Well, this is fairly unusual—it appears that yesterday, in the comments to the "Another Internet Canard" post, both Ctein and I were wrong at the same time—usually one or the other of us has the situation covered even when the other makes a mistake. (Okay, okay, usually it's me who makes the mistake and Ctein who covers me.)
With some viewfinders (or, more precisely, viewing screens), faster lenses do not result in a brighter viewfinder image. As Oren explains, "it's because the screen in an AF SLR isn't a randomly diffuse ground glass (or ground plastic), it's an engineered light pipe with microlenses designed to accept light from a narrow field of view."
Okay. Sorry. But...everyone who's saying that a good test of this is to take a fast lens and watch the viewfinder as you stop down has still designed a bad test. Slow lenses have different geometry, different nodal point positions, and collect light from different angles than fast lenses do. If you want to compare the finder image of a slow lens vs. a fast lens, the best way to do it is to look at the viewfinder image of a fast lens and a slow lens, not the image of one fast lens at different apertures. Test what you are testing, not something else.
As I mentioned yesterday, I'm starting to like the gamey taste of crow (although I still don't like the taste of my own foot in my mouth).
I have modified the post accordingly.
Mike
Send this post to a friend
Please help support TOP by patronizing our sponsors B&H Photo and Amazon
Note: Links in this post may be to our affiliates; sales through affiliate links may benefit this site. More...
Original contents copyright 2011 by Michael C. Johnston and/or the bylined author. All Rights Reserved.
Featured Comment by Gordon Lewis: "Having just finished writing a 'cheatsheet' for Canon's Digital Learning Center on EOS focusing screens, I knew that Mike's bold statement (now amended) about how ƒ/1.4 lenses make the viewfinder brighter and easier to focus would soon result in his picking bits of crow out of his teeth. I wasn't so sure about Ctein, if only because he could be a vegetarian.
"The only thing of value I can add at this point is that Live View (for cameras that have it) is an excellent way to see exactly what depth-of-field you'll get at any given aperture. Just remember that you will have to manually stop the lens down to the taking aperture to preview depth-of-field, otherwise all focusing is at maximum aperture.
"Also keep in mind that if the display is set to auto-gain it will maintain the same brightness as you stop down. Some cameras let you set the display so that it's linked with the exposure histogram and so if the image likes 'just right' on the display it should look the same way in the actual photograph.
"In either case, if you have your camera mounted on a tripod, use 10X magnification, and focus manually at the taking aperture, you're practically guaranteed tack-sharpness at the plane of focus. You can even see whether your ƒ/1.4 lens has focus-shift. As for whether it's a good idea to use an ƒ/1.4 lens at maximum aperture, well, that's purely up to you."
Featured Comment by Ctein: "Yeah, I wrote my mea culpa under the original thread, for those who want to see my flavor of the humble pie. Based on some of the comments here, though, I think there are some people who are still a little confused. The bottom line on this is that it depends on the design of the focusing screen (and possibly the design of the viewfinder optics and lens). Some focusing screens will not show a loss of brightness until the aperture is smaller than X, where X is frequently ƒ/2.8, but that's no hard and fast rule. Others will."
Featured Comment by Bernard Piechal: "As for the focusing screen brightness, I've just checked it. On the camera that Mike sold just recently—Olympus OM-4. The simple result is: on the now-obsolete, traditional screen, ƒ/1.4 gives brighter image than ƒ/2. On the contrary, if I changed the screen for the 'modern' (#2–4, for OM users), the screen at ƒ/1.4 is not any brighter than at ƒ/2. The explanation is simple: the modern focusing screen is not groundglass, but the series of microlenses (on OM screens the lenses are even hexagonally spaced, not randomly). And these microlenses have their own numerical aperture (which defines the angle of acceptance of light rays, the more general 'version' of the f-number). Any rays coming on such focussing screen at bigger angle than angle of acceptance doesn't make it to the eye. When one opens the aperture all the additional rays comes at bigger angle of incidence, because that's what numerical aperture in optics defines. Also, yes, the optical situation of the ƒ/1.4 lens set at ƒ/2 is the same, as ƒ/2 lens wide open (as for the checking the difference in finder brightness between the two)."
