What is the resolution difference between a RAW image taken with a monochrome sensor vs a RAW image taken with a sensor with a Color Filter Array?

With a color sensor, I know that sensitivity and QE are affected due to filtering and the resulting unprocessed image from the CFA sensor has somewhat of a checker pattern effect, but is there any impact to 'resolution'?

I'm not sure if I should clarify, but I'm asking from a scientific imaging perspective, not photography or artistic.


closed as off-topic by Michael C, mattdm, inkista, Olivier, StephenG Sep 21 '17 at 19:39

  • This question does not appear to be about photography within the scope defined in the help center.
If this question can be reworded to fit the rules in the help center, please edit the question.

  • If you are not asking this from a photography perspective, this is prima facie off-topic for this site. Can you reword your question so it is within that point of view? Otherwise you may better off on an engineering-focused site. – mattdm Aug 31 '17 at 14:54
  • 5
    I'm voting to close this question as off-topic because it is from a scientific imaging perspective, not photography or artistic. – Michael C Aug 31 '17 at 16:35
  • It depends on the nature of the resolution target. Is it monochrome? multichromatic? It also depends on the way the signal from the Bayer masked array is demosaiced. Some methods give better resolution than others, and some methods work better for alternating black/white lines while others work better for targets with various colored elements. – Michael C Aug 31 '17 at 16:37

The resolution would be the same. The demosaicing algorithms involve weighted summing of neighboring pixels (essentially very specific weighted-averaging of 3-by-3 (usually) pixel blocks).

But unlike other typical instances of pixel summing or averaging, sensor demosaicing does not involve a downscaling of spatial information.

Now, because the demosaicing algorithm is essentially trying to create (interpolate) data that wasn't captured, there are some noticeable effects or artifacts if the color information in the scene changes rapidly (specifically, near the resolution limit of the sensor). Because the spatial frequency of (typically) red- and blue-filtered sensor pixels is 1/4 the sensor's absolute spatial frequency, the color-specific spatial resolution of the output image is correspondingly reduced.

But in terms of absolute pixel spatial resolution, CFA sensors and pure-gray sensors, built to the same pixel pitch, spacing, and dimensions, will have the same pixel resolution1.

  1. The effective pixel area of CFA sensors is (at least) 2 pixels fewer in each dimension (1 additional pixel on each edge) than the actual pixel count on die. This is so that the edge pixels in the produced image all have neighbor values to use the same weighted-summing demosaicing. Without the "border" pixels on the sensor area, the edge pixels in the produced image won't have had data from the non-existent neighbor pixels past the edge. These extra pixels are like an unusable "picture frame" of pixels on the sensor die; they don't count in the produced megapixels of the final image.

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