The Omen -anamophic bokeh

[Mathew Collins]

I have learned that anamorphic bokeh is egg shaped or oval shaped. But I could see 2 types of bokeh in Omen in two different scenes. Could you share your insights.

[Satsuki Murashige]

Anamorphic lenses are basically normal spherical lenses with cylindrical elements added to squeeze the image on the horizontal axis. Anamorphic bokeh only looks like the first image when the cylindrical element is in front of the iris. This is the case for almost all anamorphic prime lenses. This also makes the lenses slower and heavier, since more large glass elements are used.

 

For the very large 35mm zoom lenses of the time, it was never practical to anamorphize them in this way. So the compromise was to add a small rear anamorphic element behind the iris. In this way, the lens could be kept roughly the same size and weight with no loss of functionality other than some light loss. The downside was that since the image had already been formed by the glass in front of the iris, no anamorphic artifacts would be visible. What you have essentially is an image that is squeezed 'in post.' Most long telephoto shots in old anamorphic films like the second image posted were made with rear-anamorphic zooms. So that accounts for the difference in bokeh.

[Mathew Collins]

Anamorphic lenses are basically normal spherical lenses with cylindrical elements added to squeeze the image on the horizontal axis. Anamorphic bokeh only looks like the first image when the cylindrical element is in front of the iris. This is the case for almost all anamorphic prime lenses. This also makes the lenses slower and heavier, since more large glass elements are used.

 

For the very large 35mm zoom lenses of the time, it was never practical to anamorphize them in this way. So the compromise was to add a small rear anamorphic element behind the iris. In this way, the lens could be kept roughly the same size and weight with no loss of functionality other than some light loss. The downside was that since the image had already been formed by the glass in front of the iris, no anamorphic artifacts would be visible. What you have essentially is an image that is squeezed 'in post.' Most long telephoto shots in old anamorphic films like the second image posted were made with rear-anamorphic zooms. So that accounts for the difference in bokeh.>

 

>What you have essentially is an image that is squeezed 'in post.' Most long telephoto shots in old anamorphic films like the second image posted were made with rear-anamorphic zooms. So that accounts for the difference in bokeh.

 

Would it be any creative consideration to squeezed one image in the 'post' and leave another image as it is?

Edited March 16, 2016 by Mathew Collins

[Satsuki Murashige]

Sorry, that was probably a confusing way to phrase things.

 

The point that you should take away is that lenses with anamorphic elements in front of the iris will have anamorphic artifacts, while lenses with them behind the iris will look identical to spherical lenses. So there is really no benefit to shooting with rear anamorphic lenses if you want 'the anamorphic look.'

 

These rear anamorphic lenses are from the era of 35mm shooting and photochemical printing when it was important to get a squeezed image somehow onto the original negative so the film could be contact printed without generation loss. In the digital age, they have no purpose.

[Satsuki Murashige]

I think in this case, it was most likely a practical consideration of needing to use a zoom lens for this particular shot, either because the shot has a zoom in it or because they had to shoot quickly and the zoom lens was more efficient to get the coverage required.

[Ari Michael Leeds]

Oh, no, cue the counting of how many blades on the aperture diaphragm, here B)

[Mathew Collins]

 

Thank you Satsuki.

 

[Adam Frisch FSF]

There are very few zooms that have front or middle anamorphic elements, i.e. making visible ovals. The AWZ and LWZ from Panavision are two (great lenses, btw), I think Cooke's new anamorphic zoom does it and the Hawk zoom. That's it, pretty much. None of the old "anamorphizied" Angeniuex or even the new anamorphic Optimos do it.

[Satsuki Murashige]

Interestingly, some of the recent experiments with oval iris apertures on spherical lenses by the folks at Vid-Atlantic and Richard Gale at Dogshidt Optics seem to indicate that a lot of the anamorphic look is a result of the iris shape. Which would explain why the anamorphic element needs to be in front of the iris.

[Carl Looper]

Interestingly, some of the recent experiments with oval iris apertures on spherical lenses by the folks at Vid-Atlantic and Richard Gale at Dogshidt Optics seem to indicate that a lot of the anamorphic look is a result of the iris shape. Which would explain why the anamorphic element needs to be in front of the iris.

 

Indeed. Bokeh is that which the so called "circle of confusion" describes, and this circle of confusion (or bokeh) is a function of the iris. In the case of anamorphics suitably placed it can become an oval of confusion :) The circle of confusion (or oval of confusion) used in depth of field calculations is just an arbitrary sized circle/oval of confusion, defined for whatever point spread (or amount of softness) one deems as acceptable. In other words, by defining a particular circle/oval of confusion, your bokeh, within the corresponding depth of field, will be the same size or smaller than the defined circle/oval of confusion.

 

At the precise focus distance a point of light, will produce a bokeh (or circle/oval of confusion) the diameter of zero (or as close to zero as a lens is able to achieve).

