T2A-Feedback: Improving Basic Capabilities of Text-to-Audio Generation via Fine-grained AI Feedback

2025-05-15

AI Summary

Current text-to-audio (T2A) models struggle to generate high-fidelity, multi-event audio sequences that both faithfully follow complex prompts and satisfy human perceptual preferences. Method: We propose the first fine-grained AI feedback-based learning framework for T2A, introducing an automated three-dimensional scoring system assessing event existence, temporal ordering, and acoustic harmony. We release T2A-FeedBack—the first large-scale T2A preference dataset (41K prompts, 249K audios)—and T2A-EpicBench, a long-narrative-oriented evaluation benchmark. We further design an AI-driven preference optimization method. Results: Our approach significantly improves state-of-the-art models on AudioCaps and T2A-EpicBench. The three AI-derived scores achieve strong correlation with human preferences (Spearman’s ρ > 0.82), substantially outperforming conventional metrics. This work establishes a new paradigm for principled, human-aligned evaluation and optimization of T2A systems.

Technology Category

• Humans and AI: Learning Human Values and Preferences
• Machine Learning: Learning Preferences or Rankings
• Natural Language Processing: Generation

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Text-to-audio (T2A) generation has achieved remarkable progress in generating a variety of audio outputs from language prompts. However, current state-of-the-art T2A models still struggle to satisfy human preferences for prompt-following and acoustic quality when generating complex multi-event audio. To improve the performance of the model in these high-level applications, we propose to enhance the basic capabilities of the model with AI feedback learning. First, we introduce fine-grained AI audio scoring pipelines to: 1) verify whether each event in the text prompt is present in the audio (Event Occurrence Score), 2) detect deviations in event sequences from the language description (Event Sequence Score), and 3) assess the overall acoustic and harmonic quality of the generated audio (Acoustic&Harmonic Quality). We evaluate these three automatic scoring pipelines and find that they correlate significantly better with human preferences than other evaluation metrics. This highlights their value as both feedback signals and evaluation metrics. Utilizing our robust scoring pipelines, we construct a large audio preference dataset, T2A-FeedBack, which contains 41k prompts and 249k audios, each accompanied by detailed scores. Moreover, we introduce T2A-EpicBench, a benchmark that focuses on long captions, multi-events, and story-telling scenarios, aiming to evaluate the advanced capabilities of T2A models. Finally, we demonstrate how T2A-FeedBack can enhance current state-of-the-art audio model. With simple preference tuning, the audio generation model exhibits significant improvements in both simple (AudioCaps test set) and complex (T2A-EpicBench) scenarios.

Problem

• Improving prompt-following in multi-event audio generation
• Enhancing acoustic quality via AI feedback learning
• Addressing human preference gaps in T2A model outputs

Innovation

• Fine-grained AI audio scoring pipelines
• Large audio preference dataset T2A-FeedBack
• T2A-EpicBench benchmark for advanced evaluation