HDR has proved to be something of a Pandora’s box – an appealing but rather powerful tool that can rapidly get out of hand and lead to photography on steroids. If you think about the premise of the technique, though, and keep in mind its purpose, it can be very useful in highly contrasting light conditions.
So what is it?
The dynamic range of a photograph is the extent of the image from its darkest point to its lightest. Taking a single image is often a compromise between exposing for the light areas and exposing for the darker, shadowy areas. HDR is simply combining two or more exposures to extend the dynamic range of an image: to pick up more tonal ranges in the dark areas and more subtleties in the bright areas that might otherwise burn out to white.
As I say, this can be useful in settings where there is a strong contrast between dark and light areas and an image light-balanced for a single shot would be too much of a compromise. An obvious example is the sky on a bright day, which can bleach to off-white if you have to expose for a subject in partial shadow. Another is when shooting interiors on a bright day, when light streaming in from the windows is too strong for a shady room. Do you expose for the room, the windows, or somewhere between? With HDR, you can get the interior right with one exposure, the windows right with another, then combine them to create a single harmonious image. That’s the principle; achieving it satisfactorily is a bit more of an art.
Shooting for HDR
First of all decide whether you need to shoot for HDR. In many situations it really isn’t necessary. Try taking the image with one exposure first and check on the histogram to see whether information is dropping off at the top or bottom end. If so, you might want to extend the range; but, equally, you might not. After all, there’s nothing wrong with shadows sinking into black and nothing wrong with bleached out areas either. This is the first pitfall of HDR: assuming that you have to use it because you can, or using it as a default approach instead of thinking about what you want to achieve before you click the shutter.
How many exposures should you take?
I’ve often read advice about taking seven or even nine shots as source material for the final HDR image. In my experience, this is overdoing it. To an extent, it depends on how you treat your exposures in post-production. I work manually and so want to limit the number of variations I’m juggling with in Photoshop layers. If you use automated functions, such as Merge to HDR, it’ll just take a bit longer to process if you use more frames, but there will be a massive amount of unnecessary overlap in the dynamic ranges you’re merging. If you are accustomed to using 7 images or more, try taking out a couple of them in the mid-range and see if it makes any difference. Typically I will mix three images. Occasionally I will use four or five exposures, but I can’t recall using more than that. I’d be genuinely interested to see examples where photographers have felt the need to use more.
A little aside on terminology
It’s worth mentioning at this juncture that high dynamic range could be considered a bit of a misnomer in this context, as once the multiple exposures have been merged into a single HDR image in Photoshop (that’s a 32-bit image), that image then has to be squeezed down into a 16-bit image to work on further – and that is essentially a low dynamic range version of the HDR image. I tend to think that this method could more accurately be called ‘extended dynamic range’ imaging, as the aim is precisely that: to extend the dynamic range of an image. However, the term EDR now seems to be used for fake HDR: that is, using a single frame, duping it several times, giving each a different tonal curve and then treating them like the source material to create an HDR image. This though, isn’t extending the dynamic range at all; it is merely manipulating the existing dynamic range – there is no extra data being brought to the party.
Having said that, there is a better variation of this technique called multi-RAW processing, which uses the hidden data in a single RAW file to create multiple versions of the same frame with different exposure levels. I haven’t tried this method, as I nearly always shoot with a tripod, pointing the camera at things that don’t move much, like buildings, but I will give it a go at some point.
And so to mixing the exposures
As briefly mentioned above, this can be done with an automated task such as Merge to HDR in Photoshop or with a plug-in such as Photomatix. Automated merging, though, tends to push photographs in a particular direction – very artificial, cartoonlike images. They can be impressive, certainly on first inspection, and some have a Pre-Raphaelite quality about them, but the appeal quickly palls as far as I’m concerned. If the first reaction when looking at an image is “oh, it’s one of those HDR photographs”, then you’re surely faced with a triumph of style over content.
The method I prefer is to take control over how you wish to deploy your exposure variants. And the way to do that is to work with them as layers in Photoshop, subtly mixing them through masks and channels, and deciding exactly where you want to bring in more tonal variation and where you don’t. It is more time-consuming, but the results have greater realism. To me, the manual approach simply offers a way to render an image that is closer to what we perceive in a setting – a result of the eye’s pupil rapidly reacting to the contrasts of light and dark to create a unified overall impression.
Here’s an example
This was a private pew in a church at Langley Parish in Berkshire. Think of it like a private box at a football or cricket ground. The first image shows the exposure that I felt best expressed the scene as I saw it. It was clear, however, that I would need more exposures to address the very bright areas, so I bracketed with about 7 shots, then in post-production chose just three to work with. I felt this gave me sufficient tonal values to work with for this subject.
In the image below are broadly the key three areas that required a darker exposure.
The next picture shows those corresponding areas in an image shot at a lower ev. I brought the whole of this image in as a separate layer over the main image, then made a mask in order to mix the images using the brush tool. Approximately in the areas highlighted, the darker image was overlaid at 100%; elsewhere the blend would have reduced to 5% or less, and, of course, in some parts to nothing at all. The beauty of working by hand with different exposures is that you can choose at every point on the image the levels of exposure you want to use.
By this stage I would have had a more coherent image, but there was still detail lost in a few key areas, in particular the sculpted relief visible through the open doorway on the right of the picture and the painted door in the room at the end of the corridor. So I chose an image with an even lower ev and selectively brought in three more layers. Again, for each I made a mask and subtly used the layer masks to bring in more detail to those specific areas.
It was a very wonky room, but in the final image I straightened it a little, just so that it wasn’t too prominent a factor – because photographs are usually viewed in a squared up frame, that tends to stress deviations from straight planes. The image was destined for print, so the final post-production applied was a little vibrancy and saturation, and some sharpening. An automated approach to the HDR would have brought lighter tones into the shadows, but I wanted to keep a lot of dark corners in this image and, personally, I think working manually with layer masks is the best way of achieving that.