What the imager has

Started Feb 11, 2014 | Discussions
Lord metroid
Lord metroid Regular Member • Posts: 475
Re: What the imager has

I previously watched your presentation of the Photo EDU videos from Google Tech Talk on YouTube. Very interesting and educational.

Thank you Dick Lyon, I am grateful that you who has quite a lot of knowledge of photosensor posted a very interesting and explanatory post that explained what others have only speculated.

 Lord metroid's gear list:Lord metroid's gear list
Sigma DP2s Sigma dp2 Quattro
DMillier Forum Pro • Posts: 19,865
Re: The Quattro Knows

So it definitely isn't a direct measurement of all 3 values at every pixel position any more, it's having to calculate values....

-- hide signature --
DMillier Forum Pro • Posts: 19,865
Re: What the imager has

That's the supposition I would have too, where values are calculated there lies the possibillity of uneven performance

-- hide signature --
Roland Karlsson Forum Pro • Posts: 26,226
Re: Color Resolution Zen

Kendall Helmstetter Gelner wrote:

Yes, the irritating part about bayer is varying luminance resolution, because it suddenly drops the plots as far as detail before natural factors like DOF would cause decay.

The other annoying part is the color artifacts, which bring utterly unexpected colors into view where none should be simply based on a pattern.

The Quattro should be immune to simply "made up" colors, the worst I can possibly see happening is some kind of color bleeding. But because of the top layers also getting some values the lower layers get, you can eliminate a lot of color bleeding kinds of effects simply by having a clear picture of where the boundaries are.

In the end the result should be way closer to 19MP of color resolution than 5MP. I don't even think it's possible to have any image where you would only get 5MP of color resolution.

Yes, it is possible to construct an artificial image where the color resolution is 5 MP.

Here you have it:

Lets assume that you have an image that will expose the top layer equally for all 4 pixels. But, all four pixels have different colors. You cannot then detect those 4 colors. They will look the same.

Now, repeat this property over an area, and you will have 5 MP color resolution in that area, even though there are color changes for 20 MP.

NOTE: that the luminance also is down to 5 MP in this example.

So, it is a proof that (theoretically) you can have a resolution of 5 MP for both luminance and chrominance.

Do you agree?

 Roland Karlsson's gear list:Roland Karlsson's gear list
Sigma DP3 Merrill Sigma dp2 Quattro Sony RX100 III Pentax K-3 Pentax K-1 +14 more
Roland Karlsson Forum Pro • Posts: 26,226
Re: The Quattro Knows

DMillier wrote:

So it definitely isn't a direct measurement of all 3 values at every pixel position any more, it's having to calculate values....

Yes

 Roland Karlsson's gear list:Roland Karlsson's gear list
Sigma DP3 Merrill Sigma dp2 Quattro Sony RX100 III Pentax K-3 Pentax K-1 +14 more
Roland Karlsson Forum Pro • Posts: 26,226
Re: Thanx for looking in

DickLyon wrote:

Thanx for looking in and bringing some information, instead of just idle speculations

 Roland Karlsson's gear list:Roland Karlsson's gear list
Sigma DP3 Merrill Sigma dp2 Quattro Sony RX100 III Pentax K-3 Pentax K-1 +14 more
Guenter Borgemeister Senior Member • Posts: 1,276
Re: What the imager has

DickLyon wrote:

Another good way to think of the lower levels is that they get the same four samples as the top level, and then "aggregate" or "pool" four samples into one. This is easy to simulate by processing a full-res RGB image in Photoshop or whatever.

Hi Dick, nice to hear from you after all these years. I hope you are doing fine.

Thanks for your explanations which settle my doubts about the new sensor design. I just did the simulation you suggest in Photoshop and the result looks much better than the one of my first (doubtful) one in another thread. You made my day

Original left, simulation as suggested by Dick right. See original size.

 Guenter Borgemeister's gear list:Guenter Borgemeister's gear list
Sigma 18-35mm F1.8 DC HSM Art
MOD Kendall Helmstetter Gelner Forum Pro • Posts: 19,854
Re: What the imager has

DickLyon wrote:

Kendall, long time...

