[Leumas]

I'm a gear nut....I admit it. Everytime I come across Foveon stuff I want one and am excited for the progress of the tech. I think it's the future for landscape photographers especially.
I just came across this article and they used a K5 for comparison. Just check out the image detail.

What makes Sigma Foveon sensor superior to others: Sigma DP3 Merrill camera - Photigy School Of Photography

[BigMackCam]

The Foveon sensors produce amazing detail... However, the sliding bar photo comparison in that article is very misleading. Looking at the coffee beans, for instance, there is something amiss with the focusing on the K-5 for the beans and the platform they're stood on to be so blurred. I have a K-5, and it is capable of very good detail indeed - much better than those beans would have you believe. Not as impressive as a Foveon sensor, but way better than this comparison suggests. With a traditional sensor camera like the K-5, much depends on the demosaicing and raw processing (local contrast, mild sharpening etc.) to bring out the recorded detail...

[mlt]

And even Sigma couldn’t come up with a quick, easy to use way to process Foveon raw files. It was an interesting idea 8-10 years ago. Not so much now as resolutions have increased and pixel shift got added in to equation by Pentax and others.

[Wasp]

One of those technologies that never really delivered on the initial hype?

I have been amazed by the high prices of used Sigma Foveon camera bodies in the past. Perhaps the sellers know something that the rest of us do not.

[mlt]

Probably not. Just hype, limited availability (not terrific sales when new), and weirdness.

[BigMackCam]

One of those technologies that never really delivered on the initial hype?

I have been amazed by the high prices of used Sigma Foveon camera bodies in the past. Perhaps the sellers know something that the rest of us do not.

Like the OP, I've always been tempted when I see Sigma's Foveon cameras for sale. There's no doubt that in the right application, and with the right post-processing, they can - sometimes - offer amazing image quality. But the high ISO performance is terrible, and the processing often results in some very odd artefacts. As studio cameras, or in good light where base ISO is practical, they can definitely produce beautifully detailed images. And I'd still like to own one, just for the fun of playing around with it. But I wouldn't want to rely on one as a general use camera to cover a wide range of shooting conditions. Plus, higher-resolution regular sensors combined with - as I previously implied - skilful post-processing - can yield remarkably good results that are far more competitive than the linked article would suggest...

[RobA_Oz]

And even Sigma couldn’t come up with a quick, easy to use way to process Foveon raw files. It was an interesting idea 8-10 years ago. Not so much now as resolutions have increased and pixel shift got added in to equation by Pentax and others.

So, it’s really Moveon?

[Leumas]

And even Sigma couldn’t come up with a quick, easy to use way to process Foveon raw files. It was an interesting idea 8-10 years ago. Not so much now as resolutions have increased and pixel shift got added in to equation by Pentax and others.

Pixel shift is still limited. I'm excited to see what they do with their new upcoming Full-Frame L-mount Foveon camera due in 2020

[thazooo]

Yeah Foveon is weird, limited, and takes time to use. But when you stick it on a tripod, use lowest ISO, and 2 second timer, the only thing that's close is the 645. Detail, clarity, and color are way beyond APSC and FF.
I use mine for serious shooting only because of it's limitations.

[photoptimist]

Foveon's technology was quite seductive but it was doomed by the physics of silicon.

Foveon relied on the physical fact that shorter wavelengths of light tend to be absorbed near the surface of a silicon chip and longer wavelengths of light tend to penetrate deeper. By measuring photocharge at three depths inside the chip, the sensor measured light levels in short (e.g., blue), medium (e.g., green), and long wavelengths all in the same pixel. Thus, unlike Bayer mosaics, the sensor had data on RGB at each pixel.

HOWEVER..... the operative word on the absorption depth as a function of wavelength is "tend to." In reality the "blue" pixels captured lots of green and even some red light, the "green" pixels also captured lots of red and some blue light, and the "red" pixels captured some green and a little blue light Not only was the color separation poor relative to the engineered dyes of a Bayer color filter array, but the math to estimate RGB amplified any noise

Foveon's dynamic range and high ISO performance was competitive with early CCD and CMOS sensors but not with the more modern ones.

