Luminous Landscape Forum
Equipment & Techniques => Cameras, Lenses and Shooting gear => Topic started by: Ray on January 07, 2006, 08:06:24 am
-
First let me say I think the new site is a vast improvemnt over the old and I for one will be renewing my subscription to the LL video journal which pays for this site.
As Jonathan Wienke frequently mentioned, there's no free lunch, and I don't expect to use valuable bandwidth on someone elses site for free.
The facility of now being able to download images directly from one's HD is a major upgrade which is most welcome.
With this in mind, I'm hoping to demonstrate a curious phenomenon which basically knocked my socks off.
Most of us are not scientists, and even if we were we'd have a difficult time making sense of all the contradictory data regarding lens performance, due to lack of reliable, up to date information in the public domain.
I recently bought a Canon 5D and the 24-105mm zoom lens, which I believe is a great combination. During the past 6 weeks or so, I've taken about 4,000 images in Nepal, Tailand and Cambodia using mostly the 24-105 lens, but occasionally the Sigma 15-30.
I'm now processing these images and in the process of doing so I noticed that images I'd taken at f16 were no less sharp than the same image taken at f11, in any respect, but the f16 image exhibited greater DoF.
This actually flies in the face of common understanding that f16 is softer than f11 and f11 is very slightly softer than f8.
The following images demonstrate that this is not necessarily true.
The image that brought this to my attention was a shot of the Bayon temple near Angkor Wat. I took 2 shots, one at f11 and one at f16. The shutter speed was the same (100th) in both instances (which of course meant I bumped up the ISO to 200 for the f16 shot).
Examining these 2 shots of the same scene, having focussed on the same spot in each exposure (guesstimating the hyperfocal distance as I always do), I find the f16 shot is sharper in all respects. Geez! That can't be right! What have I done! F16 sharper than f11, or at worst equally sharp!!!
So here's the proof. I should also add that the following images have had zero sharpening and zero luminance smoothing, I've also adjusted the shadows and contrast sliders to ensure no clipping of highlights and shadows. These are not images adjusted for printing, and as a matter of fact my monitor is not even calibrated because my system is 64 bit (the calibration people are seriously lagging behind technological developments) .
Here's the overview.
[attachment=135:attachment]
Here's the foreground showing a portion of the Lion's mane (sorry for another cat shot). It's a bit blurry.
[attachment=136:attachment]
Now here's the foreground of exactly the same shot but taken at f16 and ISO 200 to maintain shutter speed of 100th.
[attachment=137:attachment]
You should be able to see that it's noticeably sharper and that shooting at f16 has had its advantages. But what about the background? Well here's the background at f11 and f16.
[attachment=138:attachment]
[attachment=139:attachment]
If you examine these images you'll see that the f16 images are sharper. The f16 foreground image is quite noticeably sharper. The f16 background image is marginally sharper.
So what! I hear you say. That's what you'd expect. And yes it is. But what you'd also expect is that at some point in between the nearest point and the furthest point the f11 shot would be sharper. You have to believe me on this, it isn't.
There's no point anywhere on the f11 shot that is as sharp or sharper than the f16 shot.
Conclusion? This lens (at 85mm, did I forget to mention that) is sharpest at f16. Now that's good news because there are 2 interesting ramifications. (1) I can use f16 with impunity without agonising over any trade-off in resolution, (2) The 5D sensor (and the 1Ds before it) is really not as good as some of the lenses attached.
For those of you who are skeptical, who think maybe I am suffering from Parkinson's or dipsomaniacal tremors, I put my camera on a tripod indoors; used remote control and mirror lock-up, and produced the following results.
[attachment=140:attachment]
[attachment=141:attachment]
[attachment=142:attachment]
Essentially, maximum resolution of the 24-105 lens at 85mm is the same across f8 to f16. What differs is the DoF. I haven't wasted bandwidth by showing this. It's too obvious.
I should qualify the above statement. Maximum resolution of the 25-105/5D system is the same across f8 to f16 at 85mm. I can only speculate on the reasons. My guess is that the 5D pixel size is too big to delineate (differentiate whatever) anything smaller than the f16 Airy disc. Please feel free to dispute this.
Okay! The images are too small. They are in fact around 100th of the area of the full image. If you need to see them, I can provide bigger images. Maybe I confused centimetres with inches in PS.
Yep! PS on my 64 bit system was set to cms instead of inches. Sorry about that, but I don't expect there'll be any clamour for larger images, but I do wish you Americans would get in line with the rest of the world .
-
Your own examination and evaluation of dozens (or hundreds) of images is more important than looking at a few selected crops. If you are satisfied, that is the main thing. Whether someone can demonstrate measurable diffraction at f/11 on the test bench really doesn't matter if the results are not apparent in real world shooting.
Your conclusion seems quite reasonable and is about what I would have expected from my own experience. I have the Nikon D2X which has about the same pixel count as the 5D, but on a much smaller chip. The D2X begins to show diffraction beyond f/11--smaller apertures are still usable if you need the depth of field but you have to accept that there is a little softness by the time you get to f/16. With the full frame sensor of the 5D and therefore larger photosites, I would expect that diffraction effects would be less of an issue, so your results are not surprising.
-
Ray: On New Year's Day I spent the day shooting seascapes with a 1Ds MkII and 24~105 (most at the wide end 24 to about 35mm). I purposly used mostly f11 because overall sharpness was a goal. Next time I will shoot at both f11 and 16 at the wide end. Your post has me re-examining how I work.
Thanks for the post.
Dave
-
Just a small request...
When posting images/crops for comparison how about doing a quick "merge" and present the samples side by side in the same frame?
-
Would love to hear Michael's input on this.
-
Ray: On New Year's Day I spent the day shooting seascapes with a 1Ds MkII and 24~105 (most at the wide end 24 to about 35mm).
[a href=\"index.php?act=findpost&pid=55442\"][{POST_SNAPBACK}][/a]
Dave, I'll make a prediction. The 1Ds2 will show a greater difference going from f8 to f16 because of its higher absolute resolving power. The lens has to be better at f8 but the 5D's resolving power with its lower pixel density is not so capable of revealing it.
Having read the reviews of the 24-105 on this site it seems clear that this lens has a sweet spot around 50mm. I couldn't resist doing a comparison at all stops at 50mm. I believe the differences are more noticeable at this focal length, but there seems to be no serious degradation till f22.
This time I've pasted the images together to make it easier for Bobtrips and others .
Full image:- [attachment=147:attachment]
Composite image:- [attachment=148:attachment]
Each centre crop is a bit smaller than 100th the area of the full image (although the crops have a different aspect ratio). One should always bear such matters in mind when assessing the relevance of small differences at the 'pixel-peeping' level.
Even using this lens at its best focal length doesn't seem to result in more than a very marginal loss of image quality at f16. I think I might continue to avoid using f22 though .
-
Each centre crop is a bit smaller than 100th the area of the full image (although the crops have a different aspect ratio). One should always bear such matters in mind when assessing the relevance of small differences at the 'pixel-peeping' level.
[a href=\"index.php?act=findpost&pid=55457\"][{POST_SNAPBACK}][/a]
Actually, I seem to be out by a factor of 10. I should have written each crop is smaller than one thousandth of the area of the full image. With the ruler displayed around the 12x18" image at 240 ppi in PS, each crop is about 0.31" wide.
Viewing the enlarged crop in the above post, it appears on my screen 1.75"x3.5" in size. To see the crop this size on a print, the print would need to be 5.7 (0.31/1.75) x12" wide, ie., 68"x102". I therefore conclude that any small differences between f8 and f16 (with the 5D) is even less significant than I first imagined, or in other words, quite irrelevant.
Now I'd like to repeat this test with the Canon 85/1.2, if I can get my hands on one.
-
Actually, I'm rather surprised at the lack of interest in this topic. Am I the only one who's been inhibited in using f16 because of resolution fall-off? It apears so. Do I take it, if you all want the maximum DoF, you automatically stop down to f16, f22 and even f32, with narry a worry about resolution fall-off?
I'd like to know.
-
Actually, I'm rather surprised at the lack of interest in this topic. Am I the only one who's been inhibited in using f16 because of resolution fall-off? It apears so. Do I take it, if you all want the maximum DoF, you automatically stop down to f16, f22 and even f32, with narry a worry about resolution fall-off?
I'd like to know.
[a href=\"index.php?act=findpost&pid=55702\"][{POST_SNAPBACK}][/a]
I regularly go to f/16 but will go higher sometimes (1D Mk2 & 5D). In fact I rarely go beyond f/16 and if I do then I always double the shot (one at f/16 and one at f/22). Michael has an article about this subject but I can't find it anymore.
-
I see! So I'm the conservative one. I've spent so many years with the small format D60 and 20D that I've forgotten that f16 can have a use with full frame 35mm without resolution compromise.
-
I see! So I'm the conservative one. I've spent so many years with the small format D60 and 20D that I've forgotten that f16 can have a use with full frame 35mm without resolution compromise.
Ray,
I found the article (http://www.luminous-landscape.com/essays/stop-d.shtml).
-
I'm happy if I can do f8-f11. But I have no issues going out to f16. Beyond that I start to worry. But if you need the DOF you do it and hope the results are good enough.
-
... if you need the DOF you do it and hope the results are good enough.
[a href=\"index.php?act=findpost&pid=55749\"][{POST_SNAPBACK}][/a]
Sure, but I'd like a bit more certainty. It's supposed to be science after all.
-
Ray,
I found the article (http://www.luminous-landscape.com/essays/stop-d.shtml).
[a href=\"index.php?act=findpost&pid=55747\"][{POST_SNAPBACK}][/a]
Francois,
I remember reading that article. In fact I even remember commenting to Michael at the time that f32 on an MF camera is quite different to f32 on 35mm. In fact there's about a 2.5 stop difference, ie f32 on 6x7 is equivalent to around f13 on 35mm.
As I recall, in the days I was using 35mm film, f16 always produced rather disappointing results in terms of absolute resolution. I suspect the reason lies in the nature of MTF roll-off in film beyond 20 lp/mm.
My disappointment in 35mm film at f16 was easily carried forward to the D60 and 20D because these are smaller formats with 1.6x the DoF at the same stop and FoV. It's well known that as you go down in format, the smaller apertures become less usable. No 35mm lens supports f64 and no 2/3rds format P&S supports f16.
The fact that f16 appears to be fully usable on my 5D with no resolution fall-off of practical significance is a sort of bonus for me. I'm just surprised that no-one has been adding this feature to the list of advantages in upgrading from a 20D to a 5D.
-
Sure, but I'd like a bit more certainty. It's supposed to be science after all.
[a href=\"index.php?act=findpost&pid=55792\"][{POST_SNAPBACK}][/a]
It's magic.
-
It's magic.
[a href=\"index.php?act=findpost&pid=55799\"][{POST_SNAPBACK}][/a]
Well, let's hope it's consistent magic for you. "I wasn't feeling too well yesterday. Took a few shots at f22 though, to get maximum DoF. But the resolution's not too good. Just can't explain it. I used a fast shutter speed and IS so I don't know what went wrong. A week ago my f22 shots were just fine, as sharp as a razor."
-
It's basically an issue of pixel pitch - the smaller the pixels themselves, the sooner diffraction effects kick in. There's an excellent article on understanding diffraction and pixel pitch here: http://www.cambridgeincolour.com/tutorials...photography.htm (http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm)
From that article, a 5D will start to lose resolution to diffraction _after_ f/16. In my own tests to verify this I found that I saw improved sharpness at f/16 over f/11, but lost sharpness at f/22. With my 20D, I found that f/11 was the minimum aperture I could use before losing resolution (using the same lens, the 24-70L).