It has been a long time!

Thanks for the excellent response to help clear things up further for everyone.  As you say, the chroma being spread out a bit in the end hardly matters at all.  This should be one interesting camera to see what reviewers make of it!

 Kendall Helmstetter Gelner's gear list:Kendall Helmstetter Gelner's gear list
Sigma 70-200mm F2.8 EX DG OS HSM Sigma 8-16mm F4.5-5.6 DC HSM Sigma 24-70mm F2.8 EX DG HSM Sigma 50-500mm F4.5-6.3 DG OS HSM Sigma 85mm F1.4 EX DG HSM +4 more
Raist3d Forum Pro • Posts: 36,219
Thanks Dick, pretty much in agreement, most of what I was saying -

DickLyon wrote:

[]

In the 1:1:4 arrangement, each sample layer has this property, but at different rates -- very unlike the Bayer's red and blue planes. The large area of the lower-level pixels is the ideal anti-aliasing filter for those layers; the top layer is not compromised by the extra spatial blurring in the lower layers, so it provides the extra high frequencies needed to make a full-res image.

Another good way to think of the lower levels is that they get the same four samples as the top level, and then "aggregate" or "pool" four samples into one. This is easy to simulate by processing a full-res RGB image in Photoshop or whatever.

Yup, which goes back to what I said that this is somewhat similar to what some texture compressions schemes are done in GPU / graphics. So taken an ideal Foven Merrill sensor with none of the noise issues and diffraction - the Merrill in theory would still be better, correct? Since it samples the other layers per spatial site without the averaging.

But it's the realities of the noise/other issues what makes this (new) design pull ahead, correct?

The pooling of 4 into 1 is done most efficiently in the domain of collected photo-electrons, before converting to a voltage in the readout transistor. The result is the same read noise, but four times as much signal, so about a 4X better signal-to-noise ratio. Plus with fewer plugs, transistors, wires, etc. to service the lower levels, the pixel fill factor is closer to 100% with easier microlenses, and the readout rate doesn't have to be as high. Wins all around -- except for the chroma resolution.

This is what I said elsewhere also except for the first sentence. So we actually agree. The Chroma resolution suffers, but I also expect this design to offset that tradeoff given the current Merrill issues, some of which you mention.

The main claim of Bryce Bayer, and the fact that most TV formats and image and video compression algorithms rely on, is that the visual system doesn't care nearly as much about chroma resolution as about luma resolution.

Yes, and it's the main reason why texture compression on GPU's works reasonably well.

Unfortunately, trying to exploit that factor with a one-layer mosaic sensor has these awkward aliasing problems. Doing it with the Foveon 1:1:4 arrangement works better, requiring no AA filter, no filtering compromises. So, yes, the chroma resolution is less than the luma resolution, but you'd be hard pressed to see that in images.

I would imagine this is going to depend on the image and the level of pixel peep, and each individual. Fortunately having more resolution of a sensor helps cover this also, and for publishing mediums such as web, iPad, etc. it's pretty much a non issue. But it's worth bringing up only because at least several Foveon "looK" aficionados always do this comparison at the pixel peeping level pointing out why Foveon is better even if other solutions due to the end result have their advantages also.

If you throw out the extra luma resolution and just make 5 MP images from this new camera, you'll still have super-sharp super-clean versions of what the old DP2 or SD15 could do.

I agree Dick, and it's exactly what I mentioned here:

http://www.dpreview.com/forums/post/53093930

I want to see how much of an improvement there could be in ISO using that Low resolution raw mode.

Now imagine adding 2X resolution in each dimension, but with extra luma detail only, like in a typical JPEG encoding that encodes chroma at half the sample rate of luma. Whose eyes are going to be good enough to even tell that the chroma is less sharp than the luma? It's not impossible, but hard.

I agree it will be hard. It will, depend on the person (on average that won't matter), pixel peeping, subject being photographed. It could have some slight color crawl (depending how they manage) or blur. Similar to GPU graphics texture compression schemes.

