Improving Perceptual Audio Aesthetic Assessment via Triplet Loss and Self-Supervised Embeddings

Open in new window