Front Element Anamorphic Bokeh Aspect

[Gavin Greenwalt]

Is there a rule for how anamorphic aspect affects the bokeh aspect?

Does it remain 1:1 or does it change based on the focus? As you rack does the bokeh condense back down to a circle and expand beyond the anamorphic aspect as it falls further and further out of focus?

Does F/Stop have an effect on its aspect?

 

[Pietro Impagliazzo]

Does F/Stop have an effect on its aspect?

You mean the vertical size of the bokeh, right?

 

[David Mullen ASC]

Yes, the amount of squeezing in the background changes as it comes into focus. Even as a kid, I noticed shots in "Star Wars" where the background would suddenly get skinny-looking as it went out of focus when they rack-focused to something in the foreground.

Stopping down and thus increasing depth of field helps reduce this affect because less of the frame is so out-of-focus.

Obviously a circle would become normal-looking as you focused on it or else no one would ever look normal in an anamorphic movie.

 

[Pietro Impagliazzo]

David, I think Gavin was asking if the bokeh increased in vertical size if we opened the iris.

Something like this:

I don't know If I misunderstood his question... But If I did, well, now I'm curious myself.

Perhaps the understanding of this is directly related to understading how an anamorphic lens squeezes the image.

 

[David Mullen ASC]

Yes, as the circle goes more and more out of focus, whether due to where you set the focus or because the depth of field is getting shallower and shallower, the squeezing effect is increased up to a limit, after which it is stable.

 

[Gavin Greenwalt]

Yep that was exactly my question. And it's what I remember seeing as well but it's difficult to tell as the cof gets smaller and smaller. "I think it's getting rounder... " . Anyone know where one might find an equation to predict this change in aspect ratio? I can't seem to find any non-uniform DOF functions. Nobody seems very interested in accurately modeling this stuff.

Or is it a function of the refractive index of the front element and the distance of the internal elements? In which case it would vary pretty dramatically from lens manufacturer to lens manufacturer.

I guess I need to do more research into why it squeezes in the first place. It's a brain teeze just by its very existance in the first place. I think I need to understand the fundamentals of what's causing it to better hone my investigation.

Step 1) Build a virtual anamorphic lens to play with.

 

[David Mullen ASC]

Usually before an anamorphic show, the efx people will shoot a grid chart or something with the anamorphic lenses at different stops being used, to chart the barrel distortion and focus distortions so they can match their effects to it.

 

[Gavin Greenwalt]

Yeah that could be a problem on this one. I'm going to have cg stuff inches from the lens all the way out to infinity with almost no way to split it up into seperate focal depths (Sparks).

Usually I would have Fore/mid/background to let you 'cheat it in' to its surrounding neighborhood. Just use a custom convolution kernel for each element. I'm going to have to get the focus in the render I think for this one unfortunately.

 

[David Mullen ASC]

It's funny now for me to see the three "Star Wars" prequels and see the bokeh of the out-of-focus city lights look squeezed in "Phantom Menace" and spherical for the other two because only the first one was shot in anamorphic.

 

[Gavin Greenwalt]

I've never heard why they switched? Chickened out to make post easier? The switch to 2/3" Digital? Got bored?

---

Well... I hand crafted an anamorphic front element. And still don't understand why it works. But it does! However some weird stuff happened when I tried rack focusing that wasn't consistent with anything I've seen before. So I think my model needs a bit more tweaking before it can be trusted as an experimental source of knowledge.

I did learn something though. I had always assumed that Anamorphics expanded the image vertically instead of compressing it horizontally. Good to know.

 

[Steven Caesare]

I know David is participating in this thread already, but his post from quite some time ago in the "Ask David Anything" thread of one of the Star Trek movies illustrated the anamoprhic bokeh quite nicely.

(And don't despair about the prospect of having to read thru that ginormous thread to find it... it's one of the most fascinating and educational threads I've ever read on the subject of movie making. As a complete outsider, it's changed the way I appreciate movies.)

 

[Vico Martin]

yeeap!, David is amazing!

I heard about the bokeh effect but never work with enough close depth of field to apreciate it filming.

Gonna see Phantom menace again and get funny like David.

 

[David Mullen ASC]

I've never heard why they switched? Chickened out to make post easier? The switch to 2/3" Digital?

The second reason.

Here are those shots I posted from Star Trek 5 and 6 to show the difference between shooting the bridge set in low-light on anamorphic versus Super-35 spherical:

 

[XiaoSu Han]

wow, never realized that it looks so different...

anyone saw JCVD recently? I think that they overused soft filters quite a bit, and the anamorphic breathing while rack focusing is just ugly at times...