Audio samples

Improving performance of real-time full-band blind packet-loss concealment with predictive network

Viet-Anh Nguyen1, Anh H. T. Nguyen1, and Andy W. H. Khong2
1Crystalsound Team, NamiTech JSC, Ho Chi Minh City, Vietnam
2Nanyang Technological University, Singapore
{vietanh.nguyen, anh.nguyen}@namitech.io, andykhong@ntu.edu.sg

Audio samples of our full-band recurrent network (FRN) versus TFGAN and tPLCNet for blind packet loss concealment (PLC)

Audio files are at 48 kHz sampling rate with packet size of 20 ms. Our FRN is a causal and blind PLC model while TFGAN is non-causal and tPLC is an informed PLC model.

Clean target Lossy input TFGAN tPLCNet FRN (Ours)