That being said, as Michael says in his 'stop down!' article, if you need the depth of field, shoot at the appropriate aperture and worry about diffraction later.
It's also worth remembering that everything has to go right to get the sharpest possible shot. In a controlled environment with plenty of time and no discomfort, it's easy to get razor sharp shots. In the field, with wind, rain, soft ground, misfocussing (remember - AF still isn't as reliable as we'd like it to be for landscape work, especially) and photographer discomfort it's easy to get it wrong!
Cheers,
Peter
-
First let me say I think the new site is a vast improvemnt over the old and I for one will be renewing my subscription to the LL video journal which pays for this site.
As Jonathan Wienke frequently mentioned, there's no free lunch, and I don't expect to use valuable bandwidth on someone elses site for free.
The facility of now being able to download images directly from one's HD is a major upgrade which is most welcome.
With this in mind, I'm hoping to demonstrate a curious phenomenon which basically knocked my socks off.
Most of us are not scientists, and even if we were we'd have a difficult time making sense of all the contradictory data regarding lens performance, due to lack of reliable, up to date information in the public domain.
I recently bought a Canon 5D and the 24-105mm zoom lens, which I believe is a great combination. During the past 6 weeks or so, I've taken about 4,000 images in Nepal, Tailand and Cambodia using mostly the 24-105 lens, but occasionally the Sigma 15-30.
I'm now processing these images and in the process of doing so I noticed that images I'd taken at f16 were no less sharp than the same image taken at f11, in any respect, but the f16 image exhibited greater DoF.
This actually flies in the face of common understanding that f16 is softer than f11 and f11 is very slightly softer than f8.
The following images demonstrate that this is not necessarily true.
The image that brought this to my attention was a shot of the Bayon temple near Angkor Wat. I took 2 shots, one at f11 and one at f16. The shutter speed was the same (100th) in both instances (which of course meant I bumped up the ISO to 200 for the f16 shot).
Examining these 2 shots of the same scene, having focussed on the same spot in each exposure (guesstimating the hyperfocal distance as I always do), I find the f16 shot is sharper in all respects. Geez! That can't be right! What have I done! F16 sharper than f11, or at worst equally sharp!!!
So here's the proof. I should also add that the following images have had zero sharpening and zero luminance smoothing, I've also adjusted the shadows and contrast sliders to ensure no clipping of highlights and shadows. These are not images adjusted for printing, and as a matter of fact my monitor is not even calibrated because my system is 64 bit (the calibration people are seriously lagging behind technological developments) .
Here's the overview.
[attachment=135:attachment]
Here's the foreground showing a portion of the Lion's mane (sorry for another cat shot). It's a bit blurry.
[attachment=136:attachment]
Now here's the foreground of exactly the same shot but taken at f16 and ISO 200 to maintain shutter speed of 100th.
[attachment=137:attachment]
You should be able to see that it's noticeably sharper and that shooting at f16 has had its advantages. But what about the background? Well here's the background at f11 and f16.
[attachment=138:attachment]
[attachment=139:attachment]
If you examine these images you'll see that the f16 images are sharper. The f16 foreground image is quite noticeably sharper. The f16 background image is marginally sharper.
So what! I hear you say. That's what you'd expect. And yes it is. But what you'd also expect is that at some point in between the nearest point and the furthest point the f11 shot would be sharper. You have to believe me on this, it isn't.
There's no point anywhere on the f11 shot that is as sharp or sharper than the f16 shot.
Conclusion? This lens (at 85mm, did I forget to mention that) is sharpest at f16. Now that's good news because there are 2 interesting ramifications. (1) I can use f16 with impunity without agonising over any trade-off in resolution, (2) The 5D sensor (and the 1Ds before it) is really not as good as some of the lenses attached.
For those of you who are skeptical, who think maybe I am suffering from Parkinson's or dipsomaniacal tremors, I put my camera on a tripod indoors; used remote control and mirror lock-up, and produced the following results.
[attachment=140:attachment]
[attachment=141:attachment]
[attachment=142:attachment]
Essentially, maximum resolution of the 24-105 lens at 85mm is the same across f8 to f16. What differs is the DoF. I haven't wasted bandwidth by showing this. It's too obvious.
I should qualify the above statement. Maximum resolution of the 25-105/5D system is the same across f8 to f16 at 85mm. I can only speculate on the reasons. My guess is that the 5D pixel size is too big to delineate (differentiate whatever) anything smaller than the f16 Airy disc. Please feel free to dispute this.
Okay! The images are too small. They are in fact around 100th of the area of the full image. If you need to see them, I can provide bigger images. Maybe I confused centimetres with inches in PS.
Yep! PS on my 64 bit system was set to cms instead of inches. Sorry about that, but I don't expect there'll be any clamour for larger images, but I do wish you Americans would get in line with the rest of the world .
[a href=\"index.php?act=findpost&pid=55422\"][{POST_SNAPBACK}][/a]
Diffraction is a non varible property. In otherwords, it is the lens opening not the f number responsible. A 100mm lens at f8 has the same amount of image degrading diffraction that a 50mm lens does at f 5.6 or a 200mm lens does at f11. It is a property decided by the lens opening.
DrGary
-
Diffraction is a non varible property. In otherwords, it is the lens opening not the f number responsible. A 100mm lens at f8 has the same amount of image degrading diffraction that a 50mm lens does at f 5.6 or a 200mm lens does at f11. It is a property decided by the lens opening.
DrGary
[a href=\"index.php?act=findpost&pid=56073\"][{POST_SNAPBACK}][/a]
That's true, but for a given F stop, reducing the focal length increases the diffraction due to the smaller aperture. To achieve a given FOV, a 35mm camera requires a shorter focal length lens than a medium format camera, hence at a given F stop and FOV, diffraction is greater.
To put it another way, you can use smaller F stops on MF gear, before diffraction starts to degrade the image quality.
Leif
-
Regarding the principle, it is near enough correctly stated. Perhaps "opening" is a little ambiguous, but as for the arithmetic
focal length/ f-no. aperture
100mm/8 =12.5 mm
50mm/5.6 = 8.9 mm (should have been f4, of course)
200mm/11 = 18.2 mm (should have been f16).
(Let's not worry about the true value for f11, as the correct answer only needs the
2^n f numbers that are exact.)
Note that the "opening" depends a little on the degree of retrofocus/telephoto - it is a long time since I calculated it from scratch, but anyway it makes only a very little difference in most cases (IIRC). Also we are talking about low magnification situations: with macro, as usual, the aperture behaves as if it were smaller than it really is (in the usual way, according to the magnification), so diffraction is worse than you might expect from the usual formula.
Ken
Diffraction is a non varible property. In otherwords, it is the lens opening not the f number responsible. A 100mm lens at f8 has the same amount of image degrading diffraction that a 50mm lens does at f 5.6 or a 200mm lens does at f11. It is a property decided by the lens opening.
DrGary
[a href=\"index.php?act=findpost&pid=56073\"][{POST_SNAPBACK}][/a]
-
Oh no! lets not go there again Who wants to change format in this thread? Note it says "35mm lenses". The "format" issue was done to death elsewhere on this forum recently. Summary: it depends what question you ask what answer you get (even if the answer is right).
Ken
To put it another way, you can use smaller F stops on MF gear, before diffraction starts to degrade the image quality.
Leif
[a href=\"index.php?act=findpost&pid=56076\"][{POST_SNAPBACK}][/a]
-
Diffraction is a non varible property. In otherwords, it is the lens opening not the f number responsible. A 100mm lens at f8 has the same amount of image degrading diffraction that a 50mm lens does at f 5.6 or a 200mm lens does at f11. It is a property decided by the lens opening.
DrGary
[a href=\"index.php?act=findpost&pid=56073\"][{POST_SNAPBACK}][/a]
That might be true, but there are a few other variables missing from your statement that make it incomplete. F stop is related to both the physical opening and the focal length; field of view is related to distance from the subject; diffraction spot size on the film or sensor is related to the sensor's format.
All my comparisons were done with the same focal length; the same distance to subject; the same FoV and the same format (35mm). The only variable was the f stop or opening which can correctly be described either way, ie. diffraction spot size is inversely proportiona to f stop, or to give a concrete expample, the diffraction spot size at f16 is double the diffraction spot size at f8, all else being equal, which it was.
I think Peter has hit the nail on the head by bringing in the issue of pixel pitch. That's why I think the 1Ds2 should theoretically reveal more resolution difference between f8 and f16 with any good lens. It has supposedly greater resolving power than the 5D, although I notice that this issue seems to be hotly debated on other forums by people who own both the 5D and 1Ds2, the explanation being that the 1Ds2 has a stronger AA filter, although why it should need a stronger AA filter beats me.
I've been following the LL forum since Michael produced his controversial review of the Canon D30 several years ago and I remember well all the whinging and complaining that went on in those early forums about the D30 not being full frame.
It seems clear now, if the D30 had been full frame, maximum resolution would have been achieved at f22, with any 35mm lens no matter how good, and people like me would probably have been observing that there didn't seem to be any significant resolution difference between f22 and f32 .
-
Diffraction is a non varible property. In otherwords, it is the lens opening not the f number responsible. A 100mm lens at f8 has the same amount of image degrading diffraction that a 50mm lens does at f 5.6 or a 200mm lens does at f11. It is a property decided by the lens opening.
DrGary
[{POST_SNAPBACK}][/a] (http://index.php?act=findpost&pid=56073\")
[span style=\'font-size:21pt;line-height:100%\']NO!!![/span]
Image degrading diffraction depends only on f-stop. it does not depend on focal length.
This is a common mistake. What is correct is that the angle of diffraction is proportional to the lens opening (i.e. at constant f-stop, longer lenses have a lower angle of diffraction due to the larger lens opening), but for the longer lenses, the diffraction starts further away and the lower angle of diffraction spreads out over the longer distance. These 2 exactly cancel out, resulting in the fact that degrading diffraction is proportional only to f-stop.
From the lens FAQ: "All lenses are diffraction limited to no more than about 1500/N to 1800/N line pairs per mm". Where N is the f-Number. Note that the focal length is not a part of this formula.
See this site for a detailed description:
[a href=\"http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm#]http://www.cambridgeincolour.com/tutorials...hotography.htm#[/url]
-
[span style=\'font-size:21pt;line-height:100%\']NO!!![/span]
Image degrading diffraction depends only on f-stop. it does not depend on focal length.
See this site for a detailed description:
http://www.cambridgeincolour.com/tutorials...hotography.htm# (http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm#)
[a href=\"index.php?act=findpost&pid=56104\"][{POST_SNAPBACK}][/a]
I've got a sense of dejavu here. You are technically correct if you define an image as a unit of area, such as square mm or square inch and don't concern yourself with the actual picture or the full image.
To understand this, you have to understand the distinction between absolute resolution and picture resolution. The formula 1500-1800/ f stop (n) does not tell you how much an image (picture with height and width) is degraded. It tells you what the absolute diffraction limited resolution will be in terms of line pairs per mm on the image. If you don't know what the field of view (FoV) is, you don't know what the picture is. If you are not concerned with the FoV, the actual picture or full image, or the size of the sensor (or format of the camera), then yes, you are right.
To take a concrete example, a 400mm lens at f45 will produce a diffraction spot size (Airy disk) that is inversely proportional in diameter to the f stop. The resolution limit in line pairs per mm will be determined by the Rayleigh's formula 1500/ f stop (for green light). That applies to all lenses irrespective of focal length, but only in regard to that unit area of image. Such statements have to be qualified.
