Advantages of Foveon sensor Locked

Started Jun 23, 2013 | Discussions
This thread is locked.
OP Roland Karlsson Forum Pro • Posts: 27,539
Re: Advantages of Foveon sensor

Rudi wrote:

Wrong ! That simply means the Sensor can see more fine nuances of a color gradient, the reproduction of the photographed objekt is more accurate.

Not necessarily.

If Foveon can see things that are invisible to the human eye, then that does not increase the accuracy of the camera used as a photographic device that are used to make images that are meant to look like reality.

It might make it a more fun tool to use to be able to record color gradients that normally are invisible. Just as an IR camera and an UV camera might be fun.

Now, it would be nice to see if it is true that Foveon can separate color gradients that RGB cameras cannot. I have not seen such a comparison.

And NOTE - I am not talking about color problems at pixel level. I know those exists for Bayer cameras.

 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
bobn2
bobn2 Forum Pro • Posts: 56,769
Re: Advantages of Foveon sensor

Lin Evans wrote:

Bobn2 wrote:

Lin Evans wrote:

Bobn2 wrote:

Lin Evans wrote:

Hi Mike,

The issue with color "accuracy," as you say, is primarily one of post capture processing, however, no amount of post processing can retrieve subtle variations of hue which were not captured in the first place.

No Bob,

This is exactly the problem with the CFA sensor. If subtle variations of color are not captured - and they are often not captured with CFA sensors, then they can't be reproduced in the finished image or print.

The facts are the facts. Your subjective preference is your preference. There is nothing wrong at all with you preferring the Foveon's colour response, but it is simply not as accurate as the organic colour filter. You should be able to express your preference without needing to bend fact to make your preference somehow 'factual'.\

No Bob,

What you present as "facts" are not facts at all. They are "your" suppositions based on what you have read. When theory and practice collide and practical experience doesn't match theory, then practical experience trumps theory. This is true in science or in the real world.

It depends on the 'practical experience'. The 'practical experience' of the people who tell you that a copper bracelet reduces rheumatism, or the 'practical experience' of people who tell you that they were abducted by aliens, etc. In short, old wives tales don't equal 'practical experience'. What equals practical experience is properly conducted and rigorous experiments. And, surprise surprise those back up the theory. The CIE colour vision model is not a new thing, it has been the basis of colour reproduction technology since the 1930's. This stuff is pretty well understood, and what is clear is unless the filters match the correct three stimuli frequency response, you will get metamerism (incorrect colours).

Your 'practical experience' simply says that you prefer the colour rendering of your Sigma to the 'correct' rendering - as I said, that is your privilege, but don't pretend that your preference is a universal truth that everyone else must accept.

No Bob, What you have done is present what you "think" is actuality. It's not me who is creating a fantasy. I "know" what I get from my very large collection of both Sigma and pro-level CFA cameras. You present yourself as an "expert" on Foveon sensors, yet you have yet to demonstrate that you know anything at all about them except what you have read which fits your own quite obvious bias. You are sounding very much like a Joe W. "mini-me" here.

Joe Wisniewski, a man who has actually designed Foveon sensors into equipment. Someone who actually knows how colour reproduction works. So, if I sound like him, it's not a surprise. He's right and I am too.

I have told you that CFA sensors have very poor color resolution.

You have then told me an untruth.

It has been demonstrated "scientifically" by color resolution chart photography multiple times and right here on dPReview.

I have never seen such a scientific demonstration. Maybe you could reference once such.

If the sensor can't "resolve" the colors, all the "organic" filtering it the world can't produce the colors.

Totally meaningless. The colour filters provide the colour separation between the three colour channels. They do it more effectively than do the silicon filters of the Foveon sensor, they have a sharper cutoff which more closely approximates the required RGB primaries.

Interpolating colors simply is not as accurate as primary detection of colors.

