I'll be REAL interested in whether biasing toward the top part of the sensor "sandwich" - the blue layer - makes for better low light performance. WIth the existing Merrill series Sigmas, it's well known that you can do reasonably well in low light if you process the file in B&W - good to ISO 3200 and useable beyond. And the way you process the files to make it work is to emphasize the blue channel about 90% or more and drop the red and green channels to near zero. The top layer is more sensitive to light because the light hasn't had to travel through a layer or two of sensor before it lands on the part of the sensor in question. Sigma seems to be taking this a step farther with this model. They're putting more emphasis on the part of the sensor that was already the most sensitive and de-emphasizing the lower layers. This should clearly make low light that much better for B&W and I can't imagine they'd have done it this way if it had any real negative impact on color files either. So will color be better in low light too? The blue certainly should be but the red and green would arguably worse, both in absolute terms and relative to the blue layer. So I have no idea how that will work, but I'm sure the Sigma folks have been working on this for more than a few days, so I'm real interested in their rationale and, ultimately, how well it works in practice. These are fascinating cameras with great, but limited, performance already and seemingly all sorts of potential. I'll be real interested to see how they evolve, and what role this changed balance in the color layers will play in that evolution...
-Ray
