Canon EF 50mm f1.0 listing

FWIW, I would expect that the only thing that should suffer would
be the edges, not the center image. I can’t see how the microlenses
would diminish on-axis light, since it has a direct shot at the
central sensors. Unless there is Vaseline on the lens, the edges
should not contribute to the photons striking the center.
If you used a point source (eg what you try to acheive by stopping the lens down as far as it goes) then you would be right. At small apertures, the angle from which light hits the sensor is limited, so the edge of the sensor receives light at a shallow angle where as the centre of the image receives it bang on, which is why we tend to see vignetting due to microlens performance (and sensor dust, as it now casts a more distinct shadow)

With a wide aperture light enters from many directions at once. The edge of the sensor not only gets light from an angle but also other angles too, reducing vignetting.

Personally, I think the argument for the f1.0 lens performing badly on a digital sensor is flawed. Yes, more light gets into the lens from the edges of the glass, but that doesn't mean it all strikes the sensor microlenses at an angle.

The question is probably how much extra angled light rays does the increased aperture give over the increased oblique light rays and is it noticeable?

Compared to a film camera would the image not just be darker for any given ISO generally (but this is true for all lenses?)

That is probably all complete tosh too, but at least I'm in good company posting here ;-)
 
I'm selling mine soon privately (it is already spoken for and a check on the way), and I ordered an 85 1.2L. I expect to get more use out of the 85 1.2L than the 50 for the things that I enjoy shooting.

Let me know if you deal falls through. I'm looking for one myself.

Dave
 
FWIW, I would expect that the only thing that should suffer would
be the edges, not the center image. I can’t see how the microlenses
would diminish on-axis light, since it has a direct shot at the
central sensors. Unless there is Vaseline on the lens, the edges
should not contribute to the photons striking the center.
Your assumption (I'm addressing Stephen's paragraph here) would only be true if light waves were parallel. Light from the edge of the lens is also used to focus the center of the image. Ever see one of those "how lenses work" diagrams?
If you used a point source (eg what you try to acheive by stopping
the lens down as far as it goes) then you would be right. At small
apertures, the angle from which light hits the sensor is limited,
so the edge of the sensor receives light at a shallow angle where
as the centre of the image receives it bang on, which is why we
tend to see vignetting due to microlens performance (and sensor
dust, as it now casts a more distinct shadow)
Vignetting occurs at maximum aperture, and goes away as you stop down and hide the edges of the lens from the sensor/film.
With a wide aperture light enters from many directions at once. The
edge of the sensor not only gets light from an angle but also other
angles too, reducing vignetting.
The extra light at wide aperture tends to be concentrated in the center of the image, increasing vignetting. Oddly, this extra light comes from the edge of the lens.
Personally, I think the argument for the f1.0 lens performing badly
on a digital sensor is flawed. Yes, more light gets into the lens
from the edges of the glass, but that doesn't mean it all strikes
the sensor microlenses at an angle.
Most of the extra light does when it's an f/1.0 lens at maximum aperture.
The question is probably how much extra angled light rays does the
increased aperture give over the increased oblique light rays and
is it noticeable?
About 20 degrees, according to Joseph.
Compared to a film camera would the image not just be darker for
any given ISO generally (but this is true for all lenses?)
Darker and with more DOF.
That is probably all complete tosh too, but at least I'm in good
company posting here ;-)
Welcome to the club. :)
 
"The image sensor should convert light into photoelectrons efficiently. The measure of this efficiency is called the quantum efficiency (QE). Light rays that fall on the pixel in an oblique reduce the QE and the result is a loss of photoelectrons or in photographic words: vignetting: the darkening of the image. The use of microlenses will increase the QE but has one bad side effect: the angle roll off: light entering the micro lens at higher angles is directed away form the photodiode and gets lost."
[-snip-]
"...at angles between +20 degrees and ?20 degrees the angle roll off is hardly important. Translated in normal parlance: light rays from a lens striking the microlens at angle at less than 20 degrees will be recorded with good QE! Given the usual 1.3 to 1.5 reduction factor in viewing angle most lenses will behave properly and their light rays will strike at angels within the indicated + - 20 degrees."
http://www.imx.nl/photosite/comments/c001.html

The Kodak article the author sites can be found here (pdf):

Interesting, eh? So sensor size comes back into play?

cheers,

Chris
 
http://www.imx.nl/photosite/comments/c001.html

The Kodak article the author sites can be found here (pdf):

Interesting, eh? So sensor size comes back into play?
Nope. Only for slow lenses. With fast lenses, the extra light comes from the edge of the lens. How many times do I need to repeat it? Even the center pixel gets light from the edge of the lens. Ever use a mirror lens? There's a great big black circle in the middle of the lens, yet the center of the image is still as bright as (or usually brighter than) the corners of the frame.

