Pages: [1] 2 3 4   Go Down

Author Topic: shallow DOF limits of current MF systems  (Read 32537 times)

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
shallow DOF limits of current MF systems
« on: October 25, 2006, 03:35:37 pm »

Why do the new generation of medium format lenses rarely go beyond f/2.8, and not beyond about f/3.5 at typical portrait focal lengths? In the past there were lenses as "bright" as f/1.8 (manual focus) from Mamiya and one f/2 AF lens for Contax 645.

Even allowing for the longer focal lengths usable to get the same FOV in a larger format, f/2.8 in 645 format gives shallowest DOF comparable to only about f/1.6-1.8 in 35mm format, and with 36x48mm format sensors the limit is like f/1.8-2.1, yet I read many people insisting on the importance of lower f-stops than that for 35mm format, down to at least f/1.4 and preferably f/1.2.

Are extremes of shallow DOF of less interest in MF than smaller formats like 35mm?
Is the importance of extremely shallow DOF in 35mm (and smaller) formats exaggerated?
Are recent medium formats systems failing to deliver what customers need when it comes to shallow DOF for portraits?
Is there some other explanation for the more limited "selective focus" capabilities of the current AF MF systems?
« Last Edit: October 25, 2006, 03:36:30 pm by BJL »
Logged

Graham Mitchell

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2281
shallow DOF limits of current MF systems
« Reply #1 on: October 25, 2006, 04:53:46 pm »

I would personally like to see faster MF lenses - f2 would be great (for available light shooting as much as short DOF) and fortunately there are a few available. f2.8 is quite decent too. There are probably more fast lenses than you think. Here is a list I made which is probably missing one or two.

Rollei PQ/S: 50mm f2.8, 80mm f2, 80mm f2.8, 110 f2, 180mm f2.8

Contax 645: 80mm f2, 45mm f2.8, 140mm f2.8

Bronica ETR (645) has a 50mm f2.8, 60mm f2.8, 75mm f2.8.

Bronica SQ series (6x6): 80mm f2.8

Hasselblad V: f/2.8 80 mm 'C' lens. The rest of the C series is slower. The F lenses were faster: f/2 110 mm, f/2.8 50 mm, f/2.8 80 mm and f/2.8 150 mm

Hasselblad H: HC f/2.8 80 mm and HC f/2.2 100 mm

Mamiya 645: 80mm f1.9, 80mm f2.8, 45mm f/2.8, 55mm f/2.8, 150mm f2.8, 200mm f2.8, 300mm f2.8

Mamiya 67: 110mm f2.8

Pentax 67: 75 mm f2.8, 90mm f2.8, 105 mm f2.4, 165mm f2.8

Pentax 645: 45mm f2.8, 75mm f2.8, 150mm f2.8
« Last Edit: October 25, 2006, 05:35:57 pm by foto-z »
Logged

Sheldon N

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 828
shallow DOF limits of current MF systems
« Reply #2 on: October 25, 2006, 05:33:59 pm »

One thing I have noticed as an advantage to the larger formats when it comes to depth of field is the issue of separation of the subject from the background.

This is not about the calculated amount of depth of field or the amount of background blur involved. Rather, it has to do with the amount of detail available at the plane of sharp focus vs. the defocused areas. The idea behind this is that once an area is defocused, added resolution from a larger film format doesn't really matter, except perhaps in smooth tonalities. However, at the plane of sharp focus, the added resolution from larger formats does make a difference.

For example - I have a Canon 35mm f/1.4 L that I use on an EOS 3 and I also have a Rodenstock 150mm APO Sironar S f/5.6 that I use on 4x5. If you crop the 35mm shot to a 4x5 aspect ratio, the 35mm and 150mm lenses have very similar fields of view. Also, when shot wide open they have roughly the same calculated depth of field and background blur. Why then does the larger format give a greater percieved separation of the subject from the background?

