Fréchet Audio Distance: A Reference-Free Metric for Evaluating Music Enhancement Algorithms

  title={Fr{\'e}chet Audio Distance: A Reference-Free Metric for Evaluating Music Enhancement Algorithms},
  author={K. Kilgour and Mauricio Zuluaga and Dominik Roblek and Matthew Sharifi},
We propose the Fréchet Audio Distance (FAD), a novel, reference-free evaluation metric for music enhancement algorithms. [...] Key Method FAD is validated using a wide variety of artificial distortions and is compared to the signal based metrics signal to distortion ratio (SDR), cosine distance, and magnitude L2 distance. We show that, with a correlation coefficient of 0.52, FAD correlates more closely with human perception than either SDR, cosine distance or magnitude L2 distance, with correlation coefficients…Expand
7 Citations
Audio Inpainting based on Self-similarity for Sound Source Separation Applications
  • Highly Influenced
Learning to Denoise Historical Music
  • 3
  • PDF
Conditioned Source Separation for Music Instrument Performances
  • 6
  • PDF
On Filter Generalization for Music Bandwidth Extension Using Deep Neural Networks
  • Serkan Sulun, M. Davies
  • Engineering, Computer Science
  • IEEE Journal of Selected Topics in Signal Processing
  • 2021
  • PDF
Perceiving Music Quality with GANs
  • PDF


Performance measurement in blind audio source separation
  • 2,184
  • PDF
SDR – Half-baked or Well Done?
  • 220
  • PDF
A short-time objective intelligibility measure for time-frequency weighted noisy speech
  • 358
  • PDF
Perceptual evaluation of speech quality (PESQ)-a new method for speech quality assessment of telephone networks and codecs
  • 1,604
  • PDF
Music Source Separation Using Stacked Hourglass Networks
  • 25
  • PDF
Supervised Speech Separation Based on Deep Learning: An Overview
  • D. Wang, J. Chen
  • Computer Science, Medicine
  • IEEE/ACM Transactions on Audio, Speech, and Language Processing
  • 2018
  • 517
  • PDF
MIR_EVAL: A Transparent Implementation of Common MIR Metrics
  • 256
  • PDF