• 
• 

Tuesday, June 5, 2012

Misinformation: Camera Tech

Image resolution and pixel size are just a few of the many factors in the image equation

The common understanding is that the more megapixels you have in a camera, the more image resolution you'll have in an image, which translates not only to larger prints, but also to better imaging quality. During the megapixel wars, camera manufacturers latched onto the potential sales behind this concept and began to inflate megapixel counts by packing miniaturized photodiodes onto the sensor. While this resulted in more "megapixels," ironically, it also worsened image quality due to the smaller receptive sites for photons and the exponential gaps created between these diodes as they were added to sensors that stayed the same relative size.
Obviously, when you start to see compact cameras with the same amount of megapixels as professional DSLRs, something has gone wrong. On the same-sized sensor, more pixels equals more resolution, but the tradeoff is less bit depth and a lower signal-to-noise ratio. Conversely, a lower pixel count with larger photodiodes should offer a cleaner image at a loss in resolution. But now it's looking as if things may not be that simple. Canon's EOS 5D Mark III, for instance, contains 6.25-micron photosites, which are actually smaller than the 6.4-micron-sized photodiodes of its predecessor, the 5D Mark II. Canon says that despite this the camera delivers better imaging quality and less noise thanks to the new DIGIC 5+ processing engine (17 times faster than the DIGIC 4) and other enhancements, including improved photoelectric conversion efficiency, enhanced on-chip noise reduction and gapless microlenses over the photodiodes that collect light information from the entire surface of the sensor.

Canon claims an increase of approximately two stops of sensitivity, and DxOMark (www.dxomark.com), a company that independently measures and compares camera sensors and lenses, reflected this update in its review of the Mark III with an estimated Quantum Efficiency jump (the amount of detectable light hitting the sensor) from 33% in the Mark II to 59% in the Mark III. (This translates to roughly half a stop.) Still, on DxOMark's Sensor Overall Score, which averages sensor performance based on color depth, dynamic range and low-light ISO, the 22.3-megapixel 5D Mark III scores an overall score of 81 while the 36.3-megapixel Nikon D800 achieved a remarkable score of 95 points, a number that DxO says surpasses even the highest-scoring medium-format camera it has tested, the Phase One IQ180, by nearly a 1⁄3-stop. The Nikon D800's pixel pitch measures a diminutive 4.7 microns.

While DxOMark's proprietary measurements are open to criticism, the numbers point to a topic that we've covered many times in the past. Specs don't tell the whole story, and at the same time, a chain is only as strong as its weakest link. DxOMark is careful to explain that its Sensor Overall Score is independent of resolution, which is decided in a camera not only by the inner mechanics, but the optical abilities of both the camera and the imaging optics that it employs (i.e., lenses). The number of pixels certainly will affect output size, but large amounts of pixels don't guarantee better image quality, nor do larger photosites. The ability of a camera to collect photon information is the combination of these factors, plus numerous other points along the image-making pipeline, like sensor size, design and construction, as well as analog-to-digital conversion methods and optics. Regardless, with both cameras landing in the top 10 of DxOMark's Sensor Overall Score, 2012 is already proving to be an exciting year for camera technology.