Search papers, labs, and topics across Lattice.
This study explores the predictive power of contextualized embeddings (CEs) for spoken word duration and pitch in Mandarin monosyllabic words, analyzing 7,470 tokens from spontaneous speech. The findings reveal that CEs can predict spoken word duration with a significant degree of accuracy, both at the type and individual token levels, surpassing chance performance and permutation baselines. Additionally, the predicted f0 contours, derived from these durations, closely approximate empirical data, demonstrating the effectiveness of CEs in modeling prosodic features in Mandarin.
Contextualized embeddings can accurately predict spoken word duration in Mandarin, revealing a deeper link between linguistic representation and prosody.
Time-normalized f0 contours of Mandarin words in conversational speech have been shown to be predictable in part from their contextualized embeddings (CEs). The present study investigates whether CEs also predict spoken word duration for 7470 tokens of Mandarin monosyllabic CV words extracted from a Mandarin corpus of spontaneous speech. We show that CEs indeed are predictive for duration, above chance level, not only at the type level, but also at the level of individual tokens, as indicated by the results obtained with the type-wise and token-wise permutation baselines. We also show that the predicted durations are sufficiently precise to back-transform predicted f0 contours in [0,1] normalized time to contours on the ms time scale. The resulting predicted contours approximate empirical contours and also outperform a permutation baseline.