 

C

Edited April 10, 2016 by Carl Looper

[Carl Looper]

I'm not sure what the purpose of an oval shaped iris would be on a spherical lens. It would only make sense if it was for an anamorphic lens. The oval shaped iris would undo the oval shaped bokeh that would otherwise result when using an anamorphic lens (on the scene side of the iris). The alternative is to use a circular iris, but on the scene side of the anamorphic lens. In this way a circular bokeh (from the scene side iris) gets squeezed during capture and unsqueezed during projection, resulting in the reconstruction of a circular bokeh.

 

Unless I'm entirely mistaken of course (having rarely used anamorphics).

 

C

Edited April 10, 2016 by Carl Looper

[Satsuki Murashige]

The purpose is to alter the bokeh character of the spherical lens to mimic the visual character of an anamorphic lens without using expensive bent glass. Due to these tests, I'm beginning to question the received wisdom that all of the magic of an anamorphic lens comes from the cylinders.

 

The bokeh character of a lens is more than just the shape of out-of-focus point sources in the far background. With anamorphic lenses, there are really interesting optical artifacts going on in the fall-off from the plane of focus. The vertical stretching gradually increases as the focus falls off, so someone walking into focus starts skinnier than normal and then subtly expands horizontally to normal as they come into focus. With very shallow focus shots, the near background rolls off from 0 squeeze to increasingly more squeeze.

 

Some examples I shot recently with a rehoused 1.5x Iscorama anamorphic lens:

 

 

More images here: https://flickr.com/photos/18675976@N03/sets/72157663948275623

 

I had always assumed that these things were purely due to the magic of cylindrical glass elements (and some artifacts like flare still are), but it seems the way light bends around the shape of the iris is a lot more important than I had previously supposed. Anyway, I thought it was interesting.

[Carl Looper]

The purpose is to alter the bokeh character of the spherical lens to mimic the visual character of an anamorphic lens without using expensive bent glass. Due to these tests, I'm beginning to question the received wisdom that all of the magic of an anamorphic lens comes from the cylinders.

 

The bokeh character of a lens is more than just the shape of out-of-focus point sources in the far background. With anamorphic lenses, there are really interesting optical artifacts going on in the fall-off from the plane of focus. The vertical stretching gradually increases as the focus falls off, so someone walking into focus starts skinnier than normal and then subtly expands horizontally to normal as they come into focus. With very shallow focus shots, the near background rolls off from 0 squeeze to increasingly more squeeze.

 

Some examples I shot recently with a rehoused 1.5x Iscorama anamorphic lens:

image.png

 

image.png

 

More images here: https://flickr.com/photos/18675976@N03/sets/72157663948275623

 

I had always assumed that these things were purely due to the magic of cylindrical glass elements (and some artifacts like flare still are), but it seems the way light bends around the shape of the iris is a lot more important than I had previously supposed. Anyway, I thought it was interesting.

 

Thanks Satsuki.

 

I'm not completely familiar with anamorphics but getting my head around it.

 

My understanding of this is that there are effectively two irises in an anamorphic setup. The first iris is the circular cut of the cylindrical lens. The second is the conventional iris behind the lens. The play off in shape, as I understand it, becomes a function of both, and is inter-dependant on the focus distance used for the point source.

 

In a spherical lens setup there is no interposed cyclinder between both irises so there would no such play off in shape.

 

Regardless of the particularities of anamophic bokeh, my understanding of the term "bokeh" is that it comes from the Japanese word for blur: "boke". (Wikipedia being my source here). But my technical understanding of bokeh (ie. what we see) is that its literally just another word for "circle of confusion". That used in DOF calculations is properly called "circle of least confusion", or "least circle of confusion", but in DOF calculators it's often abbreviated to just "circle of confusion" (which can confuse things). Circle of confusion proper will refer to all the sizes of bokeh, from zero size to visible size.

.

C

Edited April 10, 2016 by Carl Looper

[Satsuki Murashige]

Nope, there's only one iris in an anamorphic lens. It appears elliptical when looking through the front element because the cylindrical anamorphic element squeezes the light horizontally. The squeeze factor depends on the curvature of the anamorphic element, 2x is most common. So the iris is actually a normal circular shape (which you can see if you look through the rear element), it just appears squeezed due to the distortion created by the glass.

 

This is what causes the vertical stretching in the bokeh. What those crazy folks at Vid-Atlantic figured out is that you can take a cheap photo lens, substitute an elliptical aperture in the form of a Waterhouse stop and get a very similar 'anamorphic' effect without using expensive cylinders. There are downsides of course.

 

*You don't get a squeezed image, so there is no gain in horizontal field-of-view.

 

You need to physically replace the stop if you want to change depth of field. There are also vignetting issues. But for $$ versus $$$$$, it's an interesting solution.

 

*Added for clarity.

[Carl Looper]

Actually the more I think about anamorphic bokeh it's not the circular cut of the cylinder but the fact that a point source becomes a spherical wave as it expands out from a point. It's circular anyway.

 

C

[Carl Looper]

What those crazy folks at Vid-Atlantic figured out is that you can take a cheap photo lens, substitute an elliptical aperture in the form of a Waterhouse stop and get a very similar 'anamorphic' effect without using expensive cylinders. There are downsides of course. You need to physically replace the stop if you want to change depth of field. There are also vignetting issues. But for $$ versus $$$$$, it's an interesting solution.

