Why does M8 gather so much light in the shadows?

[colorflow]

Originally Posted by rosuna

The M8 has a sensor with 6.8 microns fotosites. The real size of the "cells" are very near of 6.8*6.8 squares, because the sensor is a full frame transfer CCD. The same goes for other KAF Kodak sensors (used in digital backs) or Dalsa sensors. When Canon presents a camera with 6.8 microns spacing, the "cells" are actually smaller, because the sensors are of CMOS type (they have circuits at the surface). The same goes for inter-line CCDs (typical of compact cameras). Exact data from the manufacturers are very difficult to get.

Rubén Osuna Guerrero
.

I think the fact that CCD sensors have deeper 'wells' in the silicon than CMOS may have something to do with it? (CCD sensors require much deeper penetration by the dopants in silicon than CMOS) Perhaps a higher signal to noise ratio in low light situations resulting in better shadow details?

I am no expert in sensor technology but know a little about how semiconductors are made.

Alan

[deltoid1]

Originally Posted by rosuna

The LFI articles suggests, at the end, the possibility of 10-bit or 12-bit non-linear compressed files as better options. The sensor of the Leica and the A/D converter seems to be superb. It is a shame that a 8-bit compression is used for storing the files, even if it is of non-linear type. The argument is that "there is no practical difference".

There isn't explicit information about the "fill factor" variable from manufacturers. Dalsa has serveral interesting pages at their website. Kodak had an interesting "documents repository", now dissapeared. Canon has interesting (and long, detailed) white papers about sensors, etc.

The M8 has a sensor with 6.8 microns fotosites. The real size of the "cells" are very near of 6.8*6.8 squares, because the sensor is a full frame transfer CCD. The same goes for other KAF Kodak sensors (used in digital backs) or Dalsa sensors. When Canon presents a camera with 6.8 microns spacing, the "cells" are actually smaller, because the sensors are of CMOS type (they have circuits at the surface). The same goes for inter-line CCDs (typical of compact cameras). Exact data from the manufacturers are very difficult to get.

Rubén Osuna Guerrero
.

What about the importance of the area of the microlenses?

I will say that I have seen files from another 14bit Kodak sensor, the one in the Olympus E1. Excellent tonality.

[sandymc]

The microlenses do expand the effective area of the cell, although not with 100% efficiency - so even if the microlens cover the full theoretical area of the cell, it may only effectively gather 80% or whatever of the light. Microlenses also impact of the angular acceptance and have some slight anti-aliasing effect.

[wparsonsgisnet]

Originally Posted by carstenw

... Bill, I am very curious about your portrait work. Would you tell more?

Carsten, I'm going to start a thread because I have some comments about gnc flash.