The use of Voice Source Features for Sung Speech Recognition

Abstract

In this paper, we ask whether vocal source features (pitch, shimmer, jitter, etc) can improve the performance of automatic sung speech recognition, arguing that conclusions previously drawn from spoken speech studies may not be valid in the sung speech domain. We first use a parallel singing/speaking corpus (NUS-48E) to illustrate differences in sung vs spoken voicing characteristics including pitch range, syllables duration, vibrato, jitter and shimmer. We then use this analysis to inform speech recognition experiments on the sung speech DSing corpus, using a state of the art acoustic model and augmenting conventional features with various voice source parameters. Experiments are run with three standard (increasingly large) training sets, DSing1 (15.1 hours), DSing3 (44.7 hours) and DS-ing30 (149.1 hours). Pitch combined with degree of voicing produces a significant decrease in WER from 38.1% to 36.7% when training with DSing1 however smaller decreases in WER observed when training with the larger more varied DSing3 and DSing30 sets were not seen to be statistically significant. Voicing quality characteristics did not improve recognition performance although analysis suggests that they do contribute to an improved discrimination between voiced/unvoiced phoneme pairs.

Investigating self-supervised learning for lyrics recognition

“…Test set GMM + 4grams [6] 52.95 49.50 TDNN-F + 3grams [6] 26.24 22.32 TDNN-F + 4-gram [6] 23.33 19.60 Kaldi LN + VQ + 3grams [7] No Value 22.97 Kaldi LN + VQ + 4grams [7] No Value 19.60 However, for the other SSL models, the best performance they achieved was by the transformer-based model. In previous studies, some researchers show that the objective of the SSL model has a higher impact on representation similarity than the model architecture [25].…”

“…This is motivated by the fact that spoken and sung speech have the same production system and that semantic information is conveyed in the same way in both speech styles. [7]. However, there are several different acoustic features between sung and spoken speech, such as the pitch ranges, the syllable dura-tion, and the existence of vibrato in singing [7].…”

“…[7]. However, there are several different acoustic features between sung and spoken speech, such as the pitch ranges, the syllable dura-tion, and the existence of vibrato in singing [7]. These differences make sung speech difficult to recognize.…”