Go Back   Steve's Digicams Forums > Digicam Help > General Discussion

Reply
 
Thread Tools Search this Thread
Old Mar 12, 2006, 4:08 AM   #131
Member
 
Join Date: Jan 2006
Posts: 91
Default

Suppose you have a torch shining at you with smoke blowing through it so you can see the beam. You take a 200mm lens and position it so the beam occupies the entire frame of a 35 mm sensor. It is at a slight angle so you are only seeing the beam in the smoke. You set up another camera with a 15mm sensor and a 200mm lens right next to it. Now because it is a smaller sensor the beam is larger than the field of view, so you to get the same field of view, you need to move further away. Because you have moved further away, you are at a point where the torch beam is larger. So while it still fills the sensor, it does so at a point where it is less dense because it is farther away. this is the important factor, the density of the light coming from the source. The more that is coming from the source and the less ambient light the more detail.

If you were to imagine that in bright daylight, with the 200mm lens on the 35mm camera, there would be more torch light and less ambient light than with the 200mm lens on the 15mm camera. Point is that even though an object will fill a sensor, and create exposure, where that light comes from is two different sources. Some of it comes from the object, while some comes from the surrounding area. It is the amount of direct light you get from the subject that determins the detail. You imagine that torch beam after 500mm, there wont be much of it left will there?
Mark47 is offline   Reply With Quote
Old Mar 12, 2006, 4:14 AM   #132
Member
 
Join Date: Jan 2006
Posts: 91
Default

Quote:
Quote:
Quote:
you would never get any detail. the further something is away from you, the more the light will disapate and distort before it gets to you and the only way to correct that is through optical magnification, you cant do it digitally.
Thats some great theory, but it DOES NOT APPLY to the cameras on the market today. We are well within these limits, sorry. The difference from 200mm to 300mm is indescribably small.

Thasts some geat theory perhaps you should tell all the video camera manufacturers so they stop wasing all their time and money on building cameras with bigger optical zooms and have them all the same. :G
Mark47 is offline   Reply With Quote
Old Mar 12, 2006, 9:06 AM   #133
Moderator
 
Nagasaki's Avatar
 
Join Date: Jan 2004
Posts: 897
Default

Mark47 wrote:
Quote:

Thasts some geat theory perhaps you should tell all the video camera manufacturers so they stop wasing all their time and money on building cameras with bigger optical zooms and have them all the same. :G
That's not the same thing as having a smaller sensor with the same number of pixels.

You talk about the digital sensor as being a cropped 35mm sensor which would be true if the 35mm sensor had more pixels. When both have the same number of pixels then you can print to A4 for example at the same resolution, pixels per inch.

With film a smaller neg means less detail because the sensor is capturing the same number of dots per inch. With digital smaller sensors have the same number of Mp packed closer together.

If the lens is capable of reolving 2000 lines per inch and the digital sensor is only capable of resolving 1000 lines per inch then the lens out resolves the sensor so the physical size of the sensor is irrelevant to the final result if the field of view is the same.

Digital maginification is taking your 8Mp sensor cropping out a 1/4 of the frame and then printing it at A4. Then you lose resolution simple because you are not using the whole image you captured. That is then magnification without increased resolution.




Nagasaki is offline   Reply With Quote
Old Mar 12, 2006, 9:23 AM   #134
DBB
Senior Member
 
Join Date: Jan 2004
Posts: 2,483
Default

I'm going to ignore what everyone said. This seems fitting to me, and brings the thread to a new level of mediocrity. This is where I feel comfortable.

The cropping factor, with or without a good sensor or lens, "MAGNIFIES" the target. It does NOTHING to RESOLVE the target.

Only getting physically CLOSER will increase the ability to resolve.

Free the mind from the question of sensors and cameras.

It's NOT a question of sensors and cameras. It's a discussion where definitions have to be understood and then APPLIED to sensors and cameras.

Resolving is the ability to separate objects that are close together. From a distance, two close objects appear as one or are in fact INVISIBLE. If an object is invisible it CANNOT be resolved by simple magnification. Only details that are visible can be resolved.

As opposed to the cropping factor which is simply the SAME thing that you do when you crop in an image processing program.

THE DIFFERENCE between doing this in Photoshop and doing it by using the cropping factor is that in the case of PS, you are REDUCING the number of pixels on target, while when you do it with the camera, you are INCREASING the number of pixels on target.