Again meaningless. Sampling RGB values at every pixel position does not by itself produce more accurate colour, especially if the filters are a poor approximation to the standard primaries. All it means is that there are more bad primaries. In any case, when you display as a JPEG those chroma channels are going to be decimated in any case, throwing away any extra colour information you think you might have (one reason why the comparisons done here using JPEG images are meaningless).

In any case, the colours produced from a Foveon raw are far more highly 'interpolated' in the colour space than they are from a Bayer camera. Those require simple linear weighting to produce the standard primaries, while Foveon's require complicated non-linear maths to get a reasonable result.

Now you keep right on arguing and presenting your irrelevant "data" and you go right on believing what you wish to believe. But "please" stop trying to tell people who actually "use" this technology that they are blind and incompetent.

I cannot stop because I never started. I haven't told anyone who actually uses this technology that they are blind and incompetent, and it is dishonest of you to suggest that I have.

bobn2
bobn2 Forum Pro • Posts: 56,769
Re: Advantages of Foveon sensor

Rudi wrote:

"They both need an IR cut filter to take out the red. Not how much wider the Blue and Red are, ow much more overlap. That simply means the sensor cannot distinguish as many colours, and gets more of them wrong (high metamerism)."

Wrong ! That simply means the Sensor can see more fine nuances of a color gradient, the reproduction of the photographed objekt is more accurate.

What is a 'color gradient'?

To be more specific, what is a 'color'?

OP Roland Karlsson Forum Pro • Posts: 27,539
Re: Yet another advantage

One advantage is that binning is very easy with Foveon and tricky with Bayer.

Fuji had a sensor with a strange pattern called EXR

http://www.imaging-resource.com/EVENTS/PKNA08/1222727986.html

 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
bobn2
bobn2 Forum Pro • Posts: 56,769
Re: Yet another advantage

Roland Karlsson wrote:

One advantage is that binning is very easy with Foveon and tricky with Bayer.

Yes, that is an advantage. Thanks.

Fuji had a sensor with a strange pattern called EXR

http://www.imaging-resource.com/EVENTS/PKNA08/1222727986.html

Fuji has a history of them, they have their proponents.

OP Roland Karlsson Forum Pro • Posts: 27,539
Re: Advantages of Foveon sensor

Bobn2 wrote:

What is a 'color gradient'?

To be more specific, what is a 'color'?

A good question.

For an RGB image, a color is an RGB triple in some color space.

NOTE that most Bayer CFA sensors are reasonable good RGB sensors.

RAW Foveon images don't have any colors in the normal sense. But - you can use some method to convert it to RGB. Then, each pixel gets an RGB color. And - depending on method you get different colors.

Metamerism might make it impossible to do this conversion correct.

Then, of course, all is made more complicated by the need of color profiles for your devices, e.g. the camera and that you have image editors (e.g. SPP) that may transform the RGB triples, in arbitrary strange ways.

Sigma, and some Bayer camera, software do desaturation of dark areas and also soft clipping of bright areas. Both affecting color accuracy.

etc etc etc

 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
Lin Evans
Lin Evans Forum Pro • Posts: 17,390
Re: Advantages of Foveon sensor

Bobn2 wrote:

Lin Evans wrote:

Bobn2 wrote:

Lin Evans wrote:

Bobn2 wrote:

snip...

Your 'practical experience' simply says that you prefer the colour rendering of your Sigma to the 'correct' rendering - as I said, that is your privilege, but don't pretend that your preference is a universal truth that everyone else must accept.

No Bob, What you have done is present what you "think" is actuality. It's not me who is creating a fantasy. I "know" what I get from my very large collection of both Sigma and pro-level CFA cameras. You present yourself as an "expert" on Foveon sensors, yet you have yet to demonstrate that you know anything at all about them except what you have read which fits your own quite obvious bias. You are sounding very much like a Joe W. "mini-me" here.

Joe Wisniewski, a man who has actually designed Foveon sensors into equipment. Someone who actually knows how colour reproduction works. So, if I sound like him, it's not a surprise. He's right and I am too.

How does one "design Foveon sensors" into equipment? That's got to be the most convoluted use of the term "design" I've heard yet. One might "design" equipment around a Foveon sensor, but Joe has not "designed" Foveon sensors, period.

Why am I not surprised you are familiar - LOL...

I have told you that CFA sensors have very poor color resolution.

You have then told me an untruth.

It's not obvious that you not only have a deficit in your understanding of Foveon technology, but also a deficit in your knowledge of CFA sensors...

It has been demonstrated "scientifically" by color resolution chart photography multiple times and right here on dPReview.

I have never seen such a scientific demonstration. Maybe you could reference once such.

Why should I do all the work for you Bob - you're the one who admits you haven't seen such and yet tell me I'm wrong. I suggest you try a search for color resolution.

snip - irrelevant

Lin

 Lin Evans's gear list:Lin Evans's gear list
Olympus E-10
OP Roland Karlsson Forum Pro • Posts: 27,539
Re: Advantages of Foveon sensor

Lin Evans wrote:

Bobn2 wrote:

I have told you that CFA sensors have very poor color resolution.

You have then told me an untruth.

It's not obvious that you not only have a deficit in your understanding of Foveon technology, but also a deficit in your knowledge of CFA sensors...

Guys! You are Chimpanzee fighting with broad swings

Instead of claiming the other one wrong all the time - stop and think!

There is obviously a misunderstanding here.

 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
bobn2
bobn2 Forum Pro • Posts: 56,769
Re: Advantages of Foveon sensor

Roland Karlsson wrote:

Bobn2 wrote:

What is a 'color gradient'?

To be more specific, what is a 'color'?

A good question.

For an RGB image, a color is an RGB triple in some color space.

Exactly, and each of the 'R', 'G' and 'B' channels is defined as a function of light frequencies, not a single frequency.

NOTE that most Bayer CFA sensors are reasonable good RGB sensors.

RAW Foveon images don't have any colors in the normal sense.

Well they do, the Foveon 'filters' nave their own frequency response and so define a colour space of their own. It is just that the response is way different from the CIE primaries, so the colour space is very far from any CIE or other commonly used RGB space

But - you can use some method to convert it to RGB. Then, each pixel gets an RGB color. And - depending on method you get different colors.

Certainly, but I understand that it requires non-linear maths to get the Foveon colour space close to CIE.

Metamerism might make it impossible to do this conversion correct.

Yes.

Then, of course, all is made more complicated by the need of color profiles for your devices, e.g. the camera and that you have image editors (e.g. SPP) that may transform the RGB triples, in arbitrary strange ways.

It is further transforms to device specific spaces - it helps f the spaces are not oo dissimilar.

SandyF Forum Pro • Posts: 14,941
link: Dick Merrill's posts

finally located the thread and posts for which I've been searching, by  Foveon's Richard B Merrill

http://www.dpreview.com/forums/post/6553201

Best regards, Sandy
http://www.pbase.com/sandyfleischman (archival)
http://www.flickr.com/photos/sandyfleischmann (current)

 SandyF's gear list:SandyF's gear list
Sigma DP2 Sigma DP1 Sigma DP1 Merrill Sigma DP2 Merrill Sigma DP3 Merrill +7 more
bobn2
bobn2 Forum Pro • Posts: 56,769
Re: Advantages of Foveon sensor

Lin Evans wrote:

Joe Wisniewski, a man who has actually designed Foveon sensors into equipment. Someone who actually knows how colour reproduction works. So, if I sound like him, it's not a surprise. He's right and I am too.

How does one "design Foveon sensors" into equipment? That's got to be the most convoluted use of the term "design" I've heard yet. One might "design" equipment around a Foveon sensor, but Joe has not "designed" Foveon sensors, period.

Use of a designers vernacular, or have you not seen the phrase 'design in' before? Not that it is in any way relevant to this discussion.

Why am I not surprised you are familiar - LOL...

I don't know why, care to elucidate?

I have told you that CFA sensors have very poor color resolution.

