# AutoCycle-VC: Towards Bottleneck-Independent Zero-Shot Cross-Lingual Voice Conversion

arXiv: https://arxiv.org/abs/2310.06546 · HTML: https://mlsanity.com/paper/2310.06546/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 9 of its 20 references that are in the map.
- First posted on arXiv 10 Oct 2023.

## Builds on (most cited first, 9 of 9)

- [Rethinking the Inception Architecture for Computer Vision](https://mlsanity.com/paper/1512.00567/) (2015)
- [HiFi-GAN: Generative Adversarial Networks for Efficient and High Fidelity Speech Synthesis](https://mlsanity.com/paper/2010.05646/) (2020)
- [Generalized End-to-End Loss for Speaker Verification](https://mlsanity.com/paper/1710.10467/) (2017)
- [AUTOVC: Zero-Shot Voice Style Transfer with Only Autoencoder Loss](https://mlsanity.com/paper/1905.05879/) (2019)
- [Unsupervised Speech Decomposition via Triple Information Bottleneck](https://mlsanity.com/paper/2004.11284/) (2020)
- [CycleGAN-VC2: Improved CycleGAN-based Non-parallel Voice Conversion](https://mlsanity.com/paper/1904.04631/) (2019)
- [AGAIN-VC: A One-shot Voice Conversion using Activation Guidance and Adaptive Instance Normalization](https://mlsanity.com/paper/2011.00316/) (2020)
- [VQVC+: One-Shot Voice Conversion by Vector Quantization and U-Net architecture](https://mlsanity.com/paper/2006.04154/) (2020)
- [MaskCycleGAN-VC: Learning Non-parallel Voice Conversion with Filling in Frames](https://mlsanity.com/paper/2102.12841/) (2021)

## Abstract (arXiv)

This paper proposes a simple and robust zero-shot voice conversion system with a cycle structure and mel-spectrogram pre-processing. Previous works suffer from information loss and poor synthesis quality due to their reliance on a carefully designed bottleneck structure. Moreover, models relying solely on self-reconstruction loss struggled with reproducing different speakers' voices. To address these issues, we suggested a cycle-consistency loss that considers conversion back and forth between target and source speakers. Additionally, stacked random-shuffled mel-spectrograms and a label smoothing method are utilized during speaker encoder training to extract a time-independent global speaker representation from speech, which is the key to a zero-shot conversion. Our model outperforms existing state-of-the-art results in both subjective and objective evaluations. Furthermore, it facilitates cross-lingual voice conversions and enhances the quality of synthesized speech.