CycleGAN-VC: Non-parallel Voice Conversion Using Cycle-Consistent Adversarial Networks
The DOI matches its Crossref record; 60 of the 61 references it cites were found in a registry or an index.
Abstract
We propose a non-parallel voice-conversion (VC) method that can learn a mapping from source to target speech without relying on parallel data. The proposed method is particularly noteworthy in that it is general purpose and high quality and works without any extra data, modules, or alignment procedure. Our method, called CycleGAN-VC, uses a cycle-consistent adversarial network (CycleGAN) with gated convolutional neural networks (CNNs) and an identity-mapping loss. A CycleGAN learns forward and inverse mappings simultaneously using adversarial and cycle-consistency losses. This makes it possible to find an optimal pseudo pair from non-parallel data. Furthermore, the adversarial loss can bring the converted speech close to the target one on the basis of indistinguishability without explicit density estimation. This allows to avoid over-smoothing caused by statistical averaging, which occurs in many conventional statistical model-based VC methods that represent data distribution explicitly. We configure a CycleGAN with gated CNNs and train it with an identity-mapping loss. This allows the mapping function to capture sequential and hierarchical structures while preserving linguistic information. We evaluated our method on a non-parallel VC task. An objective evaluation showed that the converted feature sequence was near natural in terms of global variance and modulation spectra, which are structural indicators highly correlated with subjective evaluation. A subjective evaluation showed that the quality of the converted speech was comparable to that obtained with a Gaussian mixture model-based parallel VC method even though CycleGAN-VC is trained under disadvantageous conditions (non-parallel and half the amount of data).
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Recent research [59,60,61,62,63] in VC technologies has addressed long-standing challenges such as data scarcity, cross-lingual conversion, and real-time processing.”
From Audio Deepfake Detection: What Has Been Achieved and What Lies Ahead (2025) · cites this paper as [59]
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · cited 1×
“Inspired by a previous study [49] for neural style transfer and super-resolution, we used the network that included downsampling, residual [50], and upsampling layers, as well as incorporating instance normalization [51].”
MethodAdam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946 · cited 1×
“We trained the network using the Adam optimizer [53] with a batch size of 1.”
MethodFully convolutional networks for semantic segmentation
Long, J., Shelhamer, E., & Darrell, T. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298965 · Cited by 38,834
Unpaired image-to-image translation using cycle-consistent adversarial networks
Zhu, J.-Y., Park, T., Isola, P., & Efros, A. A. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.244 · Cited by 23,602
Labaca-Castro, R. (2023) · Machine Learning under Malware Attack · doi:10.1007/978-3-658-40442-0_9 · Cited by 19,272
GAN(Generative Adversarial Nets)
(2017) · Journal of Japan Society for Fuzzy Theory and Intelligent Informatics · doi:10.3156/jsoft.29.5_177_2 · Cited by 19,270
(2010) · International Conference on Machine Learning · Cited by 13,995
Perceptual Losses for Real-Time Style Transfer and Super-Resolution
Johnson, J., Alahi, A., & Fei-Fei, L. (2016) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-46475-6_43 · Cited by 11,283 · cited 2×
“Inspired by a previous study [49] for neural style transfer and super-resolution, we used the network that included downsampling, residual [50], and upsampling layers, as well as incorporating instance normalization [51].”
MethodShi, W., Caballero, J., Huszar, F., Totz, J., Aitken, A. P., Bishop, R., Rueckert, D., & Wang, Z. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.207 · Cited by 7,570
“Least squares generative adversarial networks,”
Mao, X., Li, Q., Xie, H., Lau, R. Y. K., Wang, Z., & Smolley, S. P. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.304 · Cited by 5,437 · cited 1×
“To stabilize training, we used a least squares GAN [52], which replaces the negative log likelihood objective in Ladv by a least squares loss.”
Method
Research works printed but not linked: 1
- maas, 2013, Rectifier nonlinearities improve neural network acoustic models, Proc ICML Workshop
Cited by
Items 101–200 of 354
Haque, K. N., Rana, R. K., Liu, J., Hansen, J. H. L., Cummins, N., Busso, C., & Schuller, B. W. (2021) · IEEE/ACM Transactions on Audio Speech and Language Processing · doi:10.1109/taslp.2021.3098764 · Cited by 14
“To obtain good quality continuous speech generation with GGAN, different techniques from the recent literature can be used [14], [6], [78], [27], [28].”
