Home

Please support mflenses.com if you need any graphic related work order it from us, click on above banner to order!
SearchSearch MemberlistMemberlist RegisterRegister ProfileProfile Log in to check your private messagesLog in to check your private messages Log inLog in

What are the best lenses when shooting fully stopped down?
Post new topic   Reply to topic
Home   Manual Focus Lenses
View previous topic :: View next topic  


PostPosted: Tue Oct 25, 2011 2:14 am    Post subject: Reply with quote
I shoot all my landscapes fully stopped down with both my NEX-3 and EOS 10D, should I be instead using f8 or f11 to avoid diffraction?


PostPosted: Tue Oct 25, 2011 2:25 am    Post subject: Reply with quote
iangreenhalgh1 wrote:
I shoot all my landscapes fully stopped down with both my NEX-3 and EOS 10D, should I be instead using f8 or f11 to avoid diffraction?


Yes, unless you need "deep focus", i.e. to have both foreground and background in focus.
In that case you fully stop down and sacrifice image quality in order to obtain deep focus.
In theory it could be also possible to take several shots at wider apertures shifting the focus point progressively then
combine them in dedicated software, but that is usually an overkill unless it's a well paid job.


PostPosted: Tue Oct 25, 2011 2:43 am    Post subject: Reply with quote
I mostly use my Tokina 17mm and Hexanon 3.5/28 for landscape work, I like the deep focus effect and that is indeed why I shoot fully stopped down. As I tend to do stitched HDR panos I already have a lot of PP work on my hands, combining 20-50 exposures.


PostPosted: Tue Oct 25, 2011 5:50 am    Post subject: Reply with quote
It just sounds not right for me.

F2.8 for 4/3, F4 for crop and F8 for full frame. And I always read that the best you can get is about F5.6-11 on any platform depends on lens.

Loads of mathematics. Will read everything. But I think we need to simplify it a bit. The results could be bend by using different sensors.

Some says F4 is the max for crop, some that the max is about F16.
Quote:

the limit is f5.6 for full frame and f4 for crop,
for 4:3, the best stop is around F2.8, after diffraction is the limit

Quote:

In real life? I haven't noticed it on my 18mp 60D, in fact images seem to get sharper until about f16...weird!


Let's leave the mathematics and get some real life!

This says that every lens would be the same after F8.
Quote:

this explain why all lens look the same at f8, good lens are leveled down by the diffraction


Somehow I don't want to believe in that.


PostPosted: Tue Oct 25, 2011 6:11 am    Post subject: Reply with quote
The diffraction is always there, it is just that at larger apertures, other problems (slide film losing contrast at 100 lp/mm, crappy lens, razor DOF etc) take precedence.

The diffraction makes something like gaussian (?) blur out of a single point; formulae for estimating limits differ as to how much degradation is ok. (Often forgetting to multiply with 2 or 1.7 or anything to account for Bayer sensor)

Common wisdom says f8 for crop, f11 for 35 mm.. as above, that depends on where your other limits are! There are lenses that do 200 lp/mm and above, and film materials (notably, monochrome) that resolve this easily, for this application, the limit hits at 5.6 or 4 obviously

This is also why using larger film does not help that much, because it is diffraction-limited at the same apertures as 35 mm (a little leeway due to slightly crappier MF lenses, but not much), and DOF of f5.6 with 4x5" is basically nil. Oooo..kay that's why these are t&s cameras, but still.


PostPosted: Tue Oct 25, 2011 7:06 am    Post subject: Reply with quote
Can I propose that it doesn't matter? If you need maximum DoF, then use the smallest aperture. No point having an image that is technically superb but not pictorially what you wanted to achieve?

I have spent so many years trying to get the technical aspects right, that I never took a photo I wanted to look at. Now I just take photos and the technical things come second.

To be fair, I am not a good photographer, artistically speaking, but I know for many of you that is not the case.