For example, if I attach that 400mm lens to a 35mm camera, the image will be severely degraded by diffraction. If I attach the lens to an 8x10 field camera, I think most people would be very impressed with the over all sharpness of the image. But they will be different images. The 35mm image would be a small crop or subset of the larger image. It makes a difference to the photographer. I never use f45 on my 100-400 IS zoom (it does have an f45). LF photographers use f45 frequently, producing images much sharper than any image from 35mm at any f stop.
I usually find I have an afterthought with pieces such as this. Here's something to consider. If image degrading diffraction depends only on f stop, then how is it that f stop is a relationship between the physical lens opening and focal length?
Aperture Diameter= Focal length/f stop; F stop = Focal length/aperture diameter. Rayleigh's derived formula for diffraction resolution limits, ie. 1500/ f stop, could also be expressed as 1500 x Aperture Diam/ focal length (all in mm).
-
When doing the thought expirement of moving a 400mm lens from one format to another, what impact is there on the image circle the lens is designed for? When you put an ef-s lens on a ff canon, you just get vignetting, not a wider fov.
-
Unless I've misinterpreted what has been written, I disagree with what has been said. In the absence of aberrations, the resolution of a lens is defined by the aperture. Such a lens is said to be diffraction limited, and some high end and very expensive astronomical refractors are not far off being diffraction limited. The larger the aperture, the higher the resolution. That is one reason astronomers always want bigger telescopes. The other reason is of course to collect more light.
In practice few lenses are diffraction limited, especially camera lenses, and aberrations (coma, CA, spherical aberration, field curvature, astigmatism, and others ?) reduce the resolution.
Anyway, a diaphragm acts as a stop in the system, and introducing a stop is equivalent to reducing the diameter of the objective, for example by masking it off.
So an 600mm F5 lens has a higher diffraction limited resolution than a 200mm F5 lens. So if we were to photograph the same image at the same distance with both lenses, the longer lens would show finer details but would only show a crop of the image in the shorder lens.
However, the 600mm lens will magnify the image more than the 200mm lens. So if you were to photograph an image using both lenses, but move closer when using the 200mm lens so as to get the same image size, which lens would show more detail in the image? I suspect the amount of detail (lines per millimetre) would be the same. Is that the point that is being made in the preceding posts? I suppose that way of looking at it makes sense for photographers, but is very confusing for skywatchers.
And of course a medium format camera has a larger image circle to confuse matter more. (With apologies to Kenstrain who gets the heeby jeebies at the mention of other formats.)
Leif
-
"I disagree with what has been said."
All of it or just a little?
Let me disagree with much of what you said, but not all (I did not read it all).
"In practice few lenses are diffraction limited, especially camera lenses"
except at, for example, f22 when most are, very strongly, for 35mm lenses anyway.
"Anyway, a diaphragm acts as a stop in the system"
hence the phrase "stopping down" meaning to close the aperture, and the unit of f-stops too. This is not a revelation.
"So an 600mm F5 lens has a higher diffraction limited resolution than a 200mm F5 lens. "
I can't believe you wrote that!
Not in the conventional definition of the terms you are using. If you want to redefine it that way you will have to understand that some others will not know what you are talking about.
Anyway by a conventional definition ....
Resolution indicates the finest detail measurable at the focal plane. Diffraction limits the resolution of your example LENSES to about 1600/5 = 320 lp/mm approximately (lenses focussed at infinity, white light, uncoloured images, at some particular choice of contrast). The formula was given in a post above.
Formulae, with suitable definitions of the variables and parameters tell us all we need to know here. Of course it is hard to find lenses that are diffraction-limited at f5 (for 35mm format).
There is no need to invoke focal length or image size or format size or time of day or any other irrelevant quantities in this discussion.
I decided not to read the following paragraphs of your post, but spotted my user name at the end....
"With apologies to Kenstrain who gets the heeby jeebies at the mention of other formats."
Apologies accepted, thank you.
BTW I think the OP reminded us of something very useful: THINK whether the sharpness over depth in the image will fit your needs for the image you are visualising. Know your sensor resolution expressed as an equivalent f-number. Then you can decide when to trade overall sharpness for depth of field. Now that is where focal length comes in (squared). Of course ignore this if you take portaits at f1.4.
But hey - the useful stuff has already been said, unfortunately along with all the waffle.
Norman Koren has some nice tutorials on the topic too, IIRC.
Take care
Ken
-
ken
I found your original posting rather rude, but decided not to comment. My comment on you having the beeby jeebies was a light hearted attempt to deflect your aggression. Now your follow up is extremely unpleasant. If you have some sort of personal problem, then maybe you should sort it out rather than take it out on me.
I am familiar with some optical theory and I do know what I am talking about when it comes to resolution of optics (Dawes limit and all that) which is why I'm not going to discuss the technical content of your posting. What I will say is that there are one or two valid points in your posting and that they could have been made without being so obnoxious. Some of your comments suggest that you did not read my posting properly.
If someone with some manners cares to comment, then I will respond. I happen to find the postings in this thread - Ken's rude ones excepted - extremely interesting. These issues are not at all simple which is why I explained myself as clearly as possible at the risk of sounding condescending which was not my intention.
I thought my original postings were polite and reasonable. If others disagree, then maybe they could politely say so.
Leif
BTW Are you by any chance Ken Rockwell? I'm trying to think why you would be so aggressive.
-
Dear Leif,
may I sincerely apologise for the offense that I have caused, without reserve. Please accept that I did not indend to be rude in the way which my second message clearly came over to you. It is neither my habit nor my will to do so.
Actually it horrifies me that I could be thought of as being rude in that way to the extent which you state. I am truely very sorry about that.
As explanation: I was trying to be a little bit light hearted in the face of what I sincerely believe to be incorrect information that was posted. I do not like to see, what I believe to be, errors being perpetuated. I read the parts of your post I commented on very carefully before replying.
When I saw your reference to heeby jeebies I assumed that you were also being lighthearted about it, and never guessed your ire at my first post, or I would not have continued. I also thought my posts were both inoffensive, if perhaps with a slightly exasperated tone in the second one, but I clearly misjudged that. There was no aggression in the first post, if my language conveyed such that was utterly unintentional.
I am, in fact Ken Strain, and although that is not my full name, it is my only one, so please accept the apology as personal.
Ken
ken
I found your original posting rather rude, but decided not to comment. My comment on you having the beeby jeebies was a light hearted attempt to deflect your aggression. Now your follow up is extremely unpleasant. If you have some sort of personal problem, then maybe you should sort it out rather than take it out on me.
I am familiar with some optical theory and I do know what I am talking about when it comes to resolution of optics (Dawes limit and all that) which is why I'm not going to discuss the technical content of your posting. What I will say is that there are one or two valid points in your posting and that they could have been made without being so obnoxious. Some of your comments suggest that you did not read my posting properly.
If someone with some manners cares to comment, then I will respond. I happen to find the postings in this thread - Ken's rude ones excepted - extremely interesting. These issues are not at all simple which is why I explained myself as clearly as possible at the risk of sounding condescending which was not my intention.
I thought my original postings were polite and reasonable. If others disagree, then maybe they could politely say so.
Leif
BTW Are you by any chance Ken Rockwell? I'm trying to think why you would be so aggressive.
[a href=\"index.php?act=findpost&pid=56144\"][{POST_SNAPBACK}][/a]
-
When doing the thought expirement of moving a 400mm lens from one format to another, what impact is there on the image circle the lens is designed for? When you put an ef-s lens on a ff canon, you just get vignetting, not a wider fov.
[a href=\"index.php?act=findpost&pid=56116\"][{POST_SNAPBACK}][/a]
Tim,
That's a good question. The image circle is enlarged as one stops down but it's doubtful whether any 400mm lens designed for 35mm could be stopped down sufficiently to avoid serious vignetting on a large format camera. This is the sort of question I think BJL could answer.
However, as regards the thought experiment, if there's a problem moving a 35mm lens to a larger format then one would simply use the lens designed for the larger format and transfer that to the smaller format.
-
Hi Ken: Thanks for the response. Okay, I accept that you were not trying to be rude.
I did not understand the quote that "Image degrading diffraction depends only on f-stop." because it goes against what I know about optics. I understand the effective aperture to be the deciding factor as far as diffraction limited resolution goes i.e. the upper limit to resolution from an optic.
You are of course correct that few photographic lenses are diffraction limited unless stopped down below ~F22. Hence this is largely an academic discussion.
Regarding stopping down, it is certainly not obvious to me. Did you know that the human eye acts as a stop on a binocular, and hence in 'good light' there no light gathering advantage to an 8x binocular with a larger objective than about 20mm?
BTW When I refer to a 600mm F5 lens, what I mean is a 600mm lens, stopped down to F5.
Leif
-
Having got out my college physics book, I will now prove that focal length does not affect resolution loss due to diffraction. *drum roll*
Now, we know that diffraction creates these 'Airy disks' which look like bulls-eyes - a central dot surrounded by concentric rings. The smaller the aperture, the bigger this central dot becomes - this is as described in http://www.cambridgeincolour.com/tutorials...photography.htm (http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm), to wit:
When the diameter of the airy disk's central peak becomes large relative to the pixel size in the camera (or maximum tolerable circle of confusion), it begins to have a visual impact on the image. Alternatively, if two airy disks become any closer than half their width they are also no longer resolvable (Rayleigh criterion).
Some images may help, hopefully he won't mind me linking them here:
Airy disk in 2D
(http://www.cambridgeincolour.com/tutorials/graphics/airydisk-rings.jpg)
Airy disk in 3D
(http://www.cambridgeincolour.com/tutorials/graphics/airydisk-3D.png)
In order to figure out the maximum resolution of a lens, we figure out the angular separation of two points which are barely resolved - which for the purposes of this exercise will be two points whose central dots are just touching edge-to-edge - this will be the same as the angular size of either central peak.
The equation to figure this out is:
sinθ = 1.22(λ/D)
Where θ is the angular separation, λ is the wavelength of light (considered a constant in this exercise, and set at 500nm) and D is the aperture diameter.
For simplicity's sake, we'll make sinθ just plain old θ, as θ is so small in any real world scenario to make the difference negligible. That gives us:
θ = 1.22(λ/D)
Now, the experiment: We take two lenses, of 50mm and 100mm focal lengths, and set their apertures to f/2. Take a picture of an object 10 meters away with the 100mm lens, and one of an object 5 meters away with the 50mm lens. We'll take the wavelength of light to be 500nm as a constant.
We do this because for this experiment to be valid, the size of the image on the negative/sensor must be the same for both lenses. Since a 100mm lens gives 2x the magnification of a 50mm lens, we stand back twice the distance with the 100mm lens.
For the 50mm lens, D, the aperture diameter is 50/2 = 25mm.
This gives us:
θ = 1.22(500x10⁻⁹/25x10⁻³)
so...
θ = 2.44x10⁻⁵ rad
For the 100mm lens, D, the aperture diameter is 100/2 = 50mm.
This gives us:
θ = 1.22(500x10⁻⁹/50x10⁻³)
so...
θ = 1.22x10⁻⁵ rad
So, what does that tell us? It tells us that the Airy disk of the 100mm lens is half the angular size of that of the 50mm lens. This is critical to understanding why focal length doesn't matter in terms of diffraction. Longer lenses produce a narrower 'beam' of diffracted light - remember, the number above is in 'rad', so it's an angle, not a distance.
