Tag Archives: digital photography techniques

  • High Dynamic Range Photography Techniques

    This article is adapted from a presentation I recently gave to the Southern Oregon Photographic Association.

    High Dynamic Range (HDR) photography is a type of photography that has seen a rapid growth in popularity recently. There are Flickr groups dedicated to it and it comes up regularly in photography forums across the web. In its most recent incarnation HDR photography involves combining multiple exposures of a scene using HDR imaging software, the most popular of which is Photomatix by HDRsoft. Just do a Google search for HDR photography or Photomatix and you can see thousands of examples of Photomatix HDR images.  The image below is typical of a lot of HDR images you will see out there. Some are technically excellent, some are uniquely artistic, some are captivating and some are just plain bad.

    The purpose of HDR photography is to create an image of a scene that has a greater range of light values than can be normally be captured by a single film or digital image. Depending on how it is used, HDR imaging software can allow the photographer to create an image that is closer to what the human eye perceives when looking at a high dynamic range scene. However, more and more it is being used to create surreal, whimsical or fantastical photo illustrations that don't necessarily reflect the human experience but fall into a whole new genre of photographic art (check out the Flickr group).

    I often use the basic concepts of HDR imaging in my work because I like to photograph in very extreme lighting situations that are awesome to experience in person, but impossible to capture with a simple click of the camera shutter. Sometimes I use HDR techniques to create images that are pure fantasy. Most often my goal is to attempt to recreate my personal experience in a way that is more accurate than my camera can record. Therefore, much of my HDR photography does not have the stereotypical cartoonish, haloed, surreal look that you will find in a lot of the current HDR images being made. I also do not normally use Photomatix or other HDR software to create my HDR photos. Rather, I employ a variety of other techniques first, depending on the situation, and turn to software as a final option.

    Before I get further into this discussion of high dynamic range imaging allow me to back up a bit and put a little context on the term "dynamic range". Most simply put, dynamic range, as it applies to photography, is the difference between the maximum (brightest) and minimum (darkest) measurable light intensities for a given scene, film, sensor, screen, etc. This range in photography is often referred to in terms of stops. A stop equals a doubling of the amount of light. A one stop increase in light results in double the light intensity and two stops results in four times the light and so on. The following dynamic range values are open to much debate among photographers and a lot of different values can be found depending on who you read or talk to.  So to avoid getting an argument started, let me point out that I'm using values I borrowed from other sources and only use them as simple reference points, not as the final word on what the exact values may or may not be.

    The human eye is said to have a dynamic range of as much as 24 stops. This would indicate that from the darkest shadow detail the eye can see, to the brightest highlight, the light intensity can double 24 times. That's a bit misleading however, because as we scan a scene our eyes adjust rapidly to darker and brighter areas and our brain quickly composites the various information into what seems like a single image. In reality, if a person is to look at one part of a scene without scanning (this is closer to how a camera does it and is called instantaneous dynamic range) , the dynamic range of the eye is something closer to 14 stops. However, we normally perceive a scene with the greater effective dynamic range because we naturally scan with our eyes.

    By comparison, slide film has a dynamic range of around 5 stops, print film around 7 stops and the data contained within a quality digital RAW file closer to 10 stops. The average scene you might photograph has a dynamic range of about 10 stops but the kinds of extreme lighting scenes that I like to photograph are often in that 14 to 24 stop range of human vision, and sometimes even beyond.  So the problem is that much of what we photograph is within the range of what our eyes can see, but outside the ability of a camera (film or digital) to record all at once. The result is that we often take images that either have shadows that are totally black (underexposed), highlights that are totally white (overexposed) or a combination of both. This can be really frustrating when you remember seeing an amazing sunset over a luminescent ocean, but your photos show only dark, muddy waves and a featureless white sky. Enter the need for HDR photography.

    HDR photography is basically any technique for creating an image that contains a higher dynamic range than can normally be captured in a single frame in a camera (either film or digital). There are actually many ways to extend the dynamic range of a photograph , but a great number of digital photographers are now going strait to the computer software solution. While there are some great advantages to the "Photomatix" approach, I'll point out some reasons why it might be worth it to have some other tricks in the HDR tool box as well.

    So, what are some of the different ways to create higher dynamic range images? The ones I'm most familiar with include:

    1. darkroom techniques (available to traditional film shooters with serious darkroom skills)
    2. graduated filters
    3. careful processing of a RAW image file
    4. double processing a single RAW image file and manually blending the two versions together using photo editing software
    5. manually blending multiple bracketed exposures taken in the camera
    6. using HDR imaging software to automatically blend a set of bracketed exposures.

