2018/09/12 by Suwon Shon, Hao Tang, Shon, Suwon +3 · 2 citations
Computer Science · #Audio and Speech Processing (eess.AS) #FOS: Computer and information sciences #FOS: Electrical engineering #Machine Learning (cs.LG) #Music and Audio Processing #Speech Recognition and Synthesis #Speech and Audio Processing #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.1809.04437
openalex publication_date 2018/09/12 · openalex created_date 2022/08/03 · openalex updated_date 2026/07/28
In this paper, we propose a Convolutional Neural Network (CNN) based speaker\nrecognition model for extracting robust speaker embeddings. The embedding can\nbe extracted efficiently with linear activation in the embedding layer. To\nunderstand how the speaker recognition model operates with text-independent\ninput, we modify the structure to extract frame-level speaker embeddings from\neach hidden layer. We feed utterances from the TIMIT dataset to the trained\nnetwork and use several proxy tasks to study the networks ability to represent\nspeech input and differentiate voice identity. We found that the networks are\nbetter at discriminating broad phonetic classes than individual phonemes. In\nparticular, frame-level embeddings that belong to the same phonetic classes are\nsimilar (based on cosine distance) for the same speaker. The frame level\nrepresentation also allows us to analyze the networks at the frame level, and\nhas the potential for other analyses to improve speaker recognition.\n