DSLR Sensor in P&S

[Liquid_Thought]

Well it's a good thing you're here to correct my spelling o' might English teacher.

"Algorithms," Liquid.

 

[Andrew dB]

This would be a very significant limitation for the camera: the
former would be a wonderful camera for Cartier-Bresson wannabies
( me pleads guilty)...

Yes please... can we make it say 24mm while we're at it?

Matthias

Probably not.

Wide angle lenses on SLRs are all retrofocal designs to allow for enough space between rear lens element and the sensor/film to fit the mirror box. A side effect of this is that the light hitting the sensor is usually fairly telecentric or at least more than would be the case in a simpler lens design.

Campacts and rangefinders have much smaller, simpler (and sometimes higher quality) wide lenses mainly because they don't need to be retrofocal but the flipside of this is that light will be hitting the edge of the film at fairly extreme angles. It isn't much of a problem for film but put a CCD in there and you will likely see pretty horrendous vignetting as the light is obscured by the 'well-like' nature of the photosites.

It could be alleviated perhaps by using a different lens design with a field flattener or possibly with a back thinned CCD (reverse mounted thus perfectly flat) but both of those solutions are likely to be very expensive and are rarely used outside high performance imagers.

 

[hunter]

When someone discovers a way to make pictures that does not rely on those pesky photons - and if that way is more efficient - and if the technology is economical to build - and if ................

I think I will return to my favorite liquid (Balmore on ice) - it may make some of the ideas posted here seem more possible

hunter

 

[Liquid_Thought]

Maybe that's how I came up with the ideas

When someone discovers a way to make pictures that does not rely on
those pesky photons - and if that way is more efficient - and if
the technology is economical to build - and if ................

I think I will return to my favorite liquid (Balmore on ice) - it
may make some of the ideas posted here seem more possible

hunter

 

[Joseph S Wisniewski]

Great. Well debates often wrap into circles. Here come the English
teacher cracks. GOOD GAME!

The point was not to correct your English, as I understand it's a second language for you, and allow for that in our communucations.

The point is that your basic argument is flawed. That the people who write science fiction do take care to avoit obvious violations of well known principles of physics. They were going to use the term "lasers" in StarTrek, but the people who created the show realized that the audience might know enough physics to realize that lasers simply burnt holes in stuff. If you wanted a ray gun that that could stun people or make them disappear neatly, you had to call it something different.

Got that? They understood the physics and assumed that the audience (or enough of the audience to be annoying) might also understand the physics.

They don't "redefine" lasers to suit the story, they propose another device that can perform the functions that they need for the plot, and make it plausable enough so that you can assume it's been invented in the 300 years between our time and theirs.

You're not doing as well as a typical science fiction writer, here. You're trying to redefine the physics behind photons or the devices that detect them, with not a shred of science to support how this is going to be done. None of the examples you've provided show something analogous in any othe field. Physics has not been defied in the evolution of microprocessors. Changing materials or shrinking geometires isn't going to allow you to count photons that aren't there.

And if that's all you got out of my last post, the perception that I was out to correct your English, then you really need to read more and talk less.

--
A cyberstalker told me not to post anymore...
So I'm posting even more!

Ciao!

Joe

http://www.swissarmyfork.com

 

[Joseph S Wisniewski]

Well it's a good thing you're here to correct my spelling o' might
English teacher.

You already responded to him once about this:

http://forums.dpreview.com/forums/read.asp?forum=1018&message=8983973

Is that all you've got left, repeated complaints about "English teachers" and circualr arguments?

--
A cyberstalker told me not to post anymore...
So I'm posting even more!

Ciao!

Joe

http://www.swissarmyfork.com

 

[Liquid_Thought]

Great, I have an English teacher AND an editor. This is a good forum.

Well it's a good thing you're here to correct my spelling o' might
English teacher.

You already responded to him once about this:

http://forums.dpreview.com/forums/read.asp?forum=1018&message=8983973

Is that all you've got left, repeated complaints about "English
teachers" and circualr arguments?

--
A cyberstalker told me not to post anymore...
So I'm posting even more!

Ciao!

Joe

http://www.swissarmyfork.com

 

[Borstad]

We all agree on at least one post!!!

(Emphasis added)

GREAT , I have an English teacher AND an editor. This is a GOOD
forum.

