Photography Stack Exchange is a question and answer site for professional, enthusiast and amateur photographers. It only takes a minute to sign up.
Sign up to join this community
I am trying to study basic digital image processing but i am confused by the difference between two terms "image contrast " and "image sharpening'
Please kindly explain the difference and if possible please include example photos/images with your answer to make it easier to understand
The best way to see the difference is to open an editing software like GIMP and too see how it works.
Here's the picture I'll use, it's a color gradient to simulate a blurry edge
In Filter > Enhance > Sharpen you can find the sharpening tool.
If you mess a little bit with the settings (radius, amount and threshold) you can get this picture :
As we can see the sharpening tool only increases the contrast on the edge, the brighter side of the edge becomes brighter and the darker side darker, but the constrast of the whole picture is not affected.
If you use a very low threshold, the Algorithm may recognize image noise as edges or borders, which will make the image noise more noticeable, since the constrast of the noise will be increased.
For comparison I increased the constrast of the whole picture here :
See Sharpening v Increasing Contrast, fragment:
Sharpening is a method of increasing the contrast at a border or edge. That is, where when detail in an image ends and another detail begins, sharpening (or unsharp mask as it is frequently called in some photo editing software) will darken dark edges and lighten light edges so that detail stands out. Changing contrast is simply determining how big a jump you are making as you move from one shade to another. Effectively what you are doing when you increase the contrast of an image is cutting down the range of brightnesses between black and white. In a high contrast image, something that is only a lighter brighter than middle gray will be interpreted as pure white, and something that is only a little darker than middle gray will be interpreted as pure black. Low contrast is, of course, the opposite.
Sharpening is countering the effects of blurring. Blurring comes in at least five flavors: one is defocusing blur which spreads out single spots outside of the focus plane into the shape of the aperture, with a size dependent on the amount of defocusing. Another is diffraction (mostly relevant at small apertures) that spreads out single spots into a diffraction pattern. Then there is systematic unsharpness depending on camera/lens quality that usually can be modeled as Gaussian blur. And then there is camera shake which spreads single points to some path with some statistics shared across the image, and motion blur which spreads single points of objects along the paths an object takes when shooting, with statistics being related only for single objects.
The usual sharpening technique is called "unsharp masking" and is some sort of very local contrast/edge enhancement. It works to some degree to visually counter most sorts of blur while causing a moderate amount of artifacts (usually halos around edges). For comparatively global blur alone (aperture-dependent diffraction and general camera quality), deconvolution can work rather well.
Contrast enhancement is not used for making edges more striking but for rendering differing areas with a larger contrast between them. Usually, some sort of weighted average over a larger surrounding area is made and the difference of single pixels to that average is amplified.
All of those enhancement come at the cost of amplifying image noise. Balancing their image noise amplification with gains in image quality is partly a component of those operations, partly a matter of personal choice in setting the respective parameters.
Sharpening and contrast enhancement are related. The difference is, while contrast enhancement is a global operation (ie., it is applied to the whole image, though it can of course also be restricted to areas of interest), sharpening is a local operation which affects borders and edges.
Sharpening increases the contrast at sharp transitions of brightness, ie. edges, to make them stand out more, which is perceived as increased sharpness. Parameters to choose are the range of edges to be affected (how big the difference has to be), the amount of increase/decrease of tonal values to apply, and the radius of falloff this effect should have.
Contrast is increased by applying an S-curve to the curve diagram, ie. bright values will get to be even brighter and dark values even darker.
There's even a kind of twitter by using sharpening to achieve a local contrast enhancement: unsharp masking with a low amount but large radius (in the hundreds of pixels typically) produces an interesting "clarity" effect.
