Home · Register · Join Upload & Sell

Moderated by: Fred Miranda
Username  

  New fredmiranda.com Mobile Site
  New Feature: SMS Notification alert
  New Feature: Buy & Sell Watchlist
  

FM Forums | Fuji Forum | Join Upload & Sell

1       2       3              5      
6
       end
  

Can you get shallower Depth of Field on FF 35mm than 44 X 33 sensors?

  
 
gdanmitchell
Offline
• • • • • • •
Upload & Sell: Off
p.6 #1 · p.6 #1 · Can you get shallower Depth of Field on FF 35mm than 44 X 33 sensors?


alundeb wrote:
That is known stuff. Smaller microlenses giving lower fill factor and higher MTF. Reducing fill factor and thus sensitivity gives sharper images at all apertures, at the cost of lower sensitivity and then giving away some of the initial benefit of the larger sensor. This has nothing to do with diffraction and depth of field though.


Kind of what I was thinking...



Oct 31, 2018 at 02:18 PM
alundeb
Offline
• • • • •
Upload & Sell: Off
p.6 #2 · p.6 #2 · Can you get shallower Depth of Field on FF 35mm than 44 X 33 sensors?


gdanmitchell wrote:
Kind of what I was thinking...


We could compare the effect of smaller microlenses to the effect of a smaller airy disc. In the case of 70% area fill factor vs 100% fill factor over 5.3 um pixels, we get roughly that it corresponds to diffraction blur from aperture f/6.7 contra f/8. Hardly relevant at apertures f/16 to f/22 which have been used as examples to prove how the GFX is so-called less sensitive to diffraction.



Nov 01, 2018 at 06:17 AM
Steve Spencer
Offline
• • • • • •
Upload & Sell: On
p.6 #3 · p.6 #3 · Can you get shallower Depth of Field on FF 35mm than 44 X 33 sensors?


alundeb wrote:
We could compare the effect of smaller microlenses to the effect of a smaller airy disc. In the case of 70% area fill factor vs 100% fill factor over 5.3 um pixels, we get roughly that it corresponds to diffraction blur from aperture f/6.7 contra f/8. Hardly relevant at apertures f/16 to f/22 which have been used as examples to prove how the GFX is so-called less sensitive to diffraction.


But surely you can see how reducing fill factor with smaller micro lenses which gives sharper images at all apertures leads to people looking at images at f/16 and f/22 and seeing sharper images makes the impression that diffraction is less of an issue. At the small apertures like all apertures the smaller micro lenses give sharper images and it isn't unreasonable to think of that as a form of mitigation of diffraction. Comparing images with a sensor with the same microlenses there is still the same drop in sharpness from diffraction, but in comparing between two sensors one with smaller and one with larger microlenses the images from the sensor with smaller microlenses will be sharper even at small apertures like f/16 and f/22 and one could see this as mitigating the negative effects of distraction on sharpness in that comparison.



Nov 01, 2018 at 06:33 AM
alundeb
Offline
• • • • •
Upload & Sell: Off
p.6 #4 · p.6 #4 · Can you get shallower Depth of Field on FF 35mm than 44 X 33 sensors?


Steve Spencer wrote:
But surely you can see how reducing fill factor with smaller micro lenses which gives sharper images at all apertures leads to people looking at images at f/16 and f/22 and seeing sharper images makes the impression that diffraction is less of an issue. At the small apertures like all apertures the smaller micro lenses give sharper images and it isn't unreasonable to think of that as a form of mitigation of diffraction. Comparing images with a sensor with the same microlenses there is still the same drop in sharpness from diffraction, but in comparing between two sensors one
...Show more

At f/8, the difference of +/- 0.9 um blur circle corresponds to 1/2 stop. At f/16 the difference corresponds to 1/4 stop, and at f/22 to 1/8 stop. While it may be measureable and even visible at 200% magnification, it has no practical significance compared to sharpening.

In principle, any kind of resolution enhancement, like sharper lenses, faster shutter speeds, using a tripod etc, can then be called diffraction mitigation. It becomes a little absurd to claim that these techniques mitigate diffraction and enhance usable DOF, but I can follow the thought. The best method yet would still be higher resolution sensors. A gigapixel sensor would essentially remove all microlens and sampling blur, and let the diffraction alone be the limiting factor at all practical apertures. Then we can start talking about how to mitigate diffraction.