Even though the diffracted light has to travel further to reach the film/sensor in the 100mm lens than the 50mm lens, it's spreading out much more slowly. By the time it reaches the film/sensor it will produce a spot of light exactly as large as the 50mm lens does.
Now to prove this concretely, we want to measure what that means in terms of resolution. To do this, we want to measure the minimum distance between two points on an object that can be resolved with both lenses at the distances and aperture given.
From basic optics:
y/s = y'/s'
Where y is the separation of the points on the object, y' is the separation of the corresponding points on the film/sensor, s is the distance from lens to object and s' is the distance from lens to film/sensor.
Thus the angular separations of the object points and the corresponding image points are both equal to θ, so we get:
y/s = θ
and
y'/s' = θ
Because s is greater than the focal length of the lens, the image distance s' is approximately equal to the focal length.
For the 50mm lens, s = 5m, s' = 50mm and θ = 2.44x10⁻⁵ rad:
y/5m = 2.44x10⁻⁵
so...
y = 1.22x10⁻⁴m = 0.122mm
and
y'/50mm = 2.44x10⁻⁵
so...
y' = 1.22x10⁻³mm = 0.00122mm
For the 100mm lens, s = 10m, s' = 100mm and θ = 1.22x10⁻⁵ rad:
y/10m = 1.22x10⁻⁵
so...
y = 1.22x10⁻⁴m = 0.122mm
and
y'/100mm = 1.22x10⁻⁵
so...
y' = 1.2x10⁻³mm = 0.00122mm
Phew! So, where does that leave us? Both lenses can resolve a minimum distance of .122mm on an object, which translates to a minimum resolvable dot on the film/sensor of .00122mm at those distances and an aperture of f/2. If we change the aperture, we will get different numbers, but the important thing is that they are the same for both lenses - so we've just proved that for the same image size and the same aperture, but different focal lengths, resolution is identical.
Again, the reason for this is that the longer lens produces a narrower cone of diffraction, so even though the light has to travel further to get to the film/sensor, it ends up producing the same size spot of light in both cases.
Thus, focal length does not contribute to resolution loss from diffraction.
Cheers,
Peter
-
I've partially read through this thread with mild interest - and feel slightly as though I may have just grown a pocket protector and found tape on my glasses. I am curious about a couple things though...
I tend to be someone who simplifies things into general principles rather than understanding the theory behind them. So far, Peter (and others) have done a good job establishing that when you are dealing with a single format (35mm) the diffraction limitation remains at equivalent apertures (for any given framing/picture) regardless of whether you use a telephoto lens or a wide lens.
My question is - If you take the same picture (equivalent image framing) with a smaller format camera (a pocket digicam) and with a large format camera (4x5), will you run into the diffraction limits at equivalent apertures? I've always been under the impression that small format cameras were more sensitive to diffraction than the larger formate cameras (hence why none of the pocket digicams goes above f/8 and LF goes to f/64). I'm also only interested in the resolving power at the focal plane in the image - not how many lp/mm fit onto the film itself.
Is there anyone who could clarify this issue?
-
Sheldon -
It's basically a question of enlargement. If you look at a 35mm slide on a lightbox with the naked eye, you wouldn't be able to see any resolution loss due to diffraction. Blow the image up 20 times though, and it'll be easy to see.
Same for a 4x5 transparency - look at it on the lightbox without a loupe and you won't see diffraction problems either. Blow it up 20 times and you will.
The difference is that the 35mm frame needs to be enlarged just about 4 times to be approximately the same size as the 4x5 - that's why if you look at a 4x6 print from a 4x5 frame, and compare it with a 4x6 print from a 35mm frame the 4x5 print blows it away. Along with the fact that the 4x5 frame recorded a boatload more detail, you're also not enlarging the diffraction effects.
So diffraction effects are identical on all formats, it's just not as obvious because of the different enlargement requirements. That's why you can get away with f/64 on large format. Which is just as well, because since large format uses longer lenses, and long lenses have shallower perceived depth of field at the same aperture than shorter lenses, you need smaller apertures to increase your depth of field.
Hope that helps.
Cheers,
Peter
I've partially read through this thread with mild interest - and feel slightly as though I may have just grown a pocket protector and found tape on my glasses. I am curious about a couple things though...
I tend to be someone who simplifies things into general principles rather than understanding the theory behind them. So far, Peter (and others) have done a good job establishing that when you are dealing with a single format (35mm) the diffraction limitation remains at equivalent apertures (for any given framing/picture) regardless of whether you use a telephoto lens or a wide lens.
My question is - If you take the same picture (equivalent image framing) with a smaller format camera (a pocket digicam) and with a large format camera (4x5), will you run into the diffraction limits at equivalent apertures? I've always been under the impression that small format cameras were more sensitive to diffraction than the larger formate cameras (hence why none of the pocket digicams goes above f/8 and LF goes to f/64). I'm also only interested in the resolving power at the focal plane in the image - not how many lp/mm fit onto the film itself.
Is there anyone who could clarify this issue?
[a href=\"index.php?act=findpost&pid=56223\"][{POST_SNAPBACK}][/a]
-
It's basically a question of enlargement. If you look at a 35mm slide on a lightbox with the naked eye, you wouldn't be able to see any resolution loss due to diffraction. Blow the image up 20 times though, and it'll be easy to see.
[a href=\"index.php?act=findpost&pid=56236\"][{POST_SNAPBACK}][/a]
On the other hand, if we are using digital cameras, the enlargement depends on the number of pixels, not the size of the sensor. An 8mp camera produces the same size image whatever the format of the camera (assuming same aspect ratio). So we are back to pixel pitch as a significant factor in determining the point where diffraction kicks in to degrade 'image' quality.
As I see it, it all boils down to a semantic confusion of terms. The statement, 'the degree to which diffraction degrades an image is dependent only on f stop' is in fact poorly expressed. There's an ambiguity in the meaning of the word 'image'. Are we talking about the whole image, ie. a picture, or are we talking about a unit area of that image or picture?
-
Ray -
You are correct. With digital cameras it gets tricky because every camera is different - pixel pitch, pixel count and sensor size are all variable. Still, it's easy enough to figure out a rule of thumb.
Bigger sensors will have less problems as a rule because they will (generally) have a bigger pixel pitch than a small sensor. Even if they have the same or smaller pixel pitch, diffraction will be an issue, but in that case, you have lots more pixels and you won't need to enlarge the image that much to get a good size print.
Michael would brand the lot of us as pixel peepers in this discussion. Personally, I'm not that bothered by the whole issue - I want sharp images, and I happen to know that for the 5D, the optimum aperture is f/16 due to its pixel pitch.
That being said, I still get a lot of unsharp images - misfocus, camera shake, tripod sinking in soft sand on long exposures, etc. Any of these problems creates an unusable photograph. An image that's less than optimally sharp due to diffraction is rarely an unusable photograph if it's compositionally good and otherwise technically good, however. So, to paraphrase someone or other - be happy in your photography!
I wanted to put this argument to bed because there were a lot of 'facts' being bandied about, and it had degenerated into a bit of a shouting match. So I read up on it and laid out the real deal. I enjoyed it - forgot how much I liked physics in college.
If only selling and marketing your work was as easily distilled!
Cheers,
Peter
-
Michael would brand the lot of us as pixel peepers in this discussion. Personally, I'm not that bothered by the whole issue - I want sharp images, and I happen to know that for the 5D, the optimum aperture is f/16 due to its pixel pitch.
[a href=\"index.php?act=findpost&pid=56250\"][{POST_SNAPBACK}][/a]
Peter,
If the above statement is true (and my tests with the 24-105 zoom imply that it is true) it causes me to re-evaluate so many statements I've seen on this forum from 1Ds owners since the introduction of that camera. It almost became a mantra that any shortfalls in the quality of 1Ds images were due to lens deficiencies. This camera needs the finest lenses, I read time and again.
Some folks even went so far as to claim there would be no point in increasing pixel count in a future upgrade, because the 1Ds is as good as it gets and that without an upgrade to Canon's entire range of lenses there would be no point.
I don't recall reading once, from any source, that at least part of the problem might be due to the fact that the pixel pitch of the 1Ds does not allow it to capture any more detail than any good lens can provide at f16. I guess such comments would have been considered too negative.
But before I get jumped on, I will concede that a marginal increase in contrast is apparent at smaller f stops than f16. The contrast difference between f11 and f16 is just too marginal to be concerned about. Likewise between f8 and f11. But jump two stops and you have a difference which doesn't necessarily translate to more detail but does result in slightly better defined detail, or slightly more contrasty detail.
-
Peter,
If the above statement is true (and my tests with the 24-105 zoom imply that it is true) it causes me to re-evaluate so many statements I've seen on this forum from 1Ds owners since the introduction of that camera. It almost became a mantra that any shortfalls in the quality of 1Ds images were due to lens deficiencies. This camera needs the finest lenses, I read time and again.
Well, the 1Ds (mark II) has a smaller pixel pitch than the 5D, so somewhere between f/11 and f/16 you'll see the 'optimum' quality. The fact that it needs the finest lenses is true - we're getting to the point now where these cameras can outresolve all but the finest glass, and they will show up problems that have remained hidden even to film users before.
Even with this, the statement that you'll see the best quality at f/stop x is true, regardless of other problems. I know of no lens that gets worse in sharpness as you stop down, so you'll always get the best sharpness a lens has to offer by stopping down to the point just before diffraction becomes evident. It's just that with many lenses the optimum sharpness isn't as good as it could be, but it didn't matter in the past because the camera bodies didn't have the resolution to show the problems.
Some folks even went so far as to claim there would be no point in increasing pixel count in a future upgrade, because the 1Ds is as good as it gets and that without an upgrade to Canon's entire range of lenses there would be no point.
There is something to this. At this point, I'd prefer to see an upgrade to Canon's wide-angle selection of lenses (which as a rule are nowhere near as good as they could be) than a - say - 20mp body. I suspect we'll get the camera before the lenses though, as I imagine it's easier to cram more pixels onto a sensor than it is to engineer glass to the tolerances now required and sell it at a $1500-ish price point.
That being said, it's always possible to take some of the excellent lenses out there on the market at the moment for other mounts and use them on the Canon bodies. You'll lose autofocus and auto-exposure on most of them, but the results can be pretty stunning if used properly
I don't recall reading once, from any source, that at least part of the problem might be due to the fact that the pixel pitch of the 1Ds does not allow it to capture any more detail than any good lens can provide at f16. I guess such comments would have been considered too negative.
At f/x (where x is the optimum aperture), both the 1Ds mkII and the 5D capture bloody good detail with a good lens - certainly enough that vast numbers of photographers that used medium format film are abandoning it in favour of these bodies. Those that can't afford to go to medium format digital, anyway - which is most of us =)
If you really want excellent near-far sharpness in your photos - the sort that requires f/22 or higher for enough depth of field in an SLR - a view camera is what you need. Then you can tilt the plane of focus and achieve effectively infinite depth of field without having to worry about diffraction effects (yes, they're there, but as discussed earlier they're not nearly as much of an issue).
Cheers,
Peter
-
The fact that it needs the finest lenses is true - we're getting to the point now where these cameras can outresolve all but the finest glass, and they will show up problems that have remained hidden even to film users before.
Peter,
This is the paradox I'm referring to. A camera like the 5D and 1Ds might show up problems that were (previously) partly disguised by film grain, but I'm very dubious about this notion that these cameras can 'outresolve' all but the finest glass. I would say these cameras can only outresolve mediocre lenses, lenses at the edges and cheap lenses at maximum aperture.
