CFMDCTCodec Demo Speech codec listening samples

Audio Demo

CFMDCTCodec: A Low-Bitrate Neural Speech Codec with Noise-Prior-aware Conditional Flow Matching for MDCT-Spectral Enhancement

Xiao-Hang Jiang, Yang Ai, Hui-Peng Du, Zhen-Hua Ling, and Ji Wu

Code is available at https://github.com/xhjiang1/CFMDCTCodec.

Figures

Architecture overview

Overview of the proposed CFMDCTCodec
Fig. 1

Overview of the proposed CFMDCTCodec.

MDCT spectral encoder and decoder structures
Fig. 2

Structures of the MDCT spectral encoder and decoder.

Conditional MDCT velocity field filter
Fig. 3

Structure of the conditional MDCT velocity-field filter.

Audio Samples

Listening samples