Sebboh;
A split screen prism isn't an on/off devise per se as you focus. The base pair distance (F2.8, F5.6 or whatever) in a split prism screen is just an indication of how fast the image halves will pass each other as you focus "through" the scene.

An F2.8 split means that the two image halves will be twice as far from each other (given the same misfocus amount) as if you had a F5.6 split. They will both pass each other perfectly at the correct focus.
.............

Thanks for the test BTW, it was both interesting in itself, and well executed. I know how time-consuming things like this can be.

If you want to know the REAL vignetting performance of your native Canon EF lenses, then you have to apply a slight "twist" to stop Canon from cheating. They correct vignetting in their own lenses by up to 2/3 Ev - as long as the lens i spresent in the camera's database. What you do is unlock the lens, and twist it towards the "relase" position by just a mm or so. This will cause the lens to lose electrical contact with the body, and you will have to shoot in full manual. To get the stopped down apertures, press "DoF preview" as you twist the lens. The lens will stay in the aperture it had when it looses the electrical feed.
Greetings from Sweden/
Joakim

theSuede wrote:
Sebboh;
A split screen prism isn't an on/off devise per se as you focus. The base pair distance (F2.8, F5.6 or whatever) in a split prism screen is just an indication of how fast the image halves will pass each other as you focus "through" the scene.

An F2.8 split means that the two image halves will be twice as far from each other (given the same misfocus amount) as if you had a F5.6 split. They will both pass each other perfectly at the correct focus.


Thank you for the clarification--and vindication. I knew this was not some figment of my imagination. The point is, it is much easier for me to see the break in alignment in the halves of the image rather than trying to use a matte, because all the matte does is make the image appear more blurry, and it's not as obvious when the image transitions from sharp to blurry or vice versa.

theSuede wrote:
Sebboh;
A split screen prism isn't an on/off devise per se as you focus. The base pair distance (F2.8, F5.6 or whatever) in a split prism screen is just an indication of how fast the image halves will pass each other as you focus "through" the scene.

An F2.8 split means that the two image halves will be twice as far from each other (given the same misfocus amount) as if you had a F5.6 split. They will both pass each other perfectly at the correct focus.
.............


what does that mean for focusing an f/1.2 lens with focus shift on an f/2.8 split? in the last thread you gave the impression (to me at least) that they would not pass each at the correct focus.

Thanks for the testing effort you made!

I wonder a bit because of the highlight bokeh diameter (or width) of the Rokkor 58.

For example @ f/1.2 - roughly measured left near the middle candle:
Rokkor 58/1.2: 81 Pixel
EF 50/1.2: 65 Pixel
EF 50/1.0: 65 Pixel
That is a factor 1.24 - but the factor for the focal lengths is 1.16.

At f/2.8
Rokkor 58/1.2: 40 Pixel
EF 50/1.2: 32 Pixel
EF 50/1.0: 32 Pixel
That is a factor 1.25 - but the factor for the focal lengths is 1.16.

The EF 50/1.0 at 1.0: 75 pixel diameter.

Why is the bokeh diameter so different - more than the focal length difference?
Microlens effects not, because it is the same at 2.8.
It could be the IF (internal focus) - but I measure a factor of ~1.14 for different lengths.

So that highlight difference is due to optical design of the retrofocus system?

ZoneV wrote:
Thanks for the testing effort you made!

I wonder a bit because of the highlight bokeh diameter (or width) of the Rokkor 58.

For example @ f/1.2 - roughly measured left near the middle candle:
Rokkor 58/1.2: 81 Pixel
EF 50/1.2: 65 Pixel
EF 50/1.0: 65 Pixel
That is a factor 1.24 - but the factor for the focal lengths is 1.16.

At f/2.8
Rokkor 58/1.2: 40 Pixel
EF 50/1.2: 32 Pixel
EF 50/1.0: 32 Pixel
That is a factor 1.25 - but the factor for the focal lengths is 1.16.

The EF 50/1.0 at 1.0: 75 pixel diameter.

Why is the bokeh diameter so different - more than the focal length difference?
Microlens effects not, because it is the same at 2.8.
It could be the IF (internal focus) - but I measure a factor of ~1.14 for different lengths.

So that highlight difference is due to optical design of the retrofocus system?

