Synchronized Multi-Device Environmental Audio Dataset for Spatial and Propagation Analysis

Description Despite the availability of numerous environmental audio datasets, none fully satisfy the requirements for synchronized multi-device recording across varying distances. This dataset addresses that gap by providing audio captured simultaneously from four recording devices, all synchronized at the start of each recording session using a shared synchronization flag. Recordings were conducted across six representative acoustic scenes: jungle, café, machine/factory, street, and multi-audio-source environments. These scenes capture a wide range of real-world acoustic characteristics, from highly dynamic and layered soundscapes to more stationary and monotonic environments. Hybrid scenarios, such as overlapping speech with machinery (MultiTalk) and urban noise combined with conversations (MultiStreet), represent complex environments where multiple sound sources compete. The dataset is intended for research in environmental sound analysis, acoustic propagation, spatial audio modeling, and multi-device signal alignment. Data Acquisition and Synchronization Four audio recording devices were used simultaneously. All devices begin recording at the same time. A synchronization flag is embedded at the start of each recording to ensure precise temporal alignment across devices. This flag enables accurate cross-device comparison and post-processing synchronization. Recording Environments (Scenes) Jungle – Dense, layered natural soundscape Café – Human activity with background noise Machine/Factory – Predominantly mechanical and monotonic sounds Street – Traffic, voices, and urban ambience MultiTalk – Overlapping speech with machinery MultiStreet – Urban noise with concurrent conversations These environments collectively represent common real-world acoustic conditions and varying frequency and tonal characteristics. Audio Specifications Sampling Rate: 44,100 Hz Bit Depth: 16-bit Channels: Mono Format: WAV Recording Equipment: Standardized across all devices and scenes Post-Processing Amplitude standardization applied to normalize volume levels across recordings Segmentation into fixed-length clips of: 1 second 3 seconds 5 seconds Dataset Size Total Samples: 3,600 audio clips Number of Devices: 4 Number of Scenes: 6 Intended Use This dataset is suitable for: Environmental sound classification Multi-device audio synchronization research Acoustic propagation and spatial analysis Source separation and overlapping sound studies Robust audio modeling in real-world environments Limitations The dataset focuses on controlled synchronization and may not represent unsynchronized consumer-grade recordings. Scene diversity is broad but limited to six environment types. Keywords Environmental Audio, Multi-Device Recording, Synchronized Audio, Acoustic Propagation, Spatial Audio, Soundscape Analysis, Audio Dataset