Data Augmentation Methods for End-to-End Speech Recognition on Distant-Talk Scenarios
Tsunoo, E., Shibata, K., Narisetty, C. P., Kashiwagi, Y., & Watanabe, S. (2021) · Interspeech (USB) · doi:10.21437/interspeech.2021-958 · Cited by 13
FastS2S-VC: Streaming Non-Autoregressive Sequence-to-Sequence Voice ConversionPreprint
Kameoka, H., Tanaka, K., & Kaneko, T. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2104.06900 · Cited by 20
“These include methods based on an i-vector representation [11], restricted Boltzmann machines [12], phonetic posteriograms [13], regular/variational autoen-coders [14]–[20], generative adversarial networks [21], [22], flow-based models [23], and score-based generative models [24].”
MethodSelf-citationPathological voice adaptation with autoencoder-based voice conversion
Illa, M., Halpern, B. M., van Son, R. J. J. H., Moro-Velázquez, L., & Scharenborg, O. (2021) · ISCA Speech Synthesis Workshop (SSW) · doi:10.21437/ssw.2021-4 · Cited by 12
Voice privacy using CycleGAN and time-scale modification
Prajapati, G. P., Singh, D. K., Amin, P. P., & Patil, H. A. (2022) · Computer Speech & Language · doi:10.1016/j.csl.2022.101353 · Cited by 12
The Sound of Silence: Efficiency of First Digit Features in Synthetic Audio Detection
Mari, D., Latora, F., & Milani, S. (2022) · IEEE International Workshop on Information Forensics and Security · doi:10.1109/wifs55849.2022.9975404 · Cited by 17
“The mapping can also be performed by means of Generative Adversarial Networks (GANs) since the task is similar to image to image translation allowing similar techniques to be adopted [31], [32].”
MethodVoice Conversion with Denoising Diffusion Probabilistic GAN Models
Zhang, X., Wang, J., Cheng, N., & Xiao, J. Z. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-46674-8_11 · Cited by 14
Ho, T. V., & Akagi, M. (2021) · IEEE Access · doi:10.1109/access.2021.3063519 · Cited by 13
“Similar to a previous study [31], the MS was averaged for all modulation frequencies and all utterances as where X is a batch of test utterances, N is the number of utterances, n ∈ [0 , N ) is the utterance index, F is the number of MS frequency bins, and f ∈ [0 , F ) is the MS frequency bins.”
ResultTowards end-to-end F0 voice conversion based on Dual-GAN with convolutional wavelet kernels
Veillon, C. L. M., Obin, N., & Röebel, A. (2021) · European Signal Processing Conference (EUSIPCO) · doi:10.23919/eusipco54536.2021.9616190 · Cited by 12
CycleTransGAN-EVC: A CycleGAN-based Emotional Voice Conversion Model with TransformerPreprint
Fu, C., Liu, C., Ishi, C. T., & Ishiguro, H. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2111.15159 · Cited by 12
“In this study, we extracted F0 and spectrogram from the speech as our model’s inputs inspired by Zhou et al. [13], Ming et al. [20], and Kaneko et al. [21].”
MethodGenerative Adversarial Training Data Adaptation for Very Low-Resource Automatic Speech Recognition
Matsuura, K., Mimura, M., Sakai, S., & Kawahara, T. (2020) · Interspeech (USB) · doi:10.21437/interspeech.2020-1195 · Cited by 11
Non-Parallel Voice Conversion with Autoregressive Conversion Model and Duration Adjustment
Liu, L., Chen, Y., Zhang, J., Jiang, Y., Hu, Y.-J., Ling, Z.-H., & Dai, L.-R. (2020) · doi:10.21437/vcc_bc.2020-17 · Cited by 18
“These methods include phonetic posteriorgrams (PPGs) based method [8], generative adversarial network (GAN) based method [9], variational auto-encoder (VAE) based method [10] and sequence-to-sequence based method [11].”
MethodAziz, S., & Shahnawazuddin, S. (2023) · Applied Acoustics · doi:10.1016/j.apacoust.2023.109420 · Cited by 11
Saito, Y., Takamichi, S., & Saruwatari, H. (2019) · Computer Speech & Language · doi:10.1016/j.csl.2019.05.008 · Cited by 11
MASS: Multi-task anthropomorphic speech synthesis framework
Chen, J., Ye, L., & Ming, Z. (2021) · Computer Speech & Language · doi:10.1016/j.csl.2021.101243 · Cited by 12
Manipulating Voice Attributes by Adversarial Learning of Structured Disentangled Representations
Benaroya, L., Obin, N., & Röebel, A. (2023) · Entropy · doi:10.3390/e25020375 · Cited by 14
CycleGAN-VC2: Improved CycleGAN-based Non-parallel Voice ConversionPreprint
Kaneko, T., Kameoka, H., Tanaka, K., & Hojo, N. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1904.04631 · Cited by 13
“…in contrast to some of the frame-by-frame methods [27, 30], which have difficulty in learning time dependencies, CycleGAN-VC [29] (published in [32]) makes it possible to learn a sequence-based mapping function by using CycleGAN [33, 34, 35] with a gated CNN [36] and identity-mapping loss [37].”
