2022/11/01 by Nikitaras, Karolos, Klapsas, Konstantinos, Ellinas, Nikolaos +6
#Audio and Speech Processing (eess.AS) #Computation and Language (cs.CL) #FOS: Computer and information sciences #FOS: Electrical engineering #Machine Learning (cs.LG) #Sound (cs.SD) #electronic engineering #information engineering
paper · doi:10.48550/arxiv.2211.00523
This paper proposes an Expressive Speech Synthesis model that utilizes token-level latent prosodic variables in order to capture and control utterance-level attributes, such as character acting voice and speaking style. Current works aim to explicitly factorize such fine-grained and utterance-level speech attributes into different representations extracted by modules that operate in the corresponding level. We show that the fine-grained latent space also captures coarse-grained information, which is more evident as the dimension of latent space increases in order to capture diverse prosodic representations. Therefore, a trade-off arises between the diversity of the token-level and utterance-level representations and their disentanglement. We alleviate this issue by first capturing rich speech attributes into a token-level latent space and then, separately train a prior network that given the input text, learns utterance-level representations in order to predict the phoneme-level, posterior latents extracted during the previous step. Both qualitative and quantitative evaluations are used to demonstrate the effectiveness of the proposed approach. Audio samples are available in our demo page.