In a game though the textures are much lower rez than the Quattro Foveon, usually the graphics are moving so in general it's hard to notice. In the case of the Quattro though stills, the resolution is vastly higher than the typical game texture so that counters that (sans 100%-200% pixel peeps and some photographic subjects).

Speaking of stories from the old days, Foveon's first version of Sigma Photo Pro had a minor bug in the JPEG output, as you probably recall: our calls to the jpeg-6b library defaulted to encoding with half-res chroma. It took a while, but a user did eventually find an image where he could tell something was not perfect, by comparing to TIFF output, and another user told us how to fix it, so we did. It we could have gotten that original level of JPEG quality from the SD9 with 5 million instead of 10 million pixel sensors and data values, and could have gotten cleaner color as a result, would that have been a problem? I don't think so; except for marketing, and they had enough problems already. Same way with Sigma's new one, I expect; if 30 M values gives an image that will be virtually indistinguishable from what could be done with 60 M, but with cleaner color, will someone complain?

Probably so.

I think the big huge win is that the tradeoffs of the new design will more than offset or should more than offset the Merrill's in a significant way. It's what I am expecting (say 2 stops better ISO, superb B&W).

So, it's complicated. Yes, reduced chroma resolution is a compromise; but a very good one, well matched to human perception -- not at all like the aliasing-versus-resolution compromise that the mosaic-with-AA-filter approach has to face.

Agreed, which is what I have been saying too for the most part. The Quattro would still have more data than the Bayer to do its work. What I want to see is the bottom line improvement on the ISO. 14-bit (assuming it is really accurate to sample at that level) should benefit DR and ISO.

Dick

disclaimer: I've been away from this technology too long to have any inside knowledge. And give my apologies to Laurence for my too many words.

-- hide signature --

Raist3d/Ricardo (Photographer, software dev.)- I photograph black cats in coal mines at night...
“The further a society drifts from truth the more it will hate those who speak it.” - George Orwell

Raist3d Forum Pro • Posts: 36,219
Re: What the imager has

Kendall Helmstetter Gelner wrote:

DickLyon wrote:

Kendall, long time...

It has been a long time!

Thanks for the excellent response to help clear things up further for everyone. As you say, the chroma being spread out a bit in the end hardly matters at all. This should be one interesting camera to see what reviewers make of it!

For the record, that's pretty much exactly what I have been saying on the chroma resolution.  What he said describes very well why the GPU texture compression can get away with similar (keep high on the luminance detail, but lower on the chroma).

There's other strategies that can be done like gradient direction detection, etc. depending how far it's taken. Several Bayer cameras are doing some stunts on that end also.  I see the benefit for the Quattro in still having more color data.

-- hide signature --

Raist3d/Ricardo (Photographer, software dev.)- I photograph black cats in coal mines at night...
“The further a society drifts from truth the more it will hate those who speak it.” - George Orwell

Raist3d Forum Pro • Posts: 36,219
Re: Color Resolution Zen

Kendall Helmstetter Gelner wrote:

Roland Karlsson wrote:

mike earussi wrote:
Which will vary throughout the image, dependent entirely on the value of the four top layer pixels. So some parts of the image could be as low as 5mp whereas other part could be as high as 19mp, color resolution of course.

Yes, and this is an unwanted characteristic. One of the more irritating properties of bayer sensors is its varying resolution.

Now, the Quattro will not vary in resolution, but in color resolution. Not the same thing.

Yes, the irritating part about bayer is varying luminance resolution, because it suddenly drops the plots as far as detail before natural factors like DOF would cause decay.

The other annoying part is the color artifacts, which bring utterly unexpected colors into view where none should be simply based on a pattern.

The Quattro should be immune to simply "made up" colors, the worst I can possibly see happening is some kind of color bleeding.

But color bleeding is a case were you end up with a made up color.  There's also color accuracy (for areas of color that have a higher frequency that the sensor can sample).

But because of the top layers also getting some values the lower layers get, you can eliminate a lot of color bleeding kinds of effects simply by having a clear picture of where the boundaries are.

You won't be able to make that determination like that due to the lower sampling, though there are certainly strategies that Sigma could purse based on human perception and pattern recognition (several Bayer cameras are already doing stuff like that).

