arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2110.02635eess.AS
收藏

MOS预测网络的泛化能力

Generalization Ability of MOS Prediction Networks

Erica Cooper, Wen-Chin Huang, Tomoki Toda, Junichi Yamagishi

摘要:

自动方法来预测合成语音的听众意见仍然难以捉摸,因为听众,被评估的系统,语音的特性,甚至给出的指令和评级尺度都随着测试的不同而变化。虽然自动预测指标,如平均意见分数(MOS)可以实现高预测精度的样本从同一个测试,他们通常无法推广到新的听力测试环境。在本文中,使用各种MOS预测网络,包括MOSNet和自监督语音模型,如wav2vec2,我们研究了它们在zero-shot和微调设置下对不同听力测试数据的性能。我们发现,针对MOS预测微调的wav2vec2模型对域外数据具有良好的泛化能力,即使是在zero-shot设置中最具挑战性的话语级预测情况下,对域内数据进行微调可以改善预测。我们还观察到,看不见的系统是特别具有挑战性的MOS预测模型。

英文摘要:

Automatic methods to predict listener opinions of synthesized speech remain elusive since listeners, systems being evaluated, characteristics of the speech, and even the instructions given and the rating scale all vary from test to test. While automatic predictors for metrics such as mean opinion score (MOS) can achieve high prediction accuracy on samples from the same test, they typically fail to generalize well to new listening test contexts. In this paper, using a variety of networks for MOS prediction including MOSNet and self-supervised speech models such as wav2vec2, we investigate their performance on data from different listening tests in both zero-shot and fine-tuned settings. We find that wav2vec2 models fine-tuned for MOS prediction have good generalization capability to out-of-domain data even for the most challenging case of utterance-level predictions in the zero-shot setting, and that fine-tuning to in-domain data can improve predictions. We also observe that unseen systems are especially challenging for MOS prediction models.

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。

↑