EMO-Codec: An In-Depth Look at Emotion Preservation capacity of Legacy and Neural Codec Models With Subjective and Objective Evaluations

TL;DR

This study evaluates how well neural and legacy speech codecs preserve emotional information using subjective and objective assessments, revealing limitations in emotion retention, especially across languages and certain emotions.

Contribution

It provides a comprehensive analysis of emotion preservation in codecs, highlighting the impact of training data and resynthesis on emotional content retention.

Findings

Neural codecs show limited emotion preservation, especially in Chinese.

Resynthesis degrades speech emotion recognition accuracy.

Human tests confirm emotion loss in codec processing.

Abstract

The neural codec model reduces speech data transmission delay and serves as the foundational tokenizer for speech language models (speech LMs). Preserving emotional information in codecs is crucial for effective communication and context understanding. However, there is a lack of studies on emotion loss in existing codecs. This paper evaluates neural and legacy codecs using subjective and objective methods on emotion datasets like IEMOCAP. Our study identifies which codecs best preserve emotional information under various bitrate scenarios. We found that training codec models with both English and Chinese data had limited success in retaining emotional information in Chinese. Additionally, resynthesizing speech through these codecs degrades the performance of speech emotion recognition (SER), particularly for emotions like sadness, depression, fear, and disgust. Human listening tests confirmed these findings. This work guides future speech technology developments to ensure new codecs maintain the integrity of emotional information in speech.