In the end the result should be way closer to 19MP of color resolution than 5MP.

For color I don't think so, but for perception of detail yes. There are strategies that can be pursued  but they can't make up data that doesn't exist.  I agree it should still look pretty good, but depending on subject/light/resolution of subject in those colors, it may not look as sharp.

I don't even think it's possible to have any image where you would only get 5MP of color resolution.

-- hide signature --

Raist3d/Ricardo (Photographer, software dev.)- I photograph black cats in coal mines at night...
“The further a society drifts from truth the more it will hate those who speak it.” - George Orwell

Roland Karlsson Forum Pro • Posts: 26,226
Re: What the imager has

Raist3d wrote:
There's other strategies that can be done like gradient direction detection, etc. depending how far it's taken. Several Bayer cameras are doing some stunts on that end also. I see the benefit for the Quattro in still having more color data.

Yes, that is heavily done in Bayer reconstruction.

Its a two edged sword though.

It is the main reason why Bayer images can be so sharp, but also the main reason why Bayer images have so uneven sharpness over the image. The latter is a big disadvantage and thr former a big advantage.

You alos have to understand that the gradient detection algorithms may create edges that are not really there, i.e. creating sharpness where none is found. This might be one of the reasons for the lack of 3D look - the DOF does not work reliable. Sharp things are soft and soft things are sharp.

 Roland Karlsson's gear list:Roland Karlsson's gear list
Sigma DP3 Merrill Sigma dp2 Quattro Sony RX100 III Pentax K-3 Pentax K-1 +14 more
Raist3d Forum Pro • Posts: 36,219
Re: What the imager has

Roland Karlsson wrote:

Raist3d wrote:
There's other strategies that can be done like gradient direction detection, etc. depending how far it's taken. Several Bayer cameras are doing some stunts on that end also. I see the benefit for the Quattro in still having more color data.

Yes, that is heavily done in Bayer reconstruction.

Its a two edged sword though.

It is the main reason why Bayer images can be so sharp, but also the main reason why Bayer images have so uneven sharpness over the image. The latter is a big disadvantage and thr former a big advantage.

You alos have to understand that the gradient detection algorithms may create edges that are not really there, i.e. creating sharpness where none is found. This might be one of the reasons for the lack of 3D look - the DOF does not work reliable. Sharp things are soft and soft things are sharp.

Well.. yes & yes, of course. I am not suggesting this can be done without tradeoffs And if I did, my apologies.

To me the primary reason for the lack of "3d look" is first and foremost the AA filter.  A lot of cameras with good lenses without the AA filter immediately "jump" with that- but as always, there's the tradeoff of color moire and you still have lower resolution in colors.

I find the Fuji X-trans really helps tone the color moire down, but again, it has its quirks. But I would say the benefits outweigh the cons (using a proper raw converter, for most subjects).

The math stunts that work the best of course is those that when they fail, they don't impact the image significantly but when they win and on average win more than lose, get you the gain.

I noticed something a bit weird in the latest Olympus jpeg engines- it's like they are doing outline tracing. If you look at the preview studio shots of the EM-5 generation of their m4/3rds cameras, on the bottle (older shot) that had the Woman with "radiant" set of lines behind you see the diagonals- some seem a bit too sharp and nice yet a bit ummm.. strong "outlined."  Must be one of those algorithms at work.

-- hide signature --

/Roland
X3F tools:
http://www.proxel.se/x3f.html
https://github.com/rolkar/x3f

--
Raist3d/Ricardo (Photographer, software dev.)- I photograph black cats in coal mines at night...
“The further a society drifts from truth the more it will hate those who speak it.” - George Orwell

brittonx
brittonx Senior Member • Posts: 2,438
Re: What the imager has

DickLyon wrote:

Kendall, long time...

You're right that there won't be much aliasing. A lot of people seem to have the idea that aliasing has something to do with different sampling positions or density, as in Bayer. But that's not the key issue. The problem with Bayer is that the red plane (for example) can never have more than 25% effective fill factor, because the sampling aperture is only half the size, in each direction, of the sample spacing. If you take the Fourier transform of that half-size aperture, you'll find it doesn't do much smoothing, so the response is still quite too high way past the Nyquist frequency. That's why it needs an anti-aliasing filter to do extra blurring. But if the AA filter is strong enough to remove all the aliasing in red, it also throws away the extra resolution that having twice as many green samples is supposed to give. It's a tough tradeoff.

