Creative and artistic railway photography


Home Page





The Phoenix Journal

'Photoshop' for Railway Photographers

Come Back the Front Threequarter

History of the Circle


List of Members



Members Area

How to Upload Images


Calendar 2016

Preparing Images
for the galleries




Probably everyone will want to do some sharpening to the imported image. It is very easy to overdo the sharpening and induce a lot of noise which can leave the image looking very grainy. The sharpening process is probably best left until last when all the other image adjustments and so on are complete.

Photoshop has sharpening tools under Filter> Sharpen, the only ones that you are likely to use are unsharp mask or Smart Sharpen. As a starting point for a 23mb image I'd suggest 100% @ 1 pixel with a threshold of 0. For a smaller file a lesser % might be a better start along with .75 pixel setting.

A problem with this method when a colour image is involved is that the colour information gets sharpened along with the image information, this leads to a sometimes unacceptable increse in noise, particularly in areas of similar tone, eg in skies. I find it much better to use the following procedure.

Change the mode from RGB to Lab Color Image>Mode>Lab Color. You may want to go Layer>Flatten Image before you do this.
Then in the Channels palette select the Lightness Channel by clicking on it. The image will now become monochromatic. Now apply unsharp mask
Filter> Sharpen>Unsharp Mask.until you are happy with the result. Now change back to RGB Image>Mode>RGB

If you have a Lite version of Photoshop which doesn't have Lab Color try selecting the Green channel in RGB mode and apply unsharp mask to that.

Another way is to use the High Pass filter Filter>Other>High Pass.To use this filter create a duplicate layer the apply the high pass filter with a rdius of about 10 pixels. The image will now look like a grey mess. Change the layer blending mode to soft light and the sharpened image will appear. The effect can be regulated using the layer opacity. One useful feature of this method is that a layer mask can be used so that only parts ofthe image are sharpened. It may be worth experimenting with other layer blending modes.Hard Light for instance gives a stronger sharpening effect.

The duplicate layer which looks rather like a grey bas relief can be
manipulated quite considerably the radius of the 'edge strips' can be varied
from one or two pixels up to about six or seven to vary the sharpening
effect. Altering the brightness or contrast of the layer will also have an
effect. Varying the opacity of the duplicate layer also controls the degree
of sharpening.

If however you invert the duplicate layer to create like a negative, it will
reverse the effect, dark tones become lighter and light tones darker
reducing the contrast along the 'edge strips', this has the effect of making
the image look softer and the larger the radius selected the greater the
effect. One advantage of this technique is that colours do not bleed into
each other as will occur if Gaussian Blur is applied, unfortunately it is
not terribly successful if there are patches of bright white as it does not
tone them down. Again varying the colour, contrast and opacity of the
duplicate layer will alter the effect.

Using  a layer mask on the High Pass Filter layer allows accurate control of
the parts of the image to be sharpened or softened and slight errors in the
masking are not as obvious as slight errors in selections sharpened by USM.
It is possible and I often have as many as five or six sharpening/softening
layers working on an image. Creating the layer mask is often made easier by
observing the duplicate layer in Normal mode so that the area being  masked
can easily be seen. If sharpening the image increases the grain to
unacceptable level the duplicate layer can be masked out and then just the
edges to be sharpened are painted back in on the mask - this is easily done
using Normal mode. A similar technique used with the Soft Light or Overlay
mode selected is excellent for sharpening up eyes and other detail on faces
without increasing grain in the skin texture.

Another technique which used selective contrast enhancement to increase
detail in an image is to duplicate the layer, desaturate the duplicated
layer, then in the levels pallet move the right hand end triangle to the
left until only the detail which you wish to enhance is visible, the
blending mode is then changed to Multiply.