Wednesday, April 20, 2016
Global Shutter Pixel Comparison
Point Grey publishes a nice comparison of Sony and Aptina global shutter sensor, including the latest Sony 3.45um pixel size IMX250 and IMX265. One can see them compared with 5.86um pixel in Sony IMX174 and IMX249 and Aptina's 3.75um pixel in AR0134:
Subscribe to:
Post Comments (Atom)
Great comparison based on numbers ! Unfortunately one important parameter of the GS sensors is missing : PLS or parasitic light sensitivity.
ReplyDeleteSONY should be the best since they have re-used all the techniques inside IT-CCD.
DeleteAlbert, Point Grey uses EMVA1288 for most of their measurements and PLS is part of version 3.2 that is not yet released.
DeleteA very important figure of merit is SNR, it is a curve and not a number, maximum SNR means SNR at an operating point that is not practically used (saturation). The SNR curve is very important to show the light level required to become shot noise limited and it is a pitty that it is not presented.
DeleteThe data is otherwise very interesting.
Is PLS the same as shutter efficiency?
DeletePLS is reverse of shutter efficiency. Sony never showed their PLS for smaller pixel, 5.86um pixel is 1/10-6, but for small pixel?
DeleteAnd Aptina's fullwell is very small, why?
I understand shutter efficiency as something like (Sensitivity-PLS)/Sensitivity
DeleteNot necessarily, though they are expected to be correlated. If you have your PPD overflowing with dark current into the storage area, it's bad shutter efficiency but not PLS. If you have a light sensitive storage area, it can still be shut off efficiently from the PPD and prevent (e.g.) blooming from the PPD to the storage.
DeleteSony uses always P-on-N subtrat, so the overflow absorption should be very good.
DeleteArnaud, do the math. Everything you need for shot noise limited SNR point is there...Those aren't multi exposure or other DR extended sensors ;)
ReplyDeleteThere is the data to fit a model, but the publication is about measurements and graphical comparison, i only said that the SNR plots would have been nice instead of the saturation SNR.
DeleteEric - PLS stands for Parasitic Light Sensitivity. It is basically QE of the storage node divided by QE of the photodiode. Some like to quote the inverse of that ratio when referring to PLS. Most define shutter efficiency as, GSE = 1 - PLS.
ReplyDeleteYou can also express the PLS or GSE like this : 1 : 10,000.
ReplyDeleteIt means that for 10,000 electrons generated in the Photodiode, 1 is generated in the memory node. This one is simpler I think.
Then GSE is 10,000/(10,000 + 1)or 0.9999 = 99.99%.
Delete