Home of Professional Cinematography since 1996

Mirror and Polarisation Issues

Published : 14th July 2012


I thought that you may all be interested in the pretty thorough tests we did a couple of days ago at the Gothenburg 3d workshop.

We had shot a number of outdoor set-ups and were having serious polarisation issues with the snow and ice and wet roads as well as car windows.

We decided that we needed to test for solutions.

I had a couple of 1/4 wave retarders that Schneider had given me and also some Tru-Pols.

These were all used in front of the lens and behind the mirror in all kinda of combinations.

There were 3 clear solutions to polarisation problems with the mirrors.

A 1/4 wave filter in front of everything worked really well with almost no light loss if uses at exactly the right angle.

Of course this would mean a huge custom made filter that was right with that rig in that location but may not work with other rigs or locations.

A Tru-Pol in front of everything matched the two cameras at any angle, the effect of polarisation of course varied as we turned the filter but the effect was exactly the same in each camera. So, a good solution BT again needing a huge filter and of course a 2 stop light loss.

Finally a pair of Tru-Pols one on each lens behind the mirror could be rotated to match and get the same look in each camera. Of course we could just use one filter on the least polarised lens and match but we then had a 2 stop difference between lenses.

So it seems that the approach that I was already using, a one stop pola on each lens, was the best way to go!

Geoff Boyle
Gothenburg
Teaching and Learning


Thanks for sharing this Geoff. Very interesting, informative and truly useful.

Alister Chapman
DoP - Stereographer - UK based
Office +44 1344 483748
Mobile +44 7711 152226
US Mobile 405-822-1764

www.ingenioustv.com


Geoff, we at Schneider have been making some very large 1/4 wave "windows" (10" x 23") for several current features. In most cases we can deliver these with water white clear glass with anti-reflection coating. Delivery is usually a couple of days to a couple of weeks and they tend to keep the junk off those expensive beam splitters as an added benefit. And yes, I do have a vested interest as we manufacture these.

Bob Zupka
MPTV Product Manager
Schneider Optics


>> we at Schneider have been making some very large 1/4 wave "windows" (10" >> x 23") for several current >> features.

I have just completed the Advanced 3D course at BAFTA/Twickenham and we used a 1/4 WR from Eric Deren at Dzignlight and it worked a treat (www.advanced3d.co.uk).

So, thanks Eric and onwards with the 1/4 WRs!

Phil Streather
Stereo 3D Producer and Consultant
The Old Vicarage
Leigh on Mendip
Somerset
England BA3 5QG

Tel: 44 1373 813 301
Fax: 44 1373 813 379
Mob: 44 7771 783603
Skype: philstreather
Web: www.plf.cc


I am under the impression that 21st century 3D are already dealing with that.

Their rig carries glass in front of everything and the only reason I can imagine is polarization. As a matter of fact I thought they are using a retarder, but if does not work on all set ups....Jason, would you please enlighten us?

No affiliation etc etc.

I have not worked with this rig either, only seen their pics and demo video

Argyris Theos
DoP
Athens Greece
+306944725315
Skype: Argyris.Theos


Thank you Argyris for pointing this out. 21st Century 3D was the first major 3D production company to incorporate the use of the 1/4 wave retarder and the first to commission their manufacture from Schneider Optics for stereoscopic cinematography.

Putting polarizers on the lenses may not solve the inherent problem with most beam splitters as the incoming light has already travelled through and reflected off of the mirror. It should be too late at that point as the image would already have the optical problem we are trying to prevent. I would like to know more about Geoff's research and see samples to understand his findings better. It would also seem that the light loss from the polarizers in combination with that of the mirror would be a deal killer for most applications.

The 1/4 wave is really the ideal choice when all compromises are considered. In fact, DP John Schwartzman, ASC and stereographer Rob Engle found the results of 21st Century 3D's solution so desirable they have incorporated it into the 3ality rigs being used on a major motion picture they are shooting now with 3ality and 21st Century 3D.

