Pages: 1 ... 4 5 [6] 7 8 ... 10   Go Down

Author Topic: Canon 50D @ 15MP  (Read 115790 times)

NikosR

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 622
    • http://
Canon 50D @ 15MP
« Reply #100 on: August 29, 2008, 01:32:26 am »

I will tend to agree with ejmartin. This is how I understand it.

1. Diffraction is a property of the lens aperture
2. The point where a lens becomes diffraction limited is the point where other optical aberrations do not mask the diffraction effect. Thus a perfect lens is diffraction limited at its widest aperture. No amount of closing down will increase its resolution.

3. Diffraction being a property of the lens (and independent of the kind of lens) will interact with pixel pitch according to the circle of confusion theory.

4. Because of 3. at some point while closing the aperture (thus increasing diffraction) the effect is going to be recorded by the sensor (depending on the pixel pitch). The smaller the pitch the earlier will diffraction effects be recorded (be noticable). From this and 1. above we can deduce that given a recording medium with infinite resolution, total system resolution will start to drop as soon as the perfect lens will be stopped down even a little.

5. 4 above does not mean that system resolution of the a lens x a smaller pitch sensor vs the same lens x a larger pitch sensor will be equalized when diffraction starts being noticeable for the smaller pitch sensor. Nothing in 4. leads us to believe this.

Saying that diffraction is higher with smaller pitched sensors is not only theoretically incorrect it is also misleading. It is just that diffraction starts being noticeable earlier limiting a HIGHER system resolution to start with. Diffraction at that point is not noticeable with the larger pitch sensor because the system resolution is LOWER anyways thus it is not limited by diffraction at that point.

If one uses a lens diffraction limited at its widest aperture and a theoretical sensor with pixel pitch small enough to record diffraction effects at that aperture, it is reasonable to expect that ANY amount of closing down the lens will lower the system resolution. This does not mean in any way that this lower system resolution will not be higher than using the same lens with a smaller resolution sensor at that aperture.

So, for me, increased sensor resolution is ALWAYS a good thing from a system resolution point of view. It is just that to get the max out of it one will need better and better lenses. There is no such thing as too small a pixel pitch until we reach the limits of lens design where more will provide no advantage. Of course ignoring other issues like noise etc.
« Last Edit: August 29, 2008, 01:44:15 am by NikosR »
Logged
Nikos

NLund

  • Newbie
  • *
  • Offline Offline
  • Posts: 33
    • http://flickr.com/photos/nsl/
Canon 50D @ 15MP
« Reply #101 on: August 29, 2008, 10:56:12 am »

Quote
On the other hand, perhaps I can understand why, when one poster (whom I shall not name), after objecting to the discussion, goes on to ask how many pixels would a full frame sensor have if its pixel pitch were that of the 50D, thus demonstrating that he is either mathematically challenged, or is unaware of the size difference between the Canon cropped format and FF 35mm.


[a href=\"index.php?act=findpost&pid=217977\"][{POST_SNAPBACK}][/a]


Thanks Ray, I know who I am. I am mathematically challenged, I did not feel like taking the time to perform the calculations only to pull my hair out. I wondered if someone had already done them before me. Of the 1.6x crop factor I've been using the last three years I am most certainly aware.

I does look like I was on the wrong path though.

30D - 6.4 micron pixel pitch
50D - 4.7
5D -   8.2
1ds Mk III - 6.4 (Same as the 30D...this I didn't know)

So if, as I pondered, the 5D Mk II had the same 4.7 micron pixel pitch of the 50D, it would have a fair bit more resolution than the 1Ds 3, something that is certainly not going to happen.

What are people's estimates for 5D Mk II megapixels? It has to be higher, at least in my opinion, than the 50D's 15.1 and lower than the 1Ds Mk 3's 21.1.
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Canon 50D @ 15MP
« Reply #102 on: August 29, 2008, 11:10:54 am »

Quote
Thanks Ray, I know who I am. I am mathematically challenged, I did not feel like taking the time to perform the calculations only to pull my hair out. I wondered if someone had already done them before me. Of the 1.6x crop factor I've been using the last three years I am most certainly aware.

