willie_901
Veteran
You made the best choice.
The shutter time and lens aperture settings required to avoid severe pixel underexposure of the room interior (wood door, wood frame/wall and hinges etc.) However, these parameters also caused severe pixel overexposure in the regions with light streaming through the door. The overexposure level is responsible for the blue/purple fringing.
If the shutter time was shorter, the room interior’s perceived image quality would degrade due to decreased pixel exposure levels. Eventually the darker regions’ low signal-to-noise ratio becomes unacceptable.
It is inconvenient when the difference in sensor illuminance between the the brightest and darkest regions of interest vastly exceeds the sensor’s analog dynamic range. Solutions to the problem are unfeasible, inconvenient or require some degree of image quality compromise.
Some deal with analog dynamic range problems by using the camera’s native value (base ISO setting) and auto-bracket shutter time to create set of three to five exposures. These exposure can be in +/- 1 to +/- 1/3 EV steps (e.g. 0 EV,+2/3 EV, -2/3 EV). The longest shutter time is limited by camera and, or subject movement (the minimum acceptable resolution loss due to movement). During post production the image with the best compromise between bright and dark region exposure is selected . Optimizing post-production rendering for this image using region-selective image brightness is possible using simple or sophisticated tools.
More complicated post-production techniques can be used to blend the exposures for multiple images to produce one image. Exposure blending selectively combines regions from different images such that each region has high SNR levels. The results can be difficult to distinguish from single-exposure images with even lighting.
HDR post-production uses tone-mapping methods along with exposure blending. Tone-mapping image changes hue and contrast. A diverse range of HDR rendering styles is available with the click of the mouse which often results in unrealistic renderings.
The artifacts in this image could be corrected using manageable post-production methods. Lightroom Classic or Photoshop (along with many non-Adobe products) can manipulate the regions with blue/purple fringing artifacts to render as though the door board’s edges were properly exposed.
Boring Details
Blue/purple fringing is an artifact that can occur when the number of photoelectrons produced in one or more pixels greatly exceeds the maximum analog, electrical charge capacity of those pixel(s).
As Freakscene points out, blooming occurs when low levels of pixel overexposure creates low-level artifacts (time-independent noise) in small regions of pixels. Blooming is the lowest level of selective overexposure artifacts. When the overexposure levels are high, blue/purple artifacts start to appear.
All of the overexposed pixels’ excess electrical charge has to go someplace. If the overexposure levels are low, the excess analog electrical energy has a small effect and the effects are limited to small pixel regions. At some point pixel overexposure pixels further away from the overexposed region are affected. The distant pixels’ electrical charge is no longer proportional to their exposure. Their electrical charge is altered (increased) by the excess electrical charge from acutely overexposed pixel regions. This is how the excess electrical charge from the overexposed pixels is is dissipated. If the more distant pixels happen to be overexposed too, the region renders as white. Pixels in a region that was not completely overexposed that are a similar distance from the highly exposed overexposed can render with a hue. In the extreme that hue is blue/purple.
If the light streaming through the door was from a cloudless bright sky, some of the sensor’s blue pixels in these regions could be acutely overexposed. All Bayer image rendering software assumes every pixel’s charge capacity has not exceeded the maximum level when the shutter closes. The data (integers that represent the number of photoelectrons created in each pixel) are incompatible with the image rendering algorithms.
