Multimodal Signal Processing (MSP) Podcast corpus
Resource type
Title
Multimodal Signal Processing (MSP) Podcast corpus
Abstract
The MSP-Podcast corpus contains speech segments from podcast recordings which are perceptually annotated using crowdsourcing. The collection of this corpus is an ongoing process. Version 1.11 of the corpus has 151,654 speaking turns (237 hours and 56 mins). The proposed partition attempts to create speaker-independent datasets for Train, Development, Test1, Test2, and Test3 sets.
Accessed
22/11/2024, 14:08
Citation
Multimodal Signal Processing (MSP) Podcast corpus. (n.d.). Retrieved November 22, 2024, from https://ecs.utdallas.edu/research/researchlabs/msp-lab/MSP-Podcast.html
Link to this record
Relations