Skip to content

This repository contains the trained models and some audio samples for the tPLCnet.

License

Notifications You must be signed in to change notification settings

breizhn/tPLCnet

Repository files navigation

tPLCnet: Real-time Deep Packet Loss Concealment in the Time Domain Using a Short Temporal Context

Interspeech Paper:

The paper can be found here.

When using the models please cite:

@inproceedings{westhausen22_interspeech,
  author={Nils L. Westhausen and Bernd T. Meyer},
  title={{tPLCnet: Real-time Deep Packet Loss Concealment in the Time Domain Using a Short Temporal Context}},
  year=2022,
  booktitle={Proc. Interspeech 2022},
  pages={2903--2907},
  doi={10.21437/Interspeech.2022-10157}
}

Audio Samples:

If you are looking at the README.md directly on GitHub, go this webpage to listen to the audio samples.


4607.wav

Corrupted spectrogram 4607.wav

Concealed spectrogram 4607.wav

Clean spectrogram 4607.wav


2307.wav

Corrupted spectrogram 2307.wav

Concealed spectrogram 2307.wav

Clean spectrogram 2307.wav


54.wav

Corrupted spectrogram 54.wav

Concealed spectrogram 54.wav

Clean spectrogram 54.wav


Video Presentation:

Releases

No releases published

Packages

No packages published

Languages