Organic photo sensor

The stuff about 29stops DR is utter garbage though, using the same flaw logic the 1" Nikon 1 sensor or Sony rx100 sensor would have 28 stops DR. If you actually read the details they claim something like 20% increase in efficiency, so if say their old sensor had a quantum efficiency of 75% (which was about right Roth old penny sensors) then with their organic technology it would now have a QE of 90%, so on par with Nikon and Sony.
 
Is it no possible to stack sensors to get more dynamic range, or am i talking poo?
How much dynamic range has the human got compared to the best sensor out there?
 
Stacking would be expensive.
Instead you could do something like this and have infinite DR, although it does use up some die area, which means less ISO performance and more colour noise etc.

Perhaps this technique can be combined with the above technology to effectively create a truly amazing sensor.

With this level of DR, you could effectively have a mode in the camera that allows it to expose and arrange the shadows and highlights to match what our eyes see in a single frame (works with moving subject matter), without looking all fake HDR.
Of course sometimes it's preferable to blow out highlights etc, so having it as a configurable setting would be ideal.
 
Last edited:
gary996 said:
Is it no possible to stack sensors to get more dynamic range, or am i talking poo?
How much dynamic range has the human got compared to the best sensor out there?
Click to expand...

You can make the photo diodes have different sizes and sensitivities. Fuji used sells Nikon F- mount DSLRs that co,boned very small less sensitive pixels that would capture highlight detail and larger pixels for shadow to mid-range. Tis have a DR of like 13 stops which was 2-3 more than anything Nikon or canon could do. The downside is if sensor area is used for the small pixels then that area can't be used for the bigger pixels so the bigger pixels have less surface area than a traditional sensor if you start suffering from lower efficiency, worse high ISO noise, and increased shadow noise.


One possibility is the ADc could be swapped to use a Log amplifier. These are much more complex but allow far greater DR before saturation- the log response is much similar to the way film works and is why film preserves highlights so well.


The human eyes native DR might be like 15stops but the way human's view a scene by saccades (our eyes jump to a new feature of interest and focus there before ju,ping again) we adjust the exposure for the localslly and effectively increase the DR to something like 25 stops. An HDR photo from taken multiple exposures basically achieve the same thing.
 
gary996 said:
Interesting. So how long does one think until we see a huge jump in senors capability?
Click to expand...

probably never, the huge jumps have mostly already happened. The Nikon D3/D3S sensor gave a huge jump in high ISO capability, the Sony sensors have given a huge jump in shadow performance, Canon gave huge jumps in resolution while maintaining good performance in their earlier sensors.

The latest sensors are often nearing theoretical maximize performance. In the shadow bits at high ISO the pixels are acting as individual photon counters, quantum efficiencies are north of 90% now. Something like a Fovean design might bump us a little due to improved efficiency of the color filter, the Fuji design is going to be like a half stop improvement at the most.

Where there is most space for improvement is in the highlight headroom, the linear ADCs are getting saturated very early due to the small photon wells. Improved ADCs could increase highlight headroom substantially. But in terms of sensitivities I can only see slight increases through the generations. If you look at the Canon sensors then their basic efficiency, sensitivities and performance hasn't changed since around the 300/350D back in 2005. Sensor development is slowing dues to hard physical constraints, much like the CPU developments.
 
You must log in or register to reply here.
Back
Top Bottom