mesgarani, “Hift- net: A fast high-quality neural vocoder with harmonic-plus-noise filter and inverse short time fourier transform,” CoRR, vol. abs/2309.09493, 2023

    2023 arXiv

Pith tools

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