Back to full Foveon!

Started Apr 9, 2019 | Discussions thread
xpatUSA Forum Pro • Posts: 19,261
Re: Back to full Foveon!

TN Args wrote:

xpatUSA wrote:

TN Args wrote:

xpatUSA wrote:

threw the lens wrote:

Can't help wondering from experience with the X3 whether the colour penetration/balance is the same for strong versus weak light because it is like highlights and shadows have a separate WB sometimes. Maybe it's a reflection that very little light reaches the bottom red layer and the electronics attempt a lot of "guessing" or "reconstruction".

Just because the spectrum is more continuous doesn't mean the colour as a whole is automatically better if the information from the bottom red layer is very weak and relies on algorithmic guessing.

It is a popular myth that the Foveon filters out "the light" as it passes downward through the layers and actual wavelengths are rarely mentioned. I've never seen a link where Foveon supports that view. Are there any credible references?

On the other hand, there are plenty of published Foveon graphs that show that the red channel's QE is the most efficient of the three layers ...

I did see mention of the SD14 sensor having QE of 25%, 55%, and 25% at 390, 520, and 800 nm, respectively. Is that useful?

Not useful without a reference, Arg.

Gilblom, D.L., Yoo, S.K., and Ventura, P. (2003) Operation and performance of a color image sensor with layered photodiodes. Proceedings of SPIE 5074:318–331. Referencing the Foveon Fx17-78-F13 sensor. Any more detail than that though and you have to get a copy. It was buried in some old article about in-situ UV and IR imaging of glaciers when I found it, long gone now.

I have the paper in spite of your assumption that I would "have to get a copy".

1) it is not a Foveon paper.

2) I can find no mention of the SD14 sensor having QE of 25%, 55%, and 25% at 390, 520, and 800 nm, respectively. Perhaps I have a different edition ...

3) This graph is not of the sensor alone:

4) They are using the marketing trick of adding up the QEs. If you are OK with that, our discussion is at an end.

It sounds like the Foveon curve where they add the layer QEs together and call it "total quantum efficiency", causing most engineers on the planet to LOL.

Here's a real deal:

Published by Richard Lyon and Paul Hubel in a Foveon paper, 0.35 is as high as we get and, guess what, that's the "red" layer. Please note that the curves are for just the sensor and do not include the effect of a UV/IR blocking filter.

Yes I've seen that, and you are making an important point, but I think that number is different to an SNR number for the data extracted from each layer? Are you saying the SNR is higher if the QE is higher?

Mention of SNR is a straw man argument. I have no comment.

You can hardly say "It is a popular myth that the Foveon filters out "the light" as it passes downward", when clearly, the light is filtered out as it passes downward.

Well, if the so-far un-specified "light" is at say 1150nm, it roars through the silicon virtually unimpeded, i.e. no "filtering out" whatsoever. My point here is that the popular impression of red light struggling down through the layers, losing oodles photons on the way, detracts from understanding how the Foveon works.


-- hide signature --


 xpatUSA's gear list:xpatUSA's gear list
Panasonic Lumix DMC-LX1 Sigma SD9 Panasonic Lumix DC-G9 Panasonic Leica DG Macro-Elmarit 45mm F2.8 ASPH OIS Sigma 17-50mm F2.8 EX DC OS HSM +13 more
Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow