Document detail
ID
oai:arXiv.org:2402.17735
Topic
Electrical Engineering and Systems... Computer Science - SoundAuthor
Churchwell, Cameron Morrison, Max Pardo, BryanCategory
Computer Science
Year
2024
listing date
3/6/2024
Keywords
pronunciation speechAbstract
A phonetic posteriorgram (PPG) is a time-varying categorical distribution over acoustic units of speech (e.g., phonemes).
PPGs are a popular representation in speech generation due to their ability to disentangle pronunciation features from speaker identity, allowing accurate reconstruction of pronunciation (e.g., voice conversion) and coarse-grained pronunciation editing (e.g., foreign accent conversion).
In this paper, we demonstrably improve the quality of PPGs to produce a state-of-the-art interpretable PPG representation.
We train an off-the-shelf speech synthesizer using our PPG representation and show that high-quality PPGs yield independent control over pitch and pronunciation.
We further demonstrate novel uses of PPGs, such as an acoustic pronunciation distance and fine-grained pronunciation control.
;Comment: Accepted to ICASSP 2024 Workshop on Explainable Machine Learning for Speech and Audio
Churchwell, Cameron,Morrison, Max,Pardo, Bryan, 2024, High-Fidelity Neural Phonetic Posteriorgrams
Collaborative Participatory Research with LLM Agents in South Asia: An Empirically-Grounded Methodological Initiative and Agenda from Field Evidence in Sri Lanka
community south lanka methodological sri research
Plastic rearrangement of basal forebrain parvalbumin-immunoreactive neurons in the kainite model of epilepsy
± lobe temporal animals µm cholinergic cell neurons vs epilepsy basal forebrain