Your calculation of a 1.16x factor would be correct if the images were taken at the same subject magnification, but they were not--they were taken at the same subject distance, with the longer focal length resulting in a slightly higher image magnification. If you rescaled the Rokkor image so that the subject in focus (the candle) appeared at the same size as the other 50mm images, then you will find that the background blur circle will appear closer to 1.16x larger. In fact, it is less, only about 1.1, because the background lights are not infinitely far away--if the lights were very very far away from the candle, then the ratio of the blur disks would be nearly 58/50 = 1.16.

wickerprints wrote:
theSuede wrote:
Sebboh;
A split screen prism isn't an on/off devise per se as you focus. The base pair distance (F2.8, F5.6 or whatever) in a split prism screen is just an indication of how fast the image halves will pass each other as you focus "through" the scene.

An F2.8 split means that the two image halves will be twice as far from each other (given the same misfocus amount) as if you had a F5.6 split. They will both pass each other perfectly at the correct focus.


Thank you for the clarification--and vindication. I knew this was not some figment of my imagination. The point is, it is much easier for me to see the break in alignment in the halves of the image rather than trying to use a matte, because all the matte does is make the image appear more blurry, and it's not as obvious when the image transitions from sharp to blurry or vice versa.

But have you actually manually focused a f/1.2 lens with both screens? No, obviously not.

helimat wrote:
But have you actually manually focused a f/1.2 lens with both screens? No, obviously not.

I really don't understand your hostility to this idea that I personally find it easier to see when two halves of an image are aligned, as opposed to determining when the image is sharp or blurry. I don't actually NEED to have used f/1.2 lenses with both types of screens to be able to make that claim. My personal experience with the lenses I have used supports my preference.

And for what it's worth, I use an Eg-S screen on my 5D2 with an 85/1.2L II, and there is *no* way for me to see critical focus wide open. That screen does not let me see 100% crop to determine whether I focused on the eye or the eyelash of a person. No one has that kind of visual acuity. A split prism wouldn't let me see that either.

You appear to be under this impression that your way is the only correct way to focus, and that anyone who suggests otherwise is wrong.

wickerprints wrote:
because all the matte does is make the image appear more blurry, and it's not as obvious when the image transitions from sharp to blurry or vice versa.

I believe that description is almost the opposite of how most people experience it, and also how Canon describes their Super-Precision matte:

"The Ee-S screen has finer microlenses than the Ee-A or Ee-D options, along with a steeper parabola of focus to make the image pop in and out of focus more vividly in the viewfinder"
source

I agree that it's not always as obvious as a split prism, although vertical or horizontal splits can be ineffective too in some situations (diagonal is better perhaps) -- I actually preferred microprisms when I shot a manual film body. I don't want these focusing aids in my VF though, because they're an interference when making composition. I wish I could actually get rid of the black boxes for the AF points.

wickerprints wrote:
because all the matte does is make the image appear more blurry, and it's not as obvious when the image transitions from sharp to blurry or vice versa.

AhamB wrote:
I believe that description is the opposite of how most people experience it, and also how Canon describes their Super-Precision matte:
The Ee-S screen has finer microlenses than the Ee-A or Ee-D options, along with a steeper parabola of focus to make the image pop in and out of focus more vividly in the viewfinder source

You misunderstand what I mean. I'm speaking about the way one ascertains whether a point is in focus or not is completely different between a split prism and a matte screen, whether or not it is a standard or precision matte.

A split prism shows you a sharp image but in two halves. You know it's focused if the halves are aligned. In contrast, a matte screen shows an image that can be sharp or blurry, and the extent of blur indicates the degree of focus. My point--which should not be so difficult to understand--is that aligning two halves of a broken line is perceptually easier to do (for me) than trying to see whether something is sharp or not sharp. That transition from "sharp" to "not sharp" is not sudden--there are degrees of blurriness. A precision matte makes that transition "sooner" than a standard matte, but nonetheless both do so in a continuous fashion, not in a way that allows rapid and easy identification, and certainly not in a way that indicates direction of misfocus.

wickerprints wrote:

I really don't understand your hostility to this idea that I personally find it easier to see when two halves of an image are aligned, as opposed to determining when the image is sharp or blurry. I don't actually NEED to have used f/1.2 lenses with both types of screens to be able to make that claim. My personal experience with the lenses I have used supports my preference.

And for what it's worth, I use an Eg-S screen on my 5D2 with an 85/1.2L II, and there is *no* way for me to see critical focus wide open. That screen does not let me see 100% crop to determine whether I focused on the eye or the eyelash of a person. No one has that kind of visual acuity. A split prism wouldn't let me see that either.