But in EITHER CASE, you are NOT increasing the resolution, you are increasing the Magnification.

And THIS IS MY LAST POST ON THIS TOPIC:

Dave
DBB is offline   Reply With Quote
Old Mar 12, 2006, 5:32 PM   #135
Senior Member
 
tmoreau's Avatar
 
Join Date: Dec 2005
Posts: 477
Default

Dave, Mark, your right I didnt get to your points in my post. I can only type so much, so fast, and I DO have to sleep sometime. I thought that maybe by providing you the resources and framework you could finish making your points in the scientific manner you claim so to hold so dear.

There is entirely too much time and data in my last post to ignore it completely, but I suppose if you guys dont understand this completely, what else can you do? If you ignore it [the facts], you can keep parroting your concepts with impunity.

Oy, do I have to do everything myself? If you want it done right.... Ok, don't worry guys I'll prove your points for you (scientifically)....

Mark, I'm going to continue to ignore you ideas regarding digital zoom until you explain WHAT digital zoom is and HOW it is achieved? I'd love to know how (in your head) that works, how it actually works has been pointed out to you by several people already, and you havent described the process at all. Its wearing quite thin.

************************************

Part II (The things Dave and Mark would be saying if they didn't stop short every time)

Quote:
Resolving is the ability to separate objects that are close together. From a distance, two close objects appear as one or are in fact INVISIBLE. If an object is invisible it CANNOT be resolved by simple magnification. Only details that are visible can be resolved.
I'd like to add the disambiguation that we are not refering to visible as measured by any piece of equipment, or the eye. The term visible seems to be refering to what is availible in the light stream, the natural resolution availible at that particular location, before the light enters an optical device.

Alright [In the voice of Dave], this is described scientifically as the Airy disk, which can be summarized as the way light diffraction increases with the reduction in size of the aperture is passes through.
Code:
sinθ = 1.22 (λ/d)
Where λ is the wavelength of the light and d is the diameter of the aperture. The Rayleigh criterion for barely resolving two objects is that the centre of the Airy disc for the first object occurs at the first minima of the Airy disc of the second.

Nasso fast, we've already looked at this concept, and charted it relative to a whole gamut of lenses, but its a lens property. Can we deduce any data about the light stream from the theoretical performance of the lenses? More on that later.

What we see is that a lens of the same focal length and a wider FOV can resolve greater detail at the sensor plane than one of the same focal length and narrower FOV. This is a lens limitation (so far as we can prove) and Mark should be nodding in agreement saying 'yes, larger FOV means taking in more light'.

Lets take a VERY practical approach to this, and work the numbers as we go.



Take a look at this photograph, which I took today of my television while it was displaying a white image, and consider how we can use it to illustrate everones points in this discussion. I took this from very close, several inches, with a macro lens focused to 1:1. Per Dave's argument (assuming I understand it) this image could not be captured from across the room because the light from each pixel, at that distance, has merged and is not there to be distinguished as RGB points. Per Marks argument (assuming I understand it) it could be resolved, if you had a large enough lens (though I don't particularly know what his definition of a 'large lens' would be). Per my argument, as long as your within the theoretical maximum performance of the lens, any lens, you just need a high enough resolution sensor to discern these same details. Lets pick this apart, bit by bit, and see how it stands up to scientific scrutiny.

I took the picture with the lens at 1:1 and moved the camera to focus. Using a dial caliper I measured each color pixel at about .010" wide on my standard resolution 36" screen (smaller tv's will have smaller pixels). I then calculated the size using the captured image and known physical properties of the sensor. This calculation matched very well (closer than I'd expected, given the visual nature in which I was using the caliper).

If this piques your curiosity, feel free to download the raw file.

In order to capture enough detail to discern each pixel with the cameras 1 and 2 from my above post (12 degree FOV, 3000x2000 sensor resolution), we can simply refer to the arc seconds that are captured by a single pixel and adjust camera position. In both cases, we could (just barely) capture this image at about ten feet away. (Ack, y'all arent going to make me show my work, are you? Its just simple trig). Any farther and the sensor resolution must increase, but as we know from the last post were not far from the theoretical limit of lens resolution. Bottom line, with that setup (those lenses, particularly) we can only discern the TV pixels at distances closer than about ten feet.