PostPosted: Tue Oct 25, 2011 7:24 am    Post subject: Reply with quote
The point is, when you sludged your image anyway, lens sharpness is irrelevant. You would have to select for other criteria.. minimum aperture, rectilinearity, color, color fringes although these usually go away


PostPosted: Tue Oct 25, 2011 8:58 am    Post subject: Reply with quote
what I mean is that it doesn't worth to spend big money for lens that are used after f2.8 because diffraction will kill their resolution
on this chart I collect lens resolution for Zuiko 40-150, Zuiko 14-42, Zuiko 12-60, Zuiko 35 macro and Summilux 25
you can see that no Zuiko lens can match the resolution of the Summilux at F1.4, that it is not related to diffraction but to lens quality
the Summilux have resolution limited by diffraction after f2
the other lenses have so low resolution that diffraction doesn't impact them the same
a lens like the 40-150 at 150 is so bad that it still improve at f8
if a lens on a m4:3 improve from f5.6 to f8, it just mean it is a very poor lens


PostPosted: Tue Oct 25, 2011 9:54 am    Post subject: Reply with quote
poilu wrote:
what I mean is that it doesn't worth to spend big money for lens that are used after f2.8 because diffraction will kill their resolution


What is your plot's vertical axis? I ask because at f/1.4 the Rayleigh Limit is roughly 1000 lp/mm. 2000 lp/mm is possible in air only at and below f/0.7. And the largest relative aperture possible in air is f/0.5.

Very few photographic objectives are diffraction limited near wide open. For most lenses, residual aberrations limit resolution wide open and are the reason that resolution and coverage improve on stopping down. A number of aberrations are sensitive to aperture and to distance off-axis. For a concise explanation see http://toothwalker.org/optics.html

Klaus and I have a number of lenses that are best wide open across the field covered. These are special-purpose lenses for photomacrography. Lenses sold for general use on 35 mm SLRs aren't that good. Per tests done years ago by Modern Photography, see http://www.flickr.com/photos/nesster/4424744296/sizes/o/ and http://www.flickr.com/photos/nesster/4424744224/sizes/o/ , superb fast normal lenses for 35 mm SLRs are best, in general, around f/5.6. Normal lenses for 35 mm still faster than f/1.7 or so are not superb.

When used on small-chip digital cameras these lenses may appear to be slightly better at apertures larger than f/5.6 because MP's tests measured resolution across the 24x36 camera gate. As I said, stopping down can expand the circle covered. The circle covered isn't the circle illuminated, it is the circle within which image quality exceeds the minimum standard. The minimum standard is quite arbitrary, whence come all of the squabbles about, e.g., f/6.8 Dagors' coverage.

Resolution on capture device is nice, but there's a law of nature to the effect that nothing enlarged more than around 12x from the original captured image will bear close scrutiny. What matters is resolution in the final print, not resolution on the capture device.


PostPosted: Tue Oct 25, 2011 10:45 am    Post subject: Reply with quote
Dan wrote:
What is your plot's vertical axis?

datas are from photozone.de who use imatest
vertical axis is in Line Widths per Picture Height
they used a Lumix DMC-L10 who have 3648x2736 pix for a m4:3 sensor of 17.3x13mm


PostPosted: Tue Oct 25, 2011 11:50 am    Post subject: Reply with quote
poilu wrote:
Dan wrote:
What is your plot's vertical axis?

datas are from photozone.de who use imatest
vertical axis is in Line Widths per Picture Height
they used a Lumix DMC-L10 who have 3648x2736 pix for a m4:3 sensor of 17.3x13mm


More than the comparison between the lenses, what is interesting in your scheme is that a good lens
like the 25mm Summilux when used on m4/3 starts to get worse after f/2.8, which is of course unacceptable
because a lens of that class should deliver optimal performance at least until f/5.6
This means that the m4/3 camera does really punish the good lenses, making them significantly worse.
With average lenses instead, the performance starts to decay after f/5.6 (which is still not good), but this only
because they did not perform well to start with.
My conclusion? I will never put my money in a m4/3 camera.
But I think producers also are starting to realize this, since they are more and more producing EVILs with APS-C sensors instead.