Here's my hypothesis. Let's say both lenses are capable of resolving 40 lp/mm at the plane of focus when shot wide open (a pretty conservative assumption). If you need a final print resolution of 6 lp/mm, then you can make a ~7x enlargement from the original film and still retain critical sharpness. This means at print sizes under 8x10, then the plane of sharpness for both 35mm and 4x5 will be roughly similar (though I notice a difference between 4x5 and 35mm at 8x10 size). Once you enlarge to 8x10 or larger, then the 4x5 shot has a detail advantage at the plane of focus, not really hitting the limit until over a 30 inch print. If you take 35mm up to these larger sizes, the plane of focus just doesn't hold up.

Therefore, for any print larger than 8x10, the larger film size produces greater resolution at the plane of focus, while the defocused areas remain blurred regardless of format. This creates an increase in the perceived separation between the sharpness of the subject and the blurred background.

The larger print sizes also change the circle of confusion to a more stringent value and create an effectively shallower depth of field. So, while the depth of field calculations for a smaller prints may show the advantage going to 35mm for shallow depth of field with its f/1.0 and f/1.2 lenses, the larger formats actually contain much greater potential for shallow depth of field when you go to larger print sizes.

I think I'm on track with this line of thinking, but input is welcome.
Logged
Sheldon Nalos
[url=http://www.flickr.com

Gary Ferguson

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 550
    • http://
shallow DOF limits of current MF systems
« Reply #3 on: October 25, 2006, 07:29:01 pm »

BJL, you're absolutely right that depth of field tables only give a partial explanation of the full realities of focusing. You'll get some useful insights into your question here,

http://www.bobatkins.com/photography/technical/mtf/mtf6.html

And I'd also recommend visiting Harold Merklinger's website and buying his pamphlet on the subject.

Bob Atkins illustrates that two different lenses with the same calculated depth of field can have very different visual characteristics. Harold Merklinger takes the issue further and argues that for some subjects a radically different approach to focusing can yield significantly better results.

Other authorities have made additional attacks on the orthodoxy of depth of field tables, for example Kornelius Fleischer, Zeiss's marketing manager, has pointed out that depth of field tables only apply to the very centre of the image, off axis there are different rules that apply. And even more startling is his assertion that depth of field tables are merely theoretical approximations and different lenses of the same focal length can have different depth of field characteristics.  He says for example that the Hasselblad 250mm superachromat has a shallower working depth of field at the same aperture than the regular Hasselblad 250mm lens. I also suspect that many other modern and complex lenses actually have less effective depth of field than some of their older, simpler counterparts.
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
shallow DOF limits of current MF systems
« Reply #4 on: October 26, 2006, 12:44:31 am »

As lenses get better, and particularly if metamaterials with a negative refractive index currently being produced in the lab with nanotechnology, ever become a practicable option for construction of camera lenses, then standard DoF calculators should actually be more correct, with an appropriate adjustment of CoC, since current DoF tables address degrees of OoF (out-of-focusness) at various distances from a point of perfect focus with an ideal, perfect lens.

If the CoC is chosen in relation to what the eye can see (your eye for example) on a given size print from a specified distance, it makes no difference to the perception of DoF on the print if the lens is capable of resolving detail smaller than the chosen CoC. In such a case, that additional resolution of the lens is wasted (on the smaller than need be print) and the DoF caculations should still hold.

As lenses and sensors become higher resolving, the perception of DoF should only change on appropriately larger prints, which will appear to have a shallower DoF because what's in focus will be sharper, but what's out-of-focus should be similarly fuzzy.
Logged

Steve Kerman

  • Full Member
  • ***
  • Offline Offline
  • Posts: 141
shallow DOF limits of current MF systems
« Reply #5 on: October 26, 2006, 02:32:55 am »

Quote
... and particularly if metamaterials with a negative refractive index currently being produced in the lab with nanotechnology
[a href=\"index.php?act=findpost&pid=82308\"][{POST_SNAPBACK}][/a]
Negative refractive index???!

How does (speed of light in a vacuum)/(speed of light in the material) ever become negative?  Or even become less than unity?
Logged

Morgan_Moore

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2356
    • sammorganmoore.com
shallow DOF limits of current MF systems
« Reply #6 on: October 26, 2006, 03:43:53 am »

More simply put..

