@msmoto: I get it now...thx for clarification :D

That art museum shot is a 'wow' msmoto. Nicely seen.

And, the reason I did that was because I like this look...

I remember the first fisheye...I think an 8mm f? about 3 inches across, and fairly expensive at that time... It produced a full round image on 35mm film as I remember. The effect of a "fisheye" is what I wanted. If I want a lens that covers about 150° horizontally, I can shoot the 10.5mm in DX mode.

I am not certain how this is about diffraction and teleconverters..... oh well...

And, this...not possible with a hood on the lens..

Just found the macro with this lens

http://www.flickr.com/photos/fantinesfotos/7050923197/sizes/o/in/photostream/

@spraynpray I realize that buddy. What I believe mismoto did is she physically cut out the outer preeminently installed hood in order to prevent it from showing up with such a wide angle lens. I personally, would have just switched the setting within the camera to work with DX. The pictures below is what I'm talking about:

D4 10.5 DX Fisheye 2.8 in FX mode

D4 10.5 DX Fisheye 2.8 in DX mode

Or is it something different that you did or I'm miss understanding msmoto?

Spraynpray, that one's not a hard calculation. To make the 8"x10" print of the scene taken with DX, you have to enlarge all of the fuzziness by a factor of 1.5 relative to FX print. This works out to FX having about a one-stop advantage over DX.

To illustrate by taking it even further, the diffraction fuzziness would not even need to be enlarged if you were shooting large format 8"x10" negatives to make the 8"x10" print. So the smaller the sensor format, the more important diffraction becomes.

golf007sd said:
I always wondered why you did that msmoto...why not just switch to DX mode on the D4 and keep the lens as is?

10.5 on FX is 15.75 on DX.

msmoto said:
Almost all the lenses I have are FX and I shoot these on a D90 as well as FX. And, the 10.5mm f/2.8 fisheye has been "converted" to FX by cutting off the built in lens hood.....very very carefully I might add.

I always wondered why you did that msmoto...why not just switch to DX mode on the D4 and keep the lens as is?

Almost all the lenses I have are FX and I shoot these on a D90 as well as FX. And, the 10.5mm f/2.8 fisheye has been "converted" to FX by cutting off the built in lens hood.....very very carefully I might add.

I need to know more about the DX/FX thing when you've recovered Jonny. Also FX lenses on DX bodies.....

msmoto said:
@ jonnyapple... I seem to have seen this post before somewhere...LOL

Copy and paste...I just thought it would be hard to find in the other thread.

To add to confusion...diffraction results in a loss of contrast across a sharp line...which if enough, looks like a blurred image. Loss of resolution, as in using a teleconverter on a not so great telephoto lens...the actual line definition is blurred. While the end result may appear the same, I believe the diffraction can be helped by reduction of internal reflections, while the actual loss of definition, do to the resolving power of the glass, is in the optical properties of the glass. I think this is why MTF charts have two curves...

Please understand this is not an expert opinion, just my thoughts to get any more ideas going. And to keep TTJ thinking it is in the water...:-)

I'm confused....

ahhhhhhh. And here I thought it was something in the water...

@ jonnyapple... I seem to have seen this post before somewhere...LOL

Well said, jonny. Its nice having an expert around. :]

I posted this in a D600 thread, but wanted to put it somewhere that people searching diffraction could find it. I'm not going to link to the f stoppers diffraction video.

I usually like what f stoppers do, but this video is misleading so I want to set the record straight. I admit I was a bit put off by him saying DEfraction at the beginning, but almost everything he says comes across like a third hand retelling of a bad read of a wikipedia page.

Diffraction is nothing more than the effect of waves spreading out after they hit an obstacle. Instead of the object's shadow being sharp, you always (even at f/2.8 or f/1.0) have diffraction because your aperture (or the edges of your lens in the extreme case of having no stop) blocks all but a small part of the light. The best you can do in optics is to have diffraction limited imaging.

Wide open, lens sharpness is not diffraction limited. Individual points of light in the object are imaged by the lens not as points but as small circles or comet-shaped objects. This is caused by the lens not being the perfect shape to redirect all light coming from each point of the scene to the corresponding point on the sensor. This fuzziness is caused by the limits of our ability to refract the light perfectly (or reflect it perfectly in the case of a reflecting lens or telescope). Cut the designers some slack, though. Think of how many different colors of light coming from so many different angles they are asked to redirect onto a single spot of the sensor. Getting to where lenses are today has been an engineering tour de force.