    As it turns out, HDR photography is almost as old as photography itself (there is some good historical information on Wikipedia). In the late 1800's photographers developed darkroom techniques that enabled them to create single images from two or more separately exposed negatives. This allowed them to produce a positive print that maintained the right exposure in both a bright sky and a dark foreground. Later, Ansel Adams and friends developed the Zone System which was designed to identify the best exposure for any scene that would later allow for the greatest degree of tonal information to be recovered in print. While not truly HDR because he was still working within the limits of a single exposure, his techniques did greatly improve the ability to get a high level of dynamic range from the negative to the print.

    Graduated filters are perhaps the least post production intensive way to create HDR images. A graduated filter is generally a piece of glass or plastic that has a gradient on it that fades from dark (usually neutral gray unless a color shift is desired) to clear. The dark half of the filter is usually placed over the brighter sky with the transition lining up with the horizon and the clear portion over the darker foreground. The darker portion of the filter "holds back" some of the light (usually 1, 2 or 3 stops worth) while the clear portion lets all the light through from the darker foreground. While this doesn't actually extend the dynamic range of the film or digital sensor in the camera, it does help balance a high dynamic range scene so that all the tonal information can be captured within the range of the camera (see example below). The biggest advantage to using graduated filters is that everything happens in-camera so there is no additional post processing time required. All darkroom and computer based HDR techniques require anywhere from several minutes to many hours of additional time. On the down side, graduated filters have a straight transition line that is very difficult to fit on a scene that doesn't have a straight horizon, often resulting in a dark "grad line" cutting across trees, mountains, buildings or any other dark objects that project above the horizon.

    Shooting and processing RAW image files (most current digital SLRs have the ability to shoot RAW files as well as jpegs) isn't technically an HDR technique because it only uses the light captured in a single exposure. However, there is so much data captured in the RAW format that during processing using a RAW converter (many exist but most cameras come with RAW software.   Programs like Adobe Lightroom or Adobe Photoshop also support RAW processing) it is possible to recover both shadow and highlight detail that appeared to be outside the dynamic range when the image was taken. The degree of latitude, often one or more stops in either direction, can be enough to create a final image that has a much greater dynamic range than what the camera appeared to capture (see example below / click to enlarge). The advantages to this are that it is fast, easy and all done with a single exposure. One downside is that even RAW files are limited to how much dynamic range they can contain, so it only works to a point. Another is that trying to recover too much detail from deep shadows in this way can result in a high degree of digital noise.

    Sometimes, I find that a single RAW file contains all the tonal range that is needed for a properly exposed image, but I'm not able to get the entire image to look good with a single RAW conversion. In these cases I will do what is called double processing the RAW file to create two different exposures and then manually blend the exposures using various blending techniques in Photoshop. Exposure blending is a tricky art/skill in itself that takes a lot of practice to do well. Alone, it could be the focus of a multi-day workshop (I'll have to work on offering one), so there isn't time to go into depth here. I have a very basic post on the topic HERE. In the examples below, the first image was RAW processed to have the desired sky exposure. The second is from the same RAW image file, this time processed for the desired foreground exposure.  The third is the result of blending the two exposures in Photoshop. This technique works great as long as the entire dynamic range of the scene is contained within the RAW file. For that reason this technically isn't a true HDR technique either, but the results can be a much greater dynamic range than could be achieved by single processing the RAW file and way beyond anything that could be done with a jpeg capture or film.

    When the dynamic range of a scene is too great to be reigned in with graduated filters or to contain within a single RAW capture, I apply the manually blended multiple exposure technique. The blending techniques I use here are the same as the ones I use with the double processed RAW file technique, and again are a topic for a multi-day workshop. The advantage of this approach is that I can take as many exposures in the camera as I need to contain the complete dynamic range of a scene. Using this technique I can easily take a series of exposures that can contain the full dynamic range of just about any situation, including the full 24 stops of human perception and even beyond. The blending process for combining two or more different exposures is tedious and can take hours or even days, and if not done well the result is obvious and bad. So, I only use it on images that I think are going to be great. There are many advantages. Most important to me is that this technique produces the cleanest, sharpest, artifact free, data rich final images. Another advantage is that I have complete creative and local control over how I blend the exposures and what I want the final image to look like, allowing me to get as close as possible to what I envisioned when I took the photo. In the example below, the dynamic range between the sky and the foreground was way beyond the 10 stops I could contain in a single RAW file. I didn't want to use graduated filters so I could avoid the additional flare they can cause when facing into bright light and because of the irregular shape of the lighthouse projecting above the horizon. I bracketed four different exposures, but was able to create my final image by blending together just the darkest and lightest of them. I think the final result is not only appealing and believable, but also very high quality and noise free which means it looks great as a large print as well as on the computer screen.