...wonder if this explains why I've never been able to manually focus any of the auto-focus cameras I've used since the 90's, but when I go back to my ancient Nikkormat, the image snaps in and out of focus all over the viewfinder, even with a 28mm? Everyone 'says' they can manually focus with a modern camera, but for the life of me, the difference is so subtle I don't understand how they're nailing it! Nikon seems to be better than Canon, but still difficult. Tried magnifiers and correction and...everything. But, when I go back to vintage equipment, easy as pie!
Posted by: Tom Kwas | Thursday, 21 April 2011 at 08:10 AM
"...everyone who's saying that a good test of this is to take a fast lens and watch the viewfinder as you stop down has still designed a bad test."
But that's exactly what I wanted to find out - is f/1.4 visibly brighter in the viewfinder than f/2.8 with the same lens?
If f/2.8 is different on an f/2.8 lens than it is on a faster lens, I don't necessarily care. That wouldn't tell me whether or not the viewfinder continues to get brighter as I stop down the f/1.4 lens. I might care if f/2.8 on the faster lens was darker than f/2.8 on the slower lens, but that's not what I wanted to check.
Posted by: David Bostedo | Thursday, 21 April 2011 at 08:40 AM
To add another caveat: Not all AF systems honor the superfast lenses. Depending on the design of the AF sensors and the optical path reflecting the light on it, the effective f-stop for AF measurement can be several f-stops slower than the dialed-in aperture would indicate. Some DSLR bodies like my Sony A700 can compensate this partially with an extra sensor field for lenses of f2.8 or faster.
Add to this the fact that many old lens designs have a focus shift from wide open, where the AF measurement happens, to the f-stop effective during exposure, plus the one-shot system of most AF systems, i.e. calculating the amount of correction necessary and applying it, but *not* re-checking if it's really correct, and it becomes easily imaginable that the AF measurement at f1.4 can be wrong more often than not.
Together with the microlens design of the viewing screen mentioned by Oren, that also gives the optical impression of a slower f-stop, effectively using a f1.4 lens at its widest opening can be a real challenge (aka nightmare). And that's the reason my long awaited vintage Minolta 1.4/85mm doesn't get the exercise I had anticipated.
Posted by: Markus Spring | Thursday, 21 April 2011 at 09:23 AM
Here is Chuck Westfall's take:
http://digitaljournalist.org/issue0804/tech-tips.html
I think it should also be pointed out that the *depth of field* does not change, so you cannot use the view finder to evaluate areas of focus.
Posted by: KeithB | Thursday, 21 April 2011 at 09:32 AM
Good thing you just posted your favorite cartoon a few weeks back.
Posted by: Bernd Reinhardt | Thursday, 21 April 2011 at 10:23 AM
I was under the assumption this bottoms out at f/2.8. Meaning that with the default focus screen (not true of the special screens made for manual focusing, those benefit every extra stop at any aperture) you will see a brightness difference between an f/4 lens and an f/2.8, but you won't between an f/2 and an f/1.4.
I had a 40D and bought a focusing screen. It made it a bit dimmer, but I could see a difference between f/1.4 and f/2 when pressing the DOF preview button, whereas with the default screen I could not.
Posted by: Ben Mathis | Thursday, 21 April 2011 at 11:03 AM
Fooled me, too. Canon's focusing screens for manual focus (such as the Ee-S for the 5D) are described with the warning that they are not recommended for lenses slower than f2.8 because the screens are "not very bright". That makes me presume that the faster lenses provide more light to the viewfinder screen. Perhaps Canon's viewfinders are not included in the category of "some" viewing screens.
Posted by: R. Edelman | Thursday, 21 April 2011 at 11:16 AM
I'd guess that those Canon focusing screens R. Edelman mentions are of the groundglass type instead the microlens type. This would explain why they are a) dimmer - microlenses were introduced because of the required higher brightness of the viewfinder image following the introduction of slow standard zooms and b) show a difference when stopping down the lens: On the ground glass a real image is created, showing the distribution of sharpness, whilst the microlenses set up a virtual image like a magnifying glass - it becomes visible only in the full optical system including the human eye.
Posted by: Markus Spring | Thursday, 21 April 2011 at 11:54 AM
You should think yourself lucky if you can still get your foot anywhere near your mouth at your age. I'm a couple of years behind you and I can't do it! : )
Posted by: Roger Bradbury | Thursday, 21 April 2011 at 12:00 PM
wait what?
an engineered light pipe with microlenses designed to accept light from a narrow field of view."
