The study focuses on the Colour Rendering Index, properly called the CIE R index after the International Commission on Illumination. The index reaches its maximum of 100 in daylight and incandescent light and expresses how faithfully a source renders standard colours. The Federal Agencies Digital Guidelines Initiative recommends a value above 90 for cultural heritage imaging. Brehm deliberately used the stricter CIE R e index, which is based on the matching of sixteen standard colours rather than the eight colours of the more common R a index. Spectral measurements were carried out with a JETI Blue Spec Cube spectroradiometer between 2019 and 2024 in Jena, Berlin and London, covering not only insect digitization devices but also ceiling lights and exhibition lighting, since good light seems desirable in every working area of a museum.

The results reveal deficits that are astonishing for institutions spending millions on digitization. No light source measured in the museums of London and Berlin achieved an R e index value above 90. Compared with daylight, which scores 98 to 99, the mercury vapour fluorescent tubes still widely used in reproduction systems and light boxes reached R e values of only 65 to 84, with the SatScan device in London scoring just 65. Xenon light sources, including the continuous arc lamps whose spectrum matches commercial flash units, performed better with a balanced spectrum and an R e of 87. Existing LED systems varied enormously, with R e values between 45 and 82 and sometimes very pronounced blue peaks and spectral inequalities. The Supra Quartz Scanner at the Natural History Museum in London and the Leica Z16 microscope in Berlin were particularly poor, combining strong blue overemphasis with underrepresented turquoise and red. Even ceiling lighting in the London workrooms, some of which have no windows at all, scored below 70.

Against this backdrop, modern daylight LEDs emerge as the clear winners. In the three new devices presented in the paper, they achieved R e values of 97 to 98, essentially matching natural light. Daylight LEDs were still a niche product around 2020 but are now easy and inexpensive to obtain, and manufacturer information indicates that LEDs with R e values above 90 or even 95 are readily available. Brehm recommends not relying on manufacturer specifications alone but carrying out spectral measurements when in doubt, because light quality cannot be assessed by eye. Converting existing devices to better LEDs should generally be possible, and the additional cost is insignificant compared with the higher image quality achieved.

The three lighting devices differ fundamentally in shape but share a basic principle: illuminating the object from many sides by scattering light from the inside, mimicking the perfect shadow-free conditions of an overcast day. The light cylinder, a zinc body 45 centimetres in diameter and 35 centimetres high, was built by a locksmith and serves for photographing individual moths and other insects from above and below. Eight LinearZ elements with Nichia daylight LEDs are glued to the underside of the metal lid, which doubles as a heat sink, and a suspended polyoxymethylene disc diffuses the light to avoid cast shadows. Objects rest on roughened aluminium plates and are moved by a laboratory lifting platform, while a 10-millimetre scale with reflectance standards is brought into the image independently. Three xenon flash units with their covers removed provide indirect UV illumination through the ceiling.

The light box was built from an aluminium profile system with reflective, honeycomb-structured scattering plates and is designed for photographing entire standard-format insect drawers of 51 by 42 centimetres. A Sony Alpha 7R mark IV with 61 megapixels hangs 124 centimetres above the drawer on a repro stand. The box enables evenly bright, shadow-free illumination with a maximum deviation of measured irradiance of just 4.7 percent, and in routine operation around 25 to 30 drawers can be photographed per hour. The resolution of roughly 150 pixels per centimetre allows small-print labels to be read, a capability that was outside the reach of consumer-grade cameras only a decade ago. The light hemisphere, the most inexpensive of the three, is a polystyrene hemisphere covered with aluminium foil, optimized for photographing ethanol samples whose liquid surfaces produce strong reflections under direct light. Shielding the direct LED radiation with foil enables low-reflection photography, and a sample takes about two minutes to record.

Quantitative uniformity measurements in all three devices confirmed their objectives. Deviations of irradiance across the imaging area remained below 5 percent, with the light cylinder achieving deviations as low as 1.1 percent at larger image scales. The paper also compares background materials, showing that ordinary printer paper treated with optical brighteners and matt PVC film absorb or transform part of the UV radiation and therefore appear dark under UV light, while plastics such as polyethylene and metal surfaces such as aluminium remain nearly colour-neutral across the UV and visible spectrum. Aluminium was chosen for its darker surface, although its brightness changes slightly with working distance. The study recommends manual camera settings, low ISO values, raw format archiving, standardized grey cards for exposure and a uniform 10-millimetre scale in a consistent position to facilitate automated evaluation by algorithms.

Perhaps the most provocative argument of the paper concerns the ultraviolet range. Digital colour information in digitization projects is recorded practically 100 percent in the RGB colour space, which is tailored to human vision. Brehm calls this view anthropocentric and offers a comparison: it would seem absurd to document the vocalizations of grasshoppers, bats and elephants only in the sound frequencies humans can hear, ignoring ultrasound and infrasound that are biologically essential. The same logic applies to vision. Most arthropods and many vertebrate groups, including most birds, perceive ultraviolet light, whereas humans cannot. A modified camera with its visible-pass filter removed has spectral sensitivities close to those of an insect eye or a bird’s eye, and false-colour images can recreate the viewpoints of these organisms. In male brimstone butterflies and many copper butterfly species, UV photography reveals patterns that are completely invisible to humans.

The study shows that the deficits in light quality found in insect photography very probably apply to other areas of biology and beyond. Digitization programs should fundamentally check the quality of their lighting and install better light sources where necessary, since spectral measurements provide final certainty that visual assessment cannot. Daylight LEDs represent a long-lasting, ecologically acceptable and high-quality alternative to existing systems with poor light quality. Capturing colour patterns in the UV range documents a biologically essential component of the electromagnetic spectrum and should play a far greater role in digitization programs and ecological analyses than it has until now. Future studies should examine whether and how poor light quality affects quantitative results such as lightness measurements, and whether efficient tools for evaluating UV data can be developed to move UV imaging beyond its current status as a tinkering task.

Subject of Research: Lighting quality in standardized insect photography and digitization

Article Title: Light in standardised insect photography and description of lighting devices, including the UV range

Article References: Brehm, G. (2025). Light in standardised insect photography and description of lighting devices, including the UV range. Nota Lepidopterologica, 48, 145-164. https://doi.org/10.3897/nl.48.145444

Image Credits: AI Generated

DOI: 10.3897/nl.48.145444

Keywords: insect photography, digitization, colour rendering index, daylight LEDs, UV photography, natural history museums, entomology, collection imaging, Light, standardised, insect, photography