Excuse my ignorance ( as filmy ).... but, the SD10 has 3+ M pix in 3 colours . So image quality / resollution..... of 3+ M pix, how would it compare to a D70 ?

Confused.

Has anyone suggested housing this camera?

I think it is a bit of an apples to ornges comparison. www.dpreview.com has a good review of this camera, they talk a bit about it. Some have suggested this to be about a 10 mp camera, but I don't think it is a straightforward comparison.

Rick Decker posted a note to the UW-Photo list over the weekend saying he just started using one in an Ikelite housing. See:

http://www.pbase.com/rickdecker/underwater_with_sd910

Mike

"Some" people would be the only the manufacturer itself.

I've heard it compared favorably to a 6MP traditional sensor.

I also don't like that it's called the SD10. I have a SD10, and it's a little Canon P&S.

Right, I just think it is an oversimplification to say this. In this case the 10 mp are larger and "stacked" on top of each other in each color channel. Megapixels aren't everything. I've seen some comparisions of Nikon D2H (3 MP) and D70 (6MP) both Bayer chips with the same size senso zoomed in and you can't tell the difference. Obviously there is more to the picture.



Rick's photos look great. I particularly like the composition and color in the wreck photos. Any details on his setup? I didn't see any housings available for either of the Sigmas on digideep. It was for this reason that I didn't consider the Sigma. Maybe its an older Ikelite housing that just fit?

Also-he states that he used Ikelite and Sea and Sea strobes. How has he done this? I thought Ikelite doesn't have a cable to connect his strobes to Sea and Sea? Was he directly connecting the Ikelite strobe and using the Sea and Sea as a slave?

so, not as easy as it looks then, and it looks like its not 10 MP anyway. I mean, if you have 3 different 3+ MPcolours stacked ontop of each other, it dont make 10 MP. Might give better colour depth etc, but if you look at the sample pics against some other recent camera's, it dont seem so great either in that respect.

Hi caveman,

the truth is somewhere between.

The Sigma SD9 and SD10 deliver 3 MP. That’s it. No more. Like a Slide Scanner which gives a certain resolution with all 3 colour channels (RGB). My Nikon Scanner (> 10MP) would never be called a > 30MP Scanner because it scans true RGB. The only valid statement is: the SD10 has 10 million photo elements for capture.

A Canon 10D for example has 6MP (exactly 6,3). It also only has 6 million photo elements.

The Foveon chip (SD10) uses 3 layers of 3 megapixel sensors. One for each colour.
The others (Canon, etc.) use 1 layer for the whole sensor area. You need at least 3 pixel to calculate/interpolate true RGB. So you could say: “Well, a 10D is practically only a true 2 MP camera?”.

No. There is one important fact. The human eye is able to distinguish very small differences in luminance in a photo. And that is what the Bayer (Canon, Nikon, etc.) method is build on. The human eye is not able to distinguish colour differences in the same high resolution like luminance differences.

To get a value to compare these methods and cameras:

For colour photos you can take 2/3 of the resolution of a Bayer chip to get the effective resolution. Would give you about 4MP for a 10D, 300D, D70 and D100.
The Sigma gives exactly 3MP (but they are of course very good ones).

The Bayer advantage grows if you intend to do a black and white shot as the colour item doesn’t count anymore.

Cheers,

Julian

one more point: the 2/3 rule depends on the image content.

if half of the photo is deep big blue, the bayer chip has completely no effective colour resolution reduction. the 2/3 value is just theoretically for colour details which come to the resolution limit.

Thanks....thats clears up alot......