You have then told me an untruth.

It's not obvious that you not only have a deficit in your understanding of Foveon technology, but also a deficit in your knowledge of CFA sensors...

You have told me an untruth that CFA sensors have very poor colour resolution. They do not. Of course it depends what you mean by 'color resolution'. If by it you mean resolution in a colour space, then you are wrong, they can produce very good resolution - this is the figure that DxOmark measures as 'colour sensitivity' and Bayer cameras routinely produce over 21 bits - that is 2 million distinguishable colours. If you mean sampling frequency in the chroma channels, you are also wrong. It is adequate. Not as 'good' in relationship to the pixel scale as Foveon cameras, but then Bayer cameras are available with higher pixel counts, so it rather evens out. So, whatever what you said was an untruth.

It has been demonstrated "scientifically" by color resolution chart photography multiple times and right here on dPReview.

I have never seen such a scientific demonstration. Maybe you could reference once such.

Why should I do all the work for you Bob - you're the one who admits you haven't seen such and yet tell me I'm wrong. I suggest you try a search for color resolution.

Did the search, found the 7000 odd hits, went down the first 100, didn't find anything remotely close to a 'scientific demonstration', came to the conclusion that you are bluffing again.

bobn2
bobn2 Forum Pro • Posts: 56,769
Re: Advantages of Foveon sensor

Roland Karlsson wrote:

Lin Evans wrote:

Bobn2 wrote:

I have told you that CFA sensors have very poor color resolution.

You have then told me an untruth.

It's not obvious that you not only have a deficit in your understanding of Foveon technology, but also a deficit in your knowledge of CFA sensors...

Guys! You are Chimpanzee fighting with broad swings

Instead of claiming the other one wrong all the time - stop and think!

There is obviously a misunderstanding here.

There is no misunderstanding, what there is is Lin's insistence on conduction the discussion on an 'argumentum ad hominem' level. That is, he does not give evidence to support his point of view, or logical arguments which support it, he tries to discredit the person making the counter argument. This seems to be his modus operandi. It makes for this kind of discussion, which has entertainment value, if you like that kind of thing, but little informative value.

I don't see it as much of a problem, I can have sensible discussions with the likes of you, and provide the entertainment with the likes of Lin. That way both people who like information and those who like entertainment are happy. All it does is use a little more of DPReview's disc space.

Lin Evans
Lin Evans Forum Pro • Posts: 17,390
Re: Advantages of Foveon sensor

Bobn2 wrote:

Lin Evans wrote:

It has been demonstrated "scientifically" by color resolution chart photography multiple times and right here on dPReview.

I have never seen such a scientific demonstration. Maybe you could reference once such.

Why should I do all the work for you Bob - you're the one who admits you haven't seen such and yet tell me I'm wrong. I suggest you try a search for color resolution.

Did the search, found the 7000 odd hits, went down the first 100, didn't find anything remotely close to a 'scientific demonstration', came to the conclusion that you are bluffing again.

If you are not interested enough to find one right here on the dPReview forum, I'm not interested enough in providing further education for you.

I don't "bluff."

Lin

 Lin Evans's gear list:Lin Evans's gear list
Olympus E-10
Laurence Matson
Laurence Matson Forum Pro • Posts: 11,968
What would he have known about this stuff?

SandyF wrote:

finally located the thread and posts for which I've been searching, by Foveon's Richard B Merrill

http://www.dpreview.com/forums/post/6553201

Best regards, Sandy
http://www.pbase.com/sandyfleischman (archival)
http://www.flickr.com/photos/sandyfleischmann (current)

NT

bobn2
bobn2 Forum Pro • Posts: 56,769
Re: Advantages of Foveon sensor

Lin Evans wrote:

Bobn2 wrote:

Lin Evans wrote:

It has been demonstrated "scientifically" by color resolution chart photography multiple times and right here on dPReview.

I have never seen such a scientific demonstration. Maybe you could reference once such.

