语音表征在朗读与自发语音中是否保留地区口音？
Do Speech Representations Preserve Regional Accent Across Read and Spontaneous Speech?
浏览论文内容
中文总结 AI 辅助
本研究通过RVG1数据集比较十种语音表征在朗读与自发语音间的地区口音识别，发现Whisper匹配条件下最优但迁移时显著退化，而说话者嵌入更稳健，表明匹配性能不代表稳健性。
中文摘要 AI 辅助
地区口音线索可以在匹配条件下被捕获，但尚不清楚它们在朗读语音和自发语音之间是否持续存在。我们研究了RVG1数据集，该数据集包含来自九个地区的500名德语说话者，在匹配条件下以及说话者独立的朗读-自发迁移条件下，比较了十种语音表征在地区分类和连续地理定位上的表现。Whisper在匹配条件下表现最佳，九分类UAR达到0.489，中位地理定位误差为148公里，但在跨迁移方向上UAR降至0.11/0.18，地理定位误差增至363公里。自监督模型表现出类似的退化，而说话者嵌入在域内区分性较弱，但在迁移下更为稳健。这种对比在分类和地理定位中一致。跨表征来看，稳健性与表征在风格间变化程度（风格不变性）相关，跨风格说话者检索是可解释的代理指标。年龄、性别、句子重叠和时长控制不能解释这一差距，尽管信道特征有所贡献。这些结果表明，强匹配条件下的性能并不代表稳健的地区信息。
英文摘要
Regional accent cues can be captured under matched conditions, but it remains unclear whether they persist between read and spontaneous speech. We study RVG1, with 500 German speakers from nine regions, comparing ten speech representations on regional classification and continuous geolocation under matched conditions and speaker-independent read--spontaneous transfer. Whisper performs best under matched conditions, reaching 0.489 nine-way UAR and 148 km median geolocation error, but drops to 0.11/0.18 UAR across transfer directions and 363 km geolocation error. Self-supervised models show a similar degradation, whereas speaker embeddings are less discriminative in-domain but more robust under transfer. This contrast is consistent across classification and geolocation. Across representations, robustness is associated with how little a representation shifts between styles (style-invariance), for which crossstyle speaker retrieval is an interpretable proxy. Age, sex, sentence-overlap, and duration controls do not account for the gap, although channel characteristics contribute. These results show that strong matched-condition performance does not indicate robust regional information.
发表机构
- Friedrich-Alexander-Universität Erlangen-Nürnberg（弗里德里希-亚历山大-埃尔朗根-纽伦堡大学）
- Leibniz Institute for Educational Trajectories (LIfBi)（莱布尼茨教育轨迹研究所）
- Otto-Friedrich-Universität Bamberg（奥托-弗里德里希-班贝格大学）
机构由 AI 辅助整理，请以论文原文为准。