I does look like I was on the wrong path though.

30D - 6.4 micron pixel pitch
50D - 4.7
5D -   8.2
1ds Mk III - 6.4 (Same as the 30D...this I didn't know)

So if, as I pondered, the 5D Mk II had the same 4.7 micron pixel pitch of the 50D, it would have a fair bit more resolution than the 1Ds 3, something that is certainly not going to happen.

What are people's estimates for 5D Mk II megapixels? It has to be higher, at least in my opinion, than the 50D's 15.1 and lower than the 1Ds Mk 3's 21.1.
[a href=\"index.php?act=findpost&pid=218092\"][{POST_SNAPBACK}][/a]

Well, I'm glad you have not taken offense    . A full frame sensor is approximately 2.6x the area of the Canon cropped format, so a full frame sensor comprised of 50D pixels would have 15.4 x 2.6 = 40mp.

Considering that at least one company will be announcing a 24mp FF DSLR within the next month or so, 40mp does not seem too far away. A year or two maybe?
Logged

Pete Ferling

  • Guest
Canon 50D @ 15MP
« Reply #103 on: August 29, 2008, 11:13:08 am »

All this information is making me defraction limited.

Very good stuff here, makes me want to test the 50D even more.
Logged

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
Canon 50D @ 15MP
« Reply #104 on: August 29, 2008, 12:34:43 pm »

Quote
The depth of field has NOTHING to do with the sensor size
[a href=\"index.php?act=findpost&pid=217988\"][{POST_SNAPBACK}][/a]
DOF and OOF effects are however deeply related to the size of the image that you form on the sensor.

If you photograph the same subject from the same distance and want an image on the sensor of twice the (linear) size in order to make full use of a larger sensor, then you need to use twice the focal length, and then with the same aperture ratio, the circles of confusion at each point of the image will have _four_ times the diameter.

Thus when you present the same sized image of the subject on the print or on-screen, so using half the degree of enlargement, the COC at each OOF point on the print will have twice the diameter. That is half the DOF and stronger OOF blurring by any measure.

Alternatively, if you use the same effective aperture diameter (entrance pupil diameter) and thus twice the aperture ratio with that doubled focal length and image size, each COC has twice the diameter, and by the way each diffraction spot has twice the Airy disc diameter.

Thus when you display at equal image size, all COC's and diffraction spots on the displayed image are of equal size: exactly the same DOF, OOF blurring and diffraction blurring.


Thus 15MP in any format is equally useful for landscape photography so long as diffraction control does not force you to use f-stops so low that lens aberrations impair image quality. You just have to choose a different f-stop (same effective aperture diameter) to get a given combination of OOF and diffraction effects when filling a different frame size with the desired image.


With good SLR lenses giving best resolution at f/5.6 or even f/4 or below, diffraction/aberration trade-offs still allow resolution well beyond what any curent DLSR pixel size offers. Three micron pixel spacing might start pushing the limits of the highest resolving SLR lenses, but that would be 25MP in 4/3, 37MP in EF-S, etc.


P. S. NikonsR summarizes my point very nicely in a different way, particularly in his final paragraph.

Diffraction like lens aberrations lead only to a case of the law of diminishing returns for the increases in overall resolution given by increasing sensor resolution; never a loss, and so far we are not at the point beyond which there is no significant improvement with further sensor resolution increase, so long as good lenses are used.
« Last Edit: August 29, 2008, 12:47:09 pm by BJL »
Logged

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
Canon 50D @ 15MP
« Reply #105 on: August 29, 2008, 12:36:36 pm »

Quote
I will tend to agree with ejmartin. This is how I understand it.