In the Foveon sensor, the reason no AA filter is needed is not because of where the samples are, or what the different spatial sampling densities are. It's because each sample is through an aperture of nearly 100% fill factor, that is, as wide each way as the sample pitch. The Fourier transform of this aperture has a null at the spatial frequencies that would alias to low frequencies; this combined with a tiny bit more blur from the lens psf is plenty to keep aliasing to a usually invisible level, while keeping the image sharp and high-res.

In the 1:1:4 arrangement, each sample layer has this property, but at different rates -- very unlike the Bayer's red and blue planes. The large area of the lower-level pixels is the ideal anti-aliasing filter for those layers; the top layer is not compromised by the extra spatial blurring in the lower layers, so it provides the extra high frequencies needed to make a full-res image.

Another good way to think of the lower levels is that they get the same four samples as the top level, and then "aggregate" or "pool" four samples into one. This is easy to simulate by processing a full-res RGB image in Photoshop or whatever.

The pooling of 4 into 1 is done most efficiently in the domain of collected photo-electrons, before converting to a voltage in the readout transistor. The result is the same read noise, but four times as much signal, so about a 4X better signal-to-noise ratio. Plus with fewer plugs, transistors, wires, etc. to service the lower levels, the pixel fill factor is closer to 100% with easier microlenses, and the readout rate doesn't have to be as high. Wins all around -- except for the chroma resolution.

The main claim of Bryce Bayer, and the fact that most TV formats and image and video compression algorithms rely on, is that the visual system doesn't care nearly as much about chroma resolution as about luma resolution. Unfortunately, trying to exploit that factor with a one-layer mosaic sensor has these awkward aliasing problems. Doing it with the Foveon 1:1:4 arrangement works better, requiring no AA filter, no filtering compromises. So, yes, the chroma resolution is less than the luma resolution, but you'd be hard pressed to see that in images.

If you throw out the extra luma resolution and just make 5 MP images from this new camera, you'll still have super-sharp super-clean versions of what the old DP2 or SD15 could do. Now imagine adding 2X resolution in each dimension, but with extra luma detail only, like in a typical JPEG encoding that encodes chroma at half the sample rate of luma. Whose eyes are going to be good enough to even tell that the chroma is less sharp than the luma? It's not impossible, but hard.

Speaking of stories from the old days, Foveon's first version of Sigma Photo Pro had a minor bug in the JPEG output, as you probably recall: our calls to the jpeg-6b library defaulted to encoding with half-res chroma. It took a while, but a user did eventually find an image where he could tell something was not perfect, by comparing to TIFF output, and another user told us how to fix it, so we did. It we could have gotten that original level of JPEG quality from the SD9 with 5 million instead of 10 million pixel sensors and data values, and could have gotten cleaner color as a result, would that have been a problem? I don't think so; except for marketing, and they had enough problems already. Same way with Sigma's new one, I expect; if 30 M values gives an image that will be virtually indistinguishable from what could be done with 60 M, but with cleaner color, will someone complain?

Probably so.

So, it's complicated. Yes, reduced chroma resolution is a compromise; but a very good one, well matched to human perception -- not at all like the aliasing-versus-resolution compromise that the mosaic-with-AA-filter approach has to face.

Dick

disclaimer: I've been away from this technology too long to have any inside knowledge. And give my apologies to Laurence for my too many words.

Thank you for jumping and giving your perspective.  It is great to hear from someone who was involved with Foveon's development, even if it was the earlier generations

-- hide signature --
 brittonx's gear list:brittonx's gear list
Sigma DP1 Sigma SD15 Sigma SD1 Sony a6300 Sigma 18-200mm F3.5-6.3 DC OS HSM +6 more
mike earussi Veteran Member • Posts: 7,080
Re: The Quattro Knows

Kendall Helmstetter Gelner wrote:

Roland Karlsson wrote:

Kendall Helmstetter Gelner wrote:
You had a good run of reasonableness there but you seem to have gone quite wrong. It gives UP TO 20MP full color images (that is unique color per pixel) which is UP TO 40MP bayer equivalent.