The further you get from the center of the lens, the worse these aberrations normally are. Closing down the aperture eliminates some of the light collected by the edges of the lens, which minimizes the aberrations but also eliminates useful light. You might think you can just keep making the aperture smaller, but then this is where diffraction rears its ugly head. While it's true that for an impossibly tiny hole you would have the entire world in focus and (get this!) your lens wouldn't even need to exist, the image formed would get fuzzy due to diffraction before you got there.

I think people often forget that the first cameras had no lenses. People were making images using just a hole in a dark room (a camera oscura for the illuminati out there). To form a meaningful image, you block all the stray light and only let it come in from one direction, like so:

It was only when people understood refraction and realized that there is all of that wasted light that doesn't make it into the camera that could be collected and redirected that lenses entered the picture:

Ironically, even with a lens, the smaller you make the aperture that blocks the light, the bigger the effect of diffraction! The smaller the hole, the more spread out the diffracted light gets. It's not, as this video implies, that the light arrives perpendicular to your sensor when shooting wide open. In fact, the wider the aperture, the greater the range of angles with which the incoming light hits your sensor. Just have a look at the dragonfly picture above if you don't believe me. The problem is what I said to begin this paragraph: smaller hole = bigger spread (more fuzziness) after light passes through the aperture.

There is a sweet spot where you have minimized the aberrations from the refractive imperfections of your lens while keeping diffraction under control. But what does "under control" mean? This is where TTJ and I are taking issue with the way people are analyzing diffraction as sensor tech progresses. You absolutely have to consider the format the image will be delivered to know what an acceptable loss of sharpness is. For 800-pixel screen-ready images, I don't think you can see diffraction on even DX images unless your aperture is well below f/50 (maybe tomorrow I'll be in the mood to calculate this). This is one thing that the video should have mentioned. No one should be judging diffraction effects at 100% magnification unless that's about the size you'll be printing. I can't imagine more than a handful of us around here are. I'm certainly not.

The bottom line: For a given sensor size and a given print size, sensor resolution has absolutely no bearing whatsoever on diffraction limitations of our lenses! Stated in a more practical way, a D700 and a D800 will show diffraction limitations at the same aperture for an 8"x10" print of the same scene. Now, a D7000 and a D800 would be a different story; this is one of FX's advantages that isn't discussed quite as much, but I don't think it's as important as people want to make it [he said after spending 40 minutes on a post about it].

Sorry for the long post. I couldn't go to sleep with a clear conscience after watching that video if I didn't say something to try to set the record straight. Of course, TTJ summed it up with this:

TaoTeJared said:
This will be my standard line on Diffraction this year - Pixels & image pixels are not 1:1.

jonnyapple said:
Ouch, EP! My family used to make fun of me because I'd hold toast about 6 inches in front of my face to butter it. -10 means you can only see 10cm in front of your eyes! Again, ouch.

Yeah. :/ Part of why I love photography is my camera resolves better than my own eyes! Even with good glasses on my face, my camera sees way better with good glass in front of it. :) Also I derped... I have prescription sunglasses so I can totally try the teleconverting thing with one of the pairs. I'll get back with results when I can science some things up!

jonnyapple said:
The fictitious aperture is actually real, Paperman—it's the front lens element. It's confusing because while that one's not adjustable, there is the separate adjustable stop (a.k.a. aperture) that is inside the lens that has all of the rounded blades to make it look more like the "real" aperture that's defined by the edges of the lens.

If you measure the front lens element of a f=400mm f/2.8 you'll find it's 143mm (400mm/2.8). You'd better use a ruler that isn't going to scratch the lens, though! The Nikkor takes 52mm filters, but not on the front of the lens.

I was always under the impression the measurement had to be taken from where the aperture blades were - not unreasonable when you think it is the place that opens/closes and limits the size of light passing through . The explanation I found after your reply was that the light rays were bent inwards once past the front element , eliminating the need to use an aperture mechanism the same size as the front element.

jonnyapple said:
Okay, for anyone who was actually curious, I tried the simple retrofocus experiment with a 50 1.8 and it worked fairly well. Eyeballing it, I'd say a bookshelf that took up about 60% of the frame was magnified to about 70% of the frame with my glasses between the camera and the lens—a ~1.2x teleconverter.

I had to really mash the lenses up against the flange on the camera. A -10 diopter lens (f=-100mm) would probably be better because it would spread the image out even further. My first try, I took off my glasses and realized I'd never know when it's in focus. I found an old pair and that solved it.

edit: Well, everyone here knows I'm a nerd, so I thought I'd add the math. The thin lens equation for object distance = -0.05m (virtual object) and f=-0.2m gives an image distance of 0.066m—a real image 1.5 cm behind where it was without the diverging lens. That works out to a 1.3x magnification. El_Pickerel, your -10 diopter lens should give 2x magnification in the same situation, but you'd have more stray light because you'd probably need to hold the lens about 3cm (1 inch) off the flange.