I've included a few links that show the 1/4 wave retarder in action. Here is a Mini Cooper commercial we recently shot on RED ONE MX and Phantom HD Gold with two of our BX3 beam splitters.

http://www.youtube.com/watch?v=5MCcwq_iZOg

If you don't have a 3D laptop or TV that is compatible with the YouTube 3D player, set it to cross eye or parallel view to watch in full colour (or just do that to compare the two images side by side).

The commercial features lots of chrome and shiny surfaces, liquids, black actors with hot specular highlights and other elements that commonly exhibit unacceptable artifacts when photographed with some 3D beam splitters. Obviously visual effects and a very costly post finish have been done on this spot, but you can easily see that we started with well matched material.

For a more "raw" comparison, here is last year's demo reel from 21st Century 3D. Virtually no colour finish has been done, other than simple stuff with the 3 way colour corrector in Final Cut Pro.

http://www.youtube.com/watch?v=Hz6fAFXOncc

There is one clip in there that was not shot on a BX series rig, it’s the aerial ski shot, that was side by side. We will be posting our 2011 demo reel later today so check back to see more examples of the successful use of 1/4 wave.

It should also be noted that 21st Century 3D utilizes a custom formulated coating for our mirrors, developed by our team in New York and manufactured in New York. In addition to the 1/4 wave front filter, this mirror mitigates the polarization and colour issues commonly associated with stereoscopic beam splitters. The mirror is available only in 21st Century 3D BX2, BX3 and BX3.5 camera rigs.

Jason Goodman
CEO
21st Century 3D New York
505 8th Avenue #1006
New York, NY 10018
212-244-8585

http://www.21c3d.com


>> I am under the impression that 21st century 3D are already dealing with that.

Jason Goodman's 21st Century 3D rig is, AFAIK, the first rig to incorporate a 1/4 wave retarder into the design.

Independently, I tested and then used a 1/4 wave retarder on a Swiss Rig for an $8M feature film last February-March. My original intent was to use
it just for reflective "problem shots", but I found that the wave plate fixed so many things that we ended up using it 100% of the time.

Additionally, the wave retarder is inexpensive, about $500, so it protects the significantly more expensive mirror behind it. Just last night, on the same major motion picture Jason mentioned, an exploding practical hit and scratched the wave retarder instead of damaging the mirror. Annoying, certainly, but not a show-stopper.

As long as the 1/4 wave retarder is mounted correctly to prevent mirror
box inter-reflections, I see no reason not to use one on every shot.

Eric Deren
Dzignlight Studios
VFX & Animation Design
www.dzignlight.com
Atlanta, Los Angeles


I would appreciate it if all of you could elaborate a bit.

Geoff which circumstances do you think a retarder might not be good for the shooting?

Eric what do you mean by " As long as the 1/4 wave retarder is mounted correctly to prevent mirror box inter-reflections, I see no reason not to use one on every shot."?

Does the retarder pose any special problems or is it just the standard trouble caused by any glass element?

Best

Argyris Theos
DoP
Athens Greece
+306944725315


>> Eric what do you mean by " As long as the 1/4 wave retarder is mounted correctly to prevent mirror box

>> inter-reflections, I see no reason not to use one on every shot."?

Most mounting solutions I have seen have the 1/4 wave retarder perpendicular to the optical axis, which in certain lighting situations can reveal the lenses in reflection, even with a good AR coating on the 'tard.
FWIW, the Swiss Rig front-glass mount is at an angle, which solves that problem, but the particular angle that was chosen creates some other problems. We solved those problems on "Hidden 3D" last year, but a more
elegant solution (the correct angle) has yet to be executed. I believe at least one rig manufacturer is working on this.

There is also some discussion regarding the effectiveness of a 1/4 wave retarder at anything other than a perpendicular angle, and extensive testing should be done. From the limited testing that Jason Goodman and I
have done, a paper could probably be written about it, because our empirical observations were surprising. Any lurking physics students might find this to be an interesting study.

