Across five text-to-music models, aesthetic predictor scores, pairwise human preferences, and reference-based distribution metrics produce inconsistent rankings, so the choice of evaluation metric changes the winner.
We examine reference-based evaluation metrics on the generated dataset computed in Section 4, offering a complementary perspective to subjective assessments
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
eess.AS 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
From Aesthetics to Human Preferences: Comparative Perspectives of Evaluating Text-to-Music Systems
Across five text-to-music models, aesthetic predictor scores, pairwise human preferences, and reference-based distribution metrics produce inconsistent rankings, so the choice of evaluation metric changes the winner.