Enhancing Soundscape Characterization and Pattern Analysis Using Low-Dimensional Deep Embeddings on a Large-Scale Dataset

Soundscape monitoring has become an increasingly important tool for studying ecological processes and supporting habitat conservation. While many recent advances focus on identifying species through supervised learning, there is growing interest in understanding the soundscape as a whole considering patterns that go beyond individual vocalizations. This broader view requires unsupervised approaches capable of capturing meaningful structures related to temporal dynamics, frequency content, spatial distribution, and ecological variability. In this study, we present a fully unsupervised framework for analyzing large-scale soundscape data using deep learning. We applied a convolutional autoencoder (Soundscape-Net) to extract acoustic representations from over 60,000 recordings collected across a grid-based sampling design in the Rey Zamuro Reserve, Colombia. Dimensionality reduction methods (UMAP and PaCMAP) were used to project the learned features, followed by clustering with KMeans and DBSCAN to explore latent acoustic structures. To interpret and validate the resulting clusters, we combined multiple strategies: spatial mapping through interpolation, analysis of acoustic index variance to understand cluster structure, and graph-based connectivity analysis to identify ecological relationships between recording sites. Our results demonstrate that this approach can uncover both local and broad-scale patterns in the soundscape, providing a flexible and interpretable pathway for unsupervised ecological monitoring.