 

What?

 

Changing the shape of the iris and not the lens, squeezes the image? Is that what you are saying? Or it just produces oval shaped bokeh, and the image remains unsqueezed. I can understand the latter, or are you suggesting the former? I must investigate that. Its a hole in my theoretical grasp.

 

C

Edited April 10, 2016 by Carl Looper

[Satsuki Murashige]

Btw, I can vouch for the origin of the term 'bokeh.' My mother always used to say 'watashi wa boketeru' when she forgot something important. We understood it to mean that she was unfocused and distracted.

[Carl Looper]

Btw, I can vouch for the origin of the term 'bokeh.' My mother always used to say 'watashi wa boketeru' when she forgot something important. We understood it to mean that she was unfocused and distracted.

 

I'm feeling very boketeru right now. It's back to the drawing board for me.

 

C

[Satsuki Murashige]

 

What?

 

Changing the shape of the iris and not the lens, squeezes the image? Is that what you are saying? Or it just produces oval shaped bokeh, and the image remains unsqueezed. I can understand the latter, or are you suggesting the former? I must investigate that. Its a hole in my theoretical grasp.

 

C

The in-focus areas remain normal. The out-of-focus areas take on the shape of the iris, as with a Waterhouse stop.

[Carl Looper]

The in-focus areas remain normal. The out-of-focus areas take on the shape of the iris, as with a Waterhouse stop.

 

Ah okay. Thanks. That's making more sense to me.

[Satsuki Murashige]

The in-focus areas in an anamorphic image look more or less the same as a spherical image as well.

[Satsuki Murashige]

To clarify, the cylindrical element in an anamorphic lens is a cross-section of a cylinder. The single-axis curvature of this element, as opposed to a spherical cross-section, is what causes the the image to be squeezed only on the horizontal axis. A spherical cross-section would be convex and thus distort both horizontally and vertically, as with a wide angle adapter.

[Carl Looper]

The in-focus areas in an anamorphic image look more or less the same as a spherical image as well.

 

By which you mean, when projected back though the recipriocal anamorphic lens.

 

Anyway the upshot is that as a point source becomes further away from the focus plane, the more vertically stretched (or horizontally squeezed) it becomes (in projection). Back towards the focus plane, the point source becomes circular (in projection) but also too small to see.

 

On the film itself (prior to being unsqueezed) the in-focus bokeh (so to speak) will be horizontally squeezed (but too small to see) and the out of focus bokeh even more so.

 

C

Edited April 10, 2016 by Carl Looper

[Satsuki Murashige]

By which you mean, when projected back though the recipriocal anamorphic lens.

 

Or when unsqueezed digitally by the same factor, yes.

 

 

Anyway the upshot is that as a point source becomes further away from the focus plane, the more vertically stretched (or horizontally squeezed) it becomes (in projection). Back towards the focus plane, the point source becomes circular (in projection) but also too small to see.

 

On the film itself (prior to being unsqueezed) the in-focus bokeh (so to speak) will be horizontally squeezed (but too small to see) and the out of focus bokeh even more so.

 

C

Yes.

 

Also, as David Mullen and others have pointed out in the past, Panavision anamorphics have a special design where they progressively increase the squeeze factor of the lens when focused close. This is because simple anamorphic lenses actually squeeze less as they focus closer, resulting in fat faces (or 'anamorphic mumps') when desqueezed at a constant ratio. The result is that while faces and other objects in focus appear normal at all focus marks, the bokeh will appear extra squeezed. Hence why the Panavision lenses are so prized. Not sure how other modern anamorphic lens designs have gotten around this. Possibly by making their cylindrical elements irregularly shaped? I'm sure these are all carefully guarded trade secrets.

 

I get around this problem on my 'simple' Iscorama by desqueezing by a factor of 1.4x. So people look normal when focused at around 6-10', macro shots with a diopter look slightly fat, and landscapes look slightly skinny at infinity.

[Carl Looper]

Or when unsqueezed digitally by the same factor, yes.

 

 

Yes.

 

Also, as David Mullen and others have pointed out in the past, Panavision anamorphics have a special design where they progressively increase the squeeze factor of the lens when focused close. This is because simple anamorphic lenses actually squeeze less as they focus closer, resulting in fat faces (or 'anamorphic mumps') when desqueezed at a constant ratio. The result is that while faces and other objects in focus appear normal at all focus marks, the bokeh will appear extra squeezed. Hence why the Panavision lenses are so prized. Not sure how other modern anamorphic lens designs have gotten around this. Possibly by making their cylindrical elements irregularly shaped? I'm sure these are all carefully guarded trade secrets.

 

I get around this problem on my 'simple' Iscorama by desqueezing by a factor of 1.4x. So people look normal when focused at around 6-10', macro shots with a diopter look slightly fat, and landscapes look slightly skinny at infinity.

 

Thanks Satsuki. Much appreciated.

 

C