Exposure setting

Read as much - or as little - of this as you wish

Have a look at these two images.  The same stamp, and even the same scanner, but what a difference in image quality.

Do you notice how the image on the left looks a lot softer than the image on the right - it has no contrast and neither does it have much brightness.  The reason for this primarily to do with proper use of the 'exposure' control (or whatever it is called) in your scanning program.

Here in the exposure control of my scanning program you can see a graphical representation of the stamp image's range of light/dark values.

Notice how there is a bit of space with no values at either end - this is perfectly normal.  Ignore the way I have the sliders set, just look at the distribution of the brightness values.

The first scan was taken essentially with no change to the image as it was first sensed by the scanner.  The spread of color values remained similar, but not exactly the same (due to inconsistencies in the scanning process and various other things), as you can see in this Photoshop Levels display

Then I clicked on the Auto button, which moved the highlight and shadow sliders to their new positions, as you can see below, and scanned again.

This has made the picture very much clearer, because the color values are now more evenly spread from light to dark.  As you can see from the Adobe levels window with the final image, the brightness values now flow pretty much all the way from minimum to maximum

.

This is an example of an image where the 'auto' setting worked close to perfectly.  There are two situations when you might want to override the auto setting.

The first is when the auto setting 'cuts off' too much image data.  Here are two more scans to illustrate what happens when you 'cut off' image data.  I've used the 150 dpi images this time, because the effect is a bit more subtle.

Do you notice the difference?  At first you might think that the image on the left is better, but it isn't.  It is more contrasty, for sure, but you've lost some of the detail in the shading - look for example at the sky or the man's collar or his rolled up paper - some of the lighter shading has been entirely lost into the plain white.  By the same token, some of the darker shading has been lost into the solid brown or black.

So, if you have a situation where the auto setting moves the minimum and maximum into the actual area where there is detail, this can be a bad thing.  Look again up at the second 'Exposure Adjustment' image - this is close to perfect, with the minimum and maximums set almost exactly where values first start to appear on the image.  If the two pointers had moved closer together, then the image would become progressively more like the scan sample on the left immediately above.

The second situation when your image doesn't have a standard range of colors from (close to) white to (close to) black.  Maybe it is a very old stamp and the white paper has faded to brown.  Or maybe it is a stamp with a few colors printed on a predominantly black background.  Here is an example of such a stamp.

Notice how the image on the left doesn't have as strong a black as the image on the right.  The auto-exposure has tried to 'average' the intensity in the image, and so have over-compensated for the fact that it is supposed to be a very dark image, making the colors much brighter and the background much lighter.  The image on the right was scanned with the same settings as the other 'correct' images have been scanned on this page.

Now look at the graph showing brightness for this image.  As you would expect, it is all concentrated in one small area of darkness, and much of the graph has no values at all.

I have the sliders shown in the above image close to where the auto exposure setting would put them, but this makes the image too light.  You need to move the right slider out to reflect the fact that the lightest thing in the image is not as light as a white piece of paper.

Lastly, to conclude this discussion of exposure settings, here is a 'torture test' example where I scan a piece of pure black and pure white, side by side.  Here are two results.

 What do you notice about these two images?  Do you notice that the one on the left has a dark grey rather than black, and also has a light grey instead of a white?  Can you guess why that is?

Let's look at the exposure graph for this image.

As you'd expect, it shows some values on the far left (dark) and the far right (light) and nothing at all inbetween.  I've also left the sliders where the auto setting suggested.  Do you think the auto is correct or incorrect?

This is introducing you to one last subtle and hopefully not too confusing point.  These two images - the back of a Scott mount for the black and a piece of 'high white' paper for white - are, in theory, close to pure black and pure white, and this means instead of giving a range of white and dark values, they actually should just have single lines for 'all black' and 'all white'.  The wider spread of values is (sort of) 'wrong'.  If we scan with the two sliders at each end, to include all values (like I've been saying you should!) then the image on the left above is what we get - instead of a solid punchy black and a vivid pure white, we get a range of dark greys and light greys.

This is a case where you actually want to set the sliders so that all the values are outside rather than inside the area between them.  This means that all the values the scanner sees as 'reasonably black' get translated as 'full black' and all the values the scanner sees as 'reasonably white' gets translated as 'full white'.

Have you understood this page?  Please let me know so I can add to or change it as necessary.

The issues are indeed complex, but 99% of the time can be reduced to a simple three step process :

First, select the area to be scanned

Second, use the auto exposure to recommend values

Third, think about any special issues such as pure white and black or an unusual image and adjust the auto exposure setting to suit