The megapixels wars and dynamic range

click photo to enlarge
The rise in the megapixel count of compact cameras has become ridiculous. Models are routinely being sold boasting sensors with 14MP, and consumers are buying them thinking that they offer better image quality than the models with a lower number. Manufacturers know this isn't the case. Enthusiast and professional photographers do too. But, the camera makers, years ago, embarked on the "more is better" method of selling new models and seem unable (or unwilling) to depart from it. This has also started to infect DSLRs: witness the concerns over the Canon 50D (15MP) and the 7D (18MP).

Anyone who is confused by all this needs to know that the main factors in image quality are the lens (just as it always has been) and the processor that works on the image in the camera, but that possibly the most important factor is the density of pixels on the sensor. The lens on an inexpensive camera is usually fairly basic, but given that it's small is usually adequate. In-camera processors improve year on year, so that the images from this year's 10MP camera are likely to be better than the equivalent sensor/MP size of three years ago. Now, what about pixel density? Here are some examples from cameras available today:

Samsung TL240 - 14.1MP - 50MP/sq.cm.
Canon Powershot S90 - 10MP - 23MP/sq.cm.
Olympus E510 DSLR - 10MP - 4.1MP/sq.cm.
Pentax K-7 DSLR - 14.6MP - 4.0MP/sq.cm.
Nikon D3S DSLR - 12.1MP - 1.4MP/sq.cm.
Go to this page of DPReview to find out the pixel density of your camera.

The best image quality in that list will come from the Nikon (the one with the biggest sensor) and the worst from the camera with the highest megapixel count, the Samsung (which has the smallest sensor.) I could have chosen 5 different cameras and, by and large, the image quality would have been directly related to the sensor size and pixel density. Of course, when people ignore the megapixel count they tend to buy within a price range, and that's when choice and comparisons become more complicated. However, the truth is that usable detail, dynamic range, visual noise and all the other factors that make for a good image is closely related to the sensor size and pixel density, not the number of megapixels. In fact, most compact consumer digital cameras would today be producing better images if the manufacturers had stopped at 8MP.

Today's photograph was taken with my LX3 (10MP, 24MP/sq.cm.) and illustrates the second theme in today's "reflection." I'd trade dynamic range for megapixels any day. This shot was difficult for the camera because of the dark shadows of the street near the bright, sunlit clouds. I had to set the EV to -1.33 to control the highlights, and that introduced quite a bit more noise into the shadows which I had to clean up afterwards at the expense of detail. If digital camera manufacturers put less effort into increasing the number of megapixels and more into enabling their products to record the detail in both bright highlights and darker shadows, then images would be sharper and closer to what our eye/brain sees. Cameras that could do this would be easier to expose, and, most helpful of all, images would require less post-processing.

photograph & text (c) T. Boughen

Camera: Lumix LX3
Mode: Aperture Priority
Focal Length: 7.9mm (37mm/35mm equiv.)
F No: f5
Shutter Speed: 1/800
ISO: 80
Exposure Compensation: -1.33 EV
Image Stabilisation: On

Looking back No. 2

click photo to enlarge

High Dynamic Range (HDR) is one of the current buzz phrases in digital photography. Concisely put, it's a method of combining a number of shots of the same subject into a composite image with an extended range of values between light and dark areas. Usually three images are combined: one "underexposed" (to benefit from the highlight detail), one "overexposed" to use the shadow detail, and one "correctly" exposed (to supply the mid-tones). The resulting image is said to more closely approximate what the eye and the brain "sees".

In the hands of a skilled and thoughtful photographer this is broadly what happens. However, others use it, wittingly and unwittingly, to produce "hyper-realistic" images that have the unreal glow that characterises a certain type of naive painting. These can be interesting, but have become gimmicky, and any interest soon wanes when you've seen a lot of them. And over the last couple of years I've seen many! But a thought occurred to me a while ago, when I did my own HDR shot (above) - the Victorian painters of the Pre-Raphaelite Brotherhood (PRB) would have loved this technique! As I sat back and looked at my completed photograph I was reminded of the finish (not the subject necessarily) of their radiant images that lovingly depict the minutiae of the English countryside; paintings such as "A Study in March" by John William Inchbold (1830-1888), or "Our English Coasts (Strayed Sheep)", by William Holman Hunt (1827 - 1910). In the years since they flourished the PRB have ridden the roller coaster of critical acclaim, with the dips being every bit as extreme as the heights. The interest in HDR shows some signs of having reached its own summit of popularity, and I hope that instead of being dispensed with it will settle down to become a useful, and occasional (!), technique in the photographer's armoury.

My photograph shows the church of St Thomas A Becket, in the village of Aunsby, Lincolnshire, taken last autumn during my break from posting. This building, erected mainly between the 1100s and the 1500s, is one of the many medieval masterpieces that await the visitor to this eastern county. I came upon it part way through a walk that took me to a succession of such village churches. My shot was taken with the camera programmed to record three images simultaneously, each with different levels of exposure.

photograph & text (c) T. Boughen

Camera: Olympus E510
Mode: Aperture Priority
Focal Length: 14mm (28mm/35mm equiv.)
F No: f7.1
Shutter Speed: 1/400
ISO: 100
Exposure Compensation: -1.0, 0, +1.0EV
Image Stabilisation: Off