Quantifying Internal Reflections and Hotspot Severity Across Modern Mirrorless Lenses in 850nm Infrared

Why Does a Clean Daylight Lens Hotspot at 850nm?

Which anti-reflective coatings that look invisible in daylight still throw a central hotspot and eat micro-contrast once the exposure is limited to 850nm, and how does aperture change that severity?

A color photograph can look clean while the infrared frame from the same lens places a pale disk over central foliage or stone. The surprise often arrives after a converted-body outing, when careful focus cannot explain why texture looks weak in the middle. Darkening that patch during processing may restore its brightness relationship with the surroundings. It cannot recover detail separation that the capture has lost.

The optical stakes are internal reflection, veiling glare, and a local reduction in micro-contrast. Aperture and focal length can alter their appearance. A daylight inspection of the glass cannot establish which coating or interface will cause trouble in the converted camera.

I record two observations separately: central brightness and fine-detail separation. That distinction guides the chart test, because a brighter center can still retain clearly separated texture.

The Surfaces That Return an 850nm Ray

Anti-reflective coatings reduce reflection through wavelength-dependent interference. A multilayer coating designed for visible light can have higher residual reflectance near 850nm, even when the glass appears neutral in daylight. Its infrared behavior belongs to the complete optical assembly: lens, external filter, and converted sensor stack.

Consider a possible round trip. Light passes through the external filter and lens, reaches the sensor assembly, and reflects back toward a rear lens surface. A second reflection sends some of that light toward the sensor again. The sensor cover glass and microlens structures participate in this environment; the conversion has already removed or replaced the original internal hot-mirror.

Other paths can involve the external filter, front-element coating, and internal air-glass interfaces. Their spacing and curvature influence where returning light lands.

Read the Filter Specification First

The 850nm label alone does not define the transmitted spectrum. A long-pass filter admits wavelengths beyond its transition; a true bandpass filter has lower and upper transmission boundaries. Check which filter you own before describing a test as limited to an 850nm band.

A compact central disk suggests a localized ghost, while a broad veil describes a more distributed contrast wash. These appearances do not reliably identify how many surfaces contributed. Keep any explanation tied to the tested body, filter, lens copy, and lighting geometry.

How a Flat Chart Separates Disk from Veil

I begin with a matte, finely textured gray target rather than a scenic subject. It gives the center and surrounding regions comparable detail, so the inspection can separate brightness distribution from texture separation. An evenly lit interior wall can serve if its surface carries enough fine texture.

Square the camera to the target and fill the frame. Use illumination with sufficient near-infrared output, then check its distribution through the filter. A wall that looks evenly illuminated to the eye can show a different pattern in infrared. Establish this before interpreting a brighter center as an optical fault.

1. Set manual exposure and fixed ISO. Focus on the target and retain that focus through the aperture sequence.
2. Record the hood state and filter assembly. Adjust shutter time as the aperture changes.
3. Apply identical raw conversion, white balance, sharpening, noise reduction, and lens-correction settings.
4. Inspect the whole frame, then compare matching detail crops from the center and surrounding target.

First ask whether a central brightness excess appears beyond ordinary corner falloff. Then ask whether adjacent light and dark details remain separable inside it. A bright patch with intact texture differs from a patch where fine tonal differences merge.

This is a relative chart comparison within one lighting geometry. It supplies neither absolute photometry nor MTF measurements, and it cannot identify a defective coating. It also says nothing definitive about other copies or production runs.

Fast Primes and Standard Zooms Fail on Different Schedules

For a fast normal prime, aperture provides the main sequence to investigate. Start wide open and inspect both a compact central patch and a broad contrast wash. A smaller, brighter disk is a possible finding, rather than a signature guaranteed by the lens class.

A standard zoom adds another changing variable: focal length. Moving groups change surface spacing and reflection geometry, so the wide-end result cannot stand in for the whole zoom range. Repeat the aperture sequence at the wide end, midpoint, and long end. Inspect for both localized and distributed faults at every position.

A beginner can keep this manageable by completing one focal-length sequence before moving the zoom ring. Once that routine is consistent, compare how disk extent and texture separation change across focal lengths.

Keep the Mounting Hardware in the Record

Record the lens version, copy identifier, adapter, filter, and step rings. Check adapter seating and inspect for reflective interior surfaces. An added reflection from mounting hardware can otherwise become an unsupported accusation against the lens coatings.

A favorable result at 720nm does not establish performance with your nominal 850nm filter. An unconverted body also changes the comparison. Use those reports to decide what to test, then judge the assembly you intend to carry.

What Stopping Down Actually Buys You

Stopping down changes the pupil and the geometry of reflected light. The visible result needs measurement: a hotspot may shrink, become more distinct, or remain troublesome. A smaller aperture provides no automatic recovery of central micro-contrast.

For each full stop of aperture reduction, approximately double the exposure time at fixed ISO under steady illumination. Verify the target exposure in the raw files. Otherwise, a darker frame can make the bright patch look less conspicuous without demonstrating an optical improvement.

Record disk diameter, disk prominence, and center-texture separation independently. They can move in different directions. If the disk disappears while stone-like texture remains indistinct, the aperture sequence has improved brightness distribution without establishing a recovery of local detail.

Inspect Beyond the Disk

At f/8 to f/11, compare surrounding texture as well as the center. Diffraction blur at 850nm exceeds that at visible wavelengths at the same f-number, so the smallest aperture can soften detail across the frame.

Keep the conclusion specific to the current lens copy and coating production, your converted mirrorless body, the actual transmitted wavelengths, and the apertures you will use. One test cannot establish how another coating run behaves.

After selecting an aperture, make a field confirmation frame with fine detail and a bright source or bright surrounding area. For a Blue hour composition with bright lamps, test that arrangement explicitly. A Long exposure changes the practical capture conditions, and adding another filter introduces an interface absent from the chart test.

Run This Aperture Series on the Lens You Plan to Take Out

Use this worked setup for one fast normal prime whose maximum aperture is f/2. The camera is a converted full-frame mirrorless body with its internal hot-mirror removed or replaced, and the screw-in filter is a verified 850nm bandpass. Keep the target illumination unchanged throughout.

1. Establish the target. Fill the frame with a matte, finely textured mid-gray chart or evenly illuminated interior wall. Square the camera, select manual exposure and fixed ISO, focus on the surface, and retain that focus.
2. Isolate the hood. At f/2, make one hood-off frame and one hood-on frame with the same exposure. Leave the hood on for the remaining photographs.
3. Build the aperture sequence. Photograph at f/2, f/2.8, f/4, and f/8. Approximately double shutter time for each full stop, including the intervening stops between f/4 and f/8. Verify comparable target exposure in the raw files.
4. Preserve the comparison. Give every file the lens identifier, focal length, aperture, hood state, and filter identifier in its filename or accompanying metadata. Apply identical processing and compare the whole frames and matching texture crops.
5. Make the decision explicit. Suppose this example produces an obvious disk at f/2, a weaker disk at f/2.8, and no visible disk with separable central texture at f/4. Suppose f/8 also removes the disk but softens surrounding detail. Select f/4 for the planned detail-rich stone photograph.
6. Confirm it on location. Keep the same filter and hood, set f/4, and photograph stone texture with a bright surrounding area in the composition. Inspect the center at the same viewing scale as the chart crops, then record f/4 and the complete filter assembly alongside that field frame.