Although High Dynamic Range imaging tools are easily accessible these days, the techniques for using them effectively do not seem to be talked about that widely. I hesitated to print any of my own HDR images for a long time since they always seemed slightly flawed to me.
Some of this is probably due to my perception of their unrealistic appearance from enhanced lighting and hyperrealistic detail, but the occasional artifacts that show up from the default Photoshop photomerge process are a bigger problem.
This scene in a local marina just before sunset caught my eye. I had time to grab a few exposure-bracketed frames as the harbor was half lit by golden hour light.
Here is what I think will be the final version. It combines HDR and normal captures in different regions of the picture.
In this particular image, there seems to be a noticeable improvement after combining
several exposures into a higher dynamic range file and then mapping back to a 16-bit image in a creative way.
I can’t get all of this detail across to you in an 8-bit JPEG on a WordPress blog, but at least you will be able to see the larger scale differences – if not the finest ones.
The lower light level in the foreground makes for very saturated colors and contrasty lines, while the background colors fade a bit as the higher intensity light washes saturation out and luminance levels shift to the higher end of the sensor range. The feeling of a retreating background is intensified a bit more as objects move further away from sharp focus towards hyperfocal.
Here is the original HDR image conversion using a Scott5 setting in CS6. This one doesn’t include the masked version of a more normally exposed image added to the background. The severely faded colors and some posterization are noticeable. It’s also lacking a little contrast in the foreground.
To me, the added HDR detail in the foreground is more interesting after increasing mid-tone contrast a bit, but even better after applying a mask and
reverting the background back to more normal contrast and colors, which make me even more comfortable with the overall picture.
I utilized a fine-tuned 2-image merge to align a couple of the original frames so I could then blend the close to normal exposure image into the background of the final HDR version. Without this extra automated photomerge step, the close to normal exposure and the HDR stack are slightly out of alignment when added to different layers for masking. I captured these images without a tripod, so you may not need this step if you do use one.
The photo below is the original +/-0 eV capture with only a little NR and sharpening applied. It is flatter looking and there is much less detail to observe in the shadows and highlights. That is the look I went for when setting up the custom HDR merge preset on the right.
The standard image just doesn’t communicate as much to me by itself. I guess I am starting to see the point of the hyper-real look that is taking over from more traditional image processing. At least for certain photographic subjects, it’s an interesting option.
I hope this helps someone out there trying to find their own HDR style.
Update on Dec. 29,2013.
Even though this scene is not the most promising one from an overall photographic perspective, I decided to go back at a different time of day to see if changing the angle of sunlight would help. As it turned out, and maybe not unexpectedly, finding better light seems to improve a picture much more than any HDR process.