 

[Liquid_Thought]

I've noticed something quite interesting in this thread. The more that people got onboard the 'bash Liquid Thought bus', the more you became braver in your insults against me.

I wonder what this mob mentality means?

Oh well, it doesn't really matter because when the technology does arrive and I was right, you'll probably say to people here "I predicted that in 2004. I knew it was coming."

Just to clarify, my original argument was why don't they apply future research toward a technology that would allow the same qualities of a full frame sensor such as ISO & Noise and apply it to a sensor the size of the current 8mp prosumer sensors. This will be done and I will be right. It is currently impossible.

But it will not remain so. And when it does happen, and it will happen, remember how you treated me. It's the same with all people who are great visionaries. Everyone always thinks their ideas are wacko. But then they are proved to be true. What happens to those people who called them wacko? No one knows and no one cares either. They thought they had a grasp on but really all they were doing was harrassing someone who could think beyond the confines of the shell they all existed within. The earth does not revolve around the sun, it's the center of the universe. Everyone knows that. The world is not round, what kind of drugs are you taking? Photons can't fit in an area much smaller than ful-frame, it's basic physics; you should know that. It's impossible. Blah blah blah blah blah blah blah. All the while the mob pats each other on the back. Good one! Yeah, you ripped him there. 2-0, blah blah blah blah blah blah.

So when it does happen. Then will you apologize for being jerks, for not realizing that you were talking with someone who could see that the impossible is only a measure of those things which we see barriers. If we begin to believe that these things are not barriers, then we work on them because it is possible. If it is perceived impossible, no one works on it. You people are a bunch of pessemistic mean jerks. When this happens, you won't say, wow I'm surprised, I talked with this guy in 2004 and he was sure it was going to happen & I made fun of him. Nope, instead you'll lie like the honorless dogs you are and say that you knew it was going to happen all along. You'll say this because you've spent too much time trying to put up this facade that you know so much, that you're a professional. It will be impossible for you admit that you were wrong in this matter. So you'll lie about it to cover yourself, and cover any remaining consciouses you might have. So bash away. To me, it's all blah blah blah blah blah blah. I must admit that I'm greatly surprised at your lack of realization to the state of technology advancement. If you aren't going to get it, you just aren't going to get it, so I think I'll just repeat this song now instead of attempting to get it through you fat heads how things work around here now.

Great. Well debates often wrap into circles. Here come the English
teacher cracks. GOOD GAME!

The point was not to correct your English, as I understand it's a
second language for you, and allow for that in our communucations.

The point is that your basic argument is flawed. That the people
who write science fiction do take care to avoit obvious violations
of well known principles of physics. They were going to use the
term "lasers" in StarTrek, but the people who created the show
realized that the audience might know enough physics to realize
that lasers simply burnt holes in stuff. If you wanted a ray gun
that that could stun people or make them disappear neatly, you had
to call it something different.

Got that? They understood the physics and assumed that the audience
(or enough of the audience to be annoying) might also understand
the physics.

They don't "redefine" lasers to suit the story, they propose
another device that can perform the functions that they need for
the plot, and make it plausable enough so that you can assume it's
been invented in the 300 years between our time and theirs.

You're not doing as well as a typical science fiction writer, here.
You're trying to redefine the physics behind photons or the devices
that detect them, with not a shred of science to support how this
is going to be done. None of the examples you've provided show
something analogous in any othe field. Physics has not been defied
in the evolution of microprocessors. Changing materials or
shrinking geometires isn't going to allow you to count photons that
aren't there.

And if that's all you got out of my last post, the perception that
I was out to correct your English, then you really need to read
more and talk less.

--
A cyberstalker told me not to post anymore...
So I'm posting even more!

Ciao!

Joe

http://www.swissarmyfork.com

 

[UnderWater]

Don't let the nitpickers and scientist wannabes get to you.

What was evolutionary about the CD?

I suggested the idea to the head of R&D of a company in that area several years before all the patents were taken out by Phillips etc. Was scoffed at and told it was not possible. hmmm

Look at how digital communications bandwidth/distance is now exceeding "theoretical" limits of a few years ago. Yes people said they were limits in print.

Back to thread topic, what about doing signal processing on images over time to increase image information and reduce noise. This is just one obvious idea.

