Descriptor: Certus Caliber Classification Gunshot Dataset (C3GD)

A publicly accessible dataset of firearm muzzle blast sounds, containing more than 8,000 field-collected recordings from 28 firearms across 16 calibers, has been introduced on the preprint repository arXiv [1][2]. The Certus Caliber Classification Gunshot Dataset, or C3GD, was described in a paper submitted June 16 to arXiv, an open-access repository for electronic preprints that has hosted more than two million articles since its founding in 1991 [6][1]. The authors state that the dataset is designed to supply metadata beyond what is currently available, covering a range of firearms, cartridges, microphones, and microphone positions [2]. Field collection of gunshot audio is expensive, so prior research has often relied on recordings pulled from the internet, a practice the paper’s authors say increases the risk of low-quality data and label noise [2]. By gathering samples in controlled outdoor conditions, the team aimed to produce a reference that can support caliber classification as well as gunshot detection, audio separation, and broader audio signal-processing tasks [2]. The paper appears under arXiv’s “Sound” category within computer science and is accessible through the repository’s standard abstract page, which also features a series of community-built tools known as arXivLabs [1][4]. arXivLabs, launched in 2020, provides a framework for third-party developers to integrate experimental features—such as citation explorers and code-finding services—directly on article pages, under guidelines that require adherence to user-data privacy and openness principles [4][5]. While the C3GD paper itself does not include peer review—consistent with arXiv’s moderation-but-not-peer-review model—the dataset’s scale and metadata detail may offer a new baseline for researchers working on real-world firearm audio analysis [6][2]. The authors note that the collection’s diversity is intended to help models generalize beyond laboratory conditions [2].