Self-citationCycleGAN with Dual Adversarial Loss for Bone-Conducted Speech Enhancement
Pan, Q. J., Pan, Y., Zhou, J., Wang, H., Tao, L., & Kwan, H. K. (2022) · IEEE Region 10 Conference (TENCON) · doi:10.1109/tencon55691.2022.9977942 · Cited by 10
“To show the effectiveness of the proposed CycleGAN-DAL model for BC speech enhancement, the Gaussian mixture model (GMM) based BC speech enhancement [14], the BLSTM-based BC speech enhancement method [5], the starGAN-based speech enhancement method [15], and the conventional CycleGAN method [16] were implemented respe…”
MethodAn Analysis of Performance Evaluation Metrics for Voice Conversion Models
Akhter, M. T., Banerjee, P., Dhar, S., & Jana, N. D. (2022) · IEEE India Conference · doi:10.1109/indicon56171.2022.10040000 · Cited by 11
Augmented CycleGANs for Continuous Scale Normal-to-Lombard Speaking Style Conversion
Seshadri, S., Juvela, L., Alku, P., & Räsänen, O. J. (2019) · Interspeech (USB) · doi:10.21437/interspeech.2019-1681 · Cited by 10
Hierarchical disentangled representation learning for singing voice conversion
Takahashi, N., Singh, M., & Mitsufuji, Y. (2021) · IEEE International Joint Conference on Neural Network · doi:10.1109/ijcnn52387.2021.9533583 · Cited by 19
Improving Recognition-Synthesis Based any-to-one Voice Conversion with Cyclic Training
Chen, Y., Liu, L., Hu, Y.-J., Jiang, Y., & Ling, Z.-H. (2022) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp43922.2022.9747140 · Cited by 9
Sudro, P. N., Das, R. K., Sinha, R., & Prasanna, S. R. M. (2021) · IEEE Spoken Language Technology Workshop (SLT) · doi:10.1109/slt48900.2021.9383543 · Cited by 11
Toward Degradation-Robust Voice Conversion
Huang, C., Chang, K., & Lee, H.-Y. (2022) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp43922.2022.9746636 · Cited by 13
Face-Driven Zero-Shot Voice Conversion with Memory-based Face-Voice Alignment
Sheng, Z.-Y., Ai, Y., Chen, Y., & Ling, Z.-H. (2023) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3581783.3613825 · Cited by 15
“Since parallel data are not always available, many non-parallel VC techniques have been proposed, including the methods based on variational auto-encoder (VAE) [14, 31], generative adversarial network (GAN) [13, 16, 18, 44], recognition-synthesis [6, 23, 31] and disentanglement [53].”
MethodEzzine, K., Martino, J. D., & Frikha, M. (2023) · Applied Sciences · doi:10.3390/app132111988 · Cited by 9
PVR-Vocoder: A Pathological Voice Repair Vocoder for Voice Disorders
Liu, G., Zhang, T., Liu, X., Hou, X., Ding, B., Fu, D., & Pang, Z. (2023) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2023.3340738 · Cited by 8
GNNAE-AVSS: Graph Neural Network Based Autoencoders for Audio-Visual Speech Synthesis
Ghosh, S., & Jana, N. D. (2024) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn60899.2024.10651476 · Cited by 9
“StarGAN-VC [25] merges advantages from both CVAE (cycle variational autoencoder) [26] and cycleGAN [27] methodologies.”
MethodTaco-VC: A Single Speaker Tacotron based Voice Conversion with Limited Data
Leshem, R. L., & Giryes, R. (2020) · European Signal Processing Conference (EUSIPCO) · doi:10.23919/eusipco47968.2020.9287448 · Cited by 9
CASIA Voice Conversion System for the Voice Conversion Challenge 2020
Lian, Z., Tao, J., Wen, Z., & Zhong, R. (2020) · doi:10.21437/vcc_bc.2020-19 · Cited by 8
“To deal with this problem, generative adversarial network (GAN) [11] and its variants (such as Cy-cleGAN [12,13] and StarGAN [14,15]) use a discriminator to amplify this artifact in the loss function.”
