Why do sharp chart edges look jagged or noisy in a RAW file?

Asked 5/27/2026

5 views

2 answers

0

I photographed a test chart and the printed edges look very clean to my eye, but in the image the edge transition looks uneven when zoomed in. I checked the RAW data and the effect is already there. This is not the edge of the frame; it appears within the image area.

I also noticed that the RAW looks different depending on how I view it: in Bayer mode it looks different than in Gray mode. Is this normal RAW preview behavior?

Is this kind of edge roughness unavoidable because of pixel sampling, diffraction, or RAW rendering, or could it indicate a poor-quality CMOS sensor?

Originally by Landon Ham. Source · Licensed CC BY-SA 4.0

Landon Ham

15d ago

2 Answers

2

Two things are occurring. Human perception in terms of your ability to see detail/resolution, and the importance of contrast in our perception of sharpness/resolution, are both different from how a digital sensor sees.

I.e. for a human a larger detail needs less contrast (sudden change) to be perceived equally as sharp/clean as a smaller detail. And a human cannot see down to the 4-8 micron level unaided (size of a typical sensor photosite). For example, bacteria are about 3 um in size; red blood cells are 8 um. Whereas a lens could theoretically resolve down to 1 um.

Any projected/reflected light pattern (or shadow) from a diffuse source will have a falloff gradient along it's edge. The sensor is simply recording that; it is unavoidable.

Originally by Steven Kersting. Source · Licensed CC BY-SA 4.0

Steven Kersting

15d ago

0

AI Answer

Generated from our catalog & community — verify before relying on it.

This is most likely normal, not a bad CMOS sensor.

At pixel level, a camera records discrete samples on a pixel grid. A real-world edge usually does not line up perfectly with that grid, so when you zoom in, the transition can look jagged or uneven. Sensors also record the edge gradient of real light; edges are rarely infinitely sharp in the optical image.

A few factors contribute:

  • pixel sampling: sharp edges rarely align exactly to pixel boundaries
  • Bayer CFA/RAW rendering: Bayer and Gray views can look different because one shows color-filter-array sampling and the other a different preview interpretation
  • lens/diffraction limits: at high resolution, diffraction and lens resolving limits soften edges slightly
  • noise and processing: some noise is always present, and RAW/software processing changes how visible it is

So yes, some edge roughness at 100% view is unavoidable. It generally reflects sensor sampling, optics, and RAW interpretation rather than poor sensor quality.

UniqueBot

AI

15d ago

Your Answer