Because many, many of the colors can be determined uniquely they are still full color. In the end the question is the percentage of pixels that can generally be determined uniquely.

Yes, you can get "up to" 20 MP of perfect color just as you can, with a Bayer CFA (without AA filter) can get "up to" 20 MP perfect color - if you are lucky.

The Bayer sensor may get lucky guesses, but it can never KNOW the correct color values. Because the bayer sensor has to guess for many many values, it is mostly wrong. Not wrong by much, but wrong.

Maybe that's a good way of phrasing it, the Quattro can only know the "true" color on a 5mp level but never on a 19mp level. Whereas the Merrill can know true color on a 14.7mp level.

The Quattro sensor can know with absolute certainly in many cases the exact colors for every pixel in each quad. So there's a high likelihood it is mostly right.

mike earussi Veteran Member • Posts: 7,080
Re: The Quattro Knows

Roland Karlsson wrote:

Kendall Helmstetter Gelner wrote:
The Bayer sensor may get lucky guesses, but it can never KNOW the correct color values. Because the bayer sensor has to guess for many many values, it is mostly wrong. Not wrong by much, but wrong.

The Quattro sensor can know with absolute certainly in many cases the exact colors for every pixel in each quad. So there's a high likelihood it is mostly right.

The Quattro can only "know" true color to a 5mp level because it averages the two lower levels together at that resolution. And though it may show some color variation at the 19mp level, those variations won't ever be "true" at that resolution, just an average.

You are very consistent. That is certain

I assume you have not read information theory?

You can show, with a quite simple reasoning, that it cannot KNOW. It is just to scale up the situation from 4 to 4 million top pixels per pixel in the other layers.

So, you throw in a "in many cases" in your assertion. Yeah! What do you mean by "many cases"? For it to work as a reliable image sensor, it has to be able to do it in almost all cases, or at least in a high percentage of the cases.

So - I do not say you are wrong about your assertion, that the Quattro is good at finding the correct color. But I do say, that you need some more evidence before I do believe that it is as good as you claim.

Personally, I think it is going to be good enough.

This 30 million detector sensor will probably easily beat a 30 million detector Bayer sensor, regarding resolution, and also regarding color artefacts.

amdme127 Regular Member • Posts: 481
Re: The Quattro Knows
1

mike earussi wrote:

Kendall Helmstetter Gelner wrote:

The Bayer sensor may get lucky guesses, but it can never KNOW the correct color values. Because the bayer sensor has to guess for many many values, it is mostly wrong. Not wrong by much, but wrong.

Maybe that's a good way of phrasing it, the Quattro can only know the "true" color on a 5mp level but never on a 19mp level. Whereas the Merrill can know true color on a 14.7mp level.

The other assumption that is going on here is that the 4.9mp layers doesn't produce better results then the two lower layers of the Merrill Sensors.  If the Merrill sensors lower layers collect enough noise and are dirty, all the extra sensor sights might be a detriment instead of an advantage due to processing the dirty data and if noisy enough makes finding the accurate color much less accurate and maybe in harder then the 4.9 mp layers in the new chip.  We may be getting to hung up on the megapixel change and don't realize this might be a very good things for the lower layers because it can collect more accurate and cleaner results.

-- hide signature --

Matt
"Knowledge, absolutely sure of its infallibility is faith" -Yevgeny Zamyatin
http://ureye-studio.com

 amdme127's gear list:amdme127's gear list
Sigma dp2 Quattro Sigma SD9 Sigma SD14 Sigma SD15 Sigma 50mm F1.4 EX DG HSM +12 more
Erik Magnuson Forum Pro • Posts: 12,247
Re: The Quattro Knows

amdme127 wrote:

We may be getting to hung up on the megapixel change and don't realize this might be a very good things for the lower layers because it can collect more accurate and cleaner results.