A larger chip has less depth of field for a given angle of view and aperture

For some this is an advantage for others it is not

This scenario is know as the MF look - loved by many - hence the desire for larger chips than more MP

--------------------------------

It is (exponentially) difficult,expensive and heavy to construct very wide aperture lenses with suitable image circle to cover a 645 chip

Given the minimal DOF rendered by current lenses wide open some would consider them pointles anyway

Others would relish the bright view rendered through the viewfinder even when not shooting at that aperture

Read again..

A larger chip has less depth of field for a given angle of view and aperture

SMM
Logged
Sam Morgan Moore Bristol UK

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
shallow DOF limits of current MF systems
« Reply #7 on: October 26, 2006, 11:38:04 am »

Quote
Negative refractive index???!

How does (speed of light in a vacuum)/(speed of light in the material) ever become negative?  Or even become less than unity?
[a href=\"index.php?act=findpost&pid=82317\"][{POST_SNAPBACK}][/a]

Don't ask me. I'm not a physicist. Just repeating what I've read   . Apparently there's work going on to produce transparent artificial materials with different properties to glass, which enable the laws of diffraction to be extended so that a lens at say f64 will have the resolution of a lens at say f8.
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
shallow DOF limits of current MF systems
« Reply #8 on: October 26, 2006, 12:11:26 pm »

Quote
Other authorities have made additional attacks on the orthodoxy of depth of field tables, for example Kornelius Fleischer, Zeiss's marketing manager, has pointed out that depth of field tables only apply to the very centre of the image, off axis there are different rules that apply. And even more startling is his assertion that depth of field tables are merely theoretical approximations and different lenses of the same focal length can have different depth of field characteristics.  He says for example that the Hasselblad 250mm superachromat has a shallower working depth of field at the same aperture than the regular Hasselblad 250mm lens. I also suspect that many other modern and complex lenses actually have less effective depth of field than some of their older, simpler counterparts.
[a href=\"index.php?act=findpost&pid=82283\"][{POST_SNAPBACK}][/a]

Gary,
I recently picked up a Canon 50/1.4 at a good price in Bangkok. I'd been meaning to buy one for some time due to the embarrassment that my sharpest lens was by far the cheapest, the 50/1.8 ll. But the fact is, although many photographers seem to drool over the ultra wide apertures of such lenses, including the f1.2 of the 85/1.2, such lenses have pretty lousy performance at full aperture, including the very expensive 85/1.2.

Checking the Phodo MTF charts for these lenses, the 50/1.4 at f8 is quite outstanding with an MTF response of around 70% at 40 lp/mm all the way to the edges (18mm). However, at f1.4, even in the centre, MTF is only 35%, falling to around 15% at the edges. The 85/1.2 is hardly better with an MTF of arounf 38% at the centre, falling to around 20% at the edges.

These are very poor performance figures for any lens, so I decided to take a few test shots from my hotel room balcony using my new 50/1.4 at all apertures. With lens focussed at infinity at f1.4, a tiled roof in the foreground (about 25m away)appeared to be noticeably OoF. However, at f1.4 the performance of the lens is so bad (worse than at f22), it's difficult to tell if the roof is more out-of-focus than the rest of the image, so one would have to conclude that DoF is just fine, as it is with a pin-hole camera.

The attached 100% crops below tell the story.

[attachment=1115:attachment]               [attachment=1116:attachment]
Logged

Sheldon N

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 828
shallow DOF limits of current MF systems
« Reply #9 on: October 26, 2006, 01:13:55 pm »

Quote
A larger chip has less depth of field for a given angle of view and aperture

For some this is an advantage for others it is not


Absolutely agreed.


Quote
This scenario is know as the MF look - loved by many - hence the desire for larger chips than more MP

It is (exponentially) difficult,expensive and heavy to construct very wide aperture lenses with suitable image circle to cover a 645 chip

I think that the idea that the MF look is exclusively because of shallow depth of field is the reason BJL asked his question.

