Canon succeeded in developing an extra-large CMOS sensor.
->
PRESS RELEASE: Canon succeeds in developing world's largest CMOS image sensor, with ultra-high sensitivity
It has 48x the surface of a 35mm full frame.
While Canon in the above press release, congratulates the arrival of nocturnal photography and video (with no doubt a great idea
), I'll have two comments to share:
1. Canon please stop the press of how technically challenging and expensive it is to produce a full frame sensor. It is increasingly getting ridiculous only exhibiting Canon's agenda (don't cannibalize the FF market).
The fact that the above sensor was produced (certainly by stitching the reticles) is an indication how good the yield in CMOS full frame sensor production already is (even if they wasted 1000 300mm wafers to get one good sensor (a 200mm wafer would have been too small), it must be better than 0.001^(1/48) or 86%. And if Canon could really afford to waste the equivalent of 66,000 FF sensors for this project then this is quite a statement too... More likely, they wasted only 100 wafers though. Thanks Canon for publishing these figures
2. One should realize that this great monster sensor would have to go with fast lenses in order not to kill it's light advantage. The appropriate (not equivalent!) lens to replace a 50/1.4 lens is a 300/1.4 lens with more than 200mm diameter and a 300mm image circle. Such a lens is easily designed but may actually be much more expensive to produce than the sensor
E.g., the Rodenstock Apo Sironar 300mm view camera lens for 320mm image circle (8"x10") are f/5.6. Then there is the 4 element lens Schneider-Kreuznach Xenar 300mm f/4.5 for 8"x10", basically from 1919 and maybe still for order. There seem to be no wider lenses.
Therefore, assuming f/4 is a
reasonable limit at the given image circle, the equivalent full frame lens would be 50mm f/0.67 . In a world where 50mm f/0.95 lenses do exist for full frame, Canon's monster sensor just gives half a stop more light. Hurra!!
I see a slight chance that there is military money involved here and the sensor project may have been accompanied by a large lens project yielding a solution for nocturnal field observation.
Last edited by falconeye; 08-31-2010 at 07:10 AM.