metadata

license: mit
task_categories:
  - audio-classification
language:
  - en
tags:
  - MOS
  - CMOS
  - SMOS
  - synthetic
  - audio
  - speech
  - subjective
  - ratings
  - mos_prediction
pretty_name: TTSDS Listening Test Results
size_categories:
  - 10K<n<100K
configs:
  - config_name: default
    data_files:
      - split: test
        path: parquet/*

Listening Test Results for TTSDS2

This dataset contains all 11,000+ ratings collected for 20 synthetic speech systems for the TTSDS2 study (link coming soon). The scores are MOS (Mean Opinion Score), CMOS (Comparative Mean Opinion Score) and SMOS (Speaker Similarity Mean Opinion Score). All annotators included passed three attention checks throughout the survey.