The question stems from the fact that in MF there are no equivalents to lenses such as the 35mm f/1.4, 50mm f/1.2 or 85mm f/1.2 that are available to 35mm shooters. Running calculations or using a depth of field calculator shows that there is actually less depth of field with a 50mm f/1.2 lens shot wide open on 35mm than there is with a 80mm f/2 lens shot wide open on 645. The same is true if you compare the 85mm f/1.2 L with the 110mm f/2.

  - BUT -

My thinking is that this is not the whole story. Using a larger imaging chip or film format does a couple things. It places less lp/mm demand on the MF lens for any given print size, likely gets the lens to a range where the MTF is better because of those lower demands, and allows for greater resolution at the plane of focus.

I believe that when these things are combined, you get greater perceived separation from the subject to background (contrast of sharp vs blur) when at smaller and medium print sizes, and also gain the ability to print larger with the larger format, which will actually result in ultimately shallower depth of field when the small format hits the limits of enlargement.

  - So -

My thinking is that even if 35mm can produce shallower depth of field when using the fastest lenses, MF holds the upper hand due to:

1) Better separation from subject to background (sharp vs. blur).

2) The POTENTIAL for shallower depth of field, when print sizes larger than the limits of 35mm are used.

Of course these aren't the only advantages of MF over 35mm, but I think they are pretty important in terms of the perception of depth of field.
Logged
Sheldon Nalos
[url=http://www.flickr.com

Morgan_Moore

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2356
    • sammorganmoore.com
shallow DOF limits of current MF systems
« Reply #10 on: October 26, 2006, 01:47:26 pm »

Quote
My thinking is that this is not the whole story
[a href=\"index.php?act=findpost&pid=82386\"][{POST_SNAPBACK}][/a]

Now I dont read charts  - I leave that to captains of ships

But I think it is the whole story.

My proof of this is if you stick a 645 lens on a 35mm body you get the same image - just less of it

--------------

I read glossy fashion mags A4 size all the time and the amount of MF/LF work is obvious even at that size due to the lack of DOF

At smaller apertures the less DOF of MF becomes more apparent because the background is still more fuzzy - many want that blown background but still a reasonable DOF

I was shooting food today and scratching my head on whether to control my DOF through a wide aperture or a longer lense - I have not made up my mind

A thought is that Items; people food whatever look most natural when viewed from a distance that one is used to viewing them from, once you have decided the crop you want a lense lengtth is therefore defined and aperture selected on required DOF

SMM
« Last Edit: October 26, 2006, 01:56:57 pm by Morgan_Moore »
Logged
Sam Morgan Moore Bristol UK

howiesmith

  • Guest
shallow DOF limits of current MF systems
« Reply #11 on: October 26, 2006, 03:14:47 pm »

Quote
Why do the new generation of medium format lenses rarely go beyond f/2.8, and not beyond about f/3.5 at typical portrait focal lengths?
[a href=\"index.php?act=findpost&pid=82243\"][{POST_SNAPBACK}][/a]

I think the answer is still size - which equals cost and weight.
Logged

Jonathan Ratzlaff

  • Full Member
  • ***
  • Offline Offline
  • Posts: 203
shallow DOF limits of current MF systems
« Reply #12 on: October 26, 2006, 04:11:26 pm »

Many medium format cameras have built in leaf shutters which limits the maxiumum aperture of the lens.  That is why there is an 80mm f2.8 but only a 150 f4.]
Logged

Sheldon N

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 828
shallow DOF limits of current MF systems
« Reply #13 on: October 26, 2006, 04:15:25 pm »

Quote
Now I dont read charts  - I leave that to captains of ships

But I think it is the whole story.


I read glossy fashion mags A4 size all the time and the amount of MF/LF work is obvious even at that size due to the lack of DOF

At smaller apertures the less DOF of MF becomes more apparent because the background is still more fuzzy - many want that blown background but still a reasonable DOF

SMM
[a href=\"index.php?act=findpost&pid=82392\"][{POST_SNAPBACK}][/a]

My apologies, Sam, you are right.

I was thinking just in terms of depth of field at the plane of focus, and totally overlooked the degree to which the backgound is blurred in larger formats.

Yes, if you use an 80mm f/2 on 645 and a 50mm f/1.2 on 35mm the "depth of field" wide open is very similar at the plane of focus, and the field of view is similar as well, but the 80mm on 645 does blur the background to a greater degree.