You appear to be under this impression that your way is the only correct way to focus, and that anyone who suggests otherwise is wrong.

My hostility is towards the increase in likely hood around here lately that anything one suggests will be instantly poo-pooed. I, for one, an not talking out of my ass:

EE-S Product Description:

"The Ee-S Super Precision focusing screen is designed to facilitate manual focusing with high-speed lenses (f/2.8 or faster). The Ee-S screen has finer microlenses than the Ee-A or Ee-D options, along with a steeper parabola of focus to make the image pop in and out of focus more vividly in the viewfinder; however, the Ee-S focusing screen is not recommended for slower lenses because it's not very bright."

And yes, in order to make a valid claim about something, there is a requirement to have tried it first, otherwise it is known as conjecture. Check my profile, I have several fast manual focus only lenses that I use in conjunction with the E*-S screen, and have tried them with split prism screens as well. Your profile suggests that you only own AF lenses, which also leads me to believe that you are, in fact, talking out of your ass. Frankly, I see this type of behavior out of you a lot around here. Additionally, I made a point of saying that it was my opinion more than once, as that is nothing more than what it is. And finally, it was you who initially claimed that I was wrong, so to turn it around to say that I believe my way is the only way is pure hypocrisy.

wickerprints wrote:
A split prism shows you a sharp image but in two halves. You know it's focused if the halves are aligned.

That may well be the case with a lens that is f/2.8, but certainly not at f/1.2. I HAVE TRIED IT.

I apologize to the OP and other interested parties for going down this road, this type of stuff just irks me. Maybe he was trying to goat me into this... If so, well done.

Which is easiest might completely be a personal experience thing, or something coupled to your personal eyesight impediments. We all work in different ways.

Sebboh;
The focus shift in a F1.4-F1.2 lens is roughly the same from WO> F2.8 as in WO> F4.0, so probably not much. But you can see the focus alignment better in a broader base pair. Focus shift is a function of spherical aberration, and most lenses are corrected to swing back a little in the last stop before WO, what we call a quadratic response. The SA deviation curve often looks like a crowbar, a slow increase in shift from F16 down to F2, and then the line converges back to zero shift at the rim rays of WO use. This is what gives bright outlines on behind-focus bokeh.

Trying to imagine the pictorial effect of this is quite non-trivial. The picture is made out of the integral sum of all rays that fits in the current aperture, scaled to unity. You use shorter shutterspeeds for larger apertures, and so on. So, at F2.8, the picture is made from every ray within the F2.8 limit. This includes the F4, F5.6, F8 and som on rays too... The en result, the picture is the sum of all this.

How the focus shift manifests is part of the lens configuration. Most of the time you have a rather constant "best focus distance" when you look at the 2-stop-down and 3-stop-down positions of the aperture (F2.8 and F4.0 on a F1.4 lens). The biggest shift is most of the time from 1/2 stop down to 1+1/2 stop down.

So, a F2.8 or F4.0 measurement will both be very similar in "misfocus" compared to a F1.4 shot, but the F2.8 pair will be easier to see (higher misfocus magnification). What you would really need to combat focus shift is a F2.0 measurement - and that's rather impossible with the hardware commercially available today...

BTW, I've used both high-precision splits (Contax) and the Ee-S modified to fit various cameras (my own A850, and some others). I can't say I really have a preference. The Ee-s is easier on moving targets, the splits are easier on still-life targets.

theSuede wrote:
Which is easiest might completely be a personal experience thing, or something coupled to your personal eyesight impediments. We all work in different ways.

Sebboh;
The focus shift in a F1.4-F1.2 lens is roughly the same from WO> F2.8 as in WO> F4.0, so probably not much. But you can see the focus alignment better in a broader base pair. Focus shift is a function of spherical aberration, and most lenses are corrected to swing back a little in the last stop before WO, what we call a quadratic response. The SA deviation curve often looks like a crowbar, a slow increase in shift from F16 down to F2, and then the line converges back to zero shift at the rim rays of WO use. This is what gives bright outlines on behind-focus bokeh.

Trying to imagine the pictorial effect of this is quite non-trivial. The picture is made out of the integral sum of all rays that fits in the current aperture, scaled to unity. You use shorter shutterspeeds for larger apertures, and so on. So, at F2.8, the picture is made from every ray within the F2.8 limit. This includes the F4, F5.6, F8 and som on rays too... The en result, the picture is the sum of all this.

How the focus shift manifests is part of the lens configuration. Most of the time you have a rather constant "best focus distance" when you look at the 2-stop-down and 3-stop-down positions of the aperture (F2.8 and F4.0 on a F1.4 lens). The biggest shift is most of the time from 1/2 stop down to 1+1/2 stop down.

So, a F2.8 or F4.0 measurement will both be very similar in "misfocus" compared to a F1.4 shot, but the F2.8 pair will be easier to see (higher misfocus magnification). What you would really need to combat focus shift is a F2.0 measurement - and that's rather impossible with the hardware commercially available today...

BTW, I've used both high-precision splits (Contax) and the Ee-S modified to fit various cameras (my own A850, and some others). I can't say I really have a preference. The Ee-s is easier on moving targets, the splits are easier on still-life targets.

thanks, i understand how the image forms for the actual image and how spherical aberration works. my confusion is regarding the split prism and whether rays from f/1.2 (the angle of rays that only come through when the lens is open to f/1.2) impact the alignment in the split image. if i shoot two shots, one at f/1.2 and one at f/2.8, with a lens that has noticeable focus shift from f/1.2 to f/2.8 (of which there are a few) and focus once using an f/2.8 split prism with the lens wide open, which image will be missfocused the f/1.2 shot or the f/2.8 shot?

carstenw wrote:
denoir wrote:
carstenw wrote:
Hmm, I am on the verge of ordering the 50MP, but I have also seen some great 50/1.4 results from philber and Makten, and I am wondering if the two of them could co-exist peacefully?

Had I known what I know know I would have probably bought the 50 Planar instead of the MP. The reason is that the 50 MP is similar in drawing style to the 100 MP while the 50 Planar is unique in its ways.

Had I not been using the M9 as my primary camera then I would have definitely have gotten the Planar by now (in addition to the 50 MP that I have) and I'm not ruling it out completely anyway

Hmm... so, another Gedankenexperiment: if you had both 50/1.4 and 50MP, does that mean that you would use the 50/1.4 the most, and the 50MP only in specific situations (wide open, close up, where sharpness is needed, and sharp portraits)?

Its perfectly natural to have both lenses. I have both the contax N 501/.4 and the ZE 50/2MP and they can both be used to achieve different goals. For portraits or landscapes where I want to have more separation from the background and maybe more 3-d and more definition of subject shape then I would use the 50/1.4. If I want to have maximum sharpness from front to back in a large DOF landscape shot I would use the 50/2MP. Also, the 50/2 MP makes a better landscape stitching lens with its more uniform sharpness across the frame.
The fact that my contax N 50/1.4 has AF and MF gives me another distinguishing factor, in those cases where I may not have the time to MF, moving subjects etc.

Sebboh;
You can see the flat prism part as a secondary aperture within the lens aperture. This secondary aperture has the size of the area of the prism mirror. Maybe a 4x2mm half-circle? Very small anyway.

This aperture then "sees" a 4x2mm section of the lens aperture opening. This section will be placed at [focal length / half base pair distance] mm out from the central ray in the lens. 50mm lens with a F4 base pair gives 50/8=6.25mm out from the center of the lens aperture (this is very simplified, but anyway...). ONLY light from these two opposite spots (6.25x2 = 12.5mm apart) is directed from the matte screen into the VF optics. This means that you ONLY see what the F4.0 aperture rim ray would give you. Nothing outside (larger apertures) and nothing inside (smaller apertures). So in your example, with a normal setup, the F1.2 shot would be front focused. And it would make no difference if you stopped down before focusing, the shots would be focused at the same distance.

As you may have noticed the prism stays equally bright, no matter if the lens has a F3.5 or F1.4 aperture. There's only two modes for the prism surface; blocked ray path or open ray path.

So, no the F1.2 rim rays have no impact at all at the prism projection. It looks at a very tiny portion of the aperture range, like having an onion slice aperure "ring" from F3.95 > F4.05. How well this works with regards to pictorial focus shift depends on how the lens behaves outside this area. A VERY matte screen, even more matte than the Ee-S might give you a better "adaption" to SA, but I really doubt it. I can't say I see much difference in DoF through an Ee-S scren from F2.5 down (F1.7>F2.5 is the biggest focus shift stop in most normal lenses). If I have the spare time some day I'll set up the macro "through the VF" setup shots again.

theSuede wrote:
Sebboh;
You can see the flat prism part as a secondary aperture within the lens aperture. This secondary aperture has the size of the area of the prism mirror. Maybe a 4x2mm half-circle? Very small anyway.

