# Correlation of Fr'echet Audio Distance With Human Perception of Environmental Audio Is Embedding Dependant

arXiv: https://arxiv.org/abs/2403.17508 · HTML: https://mlsanity.com/paper/2403.17508/

## Key facts

- Cited by 15 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 26 Mar 2024.

## Built on by (most cited first, 15 of 15)

- [Music2Latent: Consistency Autoencoders for Latent Audio Compression](https://mlsanity.com/paper/2408.06500/) (2024)
- [KAD: No More FAD! An Effective and Efficient Evaluation Metric for Audio Generation](https://mlsanity.com/paper/2502.15602/) (2025)
- [Video-Foley: Two-Stage Video-To-Sound Generation via Temporal Event Condition For Foley Sound](https://mlsanity.com/paper/2408.11915/) (2024)
- [Frechet Music Distance: A Metric For Generative Symbolic Music Evaluation](https://mlsanity.com/paper/2412.07948/) (2024)
- [Continuous Autoregressive Models with Noise Augmentation Avoid Error Accumulation](https://mlsanity.com/paper/2411.18447/) (2024)
- [CoDiCodec: Unifying Continuous and Discrete Compressed Representations of Audio](https://mlsanity.com/paper/2509.09836/) (2025)
- [Challenge on Sound Scene Synthesis: Evaluating Text-to-Audio Generation](https://mlsanity.com/paper/2410.17589/) (2024)
- [TA-V2A: Textually Assisted Video-to-Audio Generation](https://mlsanity.com/paper/2503.10700/) (2025)
- [Music2Latent2: Audio Compression with Summary Embeddings and Autoregressive Decoding](https://mlsanity.com/paper/2501.17578/) (2025)
- [MambaFoley: Foley Sound Generation using Selective State-Space Models](https://mlsanity.com/paper/2409.09162/) (2024)
- [Smooth-Foley: Creating Continuous Sound for Video-to-Audio Generation Under Semantic Guidance](https://mlsanity.com/paper/2412.18157/) (2024)
- [Contrastive Learning from Synthetic Audio Doppelg"angers](https://mlsanity.com/paper/2406.05923/) (2024)
- [The Name-Free Gap: Policy-Aware Stylistic Control in Music Generation](https://mlsanity.com/paper/2509.00654/) (2025)
- [Understanding Frechet Speech Distance for Synthetic Speech Quality Evaluation](https://mlsanity.com/paper/2601.21386/) (2026)
- [SCRAPL: Scattering Transform with Random Paths for Machine Learning](https://mlsanity.com/paper/2602.11145/) (2026)

## Abstract (arXiv)

This paper explores whether considering alternative domain-specific embeddings to calculate the Fr'echet Audio Distance (FAD) metric can help the FAD to correlate better with perceptual ratings of environmental sounds. We used embeddings from VGGish, PANNs, MS-CLAP, L-CLAP, and MERT, which are tailored for either music or environmental sound evaluation. The FAD scores were calculated for sounds from the DCASE 2023 Task 7 dataset. Using perceptual data from the same task, we find that PANNs-WGM-LogMel produces the best correlation between FAD scores and perceptual ratings of both audio quality and perceived fit with a Spearman correlation higher than 0.5. We also find that music-specific embeddings resulted in significantly lower results. Interestingly, VGGish, the embedding used for the original Fr'echet calculation, yielded a correlation below 0.1. These results underscore the critical importance of the choice of embedding for the FAD metric design.