pedro said:
Hi,
along with the recent anncouncement of this high ISO monster cam I would just like to ask the experts here in the forum,
what you think could trickle down into an overnext 1Dx, 5Dx body...Might we see a crippled 1 Mio ISO 1DxIII version along with a 5DV?
If so, they surely are in the concept phase of such a beast, while preparing the roll outs of the 1DxII and a later 5DiV/X whatever...maybe the rumored 5DC could get its share of some of these new specs.
Or will this tech remain within the 30K USD realm?
http://nofilmschool.com/2015/07/canons-multi-purpose-ME20F-SH-full-frame-35mm-camera-4-million-iso
as pointed out, big pixels are the key.
However if a foveon (layered RGB) arrangement of pixels is ever employed then it becomes relatively trivial to bin pixels together before readout as all pixels will be adjacent to same colour pixels. This is exactly what I do with my astro CCD camera when I'm searching for a target.. I have ~60% quantum efficiecy mono camera with 8x8 binning giving me ~45um pixels, it means I can clearly see and identify an object in a 1 second image that my eyes cannot see at all.
The reason binning works is that all the signal from multiple pixels can be added together pretty much noislessly.. it's only reading the signal out that adds noise.
Alternative routes to reduce noise also exist.
Look up EEV (UK sensor manufacturer). They do L3 sensors amongst other things, the idea is to add a photomultiplier prior to the readout amp. This reduces the effect of readout noise. Also look up ANDOR.. they joint forces with some other companies and created sCMOS.
The idea of sCMOS was to take a slow semiconductor process and optimise it for sensors, then produce an ultra low noise high readout speed sensor.. which they've done.
The sensor uses an ADC per row of pixels. As there are so many ADCs they can afford to run very slowly and still get a very high frame rate. Slow running means low analogue bandwidth, noise is proportional to bandwidth, so that means low noise.
http://www.andor.com/scientific-cameras/neo-and-zyla-scmos-cameras
Awsome doesn't come close... the cameras are out of this world.. if measured in ISO I'm sure they be into the millions.
I did get a quote for one for astronomy, but it was about £8k and I couldn't justify it.. but if I were a professional I might well.
DSLR Sensors have a VERY long way to go.. even the Sony ones..