CDPAM: Contrastive learning for perceptual audio similarity

TL;DR

This paper introduces CDPAM, a contrastive learning-based perceptual audio similarity metric that improves generalization and reduces data requirements compared to previous methods, aligning well with human judgments.

Contribution

It advances the DPAM approach by integrating contrastive learning and multi-dimensional representations, and incorporates human triplet judgments to enhance robustness across diverse audio perturbations.

Findings

CDPAM correlates strongly with human judgments across nine datasets.

Adding CDPAM improves speech synthesis and enhancement performance.

The method requires fewer annotations than previous approaches.

Abstract

Many speech processing methods based on deep learning require an automatic and differentiable audio metric for the loss function. The DPAM approach of Manocha et al. learns a full-reference metric trained directly on human judgments, and thus correlates well with human perception. However, it requires a large number of human annotations and does not generalize well outside the range of perturbations on which it was trained. This paper introduces CDPAM, a metric that builds on and advances DPAM. The primary improvement is to combine contrastive learning and multi-dimensional representations to build robust models from limited data. In addition, we collect human judgments on triplet comparisons to improve generalization to a broader range of audio perturbations. CDPAM correlates well with human responses across nine varied datasets. We also show that adding this metric to existing speech synthesis and enhancement methods yields significant improvement, as measured by objective and subjective tests.