jrista said:
3kramd5 said:
bdunbar79 said:
1. Will the back-illumination only really help at high ISO?
ISO setting is associated to how much amplification you need to apply to get a certain luminance, right (Neuro can likely quote the
standard)?
Light is light. If your pixels are blocked by circuitry, you need more amplification to achieve a given luminance than if they aren't. So, low ISO settings should have less associated noise due to less required gain. It may be approaching diminishing returns with Sony's architecture at low ISO since noise is so low already, though. I suspect Canon would see greater benefits from such an approach.
Consider all the above as an uneducated guess.
The amount of amplification required is based on how much light you gather and convert to charge. The two primary factors that affect that are quantum efficiency and pixel area. With BSI, pixel area is literally maximized. The entire sensor surface area is sensitive to light with a BSI design...fill factor would be around 99%. So yes, absolutely, BIS will have a meaningful impact to high ISO performance. It had a meaningful impact to high ISO performance with the Samsung NX1, which has scored higher than the 7D II in high ISO tests thanks to it's BSI APS-C sensor.
The benefit here, when BSI is combined with Sony's already superior sensor technology, is that you can get both excellent low ISO performance as well as excellent high ISO performance, with small pixels, in the same camera. No need to make a tradeoff for one or the other.
Interesting is also the stacked sensor on the 1 inch models, which sony calls the next step after BSI where the circuitry and the photo sensitive components are completely separate that combined with the on-chip RAM to speed up readout rates and allegedly produces less noise and less rolling shutter than standard BSI CMOS.
These are just some of the things we have to look forward in full frame. I think those out there that believe current CMOS technology is as good as it is going to get because canon hasn't lead the pack for years are failing to consider that many improvements yet to make it to the full frame format. I recall the same arguments being made years ago about how sensor development would stop... and yet we still don't have enough DR, we still have rolling shutter issues, we still have noise even at low ISOs, we still have bayer artifacts, we still have moire on video or pixel binning problems, and a host other set of issues which are areas to improve both in software and in sensor.
It seems one thing is clear: sony's RD arm isn't satisfied with "good enough CMOS" when it comes to their sensors and I for one applaud them because companies that lead down the road are companies that are always looking to push the edge.