You crack me up JP...could you please video tape it for us so we can all sit back and watch it...LoL :P

Now as for the question at hand, I must say, it sure made me stop and go "hmmm." And after ready all your guys responses, I then asked myself: "what happened with the whole argument of F-stop and "diffraction of light" when one use a filter...sucb as a CPL or ND?"

Okay, for anyone who was actually curious, I tried the simple retrofocus experiment with a 50 1.8 and it worked fairly well. Eyeballing it, I'd say a bookshelf that took up about 60% of the frame was magnified to about 70% of the frame with my glasses between the camera and the lens—a ~1.2x teleconverter.

I had to really mash the lenses up against the flange on the camera. A -10 diopter lens (f=-100mm) would probably be better because it would spread the image out even further. My first try, I took off my glasses and realized I'd never know when it's in focus. I found an old pair and that solved it.

edit: Well, everyone here knows I'm a nerd, so I thought I'd add the math. The thin lens equation for object distance = -0.05m (virtual object) and f=-0.2m gives an image distance of 0.066m—a real image 1.5 cm behind where it was without the diverging lens. That works out to a 1.3x magnification. El_Pickerel, your -10 diopter lens should give 2x magnification in the same situation, but you'd have more stray light because you'd probably need to hold the lens about 3cm (1 inch) off the flange.

Ouch, EP! My family used to make fun of me because I'd hold toast about 6 inches in front of my face to butter it. -10 means you can only see 10cm in front of your eyes! Again, ouch.

The fictitious aperture is actually real, Paperman—it's the front lens element. It's confusing because while that one's not adjustable, there is the separate adjustable stop (a.k.a. aperture) that is inside the lens that has all of the rounded blades to make it look more like the "real" aperture that's defined by the edges of the lens.

If you measure the front lens element of a f=400mm f/2.8 you'll find it's 143mm (400mm/2.8). You'd better use a ruler that isn't going to scratch the lens, though! The Nikkor takes 52mm filters, but not on the front of the lens.

Maybe a 200-400 F4 is optically more ideal than a 200mm F2 with a 2x teleconverter.

A teleconverter needs to work with many different lenses, but if you design a 200-400 F4 you don't need to work around a teleconverter when designing the thing. You'd design it without any compromises. I'm sure there are also handling and ergonomic issues with width if you make it that way too.

jonnyapple said:edit: Someone who's short-sighted should put their glasses between their lens and their camera and see if it has essentially the same effect. Actually, I'm probably a pretty good candidate for it—my prescription's power is -5.75 diopters.

Mine are -8/-10. Problem is if I did that I wouldn't be able to see the darn screen / viewfinder / controls / anything else I'd need to focus! :(

"The telephoto lens configuration combines positive and negative lens groups with the positive at the front, so as to reduce the back focal distance of the lens (the distance between the back of the lens and the image plane) to a figure shorter than the focal length. This is for practical, not optical reasons, because telephotos can then be made shorter and less cumbersome."

Well put into words . Have you seen a proper explanation anywhere about how they do the same with the aperture - why we don't see a 13 cm diameter aperture/blade diameter in a 400mm f2.8 ? I know it has got to do with additional elements creating a fictious aperture/focal length but I'm curious how it is done exactly .

You're right, Adamz. With a 2x TC, for example, you're throwing away 3/4 of the light and keeping just the central 1/4 of the image, which is where (most?) aberrations are smallest.

Good question about the 200-400 vs 200+2x TC. I don't know, but I'm pretty sure the TC is basically adding retrofocus to the 200:
(from http://en.wikipedia.org/wiki/Retrofocus )

The telephoto lens configuration combines positive and negative lens groups with the positive at the front, so as to reduce the back focal distance of the lens (the distance between the back of the lens and the image plane) to a figure shorter than the focal length. This is for practical, not optical reasons, because telephotos can then be made shorter and less cumbersome.

edit: Someone who's short-sighted should put their glasses between their lens and their camera and see if it has essentially the same effect. Actually, I'm probably a pretty good candidate for it—my prescription's power is -5.75 diopters.

adamz said:
@Tao - what didn't You like about 105vr+TC17? I have this combo and I'm satisfied with results.

Just not as "crisp" as I prefer. By no means bad and very usable. I also don't find needing the extra distance enough to warrant it.

It is a fun combo for sure!