    Finally we come to the technique that a vast majority of HDR jockeys are using these days, and that is to combine the tonal values of multiple exposures using HDR imaging software, most popularly Photomatix. The technique for capture is the same as the previous technique, namely shooting on a tripod to avoid camera movement and bracketing two or more exposures that can contain the entire dynamic range of a scene, often in one stop increments. Some HDR images contain as many as seven different exposures or more. Using Photomatix is much quicker than manually blending, usually taking about five minutes or so. The advantages are that when it works well, it can do an amazing job and can sometimes handle tricky areas that are particularly difficult to manually blend. It can be a huge time saver and, as noted earlier, can produce some amazing surreal effects and artistic styles that can't be obtained any other way. However, I usually only go to Photomatix as a last resort because, like with anything, there are trade offs. The controls in Photomatix are fairly blunt, making it difficult to make fine adjustments. Also, I find that what I see in the tone mapping (also a topic for another time) preview window is not always what the image ends up looking like once it has been rendered in its final form. Often times the resulting HDR image has strange color shifts, increased shadow noise, cartoonish colors, halos, edge fringing and an overall loss of contrast. Also, anything that moves in the scene from one exposure to the next creates weird ghost images in the final photo. In addition, the adjustments in Photomatix are only global, so any local adjusting that I need to do has to be done back in Photoshop, not saving me much time in the end.

    In the following example I first blended four exposures of an old house using Photomatix to create the resulting HDR image. The software did a pretty good job in some areas. The image certainly contains a wide dynamic range of tonal values, but I wasn't happy with the color in some areas, the overall contrast or the way that the movement of the clouds between exposures created problems in the final image. In addition, some areas of the sky were still overexposed. I fully admit that my skills using Photomatix are somewhat lacking, but I often find that no matter what I do, some images just don't succeed.

    Photomatix HDR image

    Not satisfied with the result, I decided to spend the time working on a manually blended HDR image. It took a lot of painstaking work, but I found that the resulting image was much cleaner, with better overall dynamic range, color, contrast and sky detail. The manually blended image also has close to zero digital noise, edge fringing or artifacting making it a much higher quality file for creating fine prints.

    Manual blend

    Manual blend

    The next photo (Double Falls) is of another high dynamic range scene. With a bright sky and very dark canyon I knew it couldn't be handled with a single exposure. I shot several exposures and then generated an HDR image using Photomatix. I felt that in this case the program did a pretty respectable job. However, I decided I would also try my hand at a manual blend so I could do a direct comparison. Even at web resolution, my hand processed HDR has greater tonal contrast and density than the Photomatix image. I was also able to make creative decisions about localized luminosity, color and contrast that I could not with Photomatix.

    While it is possible to get away with some digital artifacts and pixel degradation in web sized images, my large fine art prints have to be as sharp, clean and noise free as possible. Upon closer inspection, I noticed that the Photomatix image had much more shadow noise than the hand processed image, as well as strange color shifts in the water and areas that were still blown out.

    Click to enlarge

    The Photomatix HDR image also experienced edge fringing and loss of sky detail where clouds had moved between exposures.

    Click to enlarge

    So, in conclusion, I find that having multiple techniques to capture and create images when confronted with high dynamic range situations allows me to better express my visual experience and artistic vision through my photographs. Often the use of graduated filters or shooting and carefully processing my images in the RAW format is all I need to properly render the dynamic range of a scene that is just a little greater than average. As the dynamic range of the scene increases I begin to employ more agressive HDR techniques, such as double processing and blending a single RAW file or bracketing multiple exposures in the camera and blending two or more of these by hand. Sometimes I even use combinations of these techniques. Does HDR imaging software have its place? Absolutely. I choose to use it when time is a factor, when the images will only be shown at lower web resolution so noise and fringing isn't as important, when I want to create something that has that very stylized "HDR look" or when nothing else I have tried works. In addition, I'm sure that HDR imaging software will continue to improve and advance. In the not too distant future, I fully expect to see digital cameras that have HDR sensors which will be able to capture 15, 20, even 25 stops of light or more in a single frame, which will make much of this discussion obsolete.

    I'd love to hear other tips, techniques, opinions and experiences regarding HDR imaging, so please feel free to leave a reply or share this post.

1 Item(s)

per page