I've never seen an MTF chart of a 35mm lens at f16, and the reason no-one is producing them, I'm led to believe, is that all lenses (bar very cheap ones) would have a very similar performance at this f stop.
-
Ray -
I'll allow that the statement about the 5D or 1Ds outresolving lenses might have been a bit hyperbolic. However, I wasn't talking about at f/16 (although edge performance on some of Canon's wide angles does seem to be lacking at small apertures). Whereas before you could get away with relatively large apertures on less-than-world-class glass, now you really do have to stop down to get to the good stuff.
I don't really see a paradox here - there are just a lot of factors that contribute to image quality. The system is only as good as its weakest link, and the sensor - which has for so long been the weakest link in high-end digital photograhy - is now not necessarily so.
Cheers,
Peter
Peter,
This is the paradox I'm referring to. A camera like the 5D and 1Ds might show up problems that were (previously) partly disguised by film grain, but I'm very dubious about this notion that these cameras can 'outresolve' all but the finest glass. I would say these cameras can only outresolve mediocre lenses, lenses at the edges and cheap lenses at maximum aperture.
I've never seen an MTF chart of a 35mm lens at f16, and the reason no-one is producing them, I'm led to believe, is that all lenses (bar very cheap ones) would have a very similar performance at this f stop.
[a href=\"index.php?act=findpost&pid=56308\"][{POST_SNAPBACK}][/a]
-
Ray,
the pattern I have seen is that the highest aperture ratio before diffraction effects become noticeable is about twice the pixel spacing in mm. For example, Thom Hogan observes that about f/11 is the diffraction limit for the D2X with its 5.5 micron pixel pitch.
So with the 8.2 micron pixel pitch of the 5D, f/16 fits as the "diffraction limit". How does f/16 compare to f/22?
Of course f/16 with the 5D gives about the same DOF as f/10 in your 20D (viewing equal sized uncropped prints from the same viewing distance), so I do not see that you have gained or lost anything in DOF by the change from the 20D to the 5D; you just get to use slower f-stops and thus lower shutter speeds and/or higher ISO settings to get the same DOF.
P.S. Resolution comparisons between cameras in different formats are best done in terms of angular resolution of the subject, and then the physics of diffraction is simple: the smallest resolvable angle is inversely proportional to the effective aperture diameter, which is the focal length divided by the aperture ratio (f-stop).
And as I have said many times before, the DOF on equal sized prints and the speed at which the camera gathers light from the subject are also determined by the effective aperture diameter, independent of focal length and format: aperture diameter is a far more natural and convenient quantity than aperture ratio when comparing images of the same subject (including same field of view) made using different focal lengths and formats, even though f-stop is more convenient to use with ISO speed in determining exposure settings.
-
Whereas before you could get away with relatively large apertures on less-than-world-class glass, now you really do have to stop down to get to the good stuff.
Peter,
This is the paradox. There's no doubt that f16 is not the aperture at which 35mm lenses are sharpest. Most lenses are sharper at f8 and a few of the really expensive lenses are sharpest at f5.6 and even f4.
The system is only as good as its weakest link, and the sensor - which has for so long been the weakest link in high-end digital photograhy - is now not necessarily so.
Where's the evidence for this, considering that the pixel spacing of the 5D and earlier 1Ds cannot resolve greater resolution than that provided at f16 (now confirmed by BJL, our resident lens expert ). If the sensor is no longer the weakest link, I'd expect to see that increase in resolution between f4 and f8 with all good lenses. Instead, what I see is just a marginal increase in the contrast of micro detail at resolutions that are far below the resolution limits of the lenses at their sharpest apertures.
My impression is, to get just a marginal increase in image quality, we have to get a substantial increase in lens quality. The sensor is still the weakest link. An indication of this is the usual fall-off in resolution at the edges and particularly in the corners of full frame images.
If you've ever looked at the Photodo MTF charts for quite a number of Canon lenses they've tested, you'll see that resolution (or more accurately the contrast at specific resolutions) falls off dramatically usually in that part of the image outside the crop area of the APS-C size sensors. A disadvantage of moving up to full frame 35mm is this expectation that image degradation might be a problem towards the edges. It can be, but my experience so far with the 5D is that resolution fall off towards the edges is not nearly as dramatic as indicated in the Photodo charts and less than my worst fears.
Finally, at the risk of antagonising BJL , I'll use the example of the Olympus 4/3rds format. The Zuiko lenses designed for this format are by all accounts superb. Far better it would appear than their Canon counterparts. Yet, if you check out the very detailed and thorough reviews of the Olympus 8MP E-300 and E-500 at dpreview, showing comparisons with the Canon 350D and 20D, you'll see that absolute resolution is a shade less than that of the Canon 8mp cameras.
But the Zuiko lenses are, I suggest, a lot more than a shade better. Of course you could argue that the 8MP Olympus sensors (manufactured by Kodak, I believe) are just not as good as Canon sensors. In the noise department, they don't appear to be. But resolution?
Unfortunately, there are no adapters for using Zuiko lenses on the Canon 350 and 20D, so we'll never really know what a really good lens can do when coupled with a really good sensor .
-
Of course f/16 with the 5D gives about the same DOF as f/10 in your 20D (viewing equal sized uncropped prints from the same viewing distance), so I do not see that you have gained or lost anything in DOF by the change from the 20D to the 5D; you just get to use slower f-stops and thus lower shutter speeds and/or higher ISO settings to get the same DOF.
I've gained some 'picture' resolution, have I not? My images are now comprised of 12MP instead of 8MP. I've also gained the opportunity to use a slower shutter speed, not only because f16 is slower than f10 but also because of an ISO 50 option on the 5D.
If I'm shooting a waterfall and I've forgotten to bring my ND filter, and I'm not keen on the water looking like melted ice-cream, I just might be able to get the right effect using f16 and ISO 50 . This is an afterthought, (edited). To all you P&S shooters, if you want to get a good shot of a waterfall, forget it. You're going to be using a wide aperture and fast shutter speed, and the water's going to look like.... errh!.... ice cream?... errh! "S**t.
aperture diameter is a far more natural and convenient quantity than aperture ratio when comparing images of the same subject (including same field of view) made using different focal lengths and formats, even though f-stop is more convenient to use with ISO speed in determining exposure settings.
And this is exactly what we've been doing in our conversation in the Medium Format section, isn't it?
-
Ray -
It's edge performance I'm talking about when I say the sensor is no longer the weakest link =)
I stand corrected on the fact that some lenses lose sharpness when stopped down, although it seems counter-intuitive to me. I'll have to read up on it.
Cheers,
Peter
-
It's basically an issue of pixel pitch - the smaller the pixels themselves, the sooner diffraction effects kick in. There's an excellent article on understanding diffraction and pixel pitch here: http://www.cambridgeincolour.com/tutorials...photography.htm (http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm)
From that article, a 5D will start to lose resolution to diffraction _after_ f/16. In my own tests to verify this I found that I saw improved sharpness at f/16 over f/11, but lost sharpness at f/22. With my 20D, I found that f/11 was the minimum aperture I could use before losing resolution (using the same lens, the 24-70L).
That being said, as Michael says in his 'stop down!' article, if you need the depth of field, shoot at the appropriate aperture and worry about diffraction later.
It's also worth remembering that everything has to go right to get the sharpest possible shot. In a controlled environment with plenty of time and no discomfort, it's easy to get razor sharp shots. In the field, with wind, rain, soft ground, misfocussing (remember - AF still isn't as reliable as we'd like it to be for landscape work, especially) and photographer discomfort it's easy to get it wrong!
Cheers,
Peter
[{POST_SNAPBACK}][/a] (http://index.php?act=findpost&pid=56048\")
The Cambridge in Color site you mention is indeed quite good. However, one should be aware that the diffraction limited resolution calculation there is for the Raleigh limit, which corresponds to about 10% MTF. When speaking about resolution one should always give two figures: the resolution and the MTF at that resolution.
For most photographic work, the resolutiion at MTF of 50% corresponds much better to perceived sharpness. The contrast at 10% MTF is simply too low--that limit is good for resolving stars with a black background (very high contrast), but less useful for normal picture taking. To obtain the extra MTF one needs to use a a larger aperture. Roger Clark has a table on his website:
[a href=\"http://www.clarkvision.com/imagedetail/scandetail.html#diffraction]http://www.clarkvision.com/imagedetail/sca...tml#diffraction[/url]
Here is a MTF 50 plot for my 50mm f/1.8 Nikkor with the D70 produced with Imitest by Norman Koren. It shows definite image degradation at f/16 and beyond.
http://bjanes.smugmug.com/photos/17635874-M-2.jpg (http://bjanes.smugmug.com/photos/17635874-M-2.jpg)
-
Finally, at the risk of antagonising BJL ... if you check out the very detailed and thorough reviews of the Olympus 8MP E-300 and E-500 at dpreview, showing comparisons with the Canon 350D and 20D, you'll see that absolute resolution is a shade less than that of the Canon 8mp cameras.
[a href=\"index.php?act=findpost&pid=56382\"][{POST_SNAPBACK}][/a]
Differences in both sensor and approaches to in-camera JPEG processing make such comparisons useless as assessments of lens performance differences; after all, the 20D and even 350D are significantly more expensive than the E-300 and E-500.
-
Differences in both sensor and approaches to in-camera JPEG processing make such comparisons useless as assessments of lens performance differences; after all, the 20D and even 350D are significantly more expensive than the E-300 and E-500.
[a href=\"index.php?act=findpost&pid=56425\"][{POST_SNAPBACK}][/a]
BJL,
So has anyone done any E-500/350D comparisons using only RAW images? The obvious problem with such comparisons is variation in the quality of the converters used, for example C1 has a reputation for extracting marginally more detail from a shot than ACR.
But, for what it's worth, dpreview did do some in-camera jpeg and ACR RAW conversion comparisons with the E-500. See below:
The first visual difference between JPEG straight from the camera and RAW converted using Adobe Camera RAW is the lack of artifacts in the ACR image, it has a much cleaner and less processed appearance with no sharpening halos around detail. Detail wise there doesn't seem to be a big advantage from the RAW image. The next difference comes in color rendition which is noticeably different, Adobe Camera RAW isn't rendering greens correctly, we can only assume this could be to do with support of the E-500 in this Beta version being only preliminary.
As regards the E-500 being a $100 cheaper than the 350D, it's not really significant in the general scheme of things after lens purchases have been taken into consideration. I mean, those Zuiko lenses are not cheaper than sometimes excellent 3rd party lenses for the 350D.
-
For most photographic work, the resolutiion at MTF of 50% corresponds much better to perceived sharpness. The contrast at 10% MTF is simply too low--that limit is good for resolving stars with a black background (very high contrast), but less useful for normal picture taking. [a href=\"index.php?act=findpost&pid=56396\"][{POST_SNAPBACK}][/a]
Not only is any signal at 10% MTF less useful for normal picture taking, I think it would be fair to say it's of no use whatsoever if we are talking about high spatial frequencies. Even at f32, sometimes the smallest available aperture on 35mm lenses, a theoretical diffraction limited resolution of 50 lp/mm at 10% MTF would not be picked up at all by the sensor.
-
I should also have mentioned differences in AA filters as another reason that fine differences in lens performance are hard to judge from such tests.
As regards the E-500 being a $100 cheaper than the 350D, it's not really significant in the general scheme of things after lens purchases have been taken into consideration. I mean, those Zuiko lenses are not cheaper than sometimes excellent 3rd party lenses for the 350D.
