bmwzimmer: Well, so much for the Mirrorless is smaller argument. Bigger/longer/heavier pro lenses plus the extra weight of batteries negates this advantage.
… Oups: I see that “you” in my “you were discussing” is an error.
First: initially, you were discussing weight of f/2.8 vs weight of f/5.6. As I said: it was making no sense.
Second: 70-200mm F/4 is FE (as is [going to be] 70-200mm F/2.8, btw). So you do you now say that F/5.6 beats F/4?
And since NOW there is no 6300 either, is not what you are doing a strawman argument?
In the last paragraph: “we’d like to try them in APS-C lenses”…
Did I read it correctly that: “No APS-C lenses were being designed, recently”? Or (hopefully) maybe APS-C is being done by another team?
No wonder that “Panasonic's 35-100 f2.8 is relatively small”: it produces the same images as FF 70–200 f/5.6. (It is the entry pupil which governs the image created by the lens.)
Rishi Sanyal: his is really weird - does anyone see what looks like double images in the out of focus areas with the 35L II, [here](http://www.dpreview.com/reviews/image-comparison/fullscreen?attr202_0=canon_ef_35_1p4_usm_ii&attr202_1=canon_35_1p4&attr202_2=canon_ef_35_1p4_usm_ii&attr202_3=sigma_35_1p4&attr203_0=1.4&attr203_1=1.4&attr203_2=1.4&attr203_3=1.4&normalization=full&widget=295&x=0.4338051061297219&y=43.38571449999995)? Not present in the original 35L or the Sigma 35 Art... Is this odd, or unexpected? Does anyone know what might cause this? TIA.
This is why I'd ultimately like to have a controlled scene with a lot of depth and detail at many focal planes to assess focus 'fall-off' and bokeh.
Let me repeat what I already said:
I brought this point because the graphs on p.23 support MY (conjectural) explanation: that the artifacts in question are due to negative OTF, so that the resulting “false details” lead to aliasing.
Sorry, I cannot understand whether you support or distrust what I said…
Just calculate OTF of some simple cases of how phase error is distributed across the entry pupil. You will see the CALCULATED OTF changing sign; you will see how this looks on the graph of MTF.
In practice, it is really hard to measure OTF; only |OTF| is easily measured. And, of course, without knowing something about f, one cannot decide whether f changes sign or not by JUST looking at the graph of |f|. But the examples I know look exactly like those graphs on p.23: they dive down to a small positive value, and have a smoothed-out corner.
About negative values:
The curves are graphs of |f| for a certain function f (sometimes called OTF). Where f changes sign, the graph of |f| dives to 0, and has a sharp corner.
If you plot the graph of |f| with a certain step, you would miss the point where the graph dives to EXACTLY 0, and the plotting program would smooth the corner. So what you see is a dive to a small value, with what looks like opposite slopes on the sides.
(Of course, the situation is made SLIGHTLY ;—] harder by the fact that honestly, OTF is complex-valued! In fact, this has effects similar to those of having a finite step.
On what you said about the first 20 pages:
Note how this MATCHES what I said. ;—)
Thanks for the Wikipedia reference — myself, I would self-confine to Fourier-transform graphs. So I did not realize I was missing something in the “spacial” graphs (as opposed to “frequencial” ;—).
However, I STILL suspect that working with this graphs IN THE CONTEXT IN QUESTION should be combined with a rock of salt:
First, they discuss “just spherical” aberration, without qualifying which flavor they have in mind. Second, they discuss f/10 lenses.
Both lead me to suspect that they consider only the smallest-order spherical aberration (when phase errors are ∼r⁴). With f/1.4, my guts’ feelings are that with a well-designed lens, the phase errors will be dominated by VERY high order (in particular, they would be badly approximated by polynomials!).
So I do not know: “the general picture” MIGHT still be similar to these illustrations of what happens with small-order aberrations — but it MIGHT also be very distinct!
@JACS, about CLN35_Bokeh_en.pdf:
[I’ve read only the first ½, and it is hard to believe how complicated one can make a simple topic… 80% of what is contained in the first 20 pages is in the illustration on page 15…]
Anyway: see the graph for Planar 2/50 on p.23. One can clearly see that MTF (or OTF) becomes negative — but only when defocusing in “negative” direction (do not know what this means in the subject’s space).
Similarly for some other bright lenses. So this pours water on my hypothesis!