Whats that, you don't care? Fine. Lets look at it differently, in a way that supports everyones arguments all at once.

Lets try to discern this level of detail from across the room with a 28mm lens. How much sensor resolution would we need? Is it possible to construct a lens with enough resolution? We will assume that a sensor can be constructed to meet our needs.

We don't need a wide FOV, lets assume a distance of twenty feet (average room distances). Lets use three degrees FOV, which would be around an 800mm lens on a 35mm camera (i.e., super-telephoto range, real-world compatible). Even with a relatively fast aperture of f/2.8 this lens is only going to be about 10mm across (this would hold true, by the way, even if the FOV was 100 degrees. An important point, I think, that I hope to get to eventually). We know we need to resolve slightly better than .010" at this distance to duplicate my macro photo.

Three degrees at twenty feet gives a (diaganol) FOV of 25". 25/.010=2500 pixels to resolve the pixels of the television. The sensor needs a 2500 pixel diaganol resolution, at 3:2 aspect ratio thats 1387 x 2080. We know that sensor size at 28mm focal length and 3 degrees FOV will be 1.206mm x 0.804mm (thats TINY!). Photosites, therefore, are 0.58 microns. This lens (using the formulas were familiar with now) only resolves 16.5 microns. IT WONT WORK! The minimum sensor resolution required severely outresolves the lens. White pixels will be recorded instead of red, green, and blue ones.

That is to be expected, after all, since we know there are limits and that would be one small sensor. So how do we make it work? Enlarging the FOV increases the numerical aperture, so lets start there. Lets make a new lens for this theoretical camera with a 75 degree FOV, the same 28mm focal length, and keep the same sensor (i.e., a crop camera).

This new lens would resolve 0.7 micron detail, BUT we'd have to scale down the sensor since were only interested in the center of the image circle now. Thats a no-go, then (photosites would still be much smaller than the lens resolution).

Alright, back to the question. How DO we make this work? Were 20 feet away and trying to resolve .010" details in our subject. Is the detail there (RGB screen pixels) in the photons bouncing around the room, and can we capture it with a lens?

Lets just jump to the other side of the scale from the mini-cam idea I just explored. Lets try a full-frame DSLR with an ultra-telephoto lens of 825mm (3 degree FOV, how coincidental!). The lens takes in the same amount of light at the same f-stop as the 28mm mini-cam, but requires a lens 300mm across at the same f/2.8 (compared to the 10mm of the other lens).

This lens resolves 16.3 microns on the sensor (the SAME!), allowing a perfectly matched sensor to be designed for our theoretically perfect 825mm lens using 1472x2208 pixels. Thats a similar resolution lens, similar resolution sensor, but physically larger in both regards.

Is that enough pixels to "see" a pixel on my television at this distance? Remember were looking at a FOV at this distance of 25 inches, and looking for a .010" detail (1/2500th of the scene). The diaganol resolution is 2653. YES, this will work. Why? It seems that the resolving power of the lens scaled right with it, on a physical pixels per inch level. The different dimension is the number of inches, with more inches and the same pixels per inch, we have a higher resolution lens for the same FOV.

So who did I prove right/wrong? I'll leave that up to the rest of you guys to sort out, I've got other things to do. (Yes, go ahead, flame away. Try to remember I'm the only one willing to actually substantiate my point, AND YOURS while I was at it).

tmoreau is offline   Reply With Quote
Old Mar 13, 2006, 6:09 AM   #136
Senior Member
 
Aumma45's Avatar
 
Join Date: Jul 2005
Posts: 1,940
Default

I surrender! If the discussion was meant to help some of us with feeble minds, it has desastrously failed.
Aumma45 is offline   Reply With Quote
Old Mar 13, 2006, 7:37 AM   #137
Senior Member
 
tmoreau's Avatar
 
Join Date: Dec 2005
Posts: 477
Default

Aumma45 wrote:
Quote:
I surrender! If the discussion was meant to help some of us with feeble minds, it has desastrously failed.
Well, we can draw some conclusions from the data now that we have it, but I need a rest first!

Any lens can only resolve X amount of detail per Y inches, at Z° AOV. A longer focal length lens of the same AOV has a larger image circle, and therefore resolves more detail on the focal plane.


tmoreau is offline   Reply With Quote
Old Mar 13, 2006, 7:49 AM   #138
Member
 
Join Date: Jan 2006
Posts: 91
Default

tmoreau wrote:
Quote:
Aumma45 wrote:
Quote:
I surrender! If the discussion was meant to help some of us with feeble minds, it has desastrously failed.
Well, we can draw some conclusions from the data now that we have it, but I need a rest first!

