In the wake of CCD discontinuation news,
Adimec blog starts a series of posts comparing them with CMOS sensors. Adimec looks at QE/Read Noise metric and notes that today's CMOS sensors became better than CCDs, while in a similar 2011 comparison, CCDs were the winner:
 |
| QE/RN in 2011 |
 |
| QE/RN in 2015 |
I don't understand why QE/RN should be the right metric to compare imaging sensors, a 10% increase of QE is not equivalent to a 10% decrease of the read noise.
ReplyDeleteThere are 2 use cases:
High light: exposure time is limited by highlights clipping, so a 10% increase of QE only leads to a shorter exposure time, no benefit in this case, while a 10% decrease of read noise improves dynamic range.
Low light (or sports): exposure time is limited by motion blur, a 10% increase in QE will improve the shot noise SNR by 5%, while a 10% decrease in read noise will only impact the darkest features.