So, for me, increased sensor resolution is ALWAYS a good thing from a system resolution point of view. It is just that to get the max out of it one will need better and better lenses. There is no such thing as too small a pixel pitch until we reach the limits of lens design where more will provide no advantage. Of course ignoring other issues like noise etc.
[{POST_SNAPBACK}][/a]

The posts by ejmartin and NikosR clear up much of the confusion about small pixel pitch. Nikos is ignoring noise, but in the real world noise is a major limiting factor and there is much confusion concerning pixel size noise and signal to noise. On a per pixel basis if other factors are held constant, the large pixel will have a better S:N expressed as a standard deviation. However, we look at the entire image, not individual pixels. If the sensor size is held constant along with other factors and the pixel count is increased, the standard deviation of the noise will increase but the noise will be finer grained and less obvious. If one downsizes the higher resolution image to the same size as the lower resolution image, the noise will be reduced by pixel [a href=\"http://www.photomet.com/pm_solutions/library_encyclopedia/index.php]binning[/url].

EJ Martin gives an excellent analysis. He concludes that the sensor size, not the pixel size, is the critical factor here.

If you resample your D50 image in Photoshop to a smaller file size, the effects of shot noise will be reduced, but the effects on read noise will not be as good as if the binning were done in hardware as the above link points out. For example in 2x2 hardware binning, the superpixel is read with the same read noise as the individual pixels. In the case of the D50 image downsized by half, there are four read noises that add in quadrature, and the read noise would be decreased by 50% and not by 75% as with hardware binning.

Of course, with the d50, other things are not held constant and Canon has presumably improved the sensor over that of the D40. It will be interesting to examine actual photos taken with both cameras, and I would imagine that the noise for photos printed at the same lateral dimension (e.g. 16 by 20 inch) will not be that different between the two cameras.

However, one main reason for increased megapixels to print at a larger size, and the larger print would bring out the noise. A 15 MP full frame sensor would have an advantage here.

Bill
Logged

Panopeeper

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1805
Canon 50D @ 15MP
« Reply #106 on: August 29, 2008, 09:00:02 pm »

Quote
If you photograph the same subject from the same distance and want an image on the sensor of twice the (linear) size in order to make full use of a larger sensor, then you need to use twice the focal length, and then with the same aperture ratio, the circles of confusion at each point of the image will have _four_ times the diameter
1. The effect of diffraction is the airy disk; the circle of confusion is the limitation of the airy disk to avoid visible effect of the diffraction. Thus one has to calculate the airy disk diameter.

2. The diameter of the airy disk is directly proportional to the focal length, and inversely proportional to the aperture diameter. Thus it is a linear function of the quotient of focal length and aperture diameter, in other words it is a linear function of the aperture number.

When the aperture number is kept unchanged, the diffraction too (the airy disk diameter) remains unchanged, independently of the focal length.
Logged
Gabor

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Canon 50D @ 15MP
« Reply #107 on: August 30, 2008, 01:32:17 am »

Quote
The diameter of the airy disk is directly proportional to the focal length, and inversely proportional to the aperture diameter. Thus it is a linear function of the quotient of focal length and aperture diameter, in other words it is a linear function of the aperture number.

When the aperture number is kept unchanged, the diffraction too (the airy disk diameter) remains unchanged, independently of the focal length.
[a href=\"index.php?act=findpost&pid=218196\"][{POST_SNAPBACK}][/a]

No disagreement there, Gabor. What surprises me is that you seem to have ignored the relationship between the size of the Airy disc and the size of the composition or image. The effect of the absolute size of the Airy disc is pretty meaningless outside the context of a specific size of image.

The Airy disc produced by a 50mm lens at F11 is the same whether that lens is attached to a 35mm camera or a P&S camera, but you would never use F11 with a P&S camera if you wanted a sharp image. In fact, I suspect that most P&S cameras do not offer the option of F11, just as most 35mm lenses do not offer the option of F45.
Logged

NikosR

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 622
    • http://
Canon 50D @ 15MP
« Reply #108 on: August 30, 2008, 05:08:45 am »

Quote
No disagreement there, Gabor. What surprises me is that you seem to have ignored the relationship between the size of the Airy disc and the size of the composition or image. The effect of the absolute size of the Airy disc is pretty meaningless outside the context of a specific size of image.

