Monday, 18 November 2013

Merging Two Photographs Together to Create a Single Photograph with the ''Best' of Both Worlds' in Terms of Dynamic Range

Photograph 1 (Exposed for the Land)

Merging two photographs together to produce a natural composite of the two, with greater dynamic range than either photograph would have by itself, was a part of digital photography I had been interested in for a while now. This was mostly because it more closely resembled any given scene where the dynamic range was higher than considered average, which the human eye would be able to register. This was in contrast to the photographic medium where a single exposure (especially on digital) would not. This was an example of how a photograph could capture a moment in time, which resembled something easily recognisable by humans but that actually defied a camera's physics of limited dynamic range, all in an attempt to make the photograph more believable to the viewer. I would be aiming to make photographs which tried to play with reality rather than replicate it, most probably through similar kinds of techniques in future projects. Therefore I was pleased to find I would be experimenting with this kind of processing in an exercise.

Photograph 2 (Exposed for the Sky)
 
   I chose a slightly different approach to merging the lighter and darker exposures than as that suggested. This approach namely involved the use of a gradient on a layer mask. Initially, my reasoning for using this approach was that it would provide a quicker and more seamless result to that of erasing the over-exposed sky (Photograph 1) and pasting the well-exposed sky (Photograph 2) in its place.

Photograph 3 (Photograph 1 and 2 Blended)
   The gradient approach would be quicker in this instance because it involved only drawing a short gradation from the land features including the horizon to the new sky using the gradient tool on layer mask. I predicted it would work well in this instance because it was a very flat horizon; with no features ‘jutting out’ from the horizon. I predicted it would be a more seamless approach after reading about this methodology on a website article by Jean Miele on Digital Photography Review specifically concerning using gradients to achieve ‘feathered, believable transitions’ - Miele (2013). The effect was seamless because of the small (or large) gradation from black to white in the layer mask, which had the result of a grey band in between, which served the purpose of mixing the land (particularly the horizon) in Photograph 1 and sky in Photograph 2 together naturally.

   Another reason I favoured this approach was that it could be adjusted easily afterwards by either redrawing the gradient in the mask and/or painting on the layer mask to ‘mask out’ any horizon discrepancies. However, because the horizon was so flat in this case that step was not necessary.

Photograph 4 (Dynamic Sky)
   I found this worked well and was very efficient in blending the sky and land together (Photograph 3). I did have to raise the position of the transition between the sky and the land in the gradient's layer mask because although the transition was smooth, the horizon suffered a bit from too much of the brightness of the land layer coming through with the effect of it looking 'hazy'. By changing the position of the transition a bit higher this problem was quickly amended. Overall I found this blending of two photographs taken close together to be effective and it didn't pose any ethical questions for me mainly because the scene remained how I saw it at the time.

   When I was satisfied with the appearance of this composite I attempted to replace the entire sky from a sky taken within a few seconds of the landscape-exposed image to a sky captured three years ago (Photograph 4). Firstly, I was quite skeptical about the ethics and indeed the believability for any prospective viewers, that I could foresee in replacing a vast amount of an image with something taken at such a radically different moment in time. While I remained unsure about the ethics surrounding creating a composite like this, I appreciated that the end result did appear much more dramatic and even impressive than the pretty unremarkable, well-exposed sky that was taken a only few seconds after the well-exposed land segment of the image, which had come together in Photograph 3.

   The end-result (Photograph 5) was again achieved through using gradients but this time, because the (much more impressive) clouds were so low to the horizon, I had to employ an additional technique in order to make the horizon and clouds blend into one another well. This technique was namely dodging, which I performed on a separate neutral-grey layer with soft-light blending on top of the clouds layer (with a clipping mask applied so only the clouds were affected). I used this method for dodging as it provided me with the opportunity to amend the intensity of the dodging if I so wished afterwards. Another technical detail was 'masking out' some of the cloud details that were just present in the grey transition between the black and white gradient on the layer mask. This was also because the cloud formations were so low to the horizon. All this meant the process wasn't as quick and easy as adding in the first cloud-exposed image but I found the extra steps included were necessary to make the clouds from Photograph 4 fit in more believably.

Photograph 5 (Photographs 1 and 4 Blended)
   I would say this deliberate addition of one part of a completely separate image to replace another part of a different image was definitely bordering the limits as to what I found acceptable. This was mainly because I knew myself as the photographer that these two photographs had been taken with a space of three years between them. Therefore, the final scene I could see on my monitor of the landscape scene with a totally different sky (Photograph 5) was of course definitely not how I remembered it as I had taken it, which was usually one of my criteria for whether a final, processed image was in 'good taste'. Had I been the viewer of such an image and the image was purported to have been shot at the same place and roughly the same time (i.e. an High Dynamic Range image), I would probably have accepted the photograph as 'real' and have been impressed by the dynamic clouds in particular. It would be here that I would see the photographer as becoming responsible for claiming falsely the 'legitimacy' of the photograph depending on how it was intended to be used.