All HDR is the result of post-processing... the question is whether you want to let the camera decide how to process it vs. you deciding how to process it on the computer. I prefer having the control.
When you use built-in HDR or exposure bracketing (for cameras that have it) most cameras just grab 3 shots... the normal, plus 1 darker and 1 brigher exposure. But for some scenes you may find you really want more than 3. When I shot the total eclipse of the Sun this past August, I captured 12 exposure stops.
Every camera that has a manual exposure mode (which is every DSLR ever made) can capture the source data for HDR just as well as any other camera. This is because the original method to collect the HDR data was to take a number of exposures and manual adjust the exposure between each shot. Cameras with auto-bracketing do this for you... but often limit it to just 3 shots. You may want more than 3 shots.
To determine how many shots you need, you would scan the scene with your eyes and try to find the darkest area you can find, and also find the brightest area you can find. You would then meter both those areas and determine how many exposures difference there is between them. The metering can be a bit tricky because most cameras don’t let you spot meter to a particularly tight area. High-end hand-held light meters sometimes have a spot-meter tool that lets them pick off a 1° circle to meter. But most meters are quite a bit wider than that ... and with a DSLR the size of that area depends on the focal length of the lens. So the tightest area you meter could be quite wide and that means “bright” data is in that circle where you are just trying to meter “dark” data.
To get around that problem (without having to spend the money on a Sekonic L-858D series light meter) just take a test shot and check the histogram in the camera. When you’re taking the “bright” exposure it is because you want to make sure you aren’t “clipping” the dark end of the histogram (the bright area will be clipped). So you look at the in-camera histogram for the shot and make sure it doesn’t show any data up against the left side of the graph. Similarly, test the dark exposure and make sure the histogram shows that no data is clipped in the highlights... there should be no data up against the right side of the graph. If your histogram shows you got all the darks and all the brights (in two different shots) with no clipping... then your range of exposures will be adequate (even if you don’t have a $600 light meter).
It’s also generally not required that shoot every “stop” of exposure between your dark and bright end of the range... but you probably should grab every 2nd stop.
Suppose I had an extreme range where I need the “normal” exposure plus 4 stops darker and also plus 4 stops brighter. I can actually shoot the normal shot, then shoot a frame 2 stops darker, then another frame 4 stops darker (relative to the middle exposure), then 2 stops brigher and 4 stops brighter (so that’s a total of 5 images... -4, -2, 0, +2, +4... to get the range.) It’s usually not necessary to get every 1 stop of exposure difference. While most HDR software has no problem if you skip a stop and just every 2nd stop... it probably will struggle if you skip 2 stops and just grab every 3rd stop. So I don't recommend pushing it too much.
Usually you don’t need that much range... often times grabbing 3 frames (-2, 0, +2) will be sufficient ... and for those shots, the in-camera Auto-Exposure Bracketing feature will do the trick. Just be aware that it wont always do the trick and you may have to manually adjust the exposure between each frame when you need more range.
Just remember that the histogram is your go-to tool to confirm you really grabbed enough data. If you're not completely comfortable with how to read the histogram, then you'd want to study up on that topic.
Lastly, a tip...
When I started shooting HDR and keeping the source data, it started to get confusing to figure out which frames go with which other frames. I’d shoot a subject and grab my frames for HDR... then maybe change something and want to grab several more frames... and so on. You get back to your computer and you import the data and now you’ve got a ton of frames that all look the same and you’re not quite sure WHICH sets were all taken as part of the same bracketed sequence.
A cheap trick (since we no longer have to pay for film) is to grab a dummy frame between each sequence... stick your hand in front of the lens and that can serve to separate the sequences. It just makes it easier to sort through the data and grab the right frames that go together when you’re trying to run it through your HDR software.
Tim Campbell
5D III, 5D IV, 60Da