[a href=\"index.php?act=findpost&pid=56453\"][{POST_SNAPBACK}][/a]
I mentioned prices to compare sensor costs: the E-300 and E-500 seem to have a substantially cheaper sensor. Lens costs are irrelevant to that. (I also note that you choose to compare the more expensive Olympus to the far cheaper of the two Canons!)
My bottom line is simply that there are far too many uncontrolled variables in these comparisons to conclude anything about the inherent performance limits of lenses for two slightly different formats (10% vertical, 20% horizontal difference in frame size.)
-
Not only is any signal at 10% MTF less useful for normal picture taking, I think it would be fair to say it's of no use whatsoever if we are talking about high spatial frequencies. Even at f32, sometimes the smallest available aperture on 35mm lenses, a theoretical diffraction limited resolution of 50 lp/mm at 10% MTF would not be picked up at all by the sensor.
[a href=\"index.php?act=findpost&pid=56454\"][{POST_SNAPBACK}][/a]
Why would 50 lp/mm at 10% MTF not be picked up by the sensor? The Nyquist for the Canon EOS 1Ds Mark II is 69 lp/mm. For the Digital Rebel 350 it is 78 lp/mm.
-
Why would 50 lp/mm at 10% MTF not be picked up by the sensor? The Nyquist for the Canon EOS 1Ds Mark II is 69 lp/mm. For the Digital Rebel 350 it is 78 lp/mm.
[a href=\"index.php?act=findpost&pid=56486\"][{POST_SNAPBACK}][/a]
Why? I'm not setting myself up as an expert on such matters, but my understanding is a sensor is not a perfect recording device. It's not a 100% efficient and does not have infinite sensitivity. The light signal reaching it has to compete with noise from a number of sources, including noise within the light signal itself, called photon noise.
Furthermore, Bayer type sensors cannot deliver resolution up to the Nyquist limit because of their requirement to interpolate the RGB values. You'll find somewhere on Norman Koren's site a resolution test of his Canon 10D and plotted MTF response of the recorded resolution. From memory, there was no meaningful resolution to be had beyond 54 lp/mm and the MTF response at that cut-off point was 10%. The lens he used would probably have had an MTF resonse of around 40% at 54 lp/mm. If the lens had been delivering 50 lp/mm at just 10% MTF because the aperture was f32, there's no chance that such a faint signal would have been recorded.
The Rayleigh's and Dawe's limits ot 10% and 2% appear to be only useful for astronomy. Such faint signals can be detected when peering through a telescope and maybe even recorded on digital sensors using very long exposures, multiple exposures and stacked images.
I think it's probably true to say that any signal coming through your lens that has lost more than 70% of its original contrast, is pretty irrelevant.
-
Why? I'm not setting myself up as an expert on such matters, but my understanding is a sensor is not a perfect recording device. It's not a 100% efficient and does not have infinite sensitivity. The light signal reaching it has to compete with noise from a number of sources, including noise within the light signal itself, called photon noise.
Furthermore, Bayer type sensors cannot deliver resolution up to the Nyquist limit because of their requirement to interpolate the RGB values. You'll find somewhere on Norman Koren's site a resolution test of his Canon 10D and plotted MTF response of the recorded resolution. From memory, there was no meaningful resolution to be had beyond 54 lp/mm and the MTF response at that cut-off point was 10%. The lens he used would probably have had an MTF resonse of around 40% at 54 lp/mm. If the lens had been delivering 50 lp/mm at just 10% MTF because the aperture was f32, there's no chance that such a faint signal would have been recorded.
The Rayleigh's and Dawe's limits ot 10% and 2% appear to be only useful for astronomy. Such faint signals can be detected when peering through a telescope and maybe even recorded on digital sensors using very long exposures, multiple exposures and stacked images.
I think it's probably true to say that any signal coming through your lens that has lost more than 70% of its original contrast, is pretty irrelevant.
[a href=\"index.php?act=findpost&pid=56522\"][{POST_SNAPBACK}][/a]
Ray,
I checked Norman Koren's web site. The Nyquist limit for sensors is 0.5 cycles/pixel whereas 0.33 cycles/pixel is excellent performance for a Bayer camera. According to these figures, the 1Ds MII would resolve at best 46 lp/mm (cycles/mm) and the Digital Rebel around 51 lp/mm (at a relatively low MTF), pretty much in agreement with your memory.
At f/16 the lens would lay down 19, 48, and 100 lp/mm respectively on the sensor at MTF 80%, 50%, and the Rayleigh criterion respectively. The 100 lp/mm at 10% MTF is most likely not that meaningful as you say, but the 48 lp/mm at 50% is in the range that can be recorded by the camera and one can readily see why stopping down beyond f/16 with cameras of this pixel size is detrimental to image quality.
To get MTF above 80% on the sensor, one would have to open up to nearly f/5.6, which can be done with a diffraction limited lens. If one opens up more, abberations would begin to degrade the image with less than diffraction limited lenses. Thus, the sweet spot with these cameras is not all that different than were taught applied to 35 mm cameras. I don't claim to be an expert in this area either, but this is my best effort to conceptualize the physics involved. I would welcome other interpretations.
Bill
-
Thus, the sweet spot with these cameras is not all that different than were taught applied to 35 mm cameras. I don't claim to be an expert in this area either, but this is my best effort to conceptualize the physics involved. I would welcome other interpretations
Bill, here's my interpretation, at the risk of starting another film versus digital war .
50 lp/mm on film was always something that could be achieved without too much difficulty; just a fine grained film, tripod and the usual sound technique was all that was required. Reaching 100 lp/mm was the really difficult thing, probably requiring the use of high resolution B&W films like T-Max 100 which has an excellent MTF response (100% at 50 lp/mm and even as high as 60% at 100 lp/mm).
But to break the 50 lp/mm barrier with film required the use of really sharp lenses, such as the Canon 200/1.8 at f4, its sweet spot. Now that's a really expensive lens and the best that most lenses can deliver is a slightly lower resolution at f8, but still higher than f11 or f16.
Now there's no disputing that the 5D (and 1Ds) can deliver spades of high MTF detail below 50 lp/mm, which is cleaner and sharper than the same detail on film. But it can't deliver anything at all above 50 lp/mm.
One can argue till the cows come home whether detail on 35mm film above 50 lp/mm serves much purpose. You need a very high resolution scanner to retrieve it for a start, and a lot of that detail is unavoidably obscured by grain. Nevertheless, it would seem reasonable to suppose that a lens which is capable of delivering say 100 lp/mm at 30% MTF could have an impact on a film that is capable of recording 100 lp/mm at 60% MTF. The net effect is a recorded image which has 18% of its original contrast; clearly not something that's going to hit you in the eye but something which is discernible.
But just in case I've caused confusion here, I must make it clear I am not trying to say that high quality primes are a waste of time on cameras such as the 5D. I'm saying that the ability of such high quality lenses to provide high spatial detail at relatively high MTFs (eg. 70 lp/mm at 40% MTF) is an irrelevancy. The only thing that matters is the performance of the lens up to 50 lp/mm.
-
Bill, here's my interpretation, at the risk of starting another film versus digital war .
50 lp/mm on film was always something that could be achieved without too much difficulty; just a fine grained film, tripod and the usual sound technique was all that was required. Reaching 100 lp/mm was the really difficult thing, probably requiring the use of high resolution B&W films like T-Max 100 which has an excellent MTF response (100% at 50 lp/mm and even as high as 60% at 100 lp/mm).
But to break the 50 lp/mm barrier with film required the use of really sharp lenses, such as the Canon 200/1.8 at f4, its sweet spot. Now that's a really expensive lens and the best that most lenses can deliver is a slightly lower resolution at f8, but still higher than f11 or f16.
Now there's no disputing that the 5D (and 1Ds) can deliver spades of high MTF detail below 50 lp/mm, which is cleaner and sharper than the same detail on film. But it can't deliver anything at all above 50 lp/mm.
One can argue till the cows come home whether detail on 35mm film above 50 lp/mm serves much purpose. You need a very high resolution scanner to retrieve it for a start, and a lot of that detail is unavoidably obscured by grain. Nevertheless, it would seem reasonable to suppose that a lens which is capable of delivering say 100 lp/mm at 30% MTF could have an impact on a film that is capable of recording 100 lp/mm at 60% MTF. The net effect is a recorded image which has 18% of its original contrast; clearly not something that's going to hit you in the eye but something which is discernible.
But just in case I've caused confusion here, I must make it clear I am not trying to say that high quality primes are a waste of time on cameras such as the 5D. I'm saying that the ability of such high quality lenses to provide high spatial detail at relatively high MTFs (eg. 70 lp/mm at 40% MTF) is an irrelevancy. The only thing that matters is the performance of the lens up to 50 lp/mm.
[{POST_SNAPBACK}][/a] (http://index.php?act=findpost&pid=56603\")
Ray,
You are correct that reaching 100 lp/mm with black and white film is achievable with 35 mm, but with color transparency film, the MTF drops off beyond 20 lp/mm as shown by Norman Koren on his web page:
[a href=\"http://www.normankoren.com/Tutorials/MTF1A.html]http://www.normankoren.com/Tutorials/MTF1A.html[/url]
Furthermore, even with B/W film, getting 80 lp/mm in the print is not easy at all as discussed by the Leica authorty Erwin Puts (the enlarger or scanner also limits MTF):
http://www.imx.nl/photosite/technical/highres.html (http://www.imx.nl/photosite/technical/highres.html)
In their MTF charts, Leica publishes MTF at 5, 10, 20, and 40 lp/mm and does not bother to go higher, presumably because higher values are not meaningful in practical photography. They state that the 5 and 10 lp/mm data relates to large object contrast while the 20 and 40 lp/mm data relates to small object resolution.
I am not interested if a digital vs film flame war either, but nearly everyone agrees that the EOS 1Ds Mark II gives better results than film, even though it can't do much better than 50 lp/mm. It does this by giving high MTF at the lower frequencies where it counts.
The concept of reporting camera resolution at extinction (10% MTF), as done by DPreview and some other testing sites is obsolete. As Mr. Koren states, it does not make sense to test resolution where detail vanishes. The resolution at MTF of 50% is more meaningful.
Here is the MTF at various f/stops, taken from Roger Clark's web site for green light (wave length 500 nm) plotted on a graph. A black line indicates the 50 lp/mm limit that you have discussed. One should try to maintain high MTF at frequencies below this limit, and this is not possible with f/stops much below f/16 as shown on the graph. This interpretation is consistent with my Imitest results for the Nikon D70 with the 50mm f/1.8 lens. I would agree that high quality primes are not a waste with digital, since they can deliver higher MTF at low frequencies and this can be used by digital.
Bill
-
You are correct that reaching 100 lp/mm with black and white film is achievable with 35 mm, but with color transparency film, the MTF drops off beyond 20 lp/mm
I know . I checked out Fuji's MTF charts for their films some time ago. Their highly regarded Provia 100F has an MTF response barely more than 30% at 50 lp/mm.
The resolution at MTF of 50% is more meaningful.
No argument there, but one wonders why Canon still publish MTF data at only 10 lp/mm and 30 lp/mm.
Leica publishes MTF at 5, 10, 20, and 40 lp/mm and does not bother to go higher, presumably because higher values are not meaningful in practical photography.
But that's going to change, isn't it? Cameras like the 20D and D2X are already capable of greater than 50 lp/mm with a good lens. Doesn't Leica have a reputation for producing contrasty lenses rather than high resolution lenses?