[PJ1]

the sliding bar photo comparison in that article is very misleading

Interesting discussion but I too wondered about the K-5 comparison shots. I would have expected much better than that from the K-5 and A50/1.7 . But I have occasionally been intrigued (not enough to part with hard cash) by the Sigma cameras and their Foveon sensors. Maybe they would be good for landscape. A poor man's MF?

[Leumas]

The Foveon sensors produce amazing detail... However, the sliding bar photo comparison in that article is very misleading. Looking at the coffee beans, for instance, there is something amiss with the focusing on the K-5 for the beans and the platform they're stood on to be so blurred. I have a K-5, and it is capable of very good detail indeed - much better than those beans would have you believe. Not as impressive as a Foveon sensor, but way better than this comparison suggests. With a traditional sensor camera like the K-5, much depends on the demosaicing and raw processing (local contrast, mild sharpening etc.) to bring out the recorded detail...

I don't think it's misleading. They are comparing a 16mp sensor with a Foveon that resolves as much detail as a 46mp bayer.
I've seen other similar comparisons, and yes, the Foveon is pretty ridiculous.

Interesting discussion but I too wondered about the K-5 comparison shots. I would have expected much better than that from the K-5 and A50/1.7 . But I have occasionally been intrigued (not enough to part with hard cash) by the Sigma cameras and their Foveon sensors. Maybe they would be good for landscape. A poor man's MF?

Excellent for studio, and landscape from what I've read.

[stevebrot]

I think it's the future for landscape photographers especially.

Just a point of clarification and not to disparage any of the positive comments above, but the SD1 Merrill is a 15Mpx camera as far as output goes. The advantage it brings to the table over a K-5IIs is that there is no loss due to Bayer interpolation (i.e. no disappearing berries and such). The results are pixel equivalent to a crop-mode pixel-shift image from a K-1.

There were several comparison threads on this site back in 2013 for Foveon sensor cameras that showed both the strengths (full RGB color detail*) and weaknesses (reports and examples of puzzling and unpredictable color, textural artifacts, and loss of detail with some subjects) of the cameras in comparison with Pentax models of the time. I will see if I can find one comparing a Pentax w/o anti-alias filter (after all...apples to apples) to the SD1.

I don't remember, but does anyone know if the Foveon sensor line is slated to continue into this next decade? Yes, a Foveon-like sensor is rumored for the camera model pushed back to sometime in 2020, but mean something more solid, newswise.


Steve

* Not the same as full-color fidelity or accuracy

[BigMackCam]

I don't think it's misleading. They are comparing a 16mp sensor with a Foveon that resolves as much detail as a 46mp bayer.
I've seen other similar comparisons, and yes, the Foveon is pretty ridiculous.

When I say misleading, I mean the amount of detail in certain areas of the K-5 shot is unrealistically poor. There is something going on with the K-5 shot's focus IMHO, perhaps focused too far back so that the foreground items - the beans and the surface they're resting on - are not in acceptable focus. There is simply no way the K-5 loses that much detail otherwise. I own a bunch of bodies from the old 6MP *ist DL to the 24MP K-3II, including the original K-5 - and whilst it's true to say that the later bodies without AA filters render better detail, the K-5's Bayer sensor with AA filter is still capable of much better detail than shown here. I'm not saying it's as good as the Foveon - it's almost certainly not; but the K-5 is poorly demonstrated in this comparison.

[Leumas]

Just a point of clarification and not to disparage any of the positive comments above, but the SD1 Merrill is a 15Mpx camera as far as output goes. The advantage it brings to the table over a K-5IIs is that there is no loss due to Bayer interpolation (i.e. no disappearing berries and such). The results are pixel equivalent to a crop-mode pixel-shift image from a K-1.

There were several comparison threads on this site back in 2013 for Foveon sensor cameras that showed both the strengths (full RGB color detail*) and weaknesses (reports and examples of puzzling and unpredictable color, textural artifacts, and loss of detail with some subjects) of the cameras in comparison with Pentax models of the time. I will see if I can find one comparing a Pentax w/o anti-alias filter (after all...apples to apples) to the SD1.

I don't remember, but does anyone know if the Foveon sensor is slated to continue into this next decade?


Steve

* Not the same as full-color fidelity or accuracy

They are releasing a Full-Frame mirrorless Foveon for L-mount in 2020