The Airy disc produced by a 50mm lens at F11 is the same whether that lens is attached to a 35mm camera or a P&S camera, but you would never use F11 with a P&S camera if you wanted a sharp image. In fact, I suspect that most P&S cameras do not offer the option of F11, just as most 35mm lenses do not offer the option of F45.
[a href=\"index.php?act=findpost&pid=218220\"][{POST_SNAPBACK}][/a]

True. But you can't use a 50mm on a P&S either....
Logged
Nikos

Fine_Art

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1172
Canon 50D @ 15MP
« Reply #109 on: August 30, 2008, 01:14:02 pm »

Quote
I will tend to agree with ejmartin. This is how I understand it.

1. Diffraction is a property of the lens aperture
2. The point where a lens becomes diffraction limited is the point where other optical aberrations do not mask the diffraction effect. Thus a perfect lens is diffraction limited at its widest aperture. No amount of closing down will increase its resolution.

3. Diffraction being a property of the lens (and independent of the kind of lens) will interact with pixel pitch according to the circle of confusion theory.

4. Because of 3. at some point while closing the aperture (thus increasing diffraction) the effect is going to be recorded by the sensor (depending on the pixel pitch). The smaller the pitch the earlier will diffraction effects be recorded (be noticable). From this and 1. above we can deduce that given a recording medium with infinite resolution, total system resolution will start to drop as soon as the perfect lens will be stopped down even a little.

5. 4 above does not mean that system resolution of the a lens x a smaller pitch sensor vs the same lens x a larger pitch sensor will be equalized when diffraction starts being noticeable for the smaller pitch sensor. Nothing in 4. leads us to believe this.

Saying that diffraction is higher with smaller pitched sensors is not only theoretically incorrect it is also misleading. It is just that diffraction starts being noticeable earlier limiting a HIGHER system resolution to start with. Diffraction at that point is not noticeable with the larger pitch sensor because the system resolution is LOWER anyways thus it is not limited by diffraction at that point.

If one uses a lens diffraction limited at its widest aperture and a theoretical sensor with pixel pitch small enough to record diffraction effects at that aperture, it is reasonable to expect that ANY amount of closing down the lens will lower the system resolution. This does not mean in any way that this lower system resolution will not be higher than using the same lens with a smaller resolution sensor at that aperture.

So, for me, increased sensor resolution is ALWAYS a good thing from a system resolution point of view. It is just that to get the max out of it one will need better and better lenses. There is no such thing as too small a pixel pitch until we reach the limits of lens design where more will provide no advantage. Of course ignoring other issues like noise etc.
[a href=\"index.php?act=findpost&pid=218018\"][{POST_SNAPBACK}][/a]
A minor quibble - diffraction is a property of light, not the lens. A good lens handles it well.
Logged

NikosR

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 622
    • http://
Canon 50D @ 15MP
« Reply #110 on: August 30, 2008, 02:24:56 pm »

Quote
A minor quibble - diffraction is a property of light, not the lens. A good lens handles it well.
[a href=\"index.php?act=findpost&pid=218297\"][{POST_SNAPBACK}][/a]

Well, it's a property of light interacting with the aperture   Not sure that lens quality makes a difference though.
Logged
Nikos

Panopeeper

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1805
Canon 50D @ 15MP
« Reply #111 on: August 30, 2008, 03:46:41 pm »

Quote
A minor quibble - diffraction is a property of light, not the lens. A good lens handles it well.
Diffraction is the property of light, occuring when it passes the edge of an object. I wonder if there are photographic lenses without any edge; even at fully open aperture, if nothing else then the barrel causes diffraction. Perhaps fisheyes?

Is there any lens, which can "handle" this?
Logged
Gabor

ejmartin

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 575
Canon 50D @ 15MP
« Reply #112 on: August 30, 2008, 06:07:07 pm »

Quote
Diffraction is the property of light, occuring when it passes the edge of an object. I wonder if there are photographic lenses without any edge; even at fully open aperture, if nothing else then the barrel causes diffraction.