Here is the MTF at various f/stops, taken from Roger Clark's web site for green light (wave length 500 nm) plotted on a graph.
That's an interesting chart and confirms what I've understood for some time, namely that resolution increases inversely with f stop provided the lenses are diffraction limited at such f/stops, which of course they frequently aren't. The resolution and f/stop relationshsips on Roger Clark's graph apply to ideal lenses which are diffraction limited at the f/stops shown.
I'd like to believe my Canon 24-105 is diffraction limited at f16 and is able to deliver detail to the sensor at 40 lp/mm (if not 50) at 50% MTF. Even if this is true, there's no way this lens could have an MTF 50 response at 70 lp/mm at f11, and 100 lp/mm at f8, but this sort of performance will be needed if the megapixel race continues.
By the way, a test report of the Minolta Dimage Scan Elite 5400 I read some time ago claimed that the maximum resolution that was achieved scanning B&W film was 74 lp/mm, presumably high contrast test chart lines. Most 4000 dpi desktop scanners are probably not capable of more than 65 lp/mm.
-
No argument there, but one wonders why Canon still publish MTF data at only 10 lp/mm and 30 lp/mm.
But that's going to change, isn't it? Cameras like the 20D and D2X are already capable of greater than 50 lp/mm with a good lens.
[a href=\"index.php?act=findpost&pid=56622\"][{POST_SNAPBACK}][/a]
One sensible approach to resolution is saying how good the contrast and such is at length scales or relevance to typical print size and viewing conditions, and figures lie 10 and 30 lp/mm, or 240 and 720 line pairs per picture height (lp/ph) are well established standard choices. But I agree that once sensors resolve to about 50lp/mm or better and lenses are often being used with formats smaller than 24x36mm, some new lp/mm levels should be chosen. Olympus has gone to 20 and 60 lp/mm for FourThirds lenses (270 and 810 lp/ph) so it is strange that even for EF-S lenses explicitly intended for a smaller 15x22.5mm frame, Canon still uses 10 and 30, instead of say 15 and 50.
P. S. It would probably be messy to compute and publish 50% MTF graphs for a lens, which would have to give the lp/mm at which 50%MTF is achieved for each value of radius from the optical axis, so MTF as a function of radius at a few well chosen lp/mm or lp/ph values is probably the only practical approach.
-
One sensible approach to resolution is saying how good the contrast and such is at length scales or relevance to typical print size and viewing conditions, and figures lie 10 and 30 lp/mm, or 240 and 720 line pairs per picture height (lp/ph) are well established standard choices. approach.
[a href=\"index.php?act=findpost&pid=56644\"][{POST_SNAPBACK}][/a]
This reminds me of the dialog between Jonathan and Howard on the choice of DoF parameters relating to print size. Howard's view was that he would set up the shot with a specific print size in mind and that his choice of f/stop and CoC would relate to that print size. Jonathan reckoned that was bollocks . He didn't want to reshoot a scene because the client later decided he wanted a bigger print.
My own view is, I don't have a specific print size in mind when I take a shot. However, experience tells me that several years down the track, I shall quite likely want to make larger prints, because (a) wide format printers have come down in price, ( sophisticated 'fractal' type programs relying upon massive computing power will probably do a better interpolation job than current programs, and © as I get older and my eyesight gets worse, I shall simply desire larger prints.
-
Okay! I've learned that b in brackets means something else in this program .
To get back on topic (and I should add that I'm not at all averse to digressions); my situation here is that my previous experience with 35mm film, in addition to hearsay and rule-of-thumb assertions that f8 delivers the crispest image, has created an impression that f16 is a resolution compromise.
I've just returned from a big shooting expedition in Nepal, Thailand and Cambodia where there were many situations where I knew I wanted the sharpest result with regard to the subject that was in focus, but (if possible) an equally sharp result with equally interesting material some distance from the focus point.
In such situations I generally stuck to f11. It now seems, perhaps specifically in regard to the Canon 24-105 lens, that this was a misjudgment on my part. To put it bluntly, I've bungled a lot of my shots.
Shall I give an example? Excuse the fact these are not fully processed images for printing. The first is the overview, with some cropping of the foreground to remove OOF elements resulting from an f11 choice of aperture.
[attachment=180:attachment]
The next crop shows the sacrifice in resolution I got by choosing f11 instead of f16. Never mind the foreground. That can be cropped, and has been. But the fourth bas relief is an essential part of the composition and it's also OOF. In case it's not apparent, there are four bas reliefs there, in increasing order of distance from the camera.
[attachment=181:attachment]
Of course, anyone might make the comment, you've got a digital camera, why didn't you take a number of shots at different apertures and focussing distances?
Okay! I've got my excuses . My Tuk Tuk driver had been waiting about 5 hours and I'd told him I'd be just a couple of hours. Previous experience had taught me that f16 should be avoided if good resolution was desired.
I was sort of hoping when I started this thread, that I'd get feedback from some readers who happen to have some exceptionally good prime lenses and who might have carried out their own tests on a 5D to assure me that good primes will always deliver a noticeably sharper result at f8 than at f16.
It hasn't happened. Where are you people? Not interested?
-
I was sort of hoping when I started this thread, that I'd get feedback from some readers who happen to have some exceptionally good prime lenses and who might have carried out their own tests on a 5D to assure me that good primes will always deliver a noticeably sharper result at f8 than at f16.
It hasn't happened. Where are you people? Not interested?
[{POST_SNAPBACK}][/a] (http://index.php?act=findpost&pid=56656\")
Ray,
I did post a graph showing resolution of the Nikkor 50mm f/1.8 with the Nikon D70. Granted it is not of Leica quality, but it is known for excellent results when stopped down a bit from maximum aperture. To see the evaluation, click on the lenses link on the left and then normal lenses.
[a href=\"http://www.naturfotograf.com/index2.html]http://www.naturfotograf.com/index2.html[/url]
Here is another example of that lens with the D200. The results are from Norman Korens Imitest. Uncorrected means no sharpening and corrected means standardized sharpening. The results are hardly better than with the D70, but here they are:
-
I did post a graph showing resolution of the Nikkor 50mm f/1.8 with the Nikon D70.
[a href=\"index.php?act=findpost&pid=56669\"][{POST_SNAPBACK}][/a]
Bill,
Yes you did. But both the D70 and the D200 are smaller format than the 5D and try to make up for it with greater pixel density and higher resolution.
My guess is that the 5D has no meaningful resolution beyond 45 lp/mm. If the 25-105 IS lens is optimised for good performance at f16, there is no resolution advantage to be gained by opening up the aperture, there's only a slight increase in MTF at resolutions up to 45 lp/mm, and with this lens such MTF increases appear to be negligible. Simply switching from ACR to Raw Shooter probably has a greater effect on image detail, at the plane of focus, than opening up from f16 to f8.
Maybe I'll shoot some Norman Koren test charts when I have the time.
-
(Offtopic)
Furthermore, even with B/W film, getting 80 lp/mm in the print is not easy at all as discussed by the Leica authorty Erwin Puts (the enlarger or scanner also limits MTF):
http://www.imx.nl/photosite/technical/highres.html (http://www.imx.nl/photosite/technical/highres.html)
I was just reading these interesting articles on very high resolution film and the technology behind it ... The real limits of resolution are still far away as proven by the semiconductor industry. Silicon chips are created using photo-sensative materials and optical exposure - the chips are simply prints made from the the negative (mask). So, what resolutions are achievable at the limits of todays optic / photographical technology ...
Smallest technology in production today can produce structures on the chip of 0.065um which corresponds to 7700 lp/mm. Thats a good camera (you would hope so for $1M)!
Amazing, huh!
The structures on the 5D are made in a 0.5um technology - only 1000lp/mm needed to make those sensors.
http://www.chipworks.com/WebReports/ShowRe...ds=CAR-0601-805 (http://www.chipworks.com/WebReports/ShowReports.asp?taKeywords=CAR-0601-805)
(They also have a nice pic of the 20D sensor:
http://www.chipworks.com/WebReports/ShowRe...ds=PPR-0501-001 (http://www.chipworks.com/WebReports/ShowReports.asp?taKeywords=PPR-0501-001) )
-
Bill,
Yes you did. But both the D70 and the D200 are smaller format than the 5D and try to make up for it with greater pixel density and higher resolution.
My guess is that the 5D has no meaningful resolution beyond 45 lp/mm. If the 25-105 IS lens is optimised for good performance at f16, there is no resolution advantage to be gained by opening up the aperture, there's only a slight increase in MTF at resolutions up to 45 lp/mm, and with this lens such MTF increases appear to be negligible. Simply switching from ACR to Raw Shooter probably has a greater effect on image detail, at the plane of focus, than opening up from f16 to f8.
Maybe I'll shoot some Norman Koren test charts when I have the time.
[a href=\"index.php?act=findpost&pid=56674\"][{POST_SNAPBACK}][/a]
Although the 5D has more pixels than the D70, the pixel sizes are comparable. The 5D pixel size is 8.2 microns and the D70's is 7.8 microns. Diffraction should affect both cameras to about the same degree and other factors being equal I would expect the resolution in LP/mm to be about the same. Of course, the D70's crop factor would require that you move back to fill the field and the image details would be smaller.
Of course, the Imitest results would be more definitive. Why do you think that your lens is optimized for f/16? I would think that the "sweet spot" would occur at a somewhat larger aperture if it is well corrected for aberrations. You mention Raw Shooter vs ACR (I use the latter). Is there a significant difference in the resolution that the converters can extract from the file?
-
The 5D pixel size is 8.2 microns and the D70's is 7.8 microns. Diffraction should affect both cameras to about the same degree and other factors being equal I would expect the resolution in LP/mm to be about the same.
I didn't realise the pixel pitch was so close. I guess this is explained by the fact the 5D actually has 12.8MP and the D70 sensor is actually slightly bigger than the Canon D60 sensor, which was my first DSLR. You've got a crop factor of 1.5 there instead of 1.6.
I just checked the resolution charts at dpreview and it seems the D70 does have marginally greater resolution than the 5D. Eg. D70 lines per picture height (LPH) vertically = 1450. D70 sensor vertically = 15.6mm. 1450/15.6 = 93 lines.
93 lines = 46.5 line pairs.
A similar calculation for the 5D gives a result of 42 lp/mm. These are described as absolute resolution figures relatively free of aliasing and moire effects.
One could argue that resolution in the centre of the image could be greater. However, the edge of the 5D frame vertically is only 12mm from the centre. The Photodo MTF charts for the Canon 50/1.4 at f8 show a remarkably flat response at 40 lp/mm all the way to 18mm from the centre, so I think the 43 lp/mm maximum, practical resolution for the 5D is likely to be quite accurate.
What we're left with, it seems, is the practical effect of an increase in MTF across all resolutions below 42 lp/mm as we open up the aperture from f16. This is obviously going to vary with the quality of the lens and one would expect prime lenses to produce a contrastier result.
Why do you think that your lens is optimized for f/16? I would think that the "sweet spot" would occur at a somewhat larger aperture if it is well corrected for aberrations
Just my eyes. That's what my sample images posted earlier in the thread are telling me. The 'sweet spot' is probably still f8, but the differences between f8 and f16 are so small as not to matter.
I must get myself one of those Canon 50/1.4 lenses. I'd forgotten it has such a flat response at 40 lp/mm and f8.
You mention Raw Shooter vs ACR (I use the latter). Is there a significant difference in the resolution that the converters can extract from the file?
Yes. There appears to be; far greater than the 24-105 differences at f8 and f16.