PostPosted: Tue Oct 25, 2011 12:54 pm    Post subject: Reply with quote
Pii wrote:
The results could be bend by using different sensors.

Some says F4 is the max for crop, some that the max is about F16.


There are different issues to consider:

1) The aperture where diffraction limits the maximum resolution below that of the sensor. This depends on the size of the photosites (“pixels”) on the sensor, and does not depend on sensor size. Since more and more megapixels are being put into the same size of sensor, the limit becomes smaller with newer cameras. For example EOS 5D has the same size of sensor as EOS 5DmkII but less pixels, so this limit is lower for 5DmkII (smaller photosites). Meanwhile Sony A900 (full frame) and Sony A77 (APS-C) have different sizes of sensors but (roughly) the same number of pixels (24Mpix), so the limit is lower for the A77. However the Sony A900 (FF) and Sony A700 (APS-C, 12Mpix) have (roughly) the same size of pixel, so this limit is (roughly) the same for both.

What is important to understand is that having a lower diffraction limit does not mean that the camera is worse because of it, rather it just means that the camera could capture higher resolution than is possible at that aperture. Similarly if you had two films, one capable of super-high resolution and one with rather poor resolution, the high-res film would have a much lower diffraction limit in the sense that the film could capture higher resolution than diffraction allows. It would still be the better film (in terms of resolution).

(Of course diffraction isn't the only limit; the lens itself can be of a design that does not allow it to reach diffraction limit—as said before, almost no lens is diffraction limited wide open.)


2) The aperture where the effect of diffraction lowers resolution so much that it can be considered “bad”. This depends mainly on the definition of what is sufficiently good for you. If you always view your images at 100% crop, then it's roughly the same as the limit discussed above, but usually digital images are downsampled to “full screen” or a given web resolution (like max. 1024 pixels wide), which means that you can probably shoot at any aperture in (non-macro) photography without seeing diffraction. If you're printing, it depends on the size of your print (and of course your camera's resolution needs to be high enough to accommodate that size of print, which brings in viewing distance, printer resolution, paper used, etc). And if you crop the image before printing or enlarging to full screen, then you're increasing the total enlargement.

So, in some sense, it comes down to how much you enlarge the physical size of the image—not the pixel size, as long as you have enough pixels (and/or film resolution, etc) to make that enlargement. This is why, in practice, full frame is better than APS-C, and APS-C is better than 4/3. (And medium format is better than full frame, large format better than medium, etc.) But of course this changes if you crop your image before enlarging—for example, by viewing at 100% crop, which is the same as completely ignoring the size of the sensor (i.e. the most important aspect).


PostPosted: Tue Oct 25, 2011 12:55 pm    Post subject: Reply with quote
Orio wrote:

But I think producers also are starting to realize this, since they are more and more producing EVILs with APS-C sensors instead.


Except Nikon who decided to put an even smaller sensor in theirs. A really absurd move on their part…


PostPosted: Tue Oct 25, 2011 1:39 pm    Post subject: Reply with quote
Arkku wrote:
The aperture where diffraction limits the maximum resolution below that of the sensor. This depends on the size of the photosites (“pixels”) on the sensor, and does not depend on sensor size. Since more and more megapixels are being put into the same size of sensor, the limit becomes smaller with newer cameras. For example EOS 5D has the same size of sensor as EOS 5DmkII but less pixels, so this limit is lower for 5DmkII (smaller photosites). Meanwhile Sony A900 (full frame) and Sony A77 (APS-C) have different sizes of sensors but (roughly) the same number of pixels (24Mpix), so the limit is lower for the A77. However the Sony A900 (FF) and Sony A700 (APS-C, 12Mpix) have (roughly) the same size of pixel, so this limit is (roughly) the same for both

