arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2111.04040cs.SDcs.CLeess.AS
收藏

Meta-TTS:用于Few-Shot说话人自适应文语转换的元学习

Meta-TTS: Meta-Learning for Few-Shot Speaker Adaptive Text-to-Speech

National Taiwan University(台湾大学)

Sung-Feng Huang, Chyi-Jiunn Lin, Da-Rong Liu, Yi-Chen Chen, Hung-yi Lee

摘要:

个性化语音合成系统是非常期望的应用,其中系统可以利用用户的语音生成语音,而用户的语音具有罕见的登记记录。在最近的工作中,有两种主要的方法来建立这样的系统:说话人自适应和说话人编码。一方面,说话人自适应方法用很少的注册样本对训练好的多说话人文本到语音(TTS)模型进行微调。然而,它们需要至少数千个微调步骤才能实现高质量的适配,这使得它很难应用于设备。另一方面,说话人编码方法将登记话语编码到说话人嵌入中。训练好的TTS模型可以在相应的说话人嵌入的条件下合成用户的语音。然而,扬声器编码器受到可见扬声器和不可见扬声器之间的泛化间隙的影响。在本文中,我们提出了应用Meta学习算法的说话人自适应方法。更具体地说,我们使用模型不可知元学习(MAML)作为多说话者TTS模型的训练算法,其目的是找到一个很好的元初始化,以使模型快速适应任何Few-Shot说话者自适应任务。因此,我们也可以有效地将元训练的TTS模型适应于看不见的说话者。我们的实验比较所提出的方法(元TTS)与两个基线:一个扬声器自适应方法基线和扬声器编码方法基线。实验结果表明,在相同训练方案下,元文语转换系统能够以较少的训练步骤从少量的注册样本中合成出具有较高说话人相似度的语音,且性能优于说话人编码基线。当基线的扬声器编码器使用额外的8371个扬声器数据进行预训练时,Meta-TTS仍然可以在LibriTTS数据集上优于基线,并在VCTK数据集上获得相当的结果。

英文摘要:

Personalizing a speech synthesis system is a highly desired application, where the system can generate speech with the user's voice with rare enrolled recordings. There are two main approaches to build such a system in recent works: speaker adaptation and speaker encoding. On the one hand, speaker adaptation methods fine-tune a trained multi-speaker text-to-speech (TTS) model with few enrolled samples. However, they require at least thousands of fine-tuning steps for high-quality adaptation, making it hard to apply on devices. On the other hand, speaker encoding methods encode enrollment utterances into a speaker embedding. The trained TTS model can synthesize the user's speech conditioned on the corresponding speaker embedding. Nevertheless, the speaker encoder suffers from the generalization gap between the seen and unseen speakers. In this paper, we propose applying a meta-learning algorithm to the speaker adaptation method. More specifically, we use Model Agnostic Meta-Learning (MAML) as the training algorithm of a multi-speaker TTS model, which aims to find a great meta-initialization to adapt the model to any few-shot speaker adaptation tasks quickly. Therefore, we can also adapt the meta-trained TTS model to unseen speakers efficiently. Our experiments compare the proposed method (Meta-TTS) with two baselines: a speaker adaptation method baseline and a speaker encoding method baseline. The evaluation results show that Meta-TTS can synthesize high speaker-similarity speech from few enrollment samples with fewer adaptation steps than the speaker adaptation baseline and outperforms the speaker encoding baseline under the same training scheme. When the speaker encoder of the baseline is pre-trained with extra 8371 speakers of data, Meta-TTS can still outperform the baseline on LibriTTS dataset and achieve comparable results on VCTK dataset.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