BackgroundZero-shot Voice Conversion via Self-supervised Prosody Representation Learning
Wang, S., & Borth, D. (2022) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn55064.2022.9892405 · Cited by 8
“Since we use a vocoder PWGAN to generate waveform files from the synthesized Mel-Spectrograms, to make a fair comparison, we transform these 10 waveform files from the original dataset 3https://github.com/resemble-ai/Resemblyzer into spectrograms, then convert them back with the PWGAN.”
InfluentialNoise-robust voice conversion using adversarial training with multi-feature decoupling
Chen, L., Zhang, X., Li, Y., & Sun, M. (2024) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2023.107807 · Cited by 10
Guo, H., Liu, C., Ishi, C. T., & Ishiguro, H. (2023) · IEEE Workshop on Automatic Speech Recognition and Understanding (ASRU) · doi:10.1109/asru57964.2023.10389621 · Cited by 16
“Voice conversion (VC) is a technique that modifies a speaker’s voice to sound like that of another speaker while preserving the original meaning and content of the speech [1].”
BackgroundDoLP and AoP Synthesis from division of focal plane polarimeters using CycleGAN
Hao, Y., Duan, J., Liu, J., Zhan, J., & Cheng, C. (2023) · Optics Communications · doi:10.1016/j.optcom.2023.129296 · Cited by 8
StarGAN-based Emotional Voice Conversion for Japanese PhrasesPreprint
Moritani, A., Ozaki, R., Sakamoto, S., Kameoka, H., & Taniguchi, T. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2104.01807 · Cited by 8
Self-citationAn Objective Evaluation Framework for Pathological Speech SynthesisPreprint
Halpern, B. M., Fritsch, J., Hermann, E., van Son, R. J. J. H., Scharenborg, O., & Magimai.-Doss, M. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2107.00308 · Cited by 8
“A CycleGAN-VC system is trained with the augmented database [18]2 with the same hyperparameters and values (as”
MethodNoisy-to-Noisy Voice Conversion Under Variations of Noisy Condition
Xie, C., & Toda, T. (2023) · IEEE/ACM Transactions on Audio Speech and Language Processing · doi:10.1109/taslp.2023.3313426 · Cited by 9
“Many approaches, such as those based on generative adversarial networks [5], [6], [7], [8], variational autoencoders (VAEs) [9], [10], [11], [12], automatic speech recognition (ASR), and text-to-speech (TTS) [13], [14], [15], are frequently focused on in the latest VC studies.”
BackgroundReconstructing Dual Learning for Neural Voice Conversion Using Relatively Few Samples
Sun, A., Wang, J., Cheng, N., Tantrawenith, M., Wu, Z., Meng, H. M. L., Xiao, E., & Xiao, J. Z. (2021) · IEEE Workshop on Automatic Speech Recognition and Understanding (ASRU) · doi:10.1109/asru51503.2021.9687965 · Cited by 7
Dhar, S., Banerjee, P., Jana, N. D., & Das, S. (2023) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49357.2023.10095069 · Cited by 10
“Identity Loss: In this work identity loss [17] is used to preserve the loss of linguistic information during the VC process.”
MethodCycleDiffusion: Voice Conversion Using Cycle-Consistent Diffusion Models
Yook, D., Han, G., Chang, H.-P., & Yoo, I.-C. (2024) · Applied Sciences · doi:10.3390/app14209595 · Cited by 7
Voice Attribute Editing With Text Prompt
Sheng, Z.-Y., Liu, L., Ai, Y., Pan, J., & Ling, Z.-H. (2025) · IEEE Transactions on Audio Speech and Language Processing · doi:10.1109/taslpro.2025.3557193 · Cited by 18
Reference-Free Adversarial Sex Obfuscation in Speech
Qu, Y., Panariello, M., Todisco, M., & Evans, N. (2025) · Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) · doi:10.1109/apsipaasc65261.2025.11249000 · Cited by 7
“1 Traditional voice conversion methods, which rely on parallel corpora or target speaker references [4], [5], face two fundamental limitations, namely the high cost of acquiring sensitive target speech data and the incomplete suppression of sex-discriminative acoustic features (e.g., fundamental…”
Non-Parallel Many-To-Many Voice Conversion by Knowledge Transfer from a Text-To-Speech Model
Yu, X., & Mak, B. K.-W. (2021) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp39728.2021.9414757 · Cited by 7
Comparing the performance of classic voice-driven assistive systems for dysarthric speech
Zheng, W.-Z., Han, J.-Y., Cheng, H.-L., Chu, W.-C., Chen, K.-C., & Lai, Y.-H. (2022) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2022.104447 · Cited by 7
A Novel Attention-Guided Generative Adversarial Network for Whisper-to-Normal Speech Conversion
Gao, T. F., Pan, Q., Zhou, J., Wang, H., Tao, L., & Kwan, H. K. (2023) · Cognitive Computation · doi:10.1007/s12559-023-10108-9 · Cited by 7
“In addition, motivated by the work in [31], identity mapping loss is considered when training the proposed generative adversarial networks (GAN) to preserve the semantic information of whispered speech in the generated normal speech.”
