Comparing Foven and other sensors in terms of resolution

T

Tuolumne

Veteran
Local time
6:58 AM
Joined
Feb 12, 2007
Messages
3,005
I know that Foveon sensors, such as the ones that Sigma uses, capture RGB at a single sensor site (pixel). I know that non-Foveon sensors use a Bayer filter to capture RGB across multiple adjacent pixels while integrating the result into a single RGB "pixel".

How does this compare in terms of sensor resolution? Does this mean that a Sigma 14 megapixel sensor has true 14 meagpixel resolution while a Bayer filter sensor only has 14/3 = 4.67 megapixels? Or does the Bayer sensor actually have 14x3 = 42 megapixels? Or maybe something else?

/T
 
I think it is the other way round, Sigma claims 14 MP, while there are only 4 point something million of pixels in their sensor.
 
I think it is an unsolvable problem. A photosite on one simply is not analogous to a photosite on the other. It's like asking when has more pixels, a tree or a cloud?

Ultimately, what matters is the quality of the end result, printed.

I have an SD-14 and a K200D. I like them both. They each have strong and weak points. If someone wants to insist that my SD-14 is "really" only 4.7 megapixels, I'm perfectly OK with that. I don't really care at all. It takes photos that I like. The same for those who insist that my Bayer-filtered K200D is not "really" 10 megapixels. Fine, fine. No problem.

I simply won't engage in 'which one is better' or 'which one has more megapixels' wars. They are cameras. They take photos. Each has great capability when used properly in accordance with their strengths and weaknesses. I can't think of anything else that matters.
 
This has been debated since the Foveon came into existance. There are 4.something million pixels on the sensor. I think the rest is not answerable.
 
I just want to know how it's calculated and why it's calculated that way. I'm interested in the technology (that's why it's posted in this forum), not opinions about which is better or worse.

/T
 
This might be useful:

http://www.dpreview.com/reviews/sigmadp1/page19.asp

The Sigma DP1 delivers an admirable resolution considering the size of its image output of not even five megapixels. While it outresolves the downscaled Nikon D60 image and almost matches the resolution of the ten megapixel Ricoh, it cannot really compete with a modern 10 megapixel DSLR such as the Nikon D60.
Click to expand...

However it does good with the color resolution test:
http://www.dpreview.com/reviews/sigmadp1/page20.asp
 
Using both the M8 and DP2, the image displayed in Lightroom at 1:1 is much bigger for the M8, therefore I think that the Foveon is really only around 5 megapixels in terms of size (note that I have printed A3 without problem). I understood it has three layers of 4.7 megapixels, each one extracting color information on the same ray of light.
So, to sum it up, I think the Foveon produces less pixels in terms of final image size but that each pixel has been computed with more detailed information (3 layers) so it is supposed to have a better rendition of colors and more resolution.
I am not sure that was very clear ... I tried ... ;)
 
Last edited:
The foveon sensor has 4.7 million pixels – each with full rgb data. A Bayer sensor has x million photosites, which are divided in the ratio 2 green to 1 red to 1 blue. Conceptually a block of 4 photosites (RGGB) can be equated to a single full colour pixel, suggesting that 4.7 foveon is about equivalent to 18.8 millon Bayer. In practice, Bayer sensors can achieve a bit better resolution in colour – say up to 73% fo the linear resolution for a good system, based on information from a Red signal processing engineer – and so 4.7 foveon equates to around 8.8 million Bayer. This seems to match most people’s experience.

Of course, different images (and different areas of images) make different demands on resolution so some bits may do better or worse, depending on colour and luminance resolution components required.

Mike
 
Thank you all for the factual answers. I would love to see a Foveon sensor in an RD-2 or in a CV digital Bessa. I think that would the best of all possible worlds, and even acceptable with a 1.5x crop factor.

/T
 
Tuolumne said:
Thank you all for the factual answers. I would love to see a Foveon sensor in an RD-2 or in a CV digital Bessa. I think that would the best of all possible worlds, and even acceptable with a 1.5x crop factor.

/T
Click to expand...

Foveon is owned by Sigma. I sincerely doubt it will show up in another brand camera. Other manufacturers had their chance when Foveon was independent and apparently decided not to use that sensor.
 
Tuolumne said:
Can anyone just answer the question without flaming?
Click to expand...
:rolleyes:

Foveon:
RRR
GGG
BBB​

Bayer:
RGBRGBRGB​

Manufacturers of each would claim that this diagram has 9 pixels for their sensor. Any other explanation requires an opinion of the style of counting pixels, which, I take it, is the flaming to which you refer.
 
bmattock said:
Foveon is owned by Sigma. I sincerely doubt it will show up in another brand camera. Other manufacturers had their chance when Foveon was independent and apparently decided not to use that sensor.
Click to expand...

Well, then Sigma could make a body with an M bayonet mount. How's that!
 
Rob-F said:
Well, then Sigma could make a body with an M bayonet mount. How's that!
Click to expand...

Not likely, is how it is. Sigma is interesting and they are obviously willing to march to the beat of their own drummer. I suspect that the fact that most of their business is making lenses allows them to devote some time and money to playing at camera manufacturer. But there are some drawbacks too. They have managed to bring the DP1 and DP2 to market, but the SD15 is seriously delayed, as was the SD14 before it. They just can't seem to execute very quickly.

However, with Sigma, you never know. So perhaps. But my money would still go towards 'not likely'.
 
I would really like to see Sigma release their SD15 so I can get my SD14 files but from a much faster camera. Waiting for the SD14 to write raw files is like watching paint dry.

A GF1 with foveon sensor is a whole other story, though equally desirable :)
 
mabelsound said:
What I would really like to see at this point is Sigma joining m4/3 and making an m4/3 body. And m4/3 lenses for that matter.
Click to expand...

Sigma is part of the 4/3 consortium, but they make lenses for 4/3. I do not know if there is a different consortium for m4/3.

I believe the sensor size is different between the Sigma Foveon (1.6x crop) and the 4/3 (2x crop). That might be a small problem for coverage for lenses designed for the 4/3 or m4/3 standard, but I don't know.

I'd like to see it too, but I'm not holding my breath.
 
You must log in or register to reply here.
Back
Top Bottom