However, perhaps it would be more accurate to say, if one restricts oneself to using sharpening options only available from within the RAW converter, then Raw Shooter does a much better job than ACR. If one is a real whiz at using other sharpening programs and one has a perfected technique of post conversion sharpening, then who knows?
-
Offtopic
Okay! I've learned that b in brackets means something else in this program :D .
You can turn off emoticons by unchecking "Enable emoticons?" just below the text box.
See: (B)
But unfortunately, there appears to be no way to disable ( C ) -> ©
-
(Offtopic)
I was just reading these interesting articles on very high resolution film and the technology behind it ... The real limits of resolution are still far away as proven by the semiconductor industry. Silicon chips are created using photo-sensative materials and optical exposure - the chips are simply prints made from the the negative (mask). So, what resolutions are achievable at the limits of todays optic / photographical technology ...
Smallest technology in production today can produce structures on the chip of 0.065um which corresponds to 7700 lp/mm. Thats a good camera (you would hope so for $1M)!
Amazing, huh!
The structures on the 5D are made in a 0.5um technology - only 1000lp/mm needed to make those sensors.
You should read up a bit on how they're having problems with photolitography because of the wavelengths of the extreme ultraviolet they're using today (the mentioned 65 nm structures; 0.065 µm = 65 nm) ...
That they're having problems doesn't mean that they're confident about solving them; Intel have just recently started mass production of 65 nm technology, and several actors in the market are working on 45 nm prototypes.
But the usefulness for visible light, which is one order of magnitude greater in wavelength, is ... limited.
I'm sure Bjørn Rørslett would be delighted, should he ever get the lenses and sensor filters that would transmit UV light at these wavelengths, but your hopes are not based in reality.
-
Having got out my college physics book, I will now prove that focal length does not affect resolution loss due to diffraction. *drum roll*
Now, we know that diffraction creates these 'Airy disks' which look like bulls-eyes - a central dot surrounded by concentric rings. The smaller the aperture, the bigger this central dot becomes - this is as described in http://www.cambridgeincolour.com/tutorials...photography.htm (http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm), to wit:
Some images may help, hopefully he won't mind me linking them here:
Airy disk in 2D
(http://www.cambridgeincolour.com/tutorials/graphics/airydisk-rings.jpg)
Airy disk in 3D
(http://www.cambridgeincolour.com/tutorials/graphics/airydisk-3D.png)
In order to figure out the maximum resolution of a lens, we figure out the angular separation of two points which are barely resolved - which for the purposes of this exercise will be two points whose central dots are just touching edge-to-edge - this will be the same as the angular size of either central peak.
The equation to figure this out is:
sinθ = 1.22(λ/D)
Where θ is the angular separation, λ is the wavelength of light (considered a constant in this exercise, and set at 500nm) and D is the aperture diameter.
For simplicity's sake, we'll make sinθ just plain old θ, as θ is so small in any real world scenario to make the difference negligible. That gives us:
θ = 1.22(λ/D)
Now, the experiment: We take two lenses, of 50mm and 100mm focal lengths, and set their apertures to f/2. Take a picture of an object 10 meters away with the 100mm lens, and one of an object 5 meters away with the 50mm lens. We'll take the wavelength of light to be 500nm as a constant.
We do this because for this experiment to be valid, the size of the image on the negative/sensor must be the same for both lenses. Since a 100mm lens gives 2x the magnification of a 50mm lens, we stand back twice the distance with the 100mm lens.
For the 50mm lens, D, the aperture diameter is 50/2 = 25mm.
This gives us:
θ = 1.22(500x10⁻⁹/25x10⁻³)
so...
θ = 2.44x10⁻⁵ rad
For the 100mm lens, D, the aperture diameter is 100/2 = 50mm.
This gives us:
θ = 1.22(500x10⁻⁹/50x10⁻³)
so...
θ = 1.22x10⁻⁵ rad
So, what does that tell us? It tells us that the Airy disk of the 100mm lens is half the angular size of that of the 50mm lens. This is critical to understanding why focal length doesn't matter in terms of diffraction. Longer lenses produce a narrower 'beam' of diffracted light - remember, the number above is in 'rad', so it's an angle, not a distance.
Even though the diffracted light has to travel further to reach the film/sensor in the 100mm lens than the 50mm lens, it's spreading out much more slowly. By the time it reaches the film/sensor it will produce a spot of light exactly as large as the 50mm lens does.
Now to prove this concretely, we want to measure what that means in terms of resolution. To do this, we want to measure the minimum distance between two points on an object that can be resolved with both lenses at the distances and aperture given.
From basic optics:
y/s = y'/s'
Where y is the separation of the points on the object, y' is the separation of the corresponding points on the film/sensor, s is the distance from lens to object and s' is the distance from lens to film/sensor.
Thus the angular separations of the object points and the corresponding image points are both equal to θ, so we get:
y/s = θ
and
y'/s' = θ
Because s is greater than the focal length of the lens, the image distance s' is approximately equal to the focal length.
For the 50mm lens, s = 5m, s' = 50mm and θ = 2.44x10⁻⁵ rad:
y/5m = 2.44x10⁻⁵
so...
y = 1.22x10⁻⁴m = 0.122mm
and
y'/50mm = 2.44x10⁻⁵
so...
y' = 1.22x10⁻³mm = 0.00122mm
For the 100mm lens, s = 10m, s' = 100mm and θ = 1.22x10⁻⁵ rad:
y/10m = 1.22x10⁻⁵
so...
y = 1.22x10⁻⁴m = 0.122mm
and
y'/100mm = 1.22x10⁻⁵
so...
y' = 1.2x10⁻³mm = 0.00122mm
Phew! So, where does that leave us? Both lenses can resolve a minimum distance of .122mm on an object, which translates to a minimum resolvable dot on the film/sensor of .00122mm at those distances and an aperture of f/2. If we change the aperture, we will get different numbers, but the important thing is that they are the same for both lenses - so we've just proved that for the same image size and the same aperture, but different focal lengths, resolution is identical.
Again, the reason for this is that the longer lens produces a narrower cone of diffraction, so even though the light has to travel further to get to the film/sensor, it ends up producing the same size spot of light in both cases.
Thus, focal length does not contribute to resolution loss from diffraction.
Cheers,
Peter
[a href=\"index.php?act=findpost&pid=56198\"][{POST_SNAPBACK}][/a]
Peter: An interesting post.
You say that: "so we've just proved that for the same image size and the same aperture, but different focal lengths, resolution is identical."
I understand what you are saying, and I agree with the point, but I'm not happy with the use of the term resolution, as it is very confusing. In optics the resolution of a lens is a well defined quanity, and it does depend on the focal length as the equations you posted show. I think that what you mean is the sharpness of the image on the sensor (whether that be film, or a solid state device), though I'm sure there is a more precise phrase than "sharpness on the sensor".
Leif
-
You should read up a bit on how they're having problems with photolitography because of the wavelengths of the extreme ultraviolet they're using today (the mentioned 65 nm structures; 0.065 µm = 65 nm) ...
That they're having problems doesn't mean that they're confident about solving them; Intel have just recently started mass production of 65 nm technology, and several actors in the market are working on 45 nm prototypes.
But the usefulness for visible light, which is one order of magnitude greater in wavelength, is ... limited.
I'm sure Bjørn Rørslett would be delighted, should he ever get the lenses and sensor filters that would transmit UV light at these wavelengths, but your hopes are not based in reality.
[a href=\"index.php?act=findpost&pid=57297\"][{POST_SNAPBACK}][/a]
You are right that according to conventional optics that we are approaching the limit of what can be produced on a chip.
Interestingly there is a new field of optics called photonics, whereby light can image detail at scales less than the wavelength of the light, a seemingly nonsensical proposition. I seem to recall that this is based on using materials with negative dielectic constants, and until recently they were only a theoretical curiosity. But I seem to recall that someone, probably in America, has created such a device. Whether this ever escapes from the labs, and impacts on photography remains to be seen. (Well, I'm sure it will. But whether or not we will be alive to see it is another question.)
Also they are using shorter wavelengths of 'light'. Maybe they will end up using x-rays. But here the problem is that the 'wires' on the chips become so small that weird quantum effects (such as quantum tunnelling whereby current leaks between adjacent 'wires' when classical physics says it shouldn't) start to dominate, along with excessive heating and other problems.
Leif
-
Peter: An interesting post.
You say that: "so we've just proved that for the same image size and the same aperture, but different focal lengths, resolution is identical."
I understand what you are saying, and I agree with the point, but I'm not happy with the use of the term resolution, as it is very confusing. In optics the resolution of a lens is a well defined quanity, and it does depend on the focal length as the equations you posted show. I think that what you mean is the sharpness of the image on the sensor (whether that be film, or a solid state device), though I'm sure there is a more precise phrase than "sharpness on the sensor".
Leif
[a href=\"index.php?act=findpost&pid=57560\"][{POST_SNAPBACK}][/a]
Leif,
My understanding is the resolution per unit area (or unit length) at the focal plane will not change with a change of focal length provided the f stop remains the same. But this is partly due to the fact that f stop is dependent upon both focal length and physical aperture diameter. Another way of putting it would be to say, diffraction limited resolution will not change with focal length provided the physical aperture diameter changes in proportion to the change in focal length, ie. provided the f/stop remains the same.
But line pairs per mm are not all we are concerned about when making pictures. A camera like the 20D is capable of delivering substantially greater resolution in terms of lp/mm than the 5D, but the 5D can deliver more lines per picture height.
This is the distinction which Peter's proof does not address.
-
Leif and others,
it might be nice to use the optical jargon which is something like angular resolution of the subject. Roughly this is the smallest angle between different parts of the subject, as measured from the camera, that the camera (lens plus sensor!) can resolve. For me, this gets to the heart of what I usually care about; how much detail we get of a particular subject. Sensor/film/lens resolution in lp/mm are only means to that end.
For example, I believe that matching the angular resolution of the human eye requires aperture diameters no smaller than about 3-4mm due to diffraction which is about f/16 with a 50mm lens.
Angular resolution is about the same as "line per picture height" or "lp/mm on pictures of the same size", once the pictures we use to compare cover the same angular field of view.
Both diffraction and out of focus effects have the same effect on angular resolution of a subject when the effective aperture diameter (focal length divided by f-stop) is the same, regardless of the focal length and format.
-
Leif and others,
it might be nice to use the optical jargon which is something like angular resolution of the subject. Roughly this is the smallest angle between different parts of the subject, as measured from the camera, that the camera (lens plus sensor!) can resolve.
Both diffraction and out of focus effects have the same effect on angular resolution of a subject when the effective aperture diameter (focal length divided by f-stop) is the same, regardless of the focal length and format.
[{POST_SNAPBACK}][/a] (http://index.php?act=findpost&pid=57586\")
According to Widipedia, angular resolution (θ) is given by this formula:
sin (θ) = 1.22 λ / D, where λ is the wavelength of the light and D the diameter if the lens. The f/stop is not involved here and it is the diameter of the lens that matters. This is why astronomers rate a telescope by the diameter of the lens or mirror, not the focal length or f/number.
Spatial resolution ( Δl) is given by the formula Δl = 1.22 f λ/D, where f = the f focal length of the lens and lambda and D are as above.
Since f/stop = D/f, This simplifies to Δl = λ * f/stop. In this case, the diameter of the lens is not involved.
[a href=\"http://en.wikipedia.org/wiki/Angular_resolution]http://en.wikipedia.org/wiki/Angular_resolution[/url]
I hope this helps
Bill Janes