For this exercise, you’ll add one element from a different image. However, in order to make this less obviously full intervention, the exercise will be in two stages.
The aim is to take a conventional landscape view and render the sky do that it appears ideally exposed, with every detail of every cloud visible and textured. With a cloudy or partly cloudy sky this is frequently a problem, as you may already have experienced.
You’ll do this in two ways, the first being more ‘legitimate’ than the second, and for this you should find a location that allows you to make a landscape composition with a significant area of sky (at least a third of the image area). As you will later be making major changes to the sky, you will find it easier if the horizon line is clear, clean and obvious, without such fine details as branches and leaves. A clear blue sky will not do for this shot, as we need for the purposes of the exercise, a scene with a high dynamic range. You will have to wait for a cloudy day or partly cloudy sky, so that the ideal exposure for the sky alone would be significantly less than for the landscape.
Set up the camera on a tripod, so that you can make more than one exposure in perfect register. Make two different exposures, with the camera on either manual settings or using its exposure compensation system if it has obe.
One exposure should be perfect for the landscape (the sky will be over-exposed)
The second exposure should be perfect for the sky, with no highlight clipping (the landscape will be under-exposed).
The difference between these two exposures is likely to be in the region of two F-stops, possibly more.
Process these two images normally, without trying to make significant compensation- without using whatever highlight recovery of fill light controls your processing software offers. The next step is to combine these two images. If you use Photoshop or another image-editing program that allows you to make layers, so much the better. If so, do the following procedure. If not, skip to the paragraph after the next.
Copy the lighter image onto the darker image. Then, erase the over-exposed sky from the upper layer to reveal the darker sky beneath. Flatten the image and save as a copy. Then return to the original state of two layers before using the eraser brush. Now make a selection of the sky in the upper layer, using whatever selection tools you find convenient. Take care to refine the edge of the horizon. Save the selection. Now delete the upper layer’s sky, flatten and save as a copy. Making and saving a selection like this takes more time, but gives you more control over the erasure.
Alternatively, use an exposure blending program such as Photomatix to load and blend two images. Use the exposure blending procedure rather than the HDR tonemapping, as the latter involves advanced image processing beyond what we are disgussing here.
Exposure blending offers several choices of methods. The highlights and shadows- adjust option, is the default and normally the most useful, as it will give a natural-looking result and a choice of bias towards the darker or lighter image. Save the result.
Now, you have created a new image from two separate ones, but while this is clearly intervention, there is also a strong argument that this is perfectly legitimate, and that the procedure of taking more than one exposure within seconds is simply a way of overcoming the camera’s limitations of capturing a full range of brightness. Indeed, as technology progresses, specific software tools are appearing that make use of this’multi-shot’ approach. One such is photomerge scene cleaner in Photoshop elements, another is photomerge group shot.
Now use Photoshop, take this same image (the version exposed for landscape) or any other photograph with a sky. Then choose a different sky from an existing or new photograph.
In the first image, select the area of the sky and save the selection. Then paste into this area, the sky from the second photograph. The aim is to create a realistic effect, as if the new sky could really have been a part of the original image. You will need to consider such things as the direction of the sunlight if this is visible, the overall brightness and the contrast, in order to make the match look good.
One example of ‘doctoring’ an images, was in my work book. The writer included a situation in which he writes “A while ago, I was commissioned to take photographs of a bike workshop for a cycling charity. Unfortunately, my main model had just broken an ankle, and had a fresh cast on his left foot. A photograph of a bike instructor with a broken foot doesn’t inspire confidence to prospective learners, so I had to doctor the shot and clond his right foot. I took the photograph and ‘faked’ it myself using the following tools and techniques:”
1: Copying and pasting selected areas of the image
2: Cloning and Deleting
3: Refining Edges of Selectiobs
4: Transform command – Changing the shape and size of selections
The above case study, is a real example of a situation in which ‘doctoring’ a photograph is a useful solution to a particular situation. However, there are some issues with working in this way. Is it entirely ethical to do this? What sort of ethical issues do you think this raises?
WILL BE UPDATED SOON