Magic Wand tool color selection: limit one pixel color value Selection Errors

[JacB.Nimble]

In a Web tutorial a while ago I found a novel way to pick a Magic Wand Reference color that avoids an accidental color choice from a single pixel. To avoid selecting unintended single pixel colors with the Magic Wand, one can instead select an average Reference color value. To do this, first go to the Color Picker tool (shortcut = I) & select either the 3x3 or 5x5 pixel selection in the Options bar. Then switch to the Magic Wand & take the color sample by clicking an appropriate spot in the image. The selection should have the AVERAGE color value of the group of pixels under the cursor rather than a chance color pick from a single pixel. Has anyone verified this either in the PSE or the PS Magic Wand tools?

        

I did tests on a color border area that had some unexpected results. The MW Contiguous setting was set to ON & the Tolerance to 2, a small color range. In the single pixel color sample (top image), the selection looks as expected with very narrow color range centered on the crosshair guides where the samples were taken. But the 5x5 image (bottom image above) seems to have totally ignored the MW Tolerance setting but conformed to the Contiguous setting. It includes a broader range of colors perhaps centered around the average color of the Color Picker setting. I expected the two selections to likely be colored differently & the 2nd selection to be uniformly colored with an average color value.  Somehow, this process allowed a wide color range around that average. Any thoughts about these rersults?

[Mylenium · Accepted solution]

There's probably nothing wrong, given how noisy your images are. That's the crux of the process. Once you start using average settings on anything, the math will defeat what you perceive to see on the screen. The actual values may be way different than what you expect. I'm sure there's a super-scientific expalnation somewhere, but suffice it to say that based on my own experiences I think everything is working as it should.

Mylenium

[JacB.Nimble · Accepted solution]

Thanks Mylenium, I went back & checked my original notes & There is a warning that the cursor should be placed at least a half-inch away from any other areas of contrasting color to avoid contaminating the color sample.