My eyes and brain do this all the time, and they don't have a 35mm sized sensor in them. Hah, but they are exceeding the laws of physics! wow!
In this analogy, moire is just the camera's version of an optical illusion.

Someone I know does military research into image processing and thinks that the technology in current digital cameras is very primitive compared to what can be done. He actually laughs at the current approaches in consumer digital cameras.

--len

 

[Inigo Montoya]

Did you know that we're about to go to hexidecimal numbering on the
internet?

That's about the funniest thing I've read today, but please.... do elaborate, especially with regard to how that will be a better approach than say, octal numbering.

 

[Petteri Sulonen]

Andrew dB wrote:
[snip]

Probably not.

Wide angle lenses on SLRs are all retrofocal designs to allow for
enough space between rear lens element and the sensor/film to fit
the mirror box. A side effect of this is that the light hitting
the sensor is usually fairly telecentric or at least more than
would be the case in a simpler lens design.

Campacts and rangefinders have much smaller, simpler (and sometimes
higher quality) wide lenses mainly because they don't need to be
retrofocal but the flipside of this is that light will be hitting
the edge of the film at fairly extreme angles. It isn't much of a
problem for film but put a CCD in there and you will likely see
pretty horrendous vignetting as the light is obscured by the
'well-like' nature of the photosites.

It could be alleviated perhaps by using a different lens design
with a field flattener or possibly with a back thinned CCD (reverse
mounted thus perfectly flat) but both of those solutions are likely
to be very expensive and are rarely used outside high performance
imagers.

Epson says that the Cosina-Voigtländer 12/5.6 will work on the RD-1. It'll be very interesting to see how well it does. They must've managed some wizardry to get around this issue.

Petteri
--



[ http://www.prime-junta.tk ]

 

[Petteri Sulonen]

Liquid_Thought wrote:
[snip]

But it will not remain so. And when it does happen, and it will
happen, remember how you treated me. It's the same with all people
who are great visionaries. Everyone always thinks their ideas are
wacko. But then they are proved to be true. What happens to those

[snip]

"...but they also laughed at Bozo the Clown."

Which are you, O Great Visionary?

Petteri
--



[ http://www.prime-junta.tk ]

 

[Osku]

I suppose you could always coat the sides of the buckets with
Teflon. And make them a tessellating shape for near-100% area
coverage.

I recall that there is a specific paint (horribly expensive) that has even less friction with water than teflon, something they do use in competition sailing boats, but it might be that teflon is actually better.

Somehow, I feel that the analogy is starting to break down, though...

Yep, if you try to make a silly example a little less stupid, it is always still a silly example;-)

--
Osku

 

[Osku]

IOW, by desiging a sensor that'll capture quanta over a much wider
range of wavelengths than visual light (which is completely
feasible), we'll get a camera that gets a very interesting view of
the world -- but a not at all "realistic" one.

I think the idea here was that you could record all possible wavelengths from IR to UV (no filters in the sensor), and all the time have the intention to use only the visible area to create the actual final digital photo. A nex gen RAW converter might be developped so that it could take some advantage of the higher and lower frequencies, so that e.g. noise level would be lower in some detais (you can not record more visible light than does excist, but you could use IR and UV data do define sharper borders for objects and other details in a image, make it easier to decide if a pixel is noise or a detail, so those might be single color layers that the software would use to create sharper RGB image). To do this, the sensor would possibly need a visual light sensor and separate UV and IR sensors per every pixel (as Fuji has separated the sensors for visible light in SR), or four sensors per pixel (SR sensor + UV + IR).

But I think that for some time it is more usefull that they just develop bettr more mature sensors for visible light area, since that work is not done yet (less noise, better color reproduction, wider usable dynamical range).

--
Osku

 

[Inigo Montoya]

Bozo was a visionary in the use of orange hair.

Of course even that is now in dispute: http://msnbc.msn.com/id/5075932/

But it will not remain so. And when it does happen, and it will
happen, remember how you treated me. It's the same with all people
who are great visionaries. Everyone always thinks their ideas are
wacko. But then they are proved to be true. What happens to those

[snip]

"...but they also laughed at Bozo the Clown."

Which are you, O Great Visionary?

Petteri
--



[ http://www.prime-junta.tk ]

 

[Joseph S Wisniewski]

IOW, by desiging a sensor that'll capture quanta over a much wider
range of wavelengths than visual light (which is completely
feasible), we'll get a camera that gets a very interesting view of
the world -- but a not at all "realistic" one.

