After I have done all I want with multiple masked curve adjustment layers and the image is looking pretty darn good, I finally am ready do do some subtle dodging and burning. So that I can have better control of this, I flatten the image, duplicate the image layer (drag the layer to the second icon from the right on the bottom of the layers palette) and then work on the duplicate layer on top. If disaster befalls me, I have the original image underneath without having to save anything, and if I want I can use a white mask for the duplicate layer into which I can paint grey to tone down the dodging. It's been my experience that I dodge at least 10X as much as burn and almost exclusively in the highlights.

I want those highlights to pop - I want liquid sunshine. Using dodging highlights does this for me. I have to be very careful that I don’t overdo the dodging of highlights. One extra stroke of a 5% opacity dodge highlights will drive a chunk of image into unrelieved pure white. I find that it's really difficult to tell just how light they are getting on the monitor (I don't get aggressive enough) so I use threshold adjustment layers.

What you do is add an adjustment layer by selecting threshold (instead of curves) in the layer menu at the bottom of the palette. Photoshop then asks you what level you would like to set it at. For highlights, set it at approximately 250 (the number does depend a bit on your printer so you can adjust this with experience - you are looking for the point at which highlights start to separate from pure white on the paper). You then turn down opacity of the layer to about 70% so you have a hint of the underneath layer. For working on burning, you do the same thing with a threshold set to 10 or thereabouts. Save time but creating an action for each. It’s easy to create an action to create these with a single click or even a function key.

You have choices in both dodging and burning - for each you can apply the effect to highlights, midtones or shadows. I virtually never ever dodge shadows or burn highlights as both result in muddy looking images. I occasionally use midtone burning and dodging but for the most part, I work with dodging highlights. I can control the opacity of my brush. If I am working with highlights that are already near white, I will set the brush at about 5%. I didn’t mention it last time but all work on masks and dodging and burning is done with a brush set at 0% hardness so I don’t leave telltale marks on the image - if I want better control along an edge, then I use a smaller brush. The brush itself is the blurred circle. I haven’t even bothered to learn what flow does.

OK, the threshold layer (highlights, 250) is on, I select the layer just below it (the dodging and burning layer) and start work. If there are areas which are already light but don't even show after the application of the threshold layer (ie. the brightest point is still under 250), then translate this as DULL! Lighten at least some of these areas if you want your print to sparkle. If, on the other hand; there are no real highlights, this is your chance to create some. Almost white areas will step forward, dark areas will recede. Use this to your advantage. Were you working on a face ans wanted a stronger chin, you could lighten it.? A forehead too prominent, darken it a little. With rocks, follow the form of the rock so that the nearest part of the rock steps forth. Sometimes it's edges that need adjusted. In the image below, the dark rock in the foreground blended all too well with the bluff behind it and I used a combination of midtone burn and then highlight dodge to separate it from the background. In this particular case I did in fact set the brush to about 30% hardness to help define the edge.

35 mm. photographers seem to spend their early darkroom years trying to open up shadows and burn in highlights. Fred Picker used to say - take them the direction they want to go - darken shadows and lighten highlights'. Of course this presuposses a properly exposed and developed negative, or digital equivalent.

I think this does more to produce a luminous print than just about anything, but, and it's a big but - you have to do it at the end of image manipulation, you have to be careful, and you have to practice.

There is a large element of creativity here. It's possible to turn dull and flat into glowing but it's also possible to absolutely destroy an image and end up with a cartoon. Only practice and printing lots of cycles and going back to the beginning and starting over more than once will result in skilled, subtle yet effective dodging and burning. I think of 'dodging highlights' as the equivalent of potassium ferricyanide bleaching without the mess and ruined prints. Look at printing articles by Bruce Barnbaum for examples of bleaching which can be applied to digital equally well.

Sharpening

This could be a topic for an entire article and there are many different ways to sharpen the image. There are several things you are trying to do with sharpening:

a) undoing the effects of the low pass filter built into the camera and placed there to prevent moire patterns from developing from the regular grid of the pixels during the image processing. As Kodak found when they didn’t include one, it’s not nearly as effective to use anti moire software as it is to simply blur the image a little in the camera before hitting the sensor.

b) a general improvement in resolution due to any weaknesses in the lens - eg. corner softness, loss of resolution due to diffraction. (it does make an interesting point - what if we could remove the low pass filter and stop down to f32 when shooting landscapes.?)

c) depth of field - sometimes you can give the impression of a bit more depth of field with judicious additional sharpening.

d) dealing with the tendency of inks to bleed slightly on the paper thus spreading and blurring the image slightly. This is called dot gain and varies from paper to paper but regardless does need to be dealt with.

My own recipe born out of various trials is to use Photoshop's smart sharpen. The amount varies depending on the camera. For my 1Ds2, I always set the amount to 300 but vary the radius according to image quality and whether I ‘upsized’ in Camera Raw. If I did, I usually use a radius of .9 to 1.1 and very occasionally 1.6 (which few images can tolerate). With my 10D, I found the amount had to be under 200.

I sharpen soft corners and add apparent depth of field using PKSharpener creative sharpening if need be (probably only 10% of the images as it can become obvious very easily).

I do a final sharpening in all images with PKSharpener Output Sharpening (which is based on pixels per inch and image size) to deal with the dot gain. The image on screen will look oversharpened but it’s the print that counts.

At workshops I have seen a lot of images ruined by overly aggressive sharpening (usually in an attempt to make a print larger than the image could stand). The photographer would have been far better to shrink the image, give it a large white border on the paper and sharpen less aggressively. I can only suggest that you get a few opinions on your sharpening technique - it should not be visible.

Well, the image is looking perfect on the screen, its time to print, and that’s a story for another day.

Image before

Image after

Tip: Download the images and put them as layers on top of each other and you will see that the difference is not only subtle.

You can find George Barr's website here.