These red spheres anomalies have only came up in my photos a couple of times while using my Fuji Finepix-S. The spheres seem to be evenly displaced on certain images forming 3 rows of 3. One other thing I may add is these spheres only can up on 3 pictures of sun rising in the background and on a photo of the moon. I have thousands of pictures of the moon and sun in different positions and light exposures that have came out spectacular, so I have eliminated the camera as a reason behind the anomaly but it might just be a rare malfunction.

Question-Is there a name reference or a well known logical explanation behind these red spheres?

enter image description here

enter image description here

enter image description here


Yes, the problem you are experiencing is a subclass of general flare and ghosting issues, often called "red dot flare" (see also: What causes lens flare?).

The reason for the grid or array appearance is because some of the bright incoming light is reflected off the actual microlens lens array ("individual" tiny lens for each sensor pixel) back towards the lens, then reflecting off of one of the lens elements back towards the sensor.

This particular form of flare/ghosting is particular to digital photography; it is precisely the physical characteristics of the imaging sensors (an arrangement of discrete photosites, arranged in rows and columns) that leads to this phenomenon. All digital cameras are subject to this issue. Some factors contribute to the degree of the problem, and whether it is noticeable:

  • Smaller apertures (higher ƒ-number) make this issue more readily apparent, just like diffraction spikes that are more visible when a lens is stopped down. Basically, the red-dot flare is just small-aperture diffraction spikes convolved with a rectangular array.

  • Higher ratios of sensor size : distance from rear lens element to sensor (notionally, registration distance) create more pronounced red-dot flare. Cameras designed to mount the lens comparatively close to the sensor bend the light more to cover the sensor. This bending (partially corrected by more aggressive microlens focusing) creates more dispersion of the light, resulting is more visibly discrete dot flares. Example: Fuji X-series review at photographylife.com.

Note that technically, this happens for every ray of light entering the camera, but you don't see it because of the degree of attenuation for each round trip. Note how dark the red spots are compared to the light source that causes the issue (sun, or other overpowering light in the scene).

In the third image you attached, I barely see any red dot flare (it's hardly red to my eyes, and it' so diffuse it could easily be confused with, or intermixed with, regular flare around the sun). The biggest anomaly is a single red ghost. Note how it is on the other side of the image, radially reflected opposite from the sun. This is standard, run-of-the-mill ghosting.

  • It is far more likely the light is bouncing off the IR filter, which is smooth and flat, in front of the sensor and its microlenses. It is then getting separated by the refractive properties of the lens elements it is bouncing backwards through. – Michael C Apr 6 '16 at 9:40
  • @michael Clark - I have to say I agree with scottbb; look at the non-axial grid pattern, that's not caused by an IR filter sandwich. The only thing that could reflect like this would be a grid array of reflective points or surfaces, such as the microlens array specifically found in Fuji sensors. I can only assume that some unintentional optical trickery creates a massive magnification mirror effect similar to a catadioptric lens, and projects a small amount of this reflected light pattern back onto the rear elements of the lens, and in turn back onto the sensor. – HamishKL Apr 6 '16 at 10:55
  • @MichaelClark Sorry,likely not. The array appearance is from the pixel array. Here is a particularly noticeable example from this discussion at an astrophotography forum. Note that the linked example came out of a modded (IR filter removed) 6D. – scottbb Apr 6 '16 at 11:01
  • That would have to be some really massive magnification to get microlenses a mere 3-5µm wide to expand to around 1/8th the height of the sensor. – Michael C Apr 6 '16 at 16:08
  • @scottbb See posts #13 and 15 of the astrophotography discussion you linked above. That theory sounds much more plausible than reflections of individual microlenses. – Michael C Apr 6 '16 at 16:14

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.