Hi
People squabble all the time over the Foveon sensor resolution relative to conventional sensors, not helped by Sigma's marketing approach. As an owner of a SD9, SD14 and a DP1 as well a number of cameras from the traditional makers, I've been able to make my own mind up.
And the answer is.....
.... Foveon sensors compare very reasonably for resolution with Bayer sensors of around twice the pixel count. The absence of both the OLPF AA filter and the need for colour demosaicing means a Foveon is pixel-for-pixel a lot sharper. They take up-ressing very well.
- In practice, I cannot see any resolution difference even with the most extreme pixel peeping, between my 3.4MP X3 SD9 and 8MP Canon 350D.
- I sometimes think I see a tiny difference between the resolution of the 4.6MP X3 DP1 and SD14 and my 12MP 450D and GF1. But we are really talking tiny differences and it might be imagination or just differences in the rendering of different parts of the subject.
Info about the SD1 is rather incomplete at the moment but there have been some snippets from people close to the product. The upcoming SD1 sensor is completely different from previous Foveons, a whole new chip architecture built on lessons they learned from years of research trying to build tiny phone sensors. The word is Sigma seem pleased with the sensor and think they have made a breakthrough, a couple of stops better on noise. Likely still behind the best Bayers but an improvement over the early Foveons.
Based on the 3x and a bit hike in pixel count, and my previous experience, as well as the comments of those who have seen sample prints at shows, resolution will be roughly around that of a 30-35MP Bayer camera (not the 45MP Sigma claim which is really the photosite count, not pixel count). The SD1 should be the highest resolving 35mm style DSLR, for a little while at least.
We shall see, hopefully fairly soon (June maybe?). My advice would still be to take all this with a pinch of salt for now, let's wait and see what the experts think when the camera arrives.
Things to consider beyond resolution:
The Foveon sensor uses the natural filtering properties of the silicon rather than organic dye filters to separate colour. This gives it an unusual colour response that varies model to model and raw converter to raw converter (and even camera to camera) as Sigma try and find the balance that suits people. I have found with my cameras, it is best to use trial and error to find the set of adjustments that suits me and save them as a preset.
The Foveon doesn't use an AA filter. This contributes to its unusual pixel level sharpness at the cost of occasional luminance aliasing artifacts in the form of jaggies mainly. Soft lenses, diffraction, slight camera shake etc will all reduce or remove these in much the same way as they do with medium format backs but it is something to be aware of (occasionally). The extra resolution of the SD1 should help a great deal.
Noise has been a problem in the past. The new sensor is reportedly dramatically improved. I'll wait and see on that.
Past Sigma bodies have been interesting. The SD1 looks like a big step up market but a big price hike that goes with it. For the sort of money they have been hinting (7D pricing), you will probably get the best Sigma body yet but no where near normal feature sets (no dust shaker, no live view, no video, no in body anti shake, modest AF, no top plate LCD, low FPS, slowish buffering and modest buffer size). You will get a metal body, AF micro adjustment, weather seals and two new style command wheels and a somewhat better rear LCD than the older bodies
. The bodies have always had good grips and feel reasonable to hold.
Sigma pricing has always been optimistically high at intro as they milk the Foveon fundies but then drops catastrophically after a few months. Things might be different with the SD1, who can say.
