LOCATA: Acoustic Source Localization and Tracking Challenge Data

A benchmark data corpus for the IEEE-AASP Challenge on Acoustic Source Localization and Tracking. The repository contains development and evaluation datasets recorded in an enclosed acoustic environment using four different microphone arrays and an optical tracking system. The data encompasses six scenarios ranging from single static sources to multiple moving talkers with moving arrays.

Use Cases

• Benchmarking localization algorithms based on the six defined challenge scenarios.
• Training models for multi-source localization using recordings from multiple microphone arrays.
• Developing tracking algorithms for moving talkers based on ground-truth optical tracking data.
• Evaluating direction-of-arrival estimation accuracy using the provided ground-truth annotations.

Strengths

• Includes data from four distinct microphone arrays: a 32-channel Eigenmike, a 15-channel DICIT array, a 12-channel robot head, and 4-channel hearing aids.
• Provides ground-truth annotations for all source and microphone positions from an OptiTrack infrared camera system.
• Encompasses six progressively challenging scenarios defined by the LOCATA challenge.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file size are unknown, which may limit suitability assessment.
• Last update date is unknown; freshness unverified.

Provenance

Source

Imperial College London

Collection Method

Recordings made in an enclosed acoustic environment using microphone arrays and an optical tracking system.