FWIW, I am glad this 'tard solution is finally officially out of the bag (however politically incorrect the abbreviation is)... maybe now we will see some advancements in this realm and we will all have stronger tools to
shoot S3D movies with.

Eric Deren
Dzignlight Studios
VFX & Animation Design
www.dzignlight.com
Atlanta, Los Angeles


Eric Deren writes:

>> There is also some discussion regarding the effectiveness of a 1/4 wave retarder at anything other than a >> perpendicular angle

The obvious question is, can the glass retardation angle be made other than at the glass perpendicular, so that it retards properly at a pre-determined angle that's less problematic?

Tim Sassoon
SFD
Santa Monica, CA


>> The obvious question is, can the glass retardation angle be made other that at the glass perpendicular, so >> that it retards properly at a pre-determined angle that's less problematic?

Certainly, but our results with an angled glass did not imply a reduction in performance, at least in our limited test. Quite the opposite, actually. My intuition says that there has to be a pending performance hit as the angle changes... we just didn't have the time to find it and define it. Like I said, our results were interesting enough that I believe someone should spend some time fully researching it.

Eric Deren
Dzignlight Studios
VFX & Animation Design
Atlanta, Los Angeles



>> Certainly, but our results with an angled glass did not imply a reduction in performance, at least in our

>> limited test. Quite the opposite, actually.

If you think about a pola on a wideish angle lens, you can imagine that there is no reason to be perpendicular with regard to polarization...but I would think that you would want to be only as far from perpendicular as necessary to avoid the multiple reflections for other reasons.

Any idiosyncrasies or flaws that pop up are likely to be more of a problem the further from perpendicular you get... or put another way, the further from perpendicular you get, the more overall difference between the normal rays (perpendicular) and the furthest away from perpendicular.

OK, I know what I mean - can someone else say it better than that, I hope?

Mark H. Weingartner
LA-based DP & VFX Supervisor


>> If you think about a pola on a wideish angle lens, you can imagine that there is no reason to be

>> perpendicular with regard to polarization....

The quarter wave retarder not a polarizer, though this is a common misconception. The problem light is already linearly polarized by objects in the scene. We are keeping it from being cancelled by another linear
polarizer (the mirror) by converting it to circular polarized light with the quarter wave retarder. The performance of this conversion with an angled wave retarder, and the susceptibility of this converted light to the
natural polarization of the mirror, is what is under investigation... and it's a large question with a lot of variables.

>> If you think about a pola on a wideish angle lens,

We don't really even know if a wide-ish angle lens is or is not somewhat defeating the quarter wave retarder... there are certainly situations that it works less than others. I personally haven't seen a lens-length/wave-retarder correlation yet in my 50+ shooting days with the device. All we know is that, even in the worst situations, having it is better than not having it.

>> or put another way, the further from perpendicular you get, the more overall difference between the normal >> rays (perpendicular) and the furthest away from perpendicular.

I am sure that we can make educated hypotheses all day but it won't replace empirical evidence. It's not a small task... with the right equipment in place it would probably take several days to gather and analyze the data.
More time than I've got, certainly. The best I have is that I know that we don't know.

I will certainly defer if someone has exact experience with the problem we have in front of us, but something tells me that we have a unique problem that most likely hasn't been greatly explored in optics with regards to
getting an image out of the other side.

Eric Deren
Dzignlight Studios
VFX & Animation Design
www.dzignlight.com
Atlanta, Los Angeles


Geoff,

I have a question. When you set up your cameras with a polarizer on each lens, how do you align the polas exactly? Is there a trick, or do you just rotate them to match?

Gratefully,

Alex Falk

"Finally a pair of Tru-Pols one on each lens behind the mirror could be rotated to match and get the same look in each camera. Of course we could just use one filter on the least polarised lens and match but we
then had a 2 stop difference between lenses."


The 1/4 wave retarder sounds fantastic, and the demo I saw of OnSight's seemed to work a treat. But it has got me thinking.

Is it going to have any effect on polarised light coