Where Do D-SLRs Go From Here?
Where Do D-SLRs Go From Here?
by George Schaub
Having edited articles about or done tests myself on the latest round of digital single-lens reflex cameras I continue to be impressed with the strides made in technology, image quality and innovation these camera represent. They offer two modes of operation--an almost fully automatic ways to get quite good pictures, and levels of complexity that will challenge even the most accomplished photographer. Instruction books have become quite long, hinting at the capabilities they contain that few photographers will fully explore. Image processing continues to be at the forefront of these advances, with the image being massaged for noise reduction, sharpness and color quality even before it hits the memory card. But there are other matters to consider that might give a hint as to where these cameras go from here.
The steady climb in megapixel counts have not necessarily been accompanied by larger sensor sizes. The equation is generally that the more pixels you fit onto the sensor the more noisy the image might become, due to less efficiency in light gathering. Yet, we continue to see higher and higher megapixel counts accompanied by even higher ISO capabilities. The compromise has to come from somewhere, and that somewhere is in image processing. Many cameras now automatically apply noise reduction at ISOs of 400 and above, and many use it in the processing at all ISO speed settings.
Despite this seeming contradiction I can honestly report that images have become less noisy than ever before, due to advances in image processing. I have been testing digital cameras since they started and when opening older image files I am somewhat horrified at the amount of noise those "older" images contain. Having been at this since ISO 400 speed color negative film was first announced the analogy holds between how that film was improved to what's been happening in the digital realm. At one point ISO 400 film started yielding results comparable to ISO 100 film, and even faster films flowed from those improvements.
But as with film the best results still come from shooting at the lowest ISO setting you can for the lighting conditions at hand. You will not see much of a difference when the image is printed small, but when you dig in the shadows, especially if you try to open them up, you'll see the noise quite clearly at ISO 800 and above. Lots of folks go to higher ISO settings simply because they are there. They use it as a way to avoid camera shake or to go deep in depth of field. Faster lenses, tripod shooting and even image stabilization might be a better course than always reaching for that high ISO setting, despite the amazing gains in higher ISO image-processing quality.
More and more D-SLRs are offering Live View, and regular readers of my reviews know that I am not a huge fan of this procedure for general shooting. Live View is great for indoor, tripod work and close-ups. It also has an advantage when the LCD can be articulated to offer waist level or over the head shooting when finder viewing is impossible or impractical. The posture it makes you work with in general shooting, however, is a sure-fire way to add the potential for camera shake in your images. The general wisdom is that Live View was added as a way to seduce the point and shooter into D-SLR-dom.
My main gripe with Live View is that the LCD is, in many cameras, fairly unreadable in bright light or unless you figure out some way to shade it to make it more readable. There has been much chatter and rationalization on the web about poor LCD performance, with some of it implying that real men don't need to see what they are actually framing in the scene. This is nonsense, and perhaps the best improvement camera makers can make in future D-SLRs (and other cameras as well) is delivering a screen that makes Live View make sense. OLED-type screens seemed promising, but we are told that their expense, and the fact that they are more subject to screen burn and fade after a year or two of use, has held back their incorporation in digital cameras. If this has changed I am eager to stand corrected.
This fascinating technology has done much to allow photographers to make images that would have required tripod shooting in the past, something great for general shooting, although mounting a camera on a sturdy tripod or monopod will generally yield better results than without. This technology is increasingly incorporated in "budget" D-SLRs, and while laudable it is generally utilized to allow the kit lens to be quite "slow." I have seen mid-range tele zooms in kits with f/5.6 max aperture at the longest focal length (which can be as short as 80mm in 35mm format terms), which is a lens no one would accept without image stabilization to make it worthwhile even in normal light.
The debate between in-camera and in-lens image stabilization advantages is a worthy one, which I will pursue in assignments to our writers in future issues of the magazine. Either way, however, for image stabilization to really make sense, my advice is that it be used with a fast lens to begin with, and not a lens whose aperture range brings you back to where you should have started in the first place.
The curse of the ancient photographer is contrast and how to control it. Digital sensors are quite sensitive to overexposure, due to the nature of how photons are converted to electrons and how the sensor cuts off (clips) hot highlights. HDR, or high-dynamic range imaging is the hot-button topic these days, with some shooters going so far as to do high speed, continuous bracketing handheld to get all the tonal values possible recorded and then merge them later in software.
One possible solution is the automatic gain controls we first saw in some digicams earlier this year, where the processor analyzes the brightness values in a scene and automatically applies gain where needed and holds back ISO enhancement where it is not, sort of an in-camera burn and dodge. While the complexities of this have yet to be fully worked through it does point the way toward expanding the dynamic range, albeit in a way that does not count on the sensor but more on the processor to do the work. But these days, and into the future, we might not be able to differentiate between the two. More and more, lens quality, exposure and even those functions that we now do in software after exposure will be handled even before the image hits the memory card.