This aperture then "sees" a 4x2mm section of the lens aperture opening. This section will be placed at [focal length / half base pair distance] mm out from the central ray in the lens. 50mm lens with a F4 base pair gives 50/8=6.25mm out from the center of the lens aperture (this is very simplified, but anyway...). ONLY light from these two opposite spots (6.25x2 = 12.5mm apart) is directed from the matte screen into the VF optics. This means that you ONLY see what the F4.0 aperture rim ray would give you. Nothing outside (larger apertures) and nothing inside (smaller apertures). So in your example, with a normal setup, the F1.2 shot would be front focused. And it would make no difference if you stopped down before focusing, the shots would be focused at the same distance.

As you may have noticed the prism stays equally bright, no matter if the lens has a F3.5 or F1.4 aperture. There's only two modes for the prism surface; blocked ray path or open ray path.

So, no the F1.2 rim rays have no impact at all at the prism projection. It looks at a very tiny portion of the aperture range, like having an onion slice aperure "ring" from F3.95 > F4.05. How well this works with regards to pictorial focus shift depends on how the lens behaves outside this area. A VERY matte screen, even more matte than the Ee-S might give you a better "adaption" to SA, but I really doubt it. I can't say I see much difference in DoF through an Ee-S scren from F2.5 down (F1.7>F2.5 is the biggest focus shift stop in most normal lenses). If I have the spare time some day I'll set up the macro "through the VF" setup shots again.

thanks! that makes it much more clear, and i was right in my original statement. or at least right about what i meant to say (what i actually said i'll have to check), which was that a split prism will be off target shooting an f/1.2 lens with severe focus shift wide open. basically, you can't trust the split prism any more than you can trust the precision matte screen at f/1.2. obviously the two require a very different method of focusing and some will prefer one or the other, but it seems to me that the split prism would yield a more consistent replicable result while the precision matte could yield slightly more accurate results for people with very high visual acuity who take a fair bit of time focusing. in any event, this why i love liveview and am looking forward to seeing what can be done once evfs get better.

Thank you for this terrific comparison of lenses, and referencing it on another thread, else I wouldn't have seen it!

From the results and my needs, here's what I find:

The Canon 1.0 is amazing in center sharpness through f/4 and a real winner for this sort of general "effect" lens. I don't see why if someone could afford it, they wouldn't grab a copy. If the axiom is now to own two 50mm lenses, wouldn't this be the "fun" one?

The Canon 1.4 looks terrific as well. The creamy OOF light source is much better than the Zeiss 1.4 with the glowing ring. I recently sold my Canon 1.4, and now I regret it even more. It looks to have outstanding sharpness too.

The Zeiss 2.0 MP macro looks very impressive as a general lens.

If I had the money, I'd own the Zeiss macro and the Canon 1.0 as my two lens set.

Gunzorro wrote:
The Canon 1.4 looks terrific as well. The creamy OOF light source is much better than the Zeiss 1.4 with the glowing ring. I recently sold my Canon 1.4, and now I regret it even more. It looks to have outstanding sharpness too.


Jim if you want to see how the Canon 50/1.4 measures up against a really good fast 50, you might be interested in this:
http://www.fredmiranda.com/forum/topic/995012

Take a look at post #5 in the thread as well as it has a comparison with the 50 MP. You can expect the difference in rendering (colors & contrast) to be about the same with the 50 Planar.

Any thoughts on the Zeiss 50mm F1.7 vs the Zeiss 50mm F1.4 Zf.2?

I need a fifty and so far those are my two best candidates. It seems of the CY Zeiss lenses, the 1.7 is one of the sharpest they ever made.

Does it display the same pleasing 3d popping focus transitions similar to the F1.4 ZF?

50 1.7 wide open, OOC, unsharpened:


_MG_6878.jpg by pictorlucis, on Flickr

FlyPenFly wrote:
Any thoughts on the Zeiss 50mm F1.7 vs the Zeiss 50mm F1.4 Zf.2?

I need a fifty and so far those are my two best candidates. It seems of the CY Zeiss lenses, the 1.7 is one of the sharpest they ever made.

Does it display the same pleasing 3d popping focus transitions similar to the F1.4 ZF?

I'm assuming your using a Nikon body. Do Contax lenses work on Nikon bodies? If it does, the C/Y 50/1.7 is a no brainer - relatively cheap, great performance. Although the prices have really gone up lately. I think I got mine for $200 less than 2 years ago.