“HDR” stands for “High Dynamic Range” and, in simple terms, it is a method of generating image files that contain very much more information on the brightness levels in a scene than a camera’s sensor can detect in a single exposure. It allows us to capture an image in such a way that we record at least as much brightness information as your eyes can detect.
Before we talk about high dynamic range, it’s probably worth spending a minute or two defining “dynamic range”. Dynamic range refers to the ability of a sensor to differentiate different levels of a signal. The more distinct levels the sensor can measure, the higher the dynamic range of that sensor. Dynamic range can be expressed in several different ways:
EVsThis measurement is popular in photography where EVs (Exposure Values) are a measure of light level. Each change of 1EV is equivalent to a 1 stop change in your exposure which, in turn, equates to a halfing or doubling of the light intensity. In absolute terms, EV0 is a light level which will give a correctly exposed image with a 1 second exposure on a 100ISO medium with an f1.0 lens. Dynamic range expressed in EVs indicates the difference between the exposures necessary to correctly expose the darkest shadow and the brightest highlight in a scene.BitsIn the digital world, dynamic range can also be expressed in terms of bits or binary digits. The number of bits used to describe a value equates to the dynamic range since adding an extra digit doubles the total number of values you can represent. For example a 1 digit binary number can describe exactly 2 levels, 0 and 1, whereas a 2 digit number can represent 4 levels, 00, 01, 10 and 11 (or 0, 1, 2 and 3 in decimal).
The observant will notice that all three methods of measuring dynamic range all boil down to the same thing and, fortunately for us photographers, it is also pretty obvious that 1 bit in the digital definition equates to 1EV in traditional photographic terms since 1EV represents a factor of 2 change in the light level and 1 bit represents a factor of 2 change in the number of levels that can be measured.
Now that we’ve introduced some of the terms, let’s consider some real-world dynamic range examples.
Digital image sensors capture an image in terms of the brightness levels of red, green and blue light and the very best DSLR sensors just now return 14 bits of information per colour. A 14 bit value can represent 2 to the power 14 or 16384 different numbers so, theoretically a 14 bit sensor can distinguish 16384 different brightness levels and has a dynamic range of 1:16384 or 14EV. In real life, however, sensor noise reduces this a fair bit and you probably get something more like 4096 (12 bits, 12EV) or 8192 (13 bits, 13EV) levels. Remember, though, that this is the BEST sensors on the market. Most cameras top out at around 10 bits or 1024 levels and, if you are unfortunately enough to still shoot JPEG, you are limiting yourself to 8 bits or 256 levels.
So why is this important? The main issue here is that your eyes do a far better job of recording a scene than most cameras can. Without your iris having to contract or expand, you can see about 14EV of brightness levels and, allowing for the way your eye adjusts to different lighting conditions, you can handle a total range about 27 or 28EV and still be able to see. On a sunny day, you have no problem resolving detail in dark shadows at the same time as in full sun. In the same situation, your camera (with your help) has to make a decision – it can’t possibly resolve all that difference in brightness. Does it expose to capture the shadow detail and blow out all the highlights or expose for the highlights and have all the shadows detail fall into the black?
Shooting for HDR helps us round this problem. By capturing several images at different exposures we can capture all the brightness information in the image. “Underexposed” images capture the far highlight detail, “normal exposures” capture the detail in the midrange and “overexposed” images capture shadow detail. Special software is then used to merge those different exposures back together into a single image that contains 32 bits of information (4 billion levels) per colour and captures a huge amount more dynamic range (brightness levels) than you can achieve in a single exposure.
By this time, I’m sure some readers are wondering what this has to do with all those whacky looking pictures on Flickr. Creating an HDR image is only the first part of the process. No monitor or printer (OK, practically no monitor or printer – there are professional monitors that can display true HDR but these are not within the budget of your normal hobby photographer) can display the amount of brightness information in a true HDR image so you have to perform a second step to generate a viewable image from the HDR. This is called “Tone Mapping” and is the operation that gives HDR such a bad name in many people’s eyes.
Given a single HDR image, it is possible to tone map it in hundreds (thousands?) of different ways resulting in very different results. Some people tone map for artistic effect and generate images which look unnatural, others use settings which result in very natural images. The choice is, however, completely up to the person working the image. Personally, I do both though you probably only notice that an image started off as an HDR when I’ve gone the artistic route. For example, the image to the right is an HDR though most people don’t realise it (unless, of course, they read the title). The point to note here is that HDR is a technique and you can’t equate it to “weird images with cartoon colours and nasty haloes”. What you are seeing is someone’s attempt to convert an HDR back into something that can be displayed on a monitor or printed to a printer and, frankly, there are (in my humble opinion) a lot of people who do this in a way that is not particularly appealing. That said, I should point out tone mapped images do tend to polarise viewers – one of my images recently appeared on a blog called “I Hate Your HDR“. While it’s not my favourite, it was good enough to post to Flickr so it just goes to show – one person’s art is another person’s Recycle Bin fodder.
Having got some of the background out of the way, if you are interested in finding out some more practical information, please head on over to part 2 of the tutorial, “Shooting for HDR“.