BackgroundAdversarial Guitar Amplifier Modelling with Unpaired Data
Wright, A., Välimäki, V., & Juvela, L. (2023) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49357.2023.10094600 · Cited by 7
“This can be achieved, for example, using non-parallel speech data to train a Generative Adversarial Network (GAN) with a cycle-consistency loss [21].”
MethodCinC-GAN for Effective F0prediction for Whisper-to-Normal Speech Conversion
Patel, M., Purohit, M., Shah, J., & Patil, H. A. (2020) · European Signal Processing Conference (EUSIPCO) · doi:10.23919/eusipco47968.2020.9287385 · Cited by 7
Random Cycle Loss and Its Application to Voice Conversion
Sun, H., Wang, D., Li, L., Chen, C., & Zheng, T. F. (2023) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2023.3257839 · Cited by 11
“The VC models based on CycleGAN [47], [48], [49] and StartGAN [50] are examples of this paradigm.”
MethodTimbre-Reserved Adversarial Attack in Speaker Identification
Wang, Q., Yao, J., Zhang, L., Guo, P., & Xie, L. (2023) · IEEE/ACM Transactions on Audio Speech and Language Processing · doi:10.1109/taslp.2023.3306714 · Cited by 9
“With the development of VC, different kinds of VC models have been explored, and they are typically divided into three types: autoregressive based VC model [43], [44], non-autoregressive based VC model [45]–[49], and end-to-end VC model [50]– [52].”
MethodSakamoto, S., Taniguchi, A., Taniguchi, T., & Kameoka, H. (2021) · Interspeech (USB) · doi:10.21437/interspeech.2021-492 · Cited by 11
Self-citationThe NUS & NWPU system for Voice Conversion Challenge 2020
Tian, X., Wang, Z., Shan, Y., Zhou, X., Du, H., Zhou, Y. X., Zhang, M., Zhou, K., Şişman, B., Xie, L., et al. (2020) · doi:10.21437/vcc_bc.2020-26 · Cited by 13
Alaa, Y., Alfonse, M., & Aref, M. M. (2022) · International Conference on Computing and Informatics (ICCI) · doi:10.1109/icci54321.2022.9756059 · Cited by 7
Jafaryani, M., Sheikhzadeh, H., & Pourahmadi, V. (2022) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2022.105279 · Cited by 7
Deepfake Detection System for the ADD Challenge Track 3.2 Based on Score Fusion
Zhang, Y., Lu, J., Wang, X., Li, Z., Xiao, R., Wang, W., Li, M., & Zhang, P. (2022) · doi:10.1145/3552466.3556528 · Cited by 6
“Attacking an automatic speaker verification (ASV) system with spoof speech generated by text-to-speech (TTS) [22, 54] and VC [17] is defined as the logical access (LA) attack.”
Dhar, S., Akhter, M. T., Banerjee, P., Jana, N. D., & Das, S. (2023) · Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) · doi:10.1109/apsipaasc58517.2023.10317438 · Cited by 8
“Most of the preceding works on GAN-based VC [8], [9], [13] have been explored using the instance normalization technique.”
BackgroundMethodTowards General-Purpose Text-Instruction-Guided Voice Conversion
Kuan, C.-Y., Li, C.-A., Hsu, T.-Y., Lin, T.-Y., Chung, H.-L., Chang, K., Chang, S.-Y., & Lee, H.-Y. (2023) · IEEE Workshop on Automatic Speech Recognition and Understanding (ASRU) · doi:10.1109/asru57964.2023.10389672 · Cited by 17
A Style Transfer Approach to Source Separation
Venkataramani, S., Tzinis, E., & Smaragdis, P. (2019) · IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA) · doi:10.1109/waspaa.2019.8937203 · Cited by 6
Perturbation AUTOVC: Voice Conversion From Perturbation and Autoencoder Loss
Park, H.-Y., Lee, Y. H., & Chun, C. (2023) · IEEE Access · doi:10.1109/access.2023.3341434 · Cited by 7
“Unlike general generative models, it learns two-way mappings [26], [27].”
BackgroundMethodImproving Spoofing Capability for End-to-end Any-to-many Voice Conversion
Hua, H., Chen, Z., Zhang, Y., Li, M., & Zhang, P. (2022) · doi:10.1145/3552466.3556532 · Cited by 7
“Models based on generative adversarial networks (GAN) along with their variants, such as StarGAN-VC[9] and CycleGAN-VC[10] use a generator or a conditional generator that transforms the source speaker’s voice features to the target speakers’ directly.”
