HDR or High Dynamic Range photography has received a bad reputation in recent years. A quick Google search shows some pretty stinky examples of questionable HDR! Whether HDR is used to create a more natural image or for artistic purposes and effect, there’s little doubt that it can be a very useful way of including more dynamic range in photos. But before we get to Tone Mapping HDR and alternatves like Exposure Fusion, here’s a brief look at dynamic range and how light is recorded.
Latitude in Film and Digital
Modern digital cameras are great for many reasons. As much as I love film, the qualities of film, and the physicality of old cameras, digital imaging maintains certain benefits that I find useful each and every week. As great as digital sensors are today, they are still limited in terms of what they can produce. It is often said by snobbish film photographers that colour-negative film specifically has a wide latitude. But what do these film hipsters mean when they say that? To look at what this means, it’s actually useful to look at how film compares with digital sensors.
Film is made up of chemicals on a polymer substrate. These chemicals are structured as photo-sensitive grains that record light as the shutter is opened up. Digital camera sensors are composed of tiny photosites that are all crammed onto a thin wafer. When light hits the sensor, each photosite transfers an analog signal containing light and colour values to the Analog to Digital Converter (ADC). The digital information is then fed to the software in your camera and an image is made. This is a very simple summary, and I am certainly no expert on the matter!
When film enthusiasts speak of film having a wide latitude, what they really mean is that the chemical emulsion is able to record a wider range of lights and darks in a scene. This means that highlights don’t blow out as easily, and shadows don’t block out as much. In pracfice, it means that you can take your Kodak colour negative film and overexpose it by two or even three stops and still retain plenty of detail. This is great news for those that worry endlessly about highlight detail (I don’t)!
Digital sensors, on the other hand, aren’t quite so good at preserving highlight detail in particular. To illlustrate this, let us imagine each individual digital photosite as a small bucket and the light entering the camera as water. Each bucket can only contain a certain amount of water. When it fills up, the excess water gushes over the top and splashes into neighbouring buckets.
In photography, dynamic range is the difference between the lowest light intensities (shadows and blacks) and the highest light intensities (highlights). When a digital camera meters a scene, it is measuring the value of light in the scene. If you’re not using Center Weighted metering or Spot Metering, you’re probably using something like Matrix Metering which simply averages the light intensity of any given scene. Most of the time, this results in pretty accurate final exposure settings, but in scenes with lots of highlights, lots of darks, or lots of lights and darks with a large dynamic range value (very high contrast scenes; back lit scenes for example), the Meter will likely be fooled. Meters like this always try to average out the light intensity to 18% grey; or middle grey. The thinking is that this will give accurate enough results for most scenes. Of course, when it comes to scenes with lots of dynamic range, all digital camera meters struggle! This is sometimes why photographers use Center Weighted or Spot Metering.
When you combine a high contrast scene with the flaws in light metering and those little digital photosite buckets that fill up quickly, you often end up with photos that are either too light or too dark. This is why beach photos on a really sunny day often look dull if you leave the camera on auto. In cases like this, too much or too little light has hit the digital sensor and you end up with results that don’t look much like the scene your eyes saw. This is because our eyes and brain are far better at accounting for high dynamic range scenes and can evaluate areas of micro-contrast much better and adjust to them very quickly.
Here’s a photo from a really bright day. You can see the leaves are in shadow, but the building has highlight areas that look flat and blown out:
This is not what my eyes saw on the day. I remember that the foreground leaves were less shadowy and the building had more detail. The post processing is less than ideal, but it serves as a good example.
Exposure Fusion as an alternative to Tone Mapping HDR
So, given the limits of meters and sensors, how can we include plenty of dynamic range in a photo without losing too much detail? Software like Photomatix contains some complex algorithms aimed at those who practice High Dynamic Range photography. In use, one would combine several photos of the same scene made at different exposure settings (-2, 0 and +2 stops for example), and the HDR software would then cleverly combine them into one photo where the lightest parts and the darkest parts have more detail. You can see for yourself in the Google link I provided earlier just how far this process can be taken!
Whilst I am no HDR afficionado or expert, I like to dabble in all sorts of photography and processes. In this spirit, I have been experimenting with a HDR process called Exposure Fusion. What this does is combine different exposures of a single scene into a final scene where the highest and lowest light intensities have been compressed. This means that shadow and dark areas have been raised, and bright and higlight areas have been lowered. Exactly the same thing can be done with skill, time, patience and masks in Photoshop, but Exposure Fusion provides a way of doing it that saves a heck of a lot of time. Typically, it results in a file that is much easier to work with.
Since I rarely use a tripod, I don’t often have the opportunity to make several photos of a scene at different exposure values (this is called Exposure Bracketing). This means that all of my experiementation with Exposure Fusion so far has been done with single files. My workflow looks like this: Process a single RAW file in Lightroom and create two virtual copies of it. Underexposue one copy by two full stops and overexapose the other copy by two full stops. The three photos are then combined into a single TIFF with the Exposure Fusion plugin for Lightroom. The resulting TIFF is automatically re-imported to Lightroom for further tweaking (you will likely need to tweak).
As an example, I combined the following three exposures using the Exposure Fusion for Lightroom plugin:
I made sure to set the white balance of the original photo and also sharpened in Lightroom. Before overexposing and underexposing the file, Ithink it is also useful to raise shadows and drop highlights closer to the middle of the histogram without the result looking too unnatural if possible. Exposure Fusion will use the original exposure as a reference image and include areas of it if the light intensity value is either too dark or too light in the other exposures, so you will want to make sure the original looks reasonable before feeding it to the algorithms.
Once the Exposure Fusion plugin had done its job, I ended up with a re-imported TIFF that looked like this:
When comparing it with the opriginal exposure, you can see how the Exposure Fusion software has preserved sky detail, and even the details of the metal structure of the glass house inside the building! It does look a little flat because on default settings Exposure Fusion prioritises exposure over contrast.
I increased Clarity, Contrast and Vibrance, and ended up with this final image:
The image really pops now! Though one can get similar results by using masks in Photoshop or Brushes in Lightroom, in many cases the process would be painstaking. By combining the ‘best’ bits of different exposures, Exposure Fusion offers a really interesting – and more natural I think – alternative to regular HDR programs.
Remember that photo of the blown out and flat looking building I included earlier in this post? I tried Exposure Fusion on it as well and got the following great looking result: