The sunset – How can I improve on it?
This morning, a friend of mine, Bruce Itterman, e-mailed me a sunset photo that he had taken and asked me for comment. Personally, I love sunsets but hate trying to take pictures of them. The same for sunrises, except that, being more of a night person than a morning person, there is a better chance that I will be awake when the sun sets :-). Besides photography, Bruce and I share another interest, namely aquariums. Trying to photograph fish in aquariums can often be as troublesome. In the case of aquarium shots, the problem is often the significant difference in contrast between the light at the top of the aquarium and the darkness in the shadows of the piece of driftwood or even just the difference in light intensity between the top and the bottom of a tall aquarium. In the case of sunsets, the same problem arises. The contrast between the brightest spot in the picture – the sun – and the darkest spot in the picture – the shaded shoreline – is greater than the contrast range of the medium, in this case the digital sensor, or the computer monitor, or the print media.
This is not a new problem. It has been around since the very first time that man has tried to capture a photograph of the sun at that critical time when it departs below the horizon. The modern-day solution is something referred to as HDR processing (HDR = High Dynamic Range). It is not really a ‘solution’ but rather an averaging technique that provides a method for displaying more of the detail from the shadows of a high dynamic range scene along with the detail from the brightest areas of that same scene.
Through the use of multiple exposures and digital processing algorithms, photographers are now able to digitally combine ‘overexposed’ images with ‘underexposed’ images to produce end products capturing the details from a broader dynamic range than would be possible with the taking of just one shot of the scene. Wikipedia contains an excellent summary sheet on HDR processing and its history so, rather than trying to improve on that, I simply refer folks to Wikipedia’s article: High Dynamic Range Imaging
Although my friend only sent me one image, I know that he took other similar images at different exposure levels and will likely be experimenting with HDR as well. However, one of the other issues that people sometimes forget when dealing with sunsets, is just how fast the sun is moving (setting). For HDR manipulation to be most successful, all of the objects in the picture must be in the same relative location or as close to that point as possible. Therefore, to be most successful with creating a set of images to be combined into one final HDR image, it is best to have the camera automatically take more than one image and pre-program the camera to have the camera implement planned levels of exposure/overexposure/underexposure automatically in a sequence referred to as “bracketing”. Although I am referring to this sequence as exposure/overexposure/underexposure, what, in fact, is optimally happening is that the ‘overexposed’ setting is capturing a properly exposed image of the darkest areas of the photo and the ‘underexposed’ setting is capturing a properly exposed image of the brightest area of the photo. In sunset photos, it is difficult to compensate completely for the very bright sun so often the proper exposure is one that shows as much detail of the sky as is available and of interest. Since the HDR process is now a digital process, many camera manufacturers are now incorporating an HDR setting into the camera’s basic functionality and providing an HDR image output directly from the cameras, if so desired, and thus eliminating or replacing the post-processing step at the photographer’s computer level.
For sunsets and similar high contrast situations, the bracketing steps could be as low as three and as as high as the camera will allow with the highest exposure value (EV) in the bracketing sequence being set at +8EV and the lowest at -8EV or perhaps a lesser range depending on the circumstances. Of course, the quality of the camera’s sensor, the nature of the post processing software, and the number of steps or gradients in the bracketing sequence will all be factors in the quality of the final outcome. A range of +6EV to -8EV would not be too unusual in a high contrast environment such as the subject sunrise or many night shots where floodlights might be overwhelming the scene. With such HDR functionality now incorporated into various consumer devices already (iPhone4G, for instance) it is likely not too far-fetched to expect that almost all digital cameras will support HDR as an automatic option within the next couple of years.
So, that brings me back to my friend’s original one shot image and what changes might be made to it.
My friend has already told me that he is happy with the image and feels that, of the many shots that he took of the moment, the resulting jpeg that he has sent me most closely reflects the moment as he saw it. That is certainly good enough for me!
Would HDR improve it, if available? Well, I suspect that if my friend was standing there and wanted to see better what was in the shadows, he would have used a baseball cap or his hand to block out the sun and thus allow his eyes to adjust to the darkness of the shadows and allow him to peer more deeply into those shadows. This is much akin to what the photographer is photographically doing in the ‘overexposed’ phase of the HDR imaging process. He is effectively ignoring the bright section of the scene and focusing on the darker sections of the scene. Likewise, if he wanted to see more detail in the sky and couldn’t because of the overall brightness of the sun, then he would likely have donned a pair of sunglasses and muted the impact of the sun’s brightness in that manner. Again, this is akin to reducing the amount of light that is reaching the sensor, or in HDR creating an ‘underexposed’ image. While observing a sunset, most, probably all, humans will be conducting these overexposure/underexposure procedures many, many times a second by squinting or changing the direction of their head, donning or removing sunglasses. raising a hand to block out the intensity of the sun. Would application for HDR techniques produce a ‘better’ photo memory of the scene. Perhaps yes, perhaps no. If the memory is of an intense globe lighting up the sky with intense red coloration forcing the real-time viewer (the photographer) to don sunglasses or squint so hard that the shadows become darkened masses in the eye of the viewer, then the non-HDR image is likely to portray that feeling of the moment quite well and it is quite often at that level, the personal one, that most photographers must make their decisions as to whether they like their photo or not.
Since, in this instance, I am working with only one image, I turned my attention to the available image, and looked at how else the image might be manipulated in post processing depending on how it might be used or displayed without worrying about the benefits or drawbacks of HDR options. In most cameras, the ratio of the length to the width of the image is constant even though the image size might vary a bit from camera brand to camera brand. Either landscape or portrait format is the available default option based on how the camera is held but some folks will experiment by holding the camera at various angles to the horizontal or vertical axis. Almost any form of manipulated distortion or contortion can be achieved in post-processing but for this blog I’ve stuck to just landscape or portrait options.
I can remember, many years ago, when I had my own darkroom and would process photos from friends as 4×6 prints rather than the more traditional 3 1/2 x 5in size. Just that slight change and the cropping that I was required to do was enough of a change to elicit interesting responses from friends who might have been seeing their photos with cropping changes for the first time. Nowadays, when friends ask me what changes I might suggest for their photos, one of the first things that I do is consider different cropping options. What size will be required in the end, and what file size I am starting with, helps to determine what cropping options are viable. In most mid-range to higher end DSLRs sold today, image size and resolution quality is now sufficient to allow for very satisfactory print size past the 11*14 size so, for most images, significant latitude exists for cropping without the final output showing objectionable visual differences in the 5*7 or 8*10 range of finished print sizes.
Returning to the original sunset shot, some ratio adjustments and some relatively minor cropping can yield some interesting differences (whether or not these would be considered ‘improvements’ is highly subjective).
8*10 horizontal or vertical