Latest sigma foveon cameras

Lenses (particularly the centre) often apear 50% sharper on 35mm sensors compared to APSC sensors.. Why?

35mm sensors are around 1 stop less diffraction limited than APSC.. why?

Supposedly (if what luminous landscapes says is correct), Foveon technology can make use of 50% more information (resolution) from then lens. Bayer systems apparently throw away the luminance values from red or blue sites, this basically also throws away information from the lens. Due to this, I don't think the Bayer design is very efficient.

I'd imagine the confusion point for you is you seem to think that every lens and sensor combination has some untapped potential because the sensor is Bayer instead Foveon. When actually, as soon as you hit that 18MP mark on a crop, or the 30MP marks that we're hitting now on full frame cameras, you're capturing pretty much all the detail that any lens can project onto that sensor. You'll get a couple that are sharper than that, but pixel density is growing at a rate that far exceeds lens detail and as soon as 36MP or so cameras are very much the norm, there's all of a sudden no advantage whatsoever in a Foveon design, assuming all Bayer types move towards the Nikon low-pass design or similar. It's only an advantage in an equivalent low resolution situation, say up to about 15MP, as soon as you go beyond that it's pointless.
 
^^^ I agree with this as well, mostly what I was saying. you can push a Bayer design to a high pixel density and capture all the detail there is to get out of the lens, you then might want to down-sample to get suitable edge definition. But there is definitely more detail available in lenses than current sensors are recording. Just look at the pixel densities of compact camera, even with their poor lenses they can extract more detail over a given sensor area.

A lot comes down to deciding if you want the 100% crop to be pin sharp o you don't mind having a higher MP file that is not quite pin sharp at 100% view but in total contains more detail, which is visible when printed to big sizes.

Either-way the Foveon design is not a miracle cure, but quite interesting alternative.
There are other interesting design out there. One I am interested in uses near IR pixels interspersed on the Bayer filter. CMOS sensor are more sensitive to near IR than most of the colour spectrum and are less affected by atmospheric haze, therefore they record much better detail and contrast outdoors than conventional sensors.
 
wow that was a good article read on dx0mark benchmarks and some good comments. Ther where some posts on the luminous-landscape forum showing that the 3 colour channels did have different noise levels I guess due to the way light filters through it, also they where recommending to over expose as the Merrills where better at recovering highlights than pulling up shadows due to noise.

So to summaries what your saying is if you want a Foveon camera do it sooner than later before there's no point :p
 
Back
Top Bottom