Canon caused a ruckus among digital-camera enthusiasts when it announced this fall that its new PowerShot G7 lacked support for "raw" images. But the company now has offered an explanation for the move: increasing the number of megapixels led to more noise per pixel and meant raw was no better than JPEG.
The G lineage aims to please a more discerning crowd than the average point-and-shooter, and earlier members of the family could produce the raw files that many photographers like for their greater flexibility and subtle tones. When Canon dropped raw support in the G7, reviewers complained loudly, some speculating that the company was trying to protect sales and profits for its more expensive digital SLR (single-lens reflex) line.
"If the G7 had raw mode, I would buy one in a flash. As I wrote in this article's subhead--It Could Have Been a Contender!" said the Luminous Landscape, speculating that Canon was trying to protect profit margins on SLR cameras that still feature raw support. "Many features that made the G series stand out have been removed," including raw mode, said Digital Photography Review. Digital Camera Resources called the move a "dumbfounding downgrade," and CNET's own review opined, "Whoever at Canon decided to jettison raw-format support deserves a whack upside the head."
The PowerShot G7 has a 10-megapixel sensor compared to 7.1 megapixels for the G6. But the sensor itself is the same size, meaning that individual pixels in the sensor are smaller. Smaller pixels means it's harder to distinguish the signal from the incoming light from the random electronic noise in the sensor, said Chuck Westfall, Canon's director of media and customer relations.
"The net result is that even if the G7 offered raw image capture...there would be no discernible improvement in image quality compared to...superfine JPEG mode," Westfall said.
More specifically, a pixel on a Canon Rebel XTi SLR is 5.7 microns across, and on higher-end models such as the 5D with "full-frame" sensors the size of 35 mm film negatives, pixels are 8.2 microns wide, Westfall said. (A micron is a millionth of a meter.) A G7's pixels are less than 2 microns wide, in comparison, and therefore produce more noise and are worse at discerning subtle brightness differences.
"Do the math for surface area, and you'll see how much of a difference there really is," Westfall said.
The greater the surface area, the more light a pixel on the sensor can detect. For the record, here's the math: assuming pixels are circular, a 2-micron diameter means a surface area of 3 square microns. A 5.7-micron pixel has 26 square microns of surface area, able to capture 8 times the amount of light as a 2-micron pixel. And an 8.2-micron pixel has 53 square microns of area, able to capture 17 times the amount of light as a 2-micron pixel.
Update: This entry has been modified to reflect the fact that the 5D and the 1D Mark II N don't have the same sensor size. The 8.2-micron pixel size of Canon's 1D Mark II N is the same as that of the 5D, Westfall said. But the 1D Mark II N's sensor smaller than the full-frame sensor used in the 1Ds Mark II or the 5D.