I pass this spot everyday on my way to work, and often times the sun rises just above it in the Summer months. Many times these sunrises are gorgeous but I never have my camera with me. Other times, like this one when I have my camera, they are decent and hold some value for photographing. However, I really did want to kick myself for not capturing the sunrise the day before, it was perfect and there was quite a bit of fog in the area which always adds an awesome element to any photograph.
I ran a little experiment with a series of brackets. For quite sometime now I have been trying to see if it is really worth capturing 5 EV brackets for every HDR image. I shoot them in RAW and if you shoot your HDR’s in RAW then you know there is quite a bit of information held in each one. Many times enough dynamic range can be squeezed from a single RAW file and you may not even need those extra brackets. Regardless, I have been firing away for the past few years with 5 EV brackets at +/1 EV apart.
I ran an experiment on the shot below, I processed all 5 RAW files (-2 EV, -1 EV, 0 EV, +1 EV, and +2 EV) and then 3 RAW files (-2 EV, 0 EV, +2 EV) from the same bracketed series. Surprisingly, I found very little difference from one image to the next. I processed both in the exact same way from Photomatix to Photoshop and actually found the 3 bracket HDR to hold a bit more information. Look at the ripples in the water, they are much more defined in the three bracket image.
I went ahead and ran the same experiment on the image below from Union Station. Both versions were treated the exact same way, from Photomatix to Photoshop. All I could really see was a slight bit more noise in the 3 exposure version. Not by much though, and I really had to zoom in (200%) to see it.
Example of the Noise seen:
Will I find a circumstance where the 5 exposures will reign supreme, beyond the scope of a slight noise difference? I am not 100% sure, but I will keep experimenting until I find one! For over 2 years I have been shooting 5 exposure brackets because I could, I never really ran an extensive experiment to see if it was really more effective than 3. I am starting to see now that maybe 5 exposures shot in RAW may be a bit excessive. I could potentially be wasting a bunch of time and space with those extra 2 RAW files.
Photo Tip For The Week:
Continue to experiment with your processes! The best time to do this is just when you think you have everything down pat. Throw a stick in your spoke and try something new. You could be doing things just to do things. Question everything, but be sure to refine your skills from the experiments that you do. Anyway you look at it, a little experimentation could never hurt!
For Blake, it's all about the art and process synergy. He dives deep into complex topics and makes them easy to understand through his outside-the-box thinking so that you can use these tricks in your workflow today!
Hi!
I am using the HDR technique for a while, and my experience tells me that it is always better to use 1eV steps rather than 2eV steps, so 5 brackets are definitely better in this case!
Another important aspect is that in case of ghosting you have more choice (you can choose between 5 shots and take the best for de-ghosting!)
Regards,
Phil
Thanks Phil, I didn’t even think about the de-ghosting options. That is a very good point!
Just started working with photomatrix and HDR. I have been using the 5 bracketed exposure method as well. Was considering doing a similar experiment but looks like you have answered the question, Thanks. May need to consider the situation based on Philippe Schoch’s comment though…as usual.
Experimenting is always the best idea. I think a spread of +/- 2EV’s at 3 exposure is plenty if you are shooting in RAW, however, that is assuming the 0 EV exposure is a good one, not over or under exposed. I shoot 5, for the buffer for Philipe’s reason and for the fact that I may not have had the proper key image (0 EV) exposure. Thanks for stopping by Gary!