# CDPAM: Contrastive learning for perceptual audio similarity

**Authors:** Pranay Manocha, Zeyu Jin, Richard Zhang, Adam Finkelstein  
**arXiv:** [2102.05109](https://arxiv.org/abs/2102.05109)  
**Date:** 2021-02-09  
**Summary page:** https://www.summarizepaper.com/en/paper/2102.05109v1/

## Key points

- CDPAM is a contrastive learning approach for perceptual audio similarity
- It combines contrastive learning and multi-dimensional representations to build robust models from limited data
- Human judgments on triplet comparisons are collected to improve generalization to a broader range of audio perturbations
- CDPAM correlates well with human responses across nine varied datasets
- It is applied to speech synthesis and enhancement methods using the DEMUCS architecture-based speech denoiser
- Evaluation is performed on 500 randomly selected audio clips from the VCTK test set using objective measures (PESQ, STOI, CSIG) and subjective measures
- Adding CDPAM as a metric significantly improves the performance of existing speech synthesis and enhancement methods in both objective and subjective tests
- CDPAM captures perceptual audio similarity effectively and enhances various speech processing tasks
- It provides a valuable contribution to deep learning-based speech processing methods by introducing a robust metric that correlates well with human perception across different datasets.

## Summary

The paper introduces CDPAM, a contrastive learning approach for perceptual audio similarity. CDPAM builds on and advances the existing DPAM by combining contrastive learning and multi-dimensional representations to build robust models from limited data. In addition, the authors 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. To demonstrate its effectiveness, the authors apply it to speech synthesis and enhancement methods using the DEMUCS architecture-based speech denoiser. The evaluation is performed on 500 randomly selected audio clips from the VCTK test set using objective measures such as PESQ, STOI, CSIG, and subjective measures. The results show that adding CDPAM as a metric significantly improves the performance of existing speech synthesis and enhancement methods in both objective and subjective tests. This demonstrates the effectiveness of CDPAM in capturing perceptual audio similarity and its potential for improving various speech processing tasks. Overall, CDPAM provides a valuable contribution to deep learning-based speech processing methods by introducing a robust metric that correlates well with human perception across different datasets and enhances the performance of existing techniques in synthesizing and enhancing speech.