BackgroundEnd-to-End Voice Conversion with Information Perturbation
Xie, Q., Yang, S., Lei, Y., Xie, L., & Su, D. (2022) · International Symposium on Chinese Spoken Language Processing (ISCSLP) · doi:10.1109/iscslp57327.2022.10037890 · Cited by 8
Spectrum and Prosody Conversion for Cross-lingual Voice Conversion with CycleGANPreprint
Du, Z., Zhou, K., Şişman, B., & Li, H. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2008.04562 · Cited by 8
“CycleGAN, a successful implementation of GANs, was first proposed for unpaired image-to-image translation [46], and then applied to other non-parallel style transfer tasks, such as speech synthesis and mono-lingual voice conversion [25], [53], [54].”
MethodAlignSTS: Speech-to-Singing Conversion via Cross-Modal Alignment
Li, R., Huang, R., Zhang, L., Liu, J., & Zhao, Z. (2023) · Findings of the Association for Computational Linguistics: ACL (Findings of ACL) · doi:10.18653/v1/2023.findings-acl.442 · Cited by 8
One-Class Fake Speech Detection Based on Improved Support Vector Data Description
Zhang, J., Yi, X., & Zhao, X. (2023) · Security and Communication Networks · doi:10.1155/2023/8830894 · Cited by 7
CCLCap-AE-AVSS: Cycle consistency loss based capsule autoencoders for audio–visual speech synthesis
Ghosh, S., Jana, N. D., Si, T., Mallik, S., & Shah, M. A. (2024) · Journal of Intelligent Systems · doi:10.1515/jisys-2023-0171 · Cited by 5
“A particularly noteworthy milestone in this trajectory was the CycleGAN-based VC framework by Kaneko and Kameoka [32].”
Behdad, M., & Gharavian, D. (2024) · IEEE Access · doi:10.1109/access.2024.3471926 · Cited by 5
“The CycleGAN-VCs family [24], [25], [26], [27] inspired by CycleGAN for unpaired image-to-image translation [28], exemplifies this category [29].”
BackgroundMethodAudio-Based Kinship Verification Using Age Domain Conversion
Sun, Q., Akman, A., Jing, X., Milling, M., & Schuller, B. W. (2024) · IEEE Signal Processing Letters · doi:10.1109/lsp.2024.3515811 · Cited by 5
“Models such as ACVAE-VC [25], CycleGAN-VC [26], FragmentVC [27], and StarGAN [28] demonstrate efficacy in gender conversion and speaker identity conversion experiments.”
CycleGAN-Based Singing/Humming to Instrument Conversion Technique
Lai, W.-H., Wang, S.-L., & Xu, Z.-Y. (2022) · Electronics · doi:10.3390/electronics11111724 · Cited by 6
Neural Network-Based Signal Translation with Application to the ECG
Abdelmadjid, M. A., & Boukadoum, M. (2022) · IEEE Interregional NEWCAS Conference (NEWCAS) · doi:10.1109/newcas52662.2022.9842248 · Cited by 5
StyleVC: Non-Parallel Voice Conversion with Adversarial Style Generalization
Hwang, I.-S., Lee, S.-H., & Lee, S. (2022) · Proceedings - International Conference on Pattern Recognition/Proceedings/International Conference on Pattern Recognition · doi:10.1109/icpr56361.2022.9956613 · Cited by 9
“STARGAN-VC [25] and CycleGAN-VC [26] have been widely accepted as benchmark approaches.”
BackgroundAziz, S., & Shahnawazuddin, S. (2024) · Circuits Systems and Signal Processing · doi:10.1007/s00034-023-02592-z · Cited by 5
Emotional Speech Generator by using Generative Adversarial Networks
Asakura, T., Akama, S., Sato-Shimokawara, E., Yamaguchi, T., & Yamamoto, S. (2019) · International Symposium on Information and Communication Technology · doi:10.1145/3368926.3369662 · Cited by 5
Bhattacharjee, S., & Sinha, R. (2022) · International Conference on Signal Processing and Communications (SPCOM) · doi:10.1109/spcom55316.2022.9840769 · Cited by 5
Voice Conversion Based on Deep Neural Networks for Time-Variant Linear Transformations
Kotani, G., Saito, D., & Minematsu, N. (2022) · IEEE/ACM Transactions on Audio Speech and Language Processing · doi:10.1109/taslp.2022.3205755 · Cited by 6
“In particular, a recognition-synthesis-based approach, including the use of phonetic posteriorgrams (PPGs) or the combination of automatic speech recognition (ASR) and TTS, is one of the most popular ways [17], [18], [19], [20], [21].”
