Prev: SI Facescape
Next: FF camera with mirrorless design
From: Rich on 17 Apr 2010 03:05 On Apr 16, 11:05 pm, nospam <nos...(a)nospam.invalid> wrote: > In article <Q8WdnQRa9pi2vFTWnZ2dnUVZ_ugAA...(a)giganews.com>, Rich > > <n...(a)nowhere.com> wrote: > > Foveon has its issues, > > yes it does. > > > but it is prima proof that Bayer and AA filters do > > rob images of resolution and sharpness. > > nope. it's proof that false advertising and high levels of sharpening > can fool people. It's no worse than Fuji claiming the S5 was 12 megapixels when it clearly wasn't, but at least when you sharpen an image from a Foveon, it doesn't develop those horrible halos that Bayer/AA filtered sensors do. Pretty much every reviewer has stated that the actual resolution of a the 4 meg Foveon is on-par with an 8 megapixel Bayer sensor. http://www.dpreview.com/reviews/sigmadp2/page21.asp
From: nospam on 17 Apr 2010 03:52 In article <5ab4f313-32df-4173-8c58-018f330c14b9(a)c21g2000yqk.googlegroups.com>, Rich <rander3127(a)gmail.com> wrote: > It's no worse than Fuji claiming the S5 was 12 megapixels when it > clearly wasn't, yes it is. > but at least when you sharpen an image from a Foveon, > it doesn't develop those horrible halos that Bayer/AA filtered sensors > do. nonsense. if you oversharpen any image, you'll have problems. > Pretty much every reviewer has stated that the actual resolution > of a the 4 meg Foveon is on-par with an 8 megapixel Bayer sensor. which means it's competitive with cameras from 4 years ago. that's not very compelling. meanwhile, canon has an 18 megapixel slr that costs *less* than the sd15 will and it won't be stuck with sigma lenses either.
From: Chris Malcolm on 17 Apr 2010 05:10 Ray Fischer <rfischer(a)sonic.net> wrote: > Bubba <digitalrube(a)yahoo.com> wrote: >>I don't understand how a sensor can claim 14 pixels, and the photo >>itself be 4. > That's easy - Sigma lies. They pretend that you can take a single > pixel and split it up into three colors and have that be three pixels. They don't lie. It's a different kind of sensor with sufficiently different kinds of pixels and properties that the sensor megapixel numbers aren't comparable with hose of Bayer sensors. It's like trying to compare the engine capacities of ordinary car engines with cylindrical pistons and rotary Wankel engines. The numbers have to be normalised before you can make comparisons. Of course it may well be the case that Foveon exaggerate a bit in normalising their numbers. But then the megapixel numbers of Bayer sensors are exaggerated themselves. An ordinary 14MP Bayer sensor desn't give you the image resolution you'd expect from 14MP. Some gets lost in the necessary AA filter. And that's not a constant, because it's a trade off between detail resolution and aliasing artefacts. Some makers give you more detail, in effect more megapixels of detail, at the cost of more aliasing artefacts in the situations where aliasing obtrudes. So there's no simple way of normalising Foven megapixel numbers to make them comparable to Bayer megapixel numbers. You simply have to look at the results and make a subjective appraisal of comparability. Just as you have to do when comparing Bayer megapixels from cameras with different attitudes towards aliasing, such as Leica, Nikon, and Hasselblad. -- Chris Malcolm
From: nospam on 17 Apr 2010 05:36 In article <82tc7fF962U1(a)mid.individual.net>, Chris Malcolm <cam(a)holyrood.ed.ac.uk> wrote: > >>I don't understand how a sensor can claim 14 pixels, and the photo > >>itself be 4. > > > That's easy - Sigma lies. They pretend that you can take a single > > pixel and split it up into three colors and have that be three pixels. > > They don't lie. yes they do. calling it anything other than 4.6 or 4.7 megapixels is a lie (it's actually 4.65, so either one is fine). > It's a different kind of sensor with sufficiently > different kinds of pixels and properties that the sensor megapixel > numbers aren't comparable with hose of Bayer sensors. the difference is that bayer has one layer and foveon has three layers per pixel. you don't get to count each layer as a separate pixel. if you had a single pixel sensor with 1 million layers, would that be a megapixel sensor? no. > It's like trying > to compare the engine capacities of ordinary car engines with > cylindrical pistons and rotary Wankel engines. you can measure horsepower, torque and engine displacement for both. > The numbers have to > be normalised before you can make comparisons. pixel is normalized. > Of course it may well be the case that Foveon exaggerate a bit in > normalising their numbers. exaggerate a bit? they multiply it by 3. that's a lot. > But then the megapixel numbers of Bayer > sensors are exaggerated themselves. no they're not. > An ordinary 14MP Bayer sensor > desn't give you the image resolution you'd expect from 14MP. it does for me. > Some gets > lost in the necessary AA filter. And that's not a constant, because > it's a trade off between detail resolution and aliasing > artefacts. Some makers give you more detail, in effect more megapixels > of detail, at the cost of more aliasing artefacts in the situations > where aliasing obtrudes. it doesn't matter how much detail there is in a photo, the number of pixels on the sensor does not change. > So there's no simple way of normalising Foven megapixel numbers to > make them comparable to Bayer megapixel numbers. yes there is. > You simply have to > look at the results and make a subjective appraisal of comparability. > Just as you have to do when comparing Bayer megapixels from cameras > with different attitudes towards aliasing, such as Leica, Nikon, and > Hasselblad. sure, but the subjective quality doesn't change the number of pixels on the sensor.
From: Bubba on 17 Apr 2010 07:36
On Apr 17, 5:36 am, nospam <nos...(a)nospam.invalid> wrote: > > sure, but the subjective quality doesn't change the number of pixels on > the sensor. All right; I've read all the responses to this question I posted last night. If what I call a CMOS sensor in the Canon SX S1 is comparable to the Foveon, why all the glowing reviews from buyers--reviews that reckon with the cameras' (or the entire camera line's) shortcomings, among which is this fixed lens. At the risk of getting a rain of more invective on a day that's rainy enough in my parts, does this mean that you can't...focus?! If you can focus, there has to be *some* zoom (or is it just digital)? These Sigma owners, and the gentleman on another thread who recommended a camera with a Foveon sensor to me, seem to use it for what I want it for: nature photography in low light, without the artifacts that I suppose you all here are calling "aliasing." (I've called it red flare, because that's what I was told it was by the British folk on my Serif editing software sight--a nicer bunch of people, I have not met on any forum--including those on which I've tangled with nospam.) |