the limit of resolution by diffraction is in lp/mm, resolution of bigger sensor will be higher as it have more mm
a practical example to make it clear
the maximum resolution theoretically possible at f4 is 400lp/mm
ZEISS Biogon T* 2,8/25 ZM resolve 400lp/mm at f4 on the film SPUR Orthopan UR
that correspond to a FF dslr of 537Mpixels
that mean that a dslr will never have more resolution than film as the physical limit is reached
on the other hand, a APS-C at 400lp/mm is 212Mpixels
that mean that a APS-C dslr will never reach the resolution of film


PostPosted: Tue Oct 25, 2011 1:58 pm    Post subject: Reply with quote
Gentlemen, could I please draw your attention to the OP's original question? As I read it, whatever the drawbacks, he is committed to using the aperture fully stopped down. He just wants to know the best lenses to use.


PostPosted: Tue Oct 25, 2011 2:15 pm    Post subject: Reply with quote
peterqd wrote:
Gentlemen, could I please draw your attention to the OP's original question? As I read it, whatever the drawbacks, he is committed to using the aperture fully stopped down. He just wants to know the best lenses to use.


There's a story about an English criminal, in the middle ages, who was sentenced to death and offered a choice of methods. All of the methods involved torture with refinements of cruelty. His reply? They are all very bad.


PostPosted: Tue Oct 25, 2011 2:17 pm    Post subject: Reply with quote
I don't think I can answer that question Peter but I can tell you that I have had great results at min aperture with my Tokina AT-X 3.5/17, Pentacon 3.5/30 and Pentacon 1.8/50 on my EOS 10D (6mp) and my Tokina RMC 3.5/17, Hexanon 1.7/50 and Hexanon 3.5/28 on my NEX-3 (12mp). There are others I have used at min aperture but those are the ones I have used extensively recently for my landscape work where I am seeking deep focus.


PostPosted: Tue Oct 25, 2011 3:04 pm    Post subject: Reply with quote
poilu wrote:

the limit of resolution by diffraction is in lp/mm, resolution of bigger sensor will be higher as it have more mm


Viewing at 100% crop ignores the size of the sensor, it only comes down to the size of the pixel if we are concerned with the threshold where diffraction prevents the maximum pixel resolution of the sensor from being reached (i.e. adjacent photosites will be recording the same circle of confusion). This is what point #1 of my post was about, as many replies in this thread discuss this limit.

In any case, smaller photosites means that the sensor is capable of recording more lp/mm (but such resolution may not be available to record if limited by diffraction, or lens quality, or camera shake, or air turbulence, or whatever). It does not matter (in this context) how many mm the sensor has overall, since the same limit applies to every mm of the sensor (i.e. we can just consider 1mm² sensor area and get the same result). However, the sensor size affects how many lp there are overall—note that this is just “lp” not “lp/mm”.

As a simplified example, let's say that the image has 100 lp/mm of resolution available and the sensor is capable of recording 150 lp/mm. So, regardless of what the size of the sensor is, we are limited to 100 lp/mm (e.g. by diffraction). However, if the sensor is 36mm wide, the horizontal resolution will be 36mm × 100lp/mm = 3600lp (note how the mm disappears when units are included in the calculation). However, a sensor that is 24mm wide would record only 2400lp worth of data. So, yes, at the same limit the bigger sensor gives higher total resolution, meaning that the diffraction matters less (like I said in point #2 of my previous post), but the lp/mm limit is independent of sensor size.

Last edited by Arkku on Tue Oct 25, 2011 8:11 pm; edited 1 time in total


PostPosted: Tue Oct 25, 2011 3:07 pm    Post subject: Reply with quote
peterqd wrote:
Gentlemen, could I please draw your attention to the OP's original question? As I read it, whatever the drawbacks, he is committed to using the aperture fully stopped down. He just wants to know the best lenses to use.