MethodSpeech Enhancement Using Augmented SSL CycleGAN
Popovic, B. Z., Krstanović, L., Janev, M. B., Suzić, S., Nosek, T., & Galić, J. (2022) · European Signal Processing Conference (EUSIPCO) · doi:10.23919/eusipco55093.2022.9909754 · Cited by 5
Boosting StarGANs for Voice Conversion with Contrastive Discriminator
Si, S., Wang, J., Zhang, X., Qu, X., Cheng, N., & Xiao, J. Z. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-30108-7_30 · Cited by 5
Wang, Z., Wang, X., Xie, Q., Li, T., Xie, L., Tian, Q., & Wang, Y. (2023) · IEEE/ACM Transactions on Audio Speech and Language Processing · doi:10.1109/taslp.2023.3313414 · Cited by 7
“Tosolvethisproblem,inspiredbythetrainingprocessof CycleGAN[35],applyingtheexplicitconstraintmoduleallows ustodividethetrainingprocessintotwodifferentmodes,i.e., reconstruction mode , and simulation mode .”
BackgroundMethodInfluentialET-GAN: Cross-Language Emotion Transfer Based on Cycle-Consistent Generative Adversarial NetworksPreprint
Jia, X., Tai, J., Zhou, H., Li, Y., Zhang, W., Du, H., & Huang, Q. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1905.11173 · Cited by 6
“Particularly, voice conversion [20, 19, 21] is more related to emotion transfer.”
BackgroundMethodInfluentialZero-Shot Unseen Speaker Anonymization via Voice Conversion
Chang, H.-P., Yoo, I.-C., Jeong, C., & Yook, D. (2022) · IEEE Access · doi:10.1109/access.2022.3227963 · Cited by 7
“To increase the randomness of training data, we randomly cropped a segment of 128 frames from a randomly selected sentence instead of using an overall sentence directly, as in [15].”
BackgroundMethodWeng, S.-E., Shuai, H.-H., & Cheng, W.-H. (2023) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v37i11.26607 · Cited by 7
“A recent line for the voice conversion of non-parallel training data can be categorized mainly into autoencoder-based approaches (AE-based) (Qian et al. 2019; Chen et al. 2021; Lin et al. 2021) and generative-adversarial-based approaches (GAN-based) (Kaneko and Kameoka 2018; Kameoka et al. 2018).”
BackgroundMethodInfluentialMixRep: Hidden Representation Mixup for Low-Resource Speech Recognition
Xie, J., & Hansen, J. H. L. (2023) · Interspeech (USB) · doi:10.21437/interspeech.2023-1216 · Cited by 8
Aziz, S., & Shahnawazuddin, S. (2024) · Circuits Systems and Signal Processing · doi:10.1007/s00034-024-02598-1 · Cited by 4
Ghosh, S., Dhar, S., & Jana, N. D. (2024) · Communications in computer and information science · doi:10.1007/978-3-031-64070-4_19 · Cited by 5
Salim, S., & Ahmad, W. (2024) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2024.3484661 · Cited by 4
“A modified Cycle GAN VC training and conversion process for non-parallel voice conversion [11] was used.”
BackgroundMethodInfluentialU2-VC: one-shot voice conversion using two-level nested U-structure
Liu, F., Wang, H., Peng, R., Zheng, C., & Li, X. (2021) · EURASIP Journal on Audio Speech and Music Processing · doi:10.1186/s13636-021-00226-3 · Cited by 4
Speak Like a Dog: Human to Non-human creature Voice Conversion
Suzuki, K., Sakamoto, S., Taniguchi, T., & Kameoka, H. (2022) · Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) · doi:10.23919/apsipaasc55919.2022.9980306 · Cited by 6
Self-citationLong, Z., Zheng, Y., Yu, M., & Xin, J. X. (2022) · International Conference on Artificial Intelligence for Industries (AI4I) · doi:10.1109/ai4i54798.2022.00022 · Cited by 8
“Among them, CycleGAN-VC [9] (and the enhanced version [10]) is a GAN based model by configuring CycleGAN model (as widely used in image style transfer) with a gated CNN and cycle consistency loss.”