[a href=\"index.php?act=findpost&pid=218319\"][{POST_SNAPBACK}][/a]

Diffraction is not simply sourced by light passing near the edge of an obstruction (eg aperture blades); rather it arises from interference effects of light passing through the entire aperture.  All portions of the aperture opening contribute equally, and so one cannot localize its source any more finely than to say it comes from the aperture (and not just the edge of the aperture opening).

And so, any lens has a finite aperture and thus exhibits diffraction, more and more the smaller the aperture is closed down.   Ultimately, the effects of diffraction are dependent only on the f-number and not any particular design of the aperture, lens barrel, etc.
« Last Edit: August 30, 2008, 06:07:57 pm by ejmartin »
Logged
emil

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
Canon 50D @ 15MP
« Reply #113 on: August 30, 2008, 07:57:55 pm »

Quote
1. The effect of diffraction is the airy disk; the circle of confusion is the limitation of the airy disk to avoid visible effect of the diffraction. Thus one has to calculate the airy disk diameter.

2. The diameter of the airy disk is directly proportional to the focal length, and inversely proportional to the aperture diameter. Thus it is a linear function of the quotient of focal length and aperture diameter, in other words it is a linear function of the aperture number.

When the aperture number is kept unchanged, the diffraction too (the airy disk diameter) remains unchanged, independently of the focal length.
[a href=\"index.php?act=findpost&pid=218196\"][{POST_SNAPBACK}][/a]

You have misunderstood what I wrote; I was talking about two factors affecting resolution, diffraction and out of focus effects:

1. out of focus effects, as measured by the diameter of the circle of confusion in the focal plane, which is proportional to the square of the focal length and inversely proportional to aperture ratio,
and
2. diffraction effects, measured by the diameter of the Airy disc, which is independent of the focal length and proportional [added word] to the aperture ratio.

Together these mean that if the focal length and aperture ratio are increased by the same factor, the size of the image, teh Airy discs and eh circles of confusion all increase in proportio to the focal length. Thus when enlargement is chosen to give equal iamge size it als gives equal iszed Airy discs and circles of confusion.


If you insist on looking only on what happens at equal aperture ratio, doubling focal length will give equal sized Airy discs (so half the diffraction blurring on equal sized prints) but four times the circle of confusion at each OOF points, so half the DOF on equal sized prints. Since the main reason for having to stop down to "diffraction challenged" apertures is to get adequate DOF, the larger format and focal length will not be able to use the same aperture ratio; it will have to use double the f-stop to get that same desired DOF.


It amazes me that so many people keep making comparisons between formats under the assumption that a smaller format must use an equally high aperture ratio (and/or equally high ISO speed) as a larger format even when discussing situations when there is the option of changing to a lower aperture ratio (allowing a correspondingly less high ISO speed).


P.S. Before anyone says it: in the unusual situation that a high f-stop is chosen to get a long enough exposure at minimum ISO speed rather than for sufficient DOF, neutral density filters will allow a lower f-stop and thus reduced diffraction effects.
« Last Edit: August 31, 2008, 12:10:04 pm by BJL »
Logged

Dr. Gary

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 83
Canon 50D @ 15MP
« Reply #114 on: August 30, 2008, 08:17:56 pm »

The amount of diffraction relates the size of the opening the light passes through. F/ stops are a variable correlating to the ratio of the focal length of a lens to its (or its diaphragm's) diameter. In other words, a 50mm lens at f/8 has the same amount of diffraction applied to the light as a 100mm lens does at f/11.

drgary
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Canon 50D @ 15MP
« Reply #115 on: August 30, 2008, 08:21:49 pm »

Quote
True. But you can't use a 50mm on a P&S either....
[a href=\"index.php?act=findpost&pid=218231\"][{POST_SNAPBACK}][/a]

Well, I,m glad you agree with my point. However, it's not correct that you can't use a 50mm lens on a P&S. There are a number of P&S cameras with 10x or 12x zooms which reach and exceed 50mm in focal length.
Logged

Fine_Art

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1172
Canon 50D @ 15MP
« Reply #116 on: August 31, 2008, 01:46:33 am »

Quote
Diffraction is the property of light, occuring when it passes the edge of an object. I wonder if there are photographic lenses without any edge; even at fully open aperture, if nothing else then the barrel causes diffraction. Perhaps fisheyes?