(Negative MTF may be quite a frequent phenomenon, but to have it WITH HIGH ENOUGH MAGNITUDE one may need a pretty sharp lens. So these effects may be rarely-observed.)
elefteriadis alexandros: I like to split the cameras in two category's.1. Those who is the perfect tool for MY style of photography.2. Those who is perfect as a machine in THEIR category.And so because for me the depth control is very important, anything smaller than medium format is useless. And that's why i still use medium format film cameras as the digital medium format is expensive for me.I don't care if aps-c, 4/3 or what ever have 1 trillion asa, 20 stop dynamic range or 800 terapixel.Is just not the right tool because in the first place they don't give me the right control of one of the most important factor in photography.
@Neodp, e a:
Bokeh (and depth of field) depend on the entry pupil; the larger the entry pupil, the larger the bokeh, and the smaller the depth of field.
Now take into account that with MF lenses, the f-number is significantly smaller than with comparable field-of-view FF lenses. This means that IN PRACTICE, FF is superior to MF in these two regards. (Although IN THEORY, MF has higher POTENTIAL, this potential is not realized.)
“the perspective of subject is complete different weil you keep the same field of view”.
This does not make any sense. Angle of view (together with framing of the subject) determines the distance to the subject, therefore the perspective.
@J A C S
Thanks for your references to the images of diffraction+spherical aberration+defocus.
But what is so “exotic” about these images? It is 100%-plain ringing. Moreover, the ringing becomes less and less pronounced when we go further away from diffraction-bound systems.
And we are discussing f/1.4 image! It is about an order of magnitude from being diffration-bound.
By “depth control”, do you mean depth of focus?
Then I'm puzzled: MF lenses do not allow as shallow depth of focus as FF lenses…
expressivecanvas: I don't have a clue what Kodak is attempting to accomplish here, realistically speaking, but the whole idea seems like crapola.
One question I was left with at the end was about resolution... what will be the final resolution of the files Kodak delivers? Regardless, even if the digital files are in 4K or better (which I highly doubt) the total cost is absolutely absurd when compared to the rest of today's video options.
What a bunch of crapola though. Are Kodak's execs on crack and think it is 1980? (I guess that makes two questions I have...) My advice to Kodak... go ahead and make one for Spielberg and then try to develop something more marketable.... perhaps a digital Brownie?
Actually: mea culpa!
Above, I did the calculation of the noise for 6 stops below saturation (while I was CLAIMING that it is for 8 stops below saturation). So above, the actual exposure is 40 captured photons per pixel (and not 160, as I claimed!).
Which means that other numbers are ALSO 4x off. In particular, to reconstruct the performance of A7ii, one would need 33cm² of this film.
So this film used in IMAX camera can produce a similar resolution (normalized to a certain S/N ratio) to the contemporary digital sensor. (But it would still have less dynamic range.)
ilza: Hmm, I see that it is EXPLICITLY stated that the body is designed for underwater operation — but I see no trace of claiming that the LENS is corrected for underwater-operation aberrations.
The same lens cannot work ideally in air and in water. So it is either designed for air, or designed for water. (Separate distortion-corrections may help, but distortion is the lest annoying part of underwater-aberrations.)
So: did I miss it?
@Kharan: as I said, distortion is the lest of the worries.
@HAR: Nikonos lens (AFAIR) is optimized for underwater.
Hmm, I see that it is EXPLICITLY stated that the body is designed for underwater operation — but I see no trace of claiming that the LENS is corrected for underwater-operation aberrations.
thinking of this more, I may invent an explanation of figure-8-shaped “circle” of confusion AT THE BEST FOCUS POSITION (off-axis).
However, having such a circle of confusion WITH SHARP IN-FOCUS IMAGES — no, this is way weirder than anything my 2nd conjecture might have explained…
Another guess (which may be related to negative-MTF, may be not):
The “circle of confusion” in question has the shape of figure-8.
Cannot guess offhand which particular combination of standard aberrations would lead to this, but so far this looks very plausible.
What you write now does not make any more sense that what you wrote before — but at least now I suspect I understand what is the source of your confusion.
It looks like you think that tilting element has similar results to chopping an element in two halves, and moving one half forward, another one backwards. No, this has NO relationship to what actually happens.
Tilting the element has 2 effects: • It moves (and tilts) plane of focus; • It subtly changes aberrations of the element.Since elements are painstakingly designed so that their aberrations cancel each other as close as possible, even a slight tilt can increase the residual aberrations A LOT — but in a kind of predictable way. (If you read LensRentals blog, you have seen a lot of illustrations of this.)
@Rishi: I did not yet start “being pleasant”, and would GLADLY apologize (if needed) — the moment I understand what fmian is talking about.
Did you understand it? Can you paraphrase it?
All I can say is: read your words as directed to you.