Meta-AF Demos

Meta-Learning for Adaptive Filters

We release the inputs, targets, and results for five samples in the test set of each adaptive filter task for all models. These samples are all generated using the Meta-AF codebase in this GitHub repo. Please consult our paper for any experimental configuration details as well as descriptions of our baselines.

Acoustic Echo Cancellation

Single-Talk
Magnitude STFT of the First Sample* drawing
*STOI is zero since the near-end speech is silent.
Near-end Far-end Near-end Speech Speex LMS RMSProp NLMS RLS Kalman Filter Meta-AF
Double-Talk
Magnitude STFT of the First Sample drawing
Near-end Far-end Near-end Speech Speex LMS RMSProp NLMS RLS Kalman Filter Meta-AF
Double-Talk with Path Change
Magnitude STFT of the First Sample drawing
Near-end Far-end Near-end Speech Speex LMS RMSProp NLMS RLS Kalman Filter Meta-AF
Double-Talk with Nonlinearities
Magnitude STFT of the First Sample drawing
Near-end Far-end Near-end Speech Speex LMS RMSProp NLMS RLS Kalman Filter Meta-AF

Equalization

Unconstrained
Magnitude STFT of the First Sample drawing
Target Input LMS RMSProp NLMS RLS Meta-AF
Constrained
Magnitude STFT of the First Sample drawing
Target Input LMS RMSProp NLMS RLS Meta-AF

Dereverberation

One, Four, Eight Mic.
Magnitude STFT of the First Sample drawing
Reverberant Anechoic NARA 1 Mic. Meta-AF 1 Mic. NARA 4 Mic. Meta-AF 4 Mic. NARA 8 Mic. Meta-AF 8 Mic.

Beamforming

Diffuse Interferer
Magnitude STFT of the First Sample drawing
Clean Speech Mixture LMS RMSProp NLMS RLS Meta-AF
Directional Interferer
Magnitude STFT of the First Sample drawing
Clean Speech Mixture LMS RMSProp NLMS RLS Meta-AF

Note that all samples have been scaled to [-1,1] and saved as .mp3 files for playback. If you would like the raw files, please either follow the directions in the GitHub repository or contact me.