I think the idea here was that you could record all possible
wavelengths from IR to UV (no filters in the sensor), and all the
time have the intention to use only the visible area to create the
actual final digital photo. A nex gen RAW converter might be
developped so that it could take some advantage of the higher and
lower frequencies, so that e.g. noise level would be lower in some
detais (you can not record more visible light than does excist, but
you could use IR and UV data do define sharper borders for objects
and other details in a image, make it easier to decide if a pixel
is noise or a detail, so those might be single color layers that
the software would use to create sharper RGB image). To do this,
the sensor would possibly need a visual light sensor and separate
UV and IR sensors per every pixel (as Fuji has separated the
sensors for visible light in SR), or four sensors per pixel (SR
sensor + UV + IR).

Osku, this is a fun one.

I can see it being a potentially useful technique, but there are a large number of difficultises to overcome.

1) There's the problem of detecting UV and IR, without diminishing the visible light performance of the sensor.

The Fuji SR sensor works by having two different sizes of sensors. The new highlight sensors are quite small, relative to the main sensors, so they don't take much area away from the main sensors.

To add two more colors to a Bayer style matrix, large enough to contribute to image detail, would involve taking a substantial area away from the RGB sensors. The net effect would be a wash, you increase detail in the new sensors exactly as much as you decrease it in the originals.

There may be colutions to this, perhaps some improvement in the Foveon design (it would have to be a really big improvment, though).

2) Most lenses aren't corrected in the UV or IR, so if the visible part of the image is in focus, the IR and UV are soft. Lenses can be designed around this, problem, but it's not easy, and there are other tradeoffs.

3) Anti-reflective coatings that work well for visible light actually become "reflective coatings" in IR and increase flare and ghosts substantially. BBAR (broad band anti-reflective coatings) that work in UV, visible light, and IR don't have the efficiency of modern narrow band coatings, so improving the IR and UV flare adds more visible light flare.

4) Glass lenses don't pass much UV, absorbing a lot of UV from 350-400nm, and blocking pretty much everything below 350nm. There are lenses of artificial quartz and fluorite that get around this, but they're $4,000 - $15,000 like the famed 105mm UV Nikkor, the Pentax Quartz Takumar, or the more modern Coastal Optics UV macro. But that's the fun part of evolutionary science, we know it can be done, now all we need to do is figure out how to do it cheaply.

5) If you're going to use the UV and IR to increase detail, you have to realize that some of the detail is false, or wrong.

Here's a Dandelion that I shot yesterday with a UV filter and UV friendly EL-Nikkor lens.



Click here to see it full size and with all the image info.

http://www.pbase.com/image/29580142/original

That's not what a Dandelion lools like to humans, the dandelion evolved that "landing pad" pattern to attract bees, who see in UV.

IR is just as bad, is sees right through the skin to the surface veins, and through makup to whatever the makup is covering.You won't be very popular with the ladies if you do that.

6) There's the social issue. Sony got roasted in the media when they introduced the firts camdorder with "nightshot", the ability to slide the IR blocking filter out of the optic path and record near IR from 700nm to around 1000nm. It turns out that many fabrics that are opaque in the visable frequencies from 400-700nm are transparent or translucent in the near IR.

The fabric wrap in this shot, for example, is a visually opaque green flowered thing.

Warning, Nudity

http://misheli.image.pbase.com/u39/the_wiz/upload/25666333.200303290208crop.jpg

There was a big media circus, calls to boycott Sony because of their "x-ray cameras" that could see through clothes. Of course, the odds of running into clothing that the IR camera could actually see anything interesting through were so poor as to be near nonexistant, but that didn't stop the media and civic organizations, once they had something availiable rally against.

If you look close at that picture, you'll also see what I mean about UV showing the surface veins.

But I think that for some time it is more usefull that they just
develop bettr more mature sensors for visible light area, since
that work is not done yet (less noise, better color reproduction,
wider usable dynamical range).

I agree, it's definitely better for the mainstream. Even though I'd personally love a camera with IR and UV capability, so I could stop screwing around with esoteric filters and long exposures.

--
A cyberstalker told me not to post anymore...
So I'm posting even more!