Why should I do all the work for you Bob - you're the one who admits you haven't seen such and yet tell me I'm wrong. I suggest you try a search for color resolution.

Did the search, found the 7000 odd hits, went down the first 100, didn't find anything remotely close to a 'scientific demonstration', came to the conclusion that you are bluffing again.

If you are not interested enough to find one right here on the dPReview forum, I'm not interested enough in providing further education for you.

Bluff called.

I don't "bluff."

LOL

Lin Evans
Lin Evans Forum Pro • Posts: 17,390
You want to put your money...

where your mouth is and make it interesting?

Lin

Bobn2 wrote:

Lin Evans wrote:

Bobn2 wrote:

Lin Evans wrote:

It has been demonstrated "scientifically" by color resolution chart photography multiple times and right here on dPReview.

I have never seen such a scientific demonstration. Maybe you could reference once such.

Why should I do all the work for you Bob - you're the one who admits you haven't seen such and yet tell me I'm wrong. I suggest you try a search for color resolution.

Did the search, found the 7000 odd hits, went down the first 100, didn't find anything remotely close to a 'scientific demonstration', came to the conclusion that you are bluffing again.

If you are not interested enough to find one right here on the dPReview forum, I'm not interested enough in providing further education for you.

Bluff called.

I don't "bluff."

LOL

 Lin Evans's gear list:Lin Evans's gear list
Olympus E-10
Rudi Senior Member • Posts: 1,353
Re: Advantages of Foveon sensor

Talking about frequency we got to talk about bandwidth too and in this respect I suggest the lecture of Dick Lion´s discours on the Foveon sensor. With the Foveon technology we get exact voltage values for each single color of the color spectrum in each single pixel position.

The problem is to convert these values to the somehow limited 3 color RGB values and here, I think, Sigma has done a pretty good job in firmware design and SPP developement. The Foveon sensor technology is much more complex then CFA sensor technology and that is why it can´t be compared with each other on a 1:1 basis.

OP Roland Karlsson Forum Pro • Posts: 27,539
Re: Advantages of Foveon sensor

Bobn2 wrote:

There is no misunderstanding,

There actually was.

Lin meant spatial color resolution at pixel level, and there Foveon is better.

You meant color space color resolution.

And then you added spatial color resolution as a possible alternative, but scaled because Bayer imagers have more pixels.

Not that it is important - just saying

 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
bobn2
bobn2 Forum Pro • Posts: 56,769
Re: Advantages of Foveon sensor

Roland Karlsson wrote:

Bobn2 wrote:

There is no misunderstanding,

There actually was.

Lin meant spatial color resolution at pixel level, and there Foveon is better.

You meant color space color resolution.

And then you added spatial color resolution as a possible alternative, but scaled because Bayer imagers have more pixels.

Not that it is important - just saying

Even then, no misunderstanding. His statement was

'I have told you that CFA sensors have very poor color resolution.'

This statement is untrue. CFA sensors do not have 'very poor' colour resolution under either definition.

If the definition is 'spatial colour resolution', then they have adequate, not very poor resolution. Their chroma channel resolution is half that of the Foveon sensor, but since most, if not all display and transmission methods for colour images decimate the chroma channels in any case, that is adequate. In fact it could be argued that Foveon wastes raw file space by encoding colour information that will never be used. If one takes into account that Bayer sensors are available with higher absolute resolution, then the argument that they have 'very poor' colour resolution becomes even less tenable.

I am arguing with someone who chooses to argue with unsubstantiated rhetoric and argumentum ad hominem. That's not a complaint, just a recognition of what it is, and the type of discussion it will produce.

OP Roland Karlsson Forum Pro • Posts: 27,539
Its a fun old thread to read

Its a fun old thread to read

http://www.dpreview.com/forums/post/6544921

Its a thread from 2003, with Merrill himself arguing with Erik and Joseph and Phil etc.

There are some sharp and knowledgeable men, but they still don't agree

That's forums!

 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
Keyboard shortcuts:
FForum MMy threads