Or confusing an ideal implementation with a real one.  The old saw applies: "In theory, there is no difference between practice and theory.  In practice, there is."

-- hide signature --

Erik

hexxthalion Contributing Member • Posts: 538
Re: Once more with feeling

Laurence Matson wrote:

Since the "blue" layer reads a narrower set - more discrete set - it is where acuity will be defined. It's the steak or tofuburger; the "red" and "green" add the lettuce and French dressing.

Yes, this is a simple explanations. Perhaps some engineers can parse it better. I just studied acting.

-- hide signature --

Laurence

now we're talking business thank you for simplistic explanation

 hexxthalion's gear list:hexxthalion's gear list
Fujifilm FinePix X100 Sigma DP2 Merrill Fujifilm X-Pro1 Nikon AF Micro-Nikkor 60mm f/2.8D Fujifilm XF 18mm F2 R +11 more
mike earussi Veteran Member • Posts: 7,080
Re: Once more with feeling

Laurence Matson wrote:

Laurence Matson wrote:

What the imager has is 19 million spatial locations. How the pixels are counted is once again a big deal for those discussion types. I am guessing that the G and R layers each have around 5 million pixels and the top B, 19 million. Or thereabouts.

Of course, some of our favorite negativists will argue that this is not really an X3 imager. That also, is nonsense. There are 3 layers (X3) each of which collects stuff to yield a full-color reading at each spatial location. The oh-so obvious - at least to Ricardo - interpolation that has to be going on is a moot point at best. Moot on, if you want.

Finally, on the critical point of pixel-level sharpness - acuity - it is there, as always. The question really is how one could conceive it not being there. Acuity or mush, depending on the technology, is defined by the top layer. In the case of the single-layer Bayer process in all of its iterations, the mush comes from the fact that none of the pixels is acute from its neighbors. The Foveon imager pixels are. The 19 million plus acute blue pixels in the top layer define the spatial locations.

To my mind, this imager is true to the process. Just a different solution.

It may be a bit less straightforward and thus harder for some to get their heads around. Just wait for the images and duct-tape your jaws for support ahead of time.

As with all previous Foveon imagers there are layers where colors are detected. The "filtering" done in the silicon and the filtering process is really just counting electrons at each discrete location. The electrons are the "corpses" from the expired photons. Since photons carry energy in proportion to their frequency, the stronger ones will penetrate furthest and the weaker one will penetrate least. So the "blue" layer is merely a device to count how many dead photon bodies are lying around; the same goes for the "red" layer and the "green" layer.

Now comes Ricardo's famous interpolation, but not behind the tree where he thought it was hiding. Using these three discrete data points, an educated guess can be made about the color. The educated guess becomes more accurate as the imager processing is refined.

Simplistically put, this is how it works. What this means for the Quattro imager is that the data set made up of varying variables. The "blue" detectors count a narrow but smaller discrete set. The "green" and the "red" a large one.

Since the "blue" layer reads a narrower set - more discrete set - it is where acuity will be defined. It's the steak or tofuburger; the "red" and "green" add the lettuce and French dressing.

Yes, this is a simple explanations. Perhaps some engineers can parse it better. I just studied acting.

Laurence, so did I, and now you've got me thinking about photon acting jokes:

Alas, poor photon I knew it well.

To be a photon or not to be a photon, that is the question.

I've come to bury photons, not to praise them.

Is this a photon I see before me?

-- hide signature --

Laurence
laurence at appledore-farm dot com
"I thought: I read something in a book, I dream, I imagine, and it comes true. And it is exactly like this in life.
"You can dream, and it comes true, as long as you can get out of the certitudes. As long as you can get a pioneering spirit, as long as you can explore, as long as you can think off the grid. So much time we spend in our education, in our lives is spent learning certitudes, learning habits, trying to fight against the unknown, to avoid the doubts and question marks. As soon as you start to love the unknown, to love the doubts, to love the question marks, life becomes an absolutely fabulous adventure."
Bertrand Piccard, a Swiss person
http://www.pbase.com/lmatson
http://www.pbase.com/sigmadslr
http://www.howardmyerslaw.com

Keyboard shortcuts:
FForum MMy threads