Underwater imaging without color distortions requires RAW capture

In a new paper published on arXiv, researchers Derya Akkaynak and Michael S. Brown present a critical finding for aquatic sciences: consumer cameras capturing JPEGs are producing scientifically useless data for color-based research. While affordable cameras are ubiquitous for ecosystem surveys, monitoring, and animal behavior studies, the authors argue that when color is the variable of interest—such as in coral-bleaching research—JPEG format renders most images quantitatively unusable. The core problem is not JPEG compression itself, but the irreversible in-camera processing algorithms that modify colors and contrast to create visually pleasing pictures, not color-accurate data.

This processing severs the linear relationship between pixel values and the actual scene radiance, making it impossible to standardize, reproduce, or compare colors across different cameras, locations, or over time. This represents a massive, hidden data loss in fields relying on visual surveys. The paper serves as both a warning and a guide, urging scientists to shift their fundamental data collection practice. To preserve scientific integrity, the authors offer pragmatic guidance, beginning with the capture and archiving of minimally processed RAW images, which retain the necessary linear data for accurate, reproducible color analysis.