David Pogue's column in today's NY Times takes on the megapixel controversy. He conducted tests in an effort to show that an increase in megapixels doesn't necessarily yield better results. The column is aimed at consumers and the heavy emphasis on megapixels in P&S marketing. His first test was with a 13 megapixel camera. He doesn't specify which one. He took a shot of a toddler, then downsized it to 8 megapixels and 5 megapixels in PhotoShop. he then had 16 x24 prints made. They were digital C prints, printed on a Durst Lambda at 400 dpi. He displayed the prints in Union Square and asked volunteers to try to determine which was which: lo-res, medium-res and hi-res. One of twelve viewers got it right. He published the results on his blog and received a number of angry letters that claimed the results invalid because the lo-res images were derived from the hi-res by downsizing, rather than being shot lo-res. So he devised another test. He reasoned, quite correctly in my opinion, th at he couldn't use different cameras, because the results would be skewed by other factors. A Canon pro came up with another method. He suggested using a Caonon 1DS 16.7 megapixel camera to take the picture and effectively reducing megapixels by zooming out and cropping. Pogue agreed and they repeated the test. They produced three shots at different focal lengths and cropped two of them to 10 and 7 megapixels respectively. The longest focal length image was kept at 16.7 megapixels. Again they made three 16 x 24 prints on the Durst Lambda. This time they displayed them in a library and had 50 people evaluate them. Only three could differentiate the various resolutions. Pogue ackowledges the value of higher resolution for cropping and understands that on a large sensor it can be a plus. But his point is that with tiny P&S sensors, 5 megapixels may well yield results equal to or better than 8 megapixels. But we knew that. One has to wonder how many experienced photographers would be a ble to differentiate between the three prints? I guess we'll never know. Paul
-- PDML Pentax-Discuss Mail List PDML@pdml.net http://pdml.net/mailman/listinfo/pdml_pdml.net