The smaller the aperture, the more perpendicular the light striking the sensor will be. Think of a magnifying glass as your lens. The bigger the glass, the greater the angle from the edge of the glass to your image plane, whether in the center or the corner. It's exactly the same thing with compound lenses, and is the reason f/1.0 is no brighter than f/1.4 on a digital sensor.
cheers,

Chris
 
A smaller sensor reduces the distance (and angle) from one side of the exit pupil from crossing all the way across to the far side of the sensor. To me that makes sense until someone explains why it doesn't. Bigger sensor equal more distance when light crosses the sensor. That's the reason there are issues with fast, wide lenses on the FF sensors in Canon's line-up. I haven't heard any issues on "cropped" sensor bodies.

-- Greg
 
Take a look at this image:



Notice how every line converges on the center to create focus. As you close down the aperture, you start blocking the outermost (top and bottom) lines, thereby decreasing the angle the light hits the image plane. Even in the very center. The faster the lens, the larger it is, and the greater angle the light rays from the edge of the lens will hit the image plane Your sensor could be .5mm square, and it would still have the same effect.

One more time: this has nothing to do with sensor size. Only with the maximum aperture of the lens.

Vignetting with non-telecentric superwides is a completely different issue.
A smaller sensor reduces the distance (and angle) from one side of
the exit pupil from crossing all the way across to the far side of
the sensor. To me that makes sense until someone explains why it
doesn't. Bigger sensor equal more distance when light crosses the
sensor. That's the reason there are issues with fast, wide
lenses on the FF sensors in Canon's line-up. I haven't heard any
issues on "cropped" sensor bodies.

-- Greg
 
According to Joseph, the two are inextricably related. The EF lensmount was designed specifically for the 50mm f/1.0, which requires an exit pupil relatively close to the image plane. That's why our cameras have such big gaping holes compared to, say, Nikon.

In other words, f/1.0 requires the exit pupil to be close to the sensor.
Only with the maximum aperture of the lens.
What about exit pupil to sensor distance? Seems like that would
affect the angle of incidence. Or do you think that's wrong too?

-- Greg
 
The smaller the aperture, the more perpendicular the light striking
the sensor will be.
I get that, honest. What I may be wrong on, though, is my equating large aperture with the same issues with wide angle (and, by the way, with lens systems with recessed back elements like the Leica M system). But more on that below.
With fast lenses, the extra light comes
from the edge of the lens. How many times do I need to repeat it?
I get that too (although some of the other posters above don't seem to get it for some reason). I also get how that translates into an OVERALL reduction in light (i.e., even at the centre). However, it should still have more impact on the edges of the frame, no? With darker corners (vignetting). If that is an incorrect assumption, then that is where my mental block is (not with your other points).

Notice that there's a pretty consistent theme in the literature on wide angle with FF, or with Leica's problems going digital, or with Olympus move to 'telecentric' lenses: all these companies say oblique angles cause vignetting. What I'm assuming is that whatever the reason for the oblique angle -- wide angle lens, Leica's closer rear element, or wider aperture -- the result is the same: less light will fall from the corners, causing more vignetting than with film.

What is NEW, is the claim that OVERALL darkening happens. If all you are saying is that large apertures will cause both, then I understand and agree.

But then there's the issue of the sensor sizes...
Interesting, eh? So sensor size comes back into play?
Nope. Only for slow lenses.
OK, that I don't get. The smaller the aperture, the more perpendicular the light striking the sensor will be -- hey, isn't that what you just said? ;-)

cheers,

Chris
 
In other words, f/1.0 requires the exit pupil to be close to the
sensor.
OOOhhhh...Ok, I'm starting to get it now. That was the key piece of info that was missing (or obscurred in all the debate). I'll write something (long) up to see if I get it fully, though.