Is there any lens, which can "handle" this?
[a href=\"index.php?act=findpost&pid=218319\"][{POST_SNAPBACK}][/a]
Thats true. By handle it well, I mean a lens that yields high resolution close to the full aperture. A basic kit lens may be sharpest at over f9 on a f4.5-5.6 and even then its not very sharp. A good lens might be sharpest at 5.6 and still very sharp at f2.8 on a 2.8

A black hole would be a lens without an edge in the normal sense.
Logged

NikosR

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 622
    • http://
Canon 50D @ 15MP
« Reply #117 on: August 31, 2008, 02:13:49 am »

Quote
Thats true. By handle it well, I mean a lens that yields high resolution close to the full aperture. A basic kit lens may be sharpest at over f9 on a f4.5-5.6 and even then its not very sharp. A good lens might be sharpest at 5.6 and still very sharp at f2.8 on a 2.8

A black hole would be a lens without an edge in the normal sense.
[a href=\"index.php?act=findpost&pid=218408\"][{POST_SNAPBACK}][/a]

Yes. But as I said above this is not 'handling diffraction well'. It is just that other aberrations are handled well so diffraction is the limiting factor. BTW Diffraction is not caused by the 'edge'. It is caused by the hole.
Logged
Nikos

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Canon 50D @ 15MP
« Reply #118 on: August 31, 2008, 02:38:38 am »

Quote
The amount of diffraction relates the size of the opening the light passes through. F/ stops are a variable correlating to the ratio of the focal length of a lens to its (or its diaphragm's) diameter. In other words, a 50mm lens at f/8 has the same amount of diffraction applied to the light as a 100mm lens does at f/11.

drgary
[a href=\"index.php?act=findpost&pid=218371\"][{POST_SNAPBACK}][/a]

The first part is right but the last bit seems incorrect to me, but I'm no physicist. The physical aperture diameter of a 50mm lens at F8 is given by the simple formula: Focal length/F stop.

50mm/8 = 6.25mm. However, 100mm/11 = 9mm. Diffraction should be less with the 100mm lens at F11.
Logged

ejmartin

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 575
Canon 50D @ 15MP
« Reply #119 on: August 31, 2008, 08:55:06 am »

Quote
The amount of diffraction relates the size of the opening the light passes through. F/ stops are a variable correlating to the ratio of the focal length of a lens to its (or its diaphragm's) diameter. In other words, a 50mm lens at f/8 has the same amount of diffraction applied to the light as a 100mm lens does at f/11.

drgary
[a href=\"index.php?act=findpost&pid=218371\"][{POST_SNAPBACK}][/a]

That depends on what you mean by "the same amount of diffraction".  The range of angles over which the direction of the light spreads after passing through the aperture depends only on the physical size of the aperture; since aperture is (focal length/f-number), then indeed the range of angles will be the same for 50mm @ f8 and for 100mm @ f11.  

However, the relevant quantity for photography is the amount of blur (the Airy disk diameter) at the focal plane; for that, one should multiply the spread in angles by the distance from the aperture to the focal plane.  The spread in angles is proportional to 1/aperture, that times the distance gives the spot size at the focal plane proportional to (focal length/aperture)=f-number.

So the relevant quantity for photography (the diffraction spot size) depends only on the f-number and not the aperture or focal length separately.  And so that 50mm @f8 will have sqrt[2]~1.4 times less diffraction in the image than the 100mm @ f11.
« Last Edit: August 31, 2008, 08:58:01 am by ejmartin »
Logged
emil
Pages: 1 ... 4 5 [6] 7 8 ... 10   Go Up