Any lens can only resolve X amount of detail per Y inches, at Z° AOV. A longer focal length lens of the same AOV has a larger image circle, and therefore resolves more detail on the focal plane.


This is your problem, you are trying to prove something we already know, are not disputing and has nothing to do with what we are talking about, which only goes to show how you are unable to comprehend anything that we tell you. You have some idea fixed in your mind, so you dont try to comprehend what is being told to you.


Mark47 is offline   Reply With Quote
Old Mar 13, 2006, 9:19 AM   #139
Senior Member
 
tmoreau's Avatar
 
Join Date: Dec 2005
Posts: 477
Default

Mark, your concept seems to follow provable results, but every argument you have made to support your concept has been false. So, your ideas aren't wrong, but your arguments are. I'm interested in this topic, so I'm trying to understand the science behind it and share my findings as I go. While I'm at it, I'm trying to include your concept to help keep my posts "on topic".

Your all-knowing holier than thou bs, and your constant attempts to drag an otherwise technical discussion down to petty ad-hominem attacks, belongs elsewhere. Its been pointed out, and I've tried to correct my transgressions and get back on topic. If your not interested in doing the same, then why continue to post in this thread?

I'm not trying to prove you wrong. This is not personal, its interesting science. Stop being such a dick.

I find our different perspectives to the problem interesting, each poses questions relating to different areas of optics and light behaviour that I think we should explore (because its interesting!).

Quote:
Take a look at this photograph, which I took today of my television while it was displaying a white image, and consider how we can use it to illustrate everones points in this discussion. I took this from very close, several inches, with a macro lens focused to 1:1. Per Dave's argument (assuming I understand it) this image could not be captured from across the room because the light from each pixel, at that distance, has merged and is not there to be distinguished as RGB points. Per Marks argument (assuming I understand it) it could be resolved, if you had a large enough lens (though I don't particularly know what his definition of a 'large lens' would be). Per my argument, as long as your within the theoretical maximum performance of the lens, any lens, you just need a high enough resolution sensor to discern these same details. Lets pick this apart, bit by bit, and see how it stands up to scientific scrutiny.
tmoreau is offline   Reply With Quote
Old Mar 13, 2006, 10:31 AM   #140
Senior Member
 
Join Date: Dec 2002
Posts: 5,803
Default

I have only so much time to read the posts here. I will eventually read them all and comment, just so you know I'm not ignoring you.

I'm not sure you can convince me, because while you might be correct (I haven't read the last posts, remember?) you will also need to convince me how it actually effects the quality of my images. Something can be shown to be true in theory but when actually done some of those issues don't actually make enough of a difference to change the outcome. For example, I've heard someone give a very well grounded argument that showed that the AF system on Canon cameras doesn't work. That it is fundamental flawed. All the logic was sound, from what I and others could tell. But the real world shows us that the Canon AF system does work, so clearly some of the issues listed didn't really make a difference in the real work. (I'm not trying to suggest your argument is as... well, dumb sounding as the example I gave. I'm just using it as an illustrative example.)

Real world experience has shown me that if I do the test that I listed way back earlier in the thread the images would look identical to me. So while science might say that in the right situations they wouldn't be the same (which I think is your point)... I can say that in my experience the images are the same to me.

Eric

ps. You said this:
Quote:
If there were no benefit to a BIGGER OPTICAL ZOOM, they wouldnt bother, they would just all have the same optical zoom and use less CHEAPER digital zoom.
You are clearly a smart person, but this kind of statement makes you look bad. I certainly didn't say the blanket statement that there was "no benefit to a bigger optical zoom", nor did I say anything that would support that conclusion. I didn't read all the posts in this thread, but I saw no one else say it either (maybe I missed it?) To imply that is, to me, only inflammatory (implying someone said something which is clearly wrong and stupid.) It is the type of statement that will cause me to ignore you in the future... and that would be, I feel, to both our detriments. We're both smart people and intelligent debate is what makes forums fun.
eric s is offline   Reply With Quote
 
Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off



All times are GMT -5. The time now is 9:56 PM.