Ciao!

Joe

http://www.swissarmyfork.com

 

[Joseph S Wisniewski]

Don't let the nitpickers and scientist wannabes get to you.

Or even the real scientists.

What was evolutionary about the CD?

Basically, everything.

Laser discs evolved to the point where their data densities were sufficient to move from the large analog FM laser disc to a smaller disc.

Digital PCM evolved from something that needed a separate rack mounted box and could only be used in studios (like the Sony Beta based PCM audio recorders) to something you could sell at consumer prices.

I suggested the idea to the head of R&D of a company in that area
several years before all the patents were taken out by Phillips
etc. Was scoffed at and told it was not possible. hmmm

Can you back up that absurd claim?

Look at how digital communications bandwidth/distance is now
exceeding "theoretical" limits of a few years ago. Yes people said
they were limits in print.

Were the people who said there were limits "lay people", or the people actualyl versed in the technologies?

Back to thread topic, what about doing signal processing on images
over time to increase image information and reduce noise. This is
just one obvious idea.

Yes, an obvious idea for making the best use of the information you've got, but not for creating information where you don't have any.

My eyes and brain do this all the time, and they don't have a 35mm
sized sensor in them. Hah, but they are exceeding the laws of
physics! wow!

Of course they're not. But you already know this, it's basic information theory. If you fill in missing data based on assumptions, you run the risk that those assumptions are wrong.

In this analogy, moire is just the camera's version of an optical
illusion.

Someone I know does military research into image processing and
thinks that the technology in current digital cameras is very
primitive compared to what can be done. He actually laughs at the
current approaches in consumer digital cameras.

I agree completly. DSP is my field, although my specialty is psychoacoustics. I've got coworkers we "acquired" from Erim who are doing stuff well beyond what they can do today in a $500, or even a $50,000 professional camera. It takes time before $50,000,000 pieces of military hardware become consumerized.

But when they do, well, that still evolution, not revolution.

--
A cyberstalker told me not to post anymore...
So I'm posting even more!

Ciao!

Joe

http://www.swissarmyfork.com

 

[Petteri Sulonen]

Joseph S. Wisniewski wrote:
[snip]

But when they do, well, that still evolution, not revolution.

Lest we sound too curmudgeonly here, I think it is good to remember that on occasion genuine technological revolutions do happen. ICT in general and the Internet in particular are two recent examples; the internal combustion engine, the electric motor, and the steam engine somewhat earlier ones; steel, iron, and bronze earlier yet.

However, I have a feeling that even these revolutionary developments have a strong evolutionary component. It's more a matter of a genius putting existing things together in a new way to create something hitherto unseen, rather than coming up with something ad nihilo.

The people who came up with bronze were familiar with metals like copper, tin, gold, and silver; they knew how to beat them into ornaments and tools, and they knew what happens to them when you melt them. Iron required evolutionary improvements in smelting, but once people had figured out how to extract and purify copper and tin from ore rather than pure naturally occurring deposits, it was a matter of reaching a high-enough temperature and happening upon a promising-looking lump of rock. Steel was a similar story.

The steam engine built upon principles of mechanical pumps and knew knowledge about the physics of gases; the electric motor upon knew knowledge of physics of electricity and magnetism, and the internal combustion engine was basically a matter of perseverence: with it being well known that chemical energy can be converted to kinetic energy, it was only a matter of time before someone figured out how to do it without the intermediary of steam. Likewise for the newer examples.

I think the revolutionary potential of invention therefore hinges more upon the ability to recognize the significance of evolutionary changes in knowledge and technology: to realize that something we have now could be used for something quite different, with great effect.

Petteri
--



[ http://www.prime-junta.tk ]

 

[Joseph S Wisniewski]

I've noticed something quite interesting in this thread. The more
that people got onboard the 'bash Liquid Thought bus', the more you
became braver in your insults against me.

I wonder what this mob mentality means?

In your latest post, you call us "honorless dogs" and "mean jerks", among other thigns. Try that in person, see where it gets you. All the "mob mentality" means is that people reacting to your increasingly hostile conduct.

You've gone from "Why don't they reduce the pixel size on sensors the same way they make microchips smaller and smaller?"
http://forums.dpreview.com/forums/read.asp?forum=1018&message=8958335

To name calling and personal attacks. People are responding in kind.