I thought I knew everything and this line threw me for a loop.
Please tell me you'll spend a post (or three) in the near future and explain what this means as opposed to what we thought it meant.
Posted by: Jamin | Thursday, 21 April 2011 at 01:12 PM
Oh yes, modern DSLRs tend to come with such a screen out of the box in order to provide a bright viewing image using slower AF lenses. It sucks big time for MF work, so I changed the screen in my DSLR for an accessory screen of the Nikon F3 and it works very well for MF now, especially with fast lenses.
Posted by: Oskar Ojala | Thursday, 21 April 2011 at 02:34 PM
Add to all the previous observations that the so-called manual focus screens (for Canon AF DLSR's at least) do not have a split image or other focus aid because they interfere with metering. I tried the 5DMkII manual focus screen with Leica R lenses and gave it up because you never are sure of correct focus(good eyesight, recently lasik'ed to 20/10). It was like focusing on a very mini view camera ground glass. Frustrated, I bought an M8.
Posted by: Rick in CO | Thursday, 21 April 2011 at 02:38 PM
Wow, I never even knew that there was a microlens solution to the focusing screen problem... but it explains a lot as to why when I mount a slow zoom on a ground glass focusing screen the image really gets significantly darker than on a microlens screen.
Here's a great article which explains this (with some examples):
http://www.jayandwanda.com/photography/dslr_man_focus/man_focus.html
Pak
Posted by: Pak-Ming Wan | Thursday, 21 April 2011 at 03:02 PM
Ctein can blame ink vapour from the 3880. Mike can blame chocolate shakes (the effects are very long lasting) and many months of automatic transmission frustration. Oren Grad for President!
Posted by: Bahi | Thursday, 21 April 2011 at 04:28 PM
Come on folks, it's easy; just whip out your trusty depth of field calculator. I have one on my iPhone which I use when shooting 5x4 (an ironic convergence?) And if that fails you, there's always hyperfocal. No?
Posted by: Owen Silverwood | Thursday, 21 April 2011 at 09:38 PM
It does happen, however, that wider aperture lenses snap in and out of focus with more alacrity than do slower ones--whether or not the viewfinder image is apparently brighter.
Posted by: Ivan J. Eberle | Thursday, 21 April 2011 at 10:07 PM
from Gordon Lewis
"The only thing of value I can add at this point is that Live View (for cameras that have it) is an excellent way to see exactly what depth-of-field you'll get at any given aperture."
That is not quite true either unless you are going to print or only view your image at teh same size as your Live View Screen. The depth of field on your much larger print may indeed look a lot less narrow.
Posted by: Robert Hoehne | Thursday, 21 April 2011 at 10:23 PM
Robert,
You are correct about depth-of-field (Mike's bete-noire) when using Live View. That's why I mentioned using 10X magnification. It may not be the exact magnification you use when enlarging but it's a better indication than 1:1 on a 3" display.
Posted by: Gordon Lewis | Friday, 22 April 2011 at 07:16 AM
Solution to crow:
-Remove foot from mouth
-insert bar of wet soap
-massage soap in mouth
-remove soap
-insert crow, preferably deceased
-count quickly to numeral three
-removed deceased crow
-rinse mouth with fresh water
-insert bar of wet soap and massage
-rinse mouth with fresh water
-do not re-insert foot
Smile, you are good to go
Posted by: Bryce Lee | Friday, 22 April 2011 at 05:24 PM
Mike,
"everyone who's saying that a good test of this is to take a fast lens and watch the viewfinder as you stop down has still designed a bad test."
Can you explain why that is (or is not)? To my simple way of thinking, the fact that I can see no difference in the image on the screen between f1.4 and almost f2.8 on my 50mm Nikkor-S on my D700 shows I would have a bit of a problem trusting the focus if I were to shoot much at f1.4 (which I don't when critical focus is an issue).
Nor do I see any difference between the f1.4, f1.8, f2, and f2.8 (zoom) 50mm lenses I have. What am I missing?
I admit I have not shot much at all with all those other 50mm lenses in my "collection" so I am really asking if I should be taking a closer look at them.
Posted by: Doug C | Wednesday, 27 April 2011 at 06:46 PM