The low fill factor does explain why the GFX does not have lower high ISO noise than the miniature format camera A7rIII though, as it throws away 30% of the light.



Nov 01, 2018 at 06:50 AM
Steve Spencer
Offline
• • • • • •
Upload & Sell: On
p.6 #5 · p.6 #5 · Can you get shallower Depth of Field on FF 35mm than 44 X 33 sensors?


alundeb wrote:
At f/8, the difference of +/- 0.9 um blur circle corresponds to 1/2 stop. At f/16 the difference corresponds to 1/4 stop, and at f/22 to 1/8 stop. While it may be measureable and even visible at 200% magnification, it has no practical significance compared to sharpening.

In principle, any kind of resolution enhancement, like sharper lenses, faster shutter speeds, using a tripod etc, can then be called diffraction mitigation. It becomes a little absurd to claim that these techniques enhance usable DOF, but I can follow the thought. The best method yet would still be higher resolution sensors. A gigapixel
...Show more

Agree with all of this, but I would add there is still a small advantage with high ISO noise and that while the smaller micro lenses are part of the reason the advantage is smaller than would be expected because of the sensor size the micro lenses are only part of the story. The Sony 42mp FF35 sensor has the Aptina alternative analog to digital converter that decrease noise above ISO 640 which is an important part of the story as well. It will be interesting to see whether the new 100mp BSI sensor for 44 X 33 also has this Aptina analog to digital converter that kicks in at higher ISOs.



Nov 01, 2018 at 07:42 AM
 


Search in Used Dept. 

gdanmitchell
Offline
• • • • • • •
Upload & Sell: Off
p.6 #6 · p.6 #6 · Can you get shallower Depth of Field on FF 35mm than 44 X 33 sensors?


alundeb wrote:
We could compare the effect of smaller microlenses to the effect of a smaller airy disc. In the case of 70% area fill factor vs 100% fill factor over 5.3 um pixels, we get roughly that it corresponds to diffraction blur from aperture f/6.7 contra f/8. Hardly relevant at apertures f/16 to f/22 which have been used as examples to prove how the GFX is so-called less sensitive to diffraction.


I think that the "airy disc" issue is going to be a different critter here.

The airy disc is the same size no matter how large the photo sites are and how far apart those photo sites are situated. The way that digital sampling works, the average "spillage" to an adjacent photo site (as a result of optical diffraction) will be the same unless a) the disk is extremely small and b) the center of the disk lines up precisely with the center of the photo site(s).

If you are thinking ahead of that last point, you have already realized that it is an extraordinarily unlikely scenario in which the image elements whose diffraction concert us line up so precisely that the primary image element aligns exactly with a photo site or a row of photo sites. (In fact, I suppose one might propose that if there actually is space between the photo sites — and there is! — that such a sensor could entirely miss recording small elements that fall between them at optimally sharp apertures!)

(And even in the unlikely event that everything did align just exactly right... the difference would be extremely tiny. Angels-dancing-on-the-head-of-a-pin tiny. And don't forget that these supposed differences are purportedly taking place at dimensions below the half-sampling-rate scale. That poses, uh, some problems, if you get my drift. There are other issues as well, some of which are alluded to in posts before mine.)

Things are often more complex than folks imagine, especially when it comes to the strange and wonderful things that digital sampling does near the half sampling rate... ;-)

Dan

(I'll note that this is yet another of the areas in which a real-world comparative test of actual images wold be informative.)



Nov 01, 2018 at 10:24 AM
alundeb
Offline
• • • • •
Upload & Sell: Off
p.6 #7 · p.6 #7 · Can you get shallower Depth of Field on FF 35mm than 44 X 33 sensors?


gdanmitchell wrote:
The airy disc is the same size no matter how large the photo sites are and how far apart those photo sites are situated. The way that digital sampling works, the average "spillage" to an adjacent photo site (as a result of optical diffraction) will be the same unless a) the disk is extremely small and b) the center of the disk lines up precisely with the center of the photo site(s).



Not sure what you mean, this sounds wrong to me. The average spillage will depend on the analog filter before the sampling, and the analog filter in this domain is the microlens and the extension of the photo site itself.



Nov 01, 2018 at 11:11 AM
gdanmitchell
Offline
• • • • • • •
Upload & Sell: Off
p.6 #8 · p.6 #8 · Can you get shallower Depth of Field on FF 35mm than 44 X 33 sensors?