"No offense man, but you're being dense."
http://forums.dpreview.com/forums/read.asp?forum=1018&message=8971106

"Oh I get it, you are an elitist" or "you're obviously not one to have an open mind"
http://forums.dpreview.com/forums/read.asp?forum=1018&message=8972638

"Good grief you people are annoying. Haven't you a single clue at all?"
http://forums.dpreview.com/forums/read.asp?forum=1018&message=8958335

Oh well, it doesn't really matter because when the technology does
arrive and I was right, you'll probably say to people here "I
predicted that in 2004. I knew it was coming."

That's the nice thign about the internet. There are search engines and plenty of records. It's easy to look back and see who got things right.

Just to clarify, my original argument was why don't they apply
future research toward a technology that would allow the same
qualities of a full frame sensor such as ISO & Noise and apply it
to a sensor the size of the current 8mp prosumer sensors. This will
be done and I will be right.

Highly unlikely. Pretty much up there with free energy and anti-gravity.

It is currently impossible.

But it will not remain so. And when it does happen, and it will
happen, remember how you treated me.

We've all treated you much more politely than you've treated the rest of us.

It's the same with all people
who are great visionaries. Everyone always thinks their ideas are
wacko. But then they are proved to be true.

But you're not one of the great visionaries. You're not out there working to solve the problem. You're not sponsoring or supporting the people who will. You're an armchair quarterback.

If, unlikely as it is, someone does bring about the technology you insist is soming, no one will give you any of the credit for it, for you have done nothing to merrit any of the credit.

In fact, I'd go so far as to say you're actions are teh technological equivelant of what I call the "Rakolta, Kohmehni, Wildmon effect". Your support of these concepts is so distasteful that you may actually drive away or discourage people working in these ares. It's cliche, but "with friends like you, who needs enemeies".

They thought they had a grasp on but really all they were doing was
harrassing someone who could think beyond the confines of the shell
they all existed within. The earth does not revolve around the sun,
it's the center of the universe. Everyone knows that. The world is
not round, what kind of drugs are you taking?

An excellent example. In 1492, scientists and navigators were quite aware the world was round, and had a pretty good idea on its diameter, about 8000 miles. (But it wasn't a good idea to say this too loudly around powerful church members). The "visionary" Christopher Columbus made some calculations based an erroneous translation of the work of Ptolemy, and some interesting experiments in line of sight that didn't take into consideration atmospheric diffraction, to arrive at a quite wrong figure in the neighborhood of 5000 miles. If there hadn't been a "spare" continent along his westward route, he would have starved to death in the mid Atlantic.

http://www-istp.gsfc.nasa.gov/stargaze/Scolumb.htm

Photons can't fit in
an area much smaller than ful-frame, it's basic physics; you should
know that. It's impossible. Blah blah blah blah blah blah blah. All
the while the mob pats each other on the back.

And, once again, there wouldn't be a mob after you if you wern't throwing stones at everyone. This is what you sait to Osku, one of the people who started out on your side in this "You have the fact-helmet on and it is eliminating your ability to think foreward. Strictly speaking, you are exactly the type of engineer that can't make the technology happen. You have no imagination."

http://forums.dpreview.com/forums/read.asp?forum=1018&message=8984477

Or maybe ""what are you people smoking"
http://forums.dpreview.com/forums/read.asp?forum=1018&message=8984111

So when it does happen. Then will you apologize for being jerks,

Or course not. We were only "being jerks" after you insulted every single person who disagreed with you, or even questioned you.

for not realizing that you were talking with someone who could see
that the impossible is only a measure of those things which we see
barriers. If we begin to believe that these things are not
barriers, then we work on them because it is possible. If it is
perceived impossible, no one works on it. You people are a bunch of
pessemistic mean jerks. When this happens, you won't say, wow I'm
surprised, I talked with this guy in 2004 and he was sure it was
going to happen & I made fun of him. Nope, instead you'll lie like
the honorless dogs you are and say that you knew it was going to
happen all along.

Actually, as I said before, the internet is a large, long, pubically accessable memory. It makes it hard to lie about who said what, first.

So, think about that, about how long all your insults are going to be floating around out there.

--
A cyberstalker told me not to post anymore...
So I'm posting even more!

Ciao!

Joe

http://www.swissarmyfork.com