I think this is a milder version of Leica's problem. They have given up on a digital M system (at least for now), because the exit pupil is so close to the film with their recessed lenses.

In the meantime, your image doesn't show up for me. It will help greatly to see the diagram, if you can post it again.

thanks,

Chris
 
With fast lenses, the extra light comes
from the edge of the lens. How many times do I need to repeat it?
I get that too (although some of the other posters above don't seem
to get it for some reason). I also get how that translates into an
OVERALL reduction in light (i.e., even at the centre). However, it
should still have more impact on the edges of the frame, no? With
darker corners (vignetting). If that is an incorrect assumption,
then that is where my mental block is (not with your other points).
They are two separate issues. Do you have a lens with a manual diaphram you can look through? Take it and look through the back at a light source. In order to get full brightness on any part of the film (I'm going to use that term for now), the film must see the entire circle of a given aperture. With the lens wide open, tilt it away from center and you'll quickly see that the circle gets deformed (shaped like and American footbal) not very far from perpendicular.

Now stop down to f/22. You can tilt a lot more and still see the entire circle.

That's classic vignetting at work, and demonstrates why stopping down corrects it.

To get to the microlens issue, you sorta have to reverse your thinking. Looking at the diagram I posted may help. Again stopping down "corrects" it, since by f/2.0 marked aperture more or less coincides with the light transmittance.

I've only recently started to understand this myself. But it makes perfect sense, once you get your head around it.
Notice that there's a pretty consistent theme in the literature on
wide angle with FF, or with Leica's problems going digital, or with
Olympus move to 'telecentric' lenses: all these companies say
oblique angles cause vignetting. What I'm assuming is that
whatever the reason for the oblique angle -- wide angle lens,
Leica's closer rear element, or wider aperture -- the result is the
same: less light will fall from the corners, causing more
vignetting than with film.
The reason you don't read more about this is that there aren't any f/1.0 lenses in current production. For that matter, it wouldn't surprise me if the 85 f/1.2 wasn't currently being manufactured, either. I believe the 50mm f/1.0 was only produced in a single batch. When they sold them all, they discontinued it.
What is NEW, is the claim that OVERALL darkening happens. If all
you are saying is that large apertures will cause both, then I
understand and agree.
On this matter, I'm not certain. Could be, but I don't know. :)
But then there's the issue of the sensor sizes...
Interesting, eh? So sensor size comes back into play?
Nope. Only for slow lenses.
OK, that I don't get. The smaller the aperture, the more
perpendicular the light striking the sensor will be -- hey, isn't
that what you just said? ;-)

cheers,

Chris
 
Please read the links from this posts.

Your statement that digital cannot fully use available light wider
than f/2 is a huge generalisation

http://forums.dpreview.com/forums/read.asp?forum=1032&message=13151045

It is more complicated than you think.
And your calculations are incorrect. According to Joseph (and yes, I am relying on him for my facts), 85mm f/1.2 has an exit pupil of 50mm. The exit pupil of the f/1.0 lens is about 40mm.
Yiannis

-LORD VADER
-yes master
-RISE

http://www.pbase.com/ystasino
 
Please read the links from this posts.

Your statement that digital cannot fully use available light wider
than f/2 is a huge generalisation

http://forums.dpreview.com/forums/read.asp?forum=1032&message=13151045

It is more complicated than you think.
And your calculations are incorrect. According to Joseph (and yes,
I am relying on him for my facts), 85mm f/1.2 has an exit pupil of
50mm.
Do you have a reference that we can read upon other than Joseph? Do you have a definition of what the exit pupil is and how it is calculated ?
The exit pupil of the f/1.0 lens is about 40mm.
Yiannis

-LORD VADER
-yes master
-RISE

http://www.pbase.com/ystasino
--
Yiannis

-LORD VADER
-yes master
-RISE

http://www.pbase.com/ystasino
 
I have access to a 50mm f/1 so I decided to give it a try and see what happens.

The body is an EOS 20D. These are images shot in M mode and are right out of the camera then scaled down to web size. No other post processing has been applied.

f/2, 1/200 sec



f/1.4, 1/400 sec



f/1.0, 1/800 sec



Greg

--



http://www.pbase.com/dadas115/
 

Keyboard shortcuts

Back
Top