I don't have the math chops to try to explain it, but imagine three cases.

1. The airy disc is so small that it is contained within the diameter of any photo site regardless of how much space is between them.

2. The airy disc is of a size that is larger than that of the smaller photo sites (with the larger gaps between them) but not larger than the distance between any photo sites regardless of size.

3. The airy disk is larger than the distance between photo sites in both cases.

Assuming the oddball case that the image elements in question line up perfectly with the centers of photo sites or, even less likely, the centers of vertical or horizontal rows of photo sites...

In #1 there is no difference between the sensors affecting the rendering of optical image diffraction.

In #2 there is no difference in this regard either.

In #3 there is no difference.

Now, you might propose that since image elements do not align perfecting with individual photo sites or with vertical/horizontal rows of photo sites, that this "misalignment" (for lack of a better word) might produce a larger variation when the center of airy discs is off of the center of photo sites or even centered in the gap between them. But here you are obviously running into the half-sampling rate/Nyquist issues and any data differences are essentially aliasing — e.g. false data.

I suspect that the error people are making is in presuming that smaller photo sites with the same gap render a more accurate/sharper image... but forgetting that this case is in the domain beyond the half-sampling rate boundary.

And, of course, this could easily be settled by making some photographs. I quite certain that the theorized improvement will amount to nothing at all.

Dan



Nov 01, 2018 at 12:15 PM
alundeb
Offline
• • • • •
Upload & Sell: Off
p.6 #9 · p.6 #9 · Can you get shallower Depth of Field on FF 35mm than 44 X 33 sensors?


gdanmitchell wrote:
And, of course, this could easily be settled by making some photographs. I quite certain that the theorized improvement will amount to nothing at all.

Dan


Good idea, and it turns out I have a suitable experiment at hand.

Before that, I think false data is the key to where I didn't follow you, the area where I anticipated to see a difference, is where we have false data. And as we shall see, that occurs over a range a bit larger than we may expect, but still.The reason we don't recognize false data that often in normal but sharp photographs is because it is most easily understood that the data are false with certain target patterns.

Here we have two pixel sizes with a really significant difference, an order of magnitude more than the difference between 100% and 70% microlens coverage on a 5.3 um pixel.
Top row, 4.7 um pixels, at 300% magnification
Bottom row, 1.54 um pixels, at 100% magnification
The real magnification is thus the same. The pixel size at bottom is also close to the GFX

Left to right, aperture f/4 to f/22.

I think it is safe to say that at f/22, there is nothing to gain by reduced microlens size.










Nov 01, 2018 at 02:19 PM
Luis Cunha
Offline
• •
Upload & Sell: Off
p.6 #10 · p.6 #10 · Can you get shallower Depth of Field on FF 35mm than 44 X 33 sensors?


alundeb wrote:
This thread needs even some more shallow DOF images. This one would have been very difficult to take from another position, for various reasons. Sony A7r, Canon FD 85 mm f/1.2 ASPH SSC

https://farm4.staticflickr.com/3953/15526341758_6a6a6e47dc_b.jpgBWchallenge_01 by Anders Lundeby, on Flickr


If you allow me the delayed opinion ,-) in this image; what I see most interesting is the tree in the center out of focus. She is the protagonist and launches the possibility of mystery. Looks like it's coming our way. It's cinematic. The trees in focus are just the frame for the picture of the mystery tree. The option for monochrome was the right one because it eliminates distractions and synthesizes the scene. Congratulations to the author!

As we talk about the use of DOF, I take the opportunity to ask a question:
after all, is it still worth buying the GF63mm today for those who, like me, have the 3.5/50mm and 4/120mm, to have this possibility of playing with DOF as a narrative method, with a native lens?
I like her standard FoV.
I know we already have in 2022 the GF1.7/80mm (Extraordinary but with a little CA) and the GF110mm (excellent!); but they are much more expensive.
I've seen images here and elsewhere and despite being a somewhat "despised" lens, I really like what I see especially at f/2.8. Is it worth the GF 2.8/63mm in 2022? Thanks.



Jun 24, 2022 at 09:02 AM
1       2       3              5      
6
       end






FM Forums | Fuji Forum | Join Upload & Sell

1       2       3              5      
6
       end
    
 

You are not logged in. Login or Register

Username      Reset password