LongAudio-RAG: Event-Grounded Question Answering over Multi-Hour Long Audio

TL;DR

LongAudio-RAG is a hybrid system that improves long-audio question answering by grounding language models in timestamped acoustic events, enabling precise, efficient, and scalable answers for multi-hour recordings.

Contribution

We introduce a novel hybrid framework that grounds LLM outputs in timestamped acoustic events stored in a database, enhancing long-audio question answering accuracy and efficiency.

Findings

Structured event retrieval outperforms vanilla RAG and text-to-SQL methods.

The system achieves low-latency event extraction on edge devices.

Our benchmark demonstrates effective question answering over multi-hour audio.

Abstract

Long-duration audio is increasingly common in industrial and consumer settings, yet reviewing multi-hour recordings is impractical, motivating systems that answer natural-language queries with precise temporal grounding and minimal hallucination. Existing audio-language models show promise, but long-audio question answering remains difficult due to context-length limits. We introduce LongAudio-RAG (LA-RAG), a hybrid framework that grounds Large Language Model outputs in retrieved, timestamped acoustic event detections rather than raw audio. Multi-hour streams are converted into structured event records stored in an SQL database, and at inference time the system resolves natural-language time references, classifies intent, retrieves only the relevant events, and generates answers using this constrained evidence. To evaluate performance, we construct a synthetic long-audio benchmark by concatenating recordings with preserved timestamps and generating template-based question-answer pairs for detection, counting, and summarization tasks. Finally, we demonstrate the practicality of our approach by deploying it in a hybrid edge-cloud environment, where the audio grounding model runs on-device on IoT-class hardware while the LLM is hosted on a GPU-backed server. This architecture enables low-latency event extraction at the edge and high-quality language reasoning in the cloud. Experiments show that structured, event-level retrieval significantly improves accuracy compared to vanilla Retrieval-Augmented Generation (RAG) or text-to-SQL approaches.