MethodWagner, D., Baumann, I., & Bocklet, T. (2025) · Lecture notes in computer science · doi:10.1007/978-3-032-02548-7_20 · Cited by 4
Li, Y., He, Z., Zhang, Y., & Yang, Z. (2020) · Journal of Circuits Systems and Computers · doi:10.1142/s0218126621501887 · Cited by 5
Phoneme-guided Dysarthric speech conversion With non-parallel data by joint training
Chen, X., Oshiro, A., Chen, J., Takashima, R., & Takiguchi, T. (2022) · Signal Image and Video Processing · doi:10.1007/s11760-021-02119-6 · Cited by 4
Mandarin Electro-Laryngeal Speech Enhancement Using Cycle-Consistent Generative Adversarial Networks
Qian, Z., Xiao, K., & Yu, C. (2022) · Applied Sciences · doi:10.3390/app13010537 · Cited by 4
UniSinger: Unified End-to-End Singing Voice Synthesis With Cross-Modality Information Matching
Hong, Z., Cui, C., Huang, R., Zhang, L., Liu, J., He, J., & Zhao, Z. (2023) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3581783.3612150 · Cited by 17
“There are many studies of representation disentanglement for voice conversion[1, 18, 19, 31, 32, 37] in text-to-speech (TTS) systems.”
BackgroundWav2wav: Wave-to-Wave Voice Conversion
Jeong, C., Chang, H.-P., Yoo, I.-C., & Yook, D. (2024) · Applied Sciences · doi:10.3390/app14104251 · Cited by 5
Significance of Data Augmentation for Improving Cleft Lip and Palate Speech RecognitionPreprint
Sudro, P. N., Das, R. K., Sinha, R., & Prasanna, S. R. M. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2110.00797 · Cited by 11
“The CycleGAN is widely used for non-parallel voice conversion (VC) and it has shown its effectiveness for various other applications [33]– [36].”
MethodSpeaker anonymization : representation, evaluation and formal guarantees
Srivastava, B. M. L. (2021) · theses.fr (ABES) · doi:10.70675/5dfccefaz18d3z4687z8a62za1359a91858b · Cited by 6
Zero-shot voice conversion based on feature disentanglement
Guo, N., Wei, J., Li, Y., Lu, W., & Tao, J. (2024) · Speech Communication · doi:10.1016/j.specom.2024.103143 · Cited by 4
Gaussian Flow Bridges for Audio Domain Transfer with Unpaired Data
Moliner, E., Braun, S., & Gamper, H. (2024) · International Workshop on Acoustic Signal Enhancement · doi:10.1109/iwaenc61483.2024.10693958 · Cited by 6
“In the context of audio, there have been several contributions in this direction, employing techniques such as mixture-invariant training [2], or various forms of generative adversarial networks [3, 4, 5].”
MethodKE: A Knowledge Enhancing Framework for Machine Learning Models
Wang, Y., Shah, N., Soliman, A. T., Guo, D., & Rajasekaran, S. (2023) · The Journal of Physical Chemistry A · doi:10.1021/acs.jpca.3c04076 · Cited by 3
Enhancing bone-conducted speech with spectrum similarity metric in adversarial learning
Pan, Y., Zhou, J., Wang, H., Zheng, W., Tao, L., & Kwan, H. K. (2025) · Speech Communication · doi:10.1016/j.specom.2025.103223 · Cited by 3
Gallard, A., Brebion, B., Sippel, K., Zaylaa, A., Preissl, H. T., Moghimi, S., Frégier, Y., & Wallois, F. (2025) · Neurophysiologie Clinique · doi:10.1016/j.neucli.2025.103086 · Cited by 3
Virtual Phone Discovery for Speech Synthesis Without Text
Nayak, S., Kumar, C. S., Ramesh, G., Bhati, S., & Murty, K. S. R. (2019) · IEEE Global Conference on Signal and Information Processing · doi:10.1109/globalsip45357.2019.8969412 · Cited by 3
Metadata
- Authors
- Kaneko, T., & Kameoka, H.
- Affiliations
- NTT (Japan) (JP)
- Venue
- 2018 26th European Signal Processing Conference (EUSIPCO)
- Publisher
- IEEE
- Published
- 2018-09
- Pages
- 2100-2104
- OpenAlex
- W2902070858
- Cited by
- 358
- References
- 61
- Citations by year
- 2026: 19 · 2025: 36 · 2024: 50 · 2023: 66 · 2022: 58 · 2021: 63 · 2020: 47 · 2019: 18 · 2018: 1
- FWCI
- 13.16
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- European Signal Processing Conference (EUSIPCO) (publishedVersion)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Speech Recognition and Synthesis
- Physical Sciences › Computer Science › Signal Processing › Speech and Audio Processing
- Health Sciences › Health Professions › Speech and Hearing › Voice and Speech Disorders
- Keywords
- voice conversion, non-parallel voice conversion, CycleGAN, gated convolutional neural network, pseudo data generation, over-smoothing, modulation spectra
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.