I guess what people (at least myself) are trying to explain is that it doesn't really matter what lens is used when fully stopped down, because diffraction effects become so significant. Of course lenses will still have different focal lengths, different colour rendering, etc, but this doesn't depend on stopping down so this is almost the same as asking “what are the best lenses if sharpness and bokeh do not matter”—not a very useful question without additional info, IMHO.

(Well, actually, the number of aperture blades decides how many points there will be in diffraction “stars” created by bright lights in the image when stopped down, so maybe that's one thing to consider. =)


PostPosted: Tue Oct 25, 2011 5:27 pm    Post subject: Reply with quote
Quote:

If you need maximum DoF, then use the smallest aperture. No point having an image that is technically superb but not pictorially what you wanted to achieve?

True Graham, but still would be good to know when shooting at f22 we have some options to obtain better, sharper image.
I must admit I never heard about diffraction. I always thought it is all about the lens design itself.

Quote:

Gentlemen, could I please draw your attention to the OP's original question? As I read it, whatever the drawbacks, he is committed to using the aperture fully stopped down. He just wants to know the best lenses to use.

Cheers peterqd.
What I am trying to figure out is if we are loosing the same amount of sharpness stopping down on each lens. And how much we are loosing. Let's say we're using 2 lenses on a full frame camera. First lens - very sharp at every stop, second - not so sharp. At f22, are those lenses equal? Or the sharper still becomes sharper?

...


PostPosted: Tue Oct 25, 2011 6:48 pm    Post subject: Reply with quote
Pii wrote:
What I am trying to figure out is if we are loosing the same amount of sharpness stopping down on each lens. And how much we are loosing. Let's say we're using 2 lenses on a full frame camera. First lens - very sharp at every stop, second - not so sharp. At f22, are those lenses equal? Or the sharper still becomes sharper?

as the last figure was not clear enough, here is one with the f11 datas
and a f16 from the Zuiko macro
I let you extrapolate how they will look at f22


from photozone.de


PostPosted: Tue Oct 25, 2011 7:13 pm    Post subject: Reply with quote
Quote:

as the last figure was not clear enough, here is one with the f11 datas

Looks like I missed this one poilu. Or I am not convinced enough...

Quote:

poilu wrote:

for 4:3, the best stop is around F2.8, after diffraction is the limit


This is really bullshit, I did use for years 4/3 camera best result did come around F8-F11 just same than on film or APS-C. No diffraction on shoots.

..or there is something else to consider.


PostPosted: Tue Oct 25, 2011 7:34 pm    Post subject: Reply with quote
Pii wrote:
Quote:

as the last figure was not clear enough, here is one with the f11 datas

Looks like I missed this one poilu. Or I am not convinced enough...

Quote:

This is really bullshit, I did use for years 4/3 camera best result did come around F8-F11 just same than on film or APS-C. No diffraction on shoots.

..or there is something else to consider.

then I am afraid nothing will convince you Razz


PostPosted: Tue Oct 25, 2011 7:52 pm    Post subject: Reply with quote
Quote:

then I am afraid nothing will convince you

I'm afraid we need more tests Very Happy


PostPosted: Tue Oct 25, 2011 8:17 pm    Post subject: Reply with quote
Pii wrote:

I'm afraid we need more tests :D


Go to photozone.de, look at the resolution figures for each lens tested on the same camera (e.g. EOS 350D probably has the most tests). You'll note that no matter how sharp or unsharp the lens is wide open, the figures end up being nearly identical at f/11 already. The test don't usually show beyond f/11 since it would be pointless, as the testers well know, because almost all lenses would test identically at f/22, but already at f/11 you can see the direction things are headed.
Post new topic   Reply to topic
Home   Manual Focus Lenses
Goto page 1, 2, 3  Next
Page 1 of 3