Neural Speech Synthesis on a Shoestring: Improving the Efficiency of Lpcnet

Valin, Jean‐Marc; Isik, Umut; Smaragdis, Paris; Krishnaswamy, Arvindh

doi:10.1109/icassp43922.2022.9746103

Cited by 3 publications

(1 citation statement)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…LPCNet [4] successfully harmonized the linear predictive coding (LPC) and WaveRNN [5] by training the WaveRNN module to predict the excitation of the speech, i.e., the residual of the linear prediction, rather than the raw audio samples. Consequently, the complexity of an LPCNet-based decoder is as low as around 3 GFLOPS with 30 MFLOPS for encoding [6,7] or lower [8]. Their low arithmetic and spatial complexity, however, come at the cost of suboptimal sound quality compared to the WaveNet decoder.…”

Section: Introductionmentioning

confidence: 99%

Personalized Neural Speech Codec

Jang,

Yang,

Lim

et al. 2024

ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

In this paper, we propose a personalized neural speech codec, envisioning that personalization can reduce the model complexity or improve perceptual speech quality. Despite the common usage of speech codecs where only a single talker is involved on each side of the communication, personalizing a codec for the specific user has rarely been explored in the literature. First, we assume speakers can be grouped into smaller subsets based on their perceptual similarity. Then, we also postulate that a group-specific codec can focus on the group's speech characteristics to improve its perceptual quality and computational efficiency. To this end, we first develop a Siamese network that learns the speaker embeddings from the LibriSpeech dataset, which are then grouped into underlying speaker clusters. Finally, we retrain the LPCNet-based speech codec baselines on each of the speaker clusters. Subjective listening tests show that the proposed personalization scheme introduces model compression while maintaining speech quality. In other words, with the same model complexity, personalized codecs produce better speech quality.

show abstract