DNS-MOS

The 'DNS-MOS' dataset/benchmark contains perceptual quality scores used to evaluate the speech quality of text-to-speech systems.

Papers using DNS-MOS (6)

Towards Robust Generative Speech Enhancement Using Vector Quantisation-Based Neural Audio Codec2026Few-Shot and Pseudo-Label Guided Speech Quality Evaluation with Large Language Models2026When Fine-Tuning Fails and when it Generalises: Role of Data Diversity and Mixed Training in LLM-based TTS2026Reducing the Gap Between Pretrained Speech Enhancement and Recognition Models Using a Real Speech-Trained Bridging Module2025A Study on Zero-shot Non-intrusive Speech Assessment using Large Language Models2024Distil-DCCRN: A Small-footprint DCCRN Leveraging Feature-based Knowledge Distillation in Speech Enhancement2024