j4nu
Offline
• • • • •
Upload & Sell: Off
|
 p.96 #7 · which lens has the most 3D POP? | |
RustyBug wrote:
Reasonable question. 
I don't have a great resource to explain the correlation, so if you'll roll with me I'll give you my .02 on how it squares in my (divergent) mind.
In the "linear" z-axis, we are pretty much all aware of the how the optical projection gets "larger" as it passes the focal plane, and how that is relative to the CoC (Circle of Confusion ... also, this thread   ). The trigonometry of that is pretty straightforward and old news.
That said, it is a factor of "x" for a given lens. But, that is also contingent upon the subject distance to the film plane. As such, we all are pretty well versed in that when you are much closer to mfd, the rate of change to the blur is greater than if we are at a normal or infinity distance to the focal plane. Nothing new here. The rate of change is a byproduct of the optic, combined with the subject distance (and aperture, but we'll hold that static for now). This is a rate of change in the Z axis.
To your point that the rate of change in the MTF is across the X / Y axis (distance from center), it represents a rate of change in that direction. As rate of change is essentially an exponential, similar to acceleration, it is a vector quantity (rate of change in a direction). Now, we have rate of change in two different directions, but vector quantities can be applied (trig) to yield a combined (vector) rate of change.
Now, I can hear folks saying that's a bunch of hogwash / BS ... but, my point is that in order to understand the relationships between the MTF in the X / Y axis, as it pertains to the Z axis, I tend to think of the cumulative effects of DOF / CoC combined with the MTF. Yeah, it's a country mile from being the answer you're looking for, but while I can't quantify it empirically for you, I mentally emulate the trig, principally. CoC * MTF, as the (real world) angle changes in the x/y/z coordinate from center.
I'm certain that this is flawed in detail, but the combination of DOF / CoC and MTF are going to impact rate of change, even if the MTF only displays for the X/Y axis. I'd be curious to understand the exact formulaic calcs, but that escapes me, atm.
If someone has a better mathematical presentation, please share. I'd dig on the "correct" answer, too....Show more →
Thanks for providing your take on this.
However, as with the previous post on this topic, I have to say I don't agree.
I think it makes more sense to approach this from the other, simpler, side  .
Consider a part of the image, which is properly focused - let's say we have some detailed stuff there and the lens allows us to
resolve it at 50 lp/mm.
Now, we shift the focus and the same part of the image, which still has that detailed stuff in it, is so out of focus, it's "effective visible" detail drops to (let's say) 5 lp/mm. At this point it doesn't matter if the lens resolves 50 or 25 lp/mm at that part of the image, as there is no detail to resolve.
Now, for the other end of the argument, let's forget about the focus plane. We have a scene where everything is in focus and only lens MTF to affect our perception of the image. So, if I place the object that's supposed to pop in the center (where lenses are the sharpest most of the time), everything to its left and right will have to become a blur. I have never seen a photo like this, except for those taken with "toy" lenses (so again the extreme) .
Now, I do think there is something in MTF, that affects our perception of the image (not only the detail/resolution we can see when pixel peeing). It's astigmatism - when it's high enough, the unsharp parts of the image also get that "uneasy/nervous" look, which is unpleasant (to my eyes at least).
I also think that the blur resulting from DoF and the blur resulting from MTF drop is different. The former is natural and pleasant to look at (again, to me), the other I'd say is closer to bad eyesight (which I think has at least something to do with sagital / meridional unevenness, so again the point above).
So, yes - that's why I don't think MTF tells us much about depth / pop / dimensionality of an image.
|
...
! 
