See inside:

5D2 was about 2 stops better (in noise) than 5D classical (each pixel is better, and almost twice as many pixels), resulting from 3 years of research/improvement.

Now, it is almost 4 years since 5D2.  So Canon should have enough time to make the same progress that it made back when developing 5D2.  So that is "at least 2 stops".  1DX appears to be 3 stops better at the pixel level when compared to 5D2, but it has a slightly less pixels, so over all, it is less than 3 stops better (if each pixel is 3 stops better).

所有跟帖： 

• 你太樂觀了 -貪吃的豬豬- ♂ (159 bytes) () 02/27/2012 postreply 10:28:29

• 對，就是這樣。至少是我希望這樣。 -pieq314- ♀ (57 bytes) () 02/27/2012 postreply 10:50:56

• 這上頭沒有摩爾定律，一個階段的技術進步飽和很快 -U96- ♂ (633 bytes) () 02/27/2012 postreply 11:27:14

• Moore定律起作用的例子：Nokia發布41M的手機，比D800還多：） -貪吃的豬豬- ♂ (39 bytes) () 02/27/2012 postreply 11:42:49

• 那是扯O得沒邊了。郵票大小整的東西和小指甲蓋大小的有啥可比性？ -U96- ♂ (0 bytes) () 02/27/2012 postreply 11:49:06

• digitial zoom and hyperfocal distance? -貪吃的豬豬- ♂ (35 bytes) () 02/27/2012 postreply 11:54:56

• 謝 pieq 兄 -enTiger744- ♂ (0 bytes) () 02/27/2012 postreply 10:40:14

• 多支持2擋, 不一定好2擋... -華翰齋- ♂ (274 bytes) () 02/27/2012 postreply 11:17:25

• 回複：多支持2擋, 不一定好2擋... -貪吃的豬豬- ♂ (172 bytes) () 02/27/2012 postreply 11:34:30

• 不能隻看noise per pixel -pieq314- ♀ (327 bytes) () 02/27/2012 postreply 11:39:26

• 派兄這個比較靠譜! -華翰齋- ♂ (33 bytes) () 02/27/2012 postreply 12:21:53

• Come on ，華師這回來隻新煙吧 。。。 -enTiger744- ♂ (0 bytes) () 02/27/2012 postreply 12:54:20