First Report on CIS Reproducibility, Variability and Reliability
Albert Theuwissen releases the first report on "Reproducibility, Variability and Reliability of CIS" in a 5-year series. The 175-page report contains 118 figures and 98 tables with data on QE, FPN, DSNU, Qsat, DR, SNR, DC, and more.
Albert, for global shutter sensors missing important parameters: Global Shutter Efficiency (across wavelength preferable, or at least in visual and 850nm/950nm NIR) Dark current need separation to PD dark current and storage dark current since they contribute differently to image quality. Thank you, SV
Dear SV, thanks for your comments. Basically you are 100 % right, but I am working with commercially available cameras and they do not allow me to change the timing of the sensors (except for the exposure time). That makes the characterization of shutter efficiency and localization of the dark current not simple. Specifically for the dark current : in this project the absolute values of the dark current are of less importance, but the relative values are. I am investigating how the dark current is changing from device to device, over time and in harsh conditions. I do agree that it is still important to know where exactly the dark current is changing (if it does !) : PPD, SG or FD. If in the future any dark current change is significant, I will consider to put more effort (= Euro) into this part of the puzzle.
Albert, Separate dark current for PD and storage are very important. Most of the applications with GS pixels use very short PD exposure times thus generating little PD dark current. However, accumulated storage dark current is much higher, especially for bottom rows that would be readout in about frame time. FD based GS pixel arrays produce horrible image quality at elevated temperatures typical for tightly packaged camera modules. Exactly due to the fact that FD storage has high dark current and variability. Regards, SV
Hi SV, you are 100 % right. But the cameras/sensors I use in my tests do not use the FD as the storage node, but they do have a separate storage node between the PPD and the FD. That storage node has a very low dark current, I do not observe any shading in dark current, even at 0 seconds exposure time. You can convince yourself by getting the report at your desk ;-)
SV, you are right. But the sensors that I am testing do not use the FD as the storage node. They do have a separate storage node (in the charge domain) between the PPD and the FD. This storage node has a very low dark current. Even in images shot in dark at 0 seconds exposure time, I do not observe any dark current shading. You can verify yourself, but then you have to acquire the report of course ;-)
Albert, for global shutter sensors missing important parameters:
ReplyDeleteGlobal Shutter Efficiency (across wavelength preferable, or at least in visual and 850nm/950nm NIR)
Dark current need separation to PD dark current and storage dark current since they contribute differently to image quality.
Thank you,
SV
Dear SV, thanks for your comments.
ReplyDeleteBasically you are 100 % right, but I am working with commercially available cameras and they do not allow me to change the timing of the sensors (except for the exposure time). That makes the characterization of shutter efficiency and localization of the dark current not simple.
Specifically for the dark current : in this project the absolute values of the dark current are of less importance, but the relative values are. I am investigating how the dark current is changing from device to device, over time and in harsh conditions. I do agree that it is still important to know where exactly the dark current is changing (if it does !) : PPD, SG or FD. If in the future any dark current change is significant, I will consider to put more effort (= Euro) into this part of the puzzle.
Albert,
ReplyDeleteSeparate dark current for PD and storage are very important. Most of the applications with GS pixels use very short PD exposure times thus generating little PD dark current. However, accumulated storage dark current is much higher, especially for bottom rows that would be readout in about frame time. FD based GS pixel arrays produce horrible image quality at elevated temperatures typical for tightly packaged camera modules. Exactly due to the fact that FD storage has high dark current and variability.
Regards,
SV
Hi SV, you are 100 % right. But the cameras/sensors I use in my tests do not use the FD as the storage node, but they do have a separate storage node between the PPD and the FD. That storage node has a very low dark current, I do not observe any shading in dark current, even at 0 seconds exposure time. You can convince yourself by getting the report at your desk ;-)
DeleteSV, you are right. But the sensors that I am testing do not use the FD as the storage node. They do have a separate storage node (in the charge domain) between the PPD and the FD. This storage node has a very low dark current. Even in images shot in dark at 0 seconds exposure time, I do not observe any dark current shading. You can verify yourself, but then you have to acquire the report of course ;-)
Delete