I was wrongheaded in thinking that depth of field is the same as the degree of background blur. All it took was a few minutes of playing with my 35mm, 6x6, and 4x5 cameras and a normal lens on each to see the light.

Thanks for your patience with me.  
Logged
Sheldon Nalos
[url=http://www.flickr.com

Graham Mitchell

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2281
shallow DOF limits of current MF systems
« Reply #14 on: October 26, 2006, 04:18:48 pm »

Quote
Many medium format cameras have built in leaf shutters which limits the maxiumum aperture of the lens.  That is why there is an 80mm f2.8 but only a 150 f4.]
[a href=\"index.php?act=findpost&pid=82409\"][{POST_SNAPBACK}][/a]

All the Rollei lenses I listed above are leaf shutters, and they can still manage f2
Logged

Morgan_Moore

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2356
    • sammorganmoore.com
shallow DOF limits of current MF systems
« Reply #15 on: October 26, 2006, 04:49:52 pm »

Quote
My apologies, Sam, you are right.

[a href=\"index.php?act=findpost&pid=82411\"][{POST_SNAPBACK}][/a]

Rare words on these forums - appeciated

--------

The background of the 645/80/2 looks like the background not of the 35/50/1.4 but the same as as the background of a 35/80/2

strange but true

more telephoto like cut in the plane of focus

I think how you define muzzyness is subjective - I subjectively believe 645 to look nicer not nessacarily more muzzy - just nicer muzzines - in fact I am still confused as to which is more muzzy

SMM
« Last Edit: October 26, 2006, 04:52:06 pm by Morgan_Moore »
Logged
Sam Morgan Moore Bristol UK

Graham Mitchell

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2281
shallow DOF limits of current MF systems
« Reply #16 on: October 26, 2006, 05:16:26 pm »

Quote
The background of the 645/80/2 looks like the background not of the 35/50/1.4 but the same as as the background of a 35/80/2

strange but true

Yes that's true that a 35mm image is effectively a crop of a 645 image with the same lens. But when you compare similar images taken with 35mm and 645, the 645 lens will of course have to be of a longer focal length and will therefore have a more pronounced DOF at the same aperture.
Logged

Fritzer

  • Full Member
  • ***
  • Offline Offline
  • Posts: 212
shallow DOF limits of current MF systems
« Reply #17 on: October 26, 2006, 10:14:05 pm »

With all due respect, I think many replies hear repeat some of the most common misconseptions about DOF.

Our very own Michael has written a brief article about it , and here a more elaborate explanation is offered.

While it's true that not all lenses are created equal, it helps to keep in mind that basically aperture and magnification are the only values which determine DOF.

That explains how a larger format can change DOF, and the focal length lens can't; how a 35mm , 6x6 and 4x5 camera produce identical DOF once you match the focal length and apertures used.
« Last Edit: October 26, 2006, 10:14:52 pm by Fritzer »
Logged

howiesmith

  • Guest
shallow DOF limits of current MF systems
« Reply #18 on: October 26, 2006, 10:23:04 pm »

Quote
While it's true that not all lenses are created equal, it helps to keep in mind that basically aperture and magnification are the only values which determine DOF.

[a href=\"index.php?act=findpost&pid=82473\"][{POST_SNAPBACK}][/a]

I would add the focus distance - distance from camera to point of focus - and the lens' focal length.
Logged

ivan muller

  • Full Member
  • ***
  • Offline Offline
  • Posts: 245
    • Ivan Muller
shallow DOF limits of current MF systems
« Reply #19 on: October 27, 2006, 01:33:45 am »

hi Depth of field of a 150mm 35mm lens is the same as the dof of a 150mm lens for 4x5 at the same fstop. Thats why a 240mm lens on a 8x10 gives this amazing effect of very little dof but an extreme wide angle look. You cannot get this with any other 'smaller' format. Look at the work of Jock Sturgess. As far as I know he only uses one lens, the 240 on his 8x10.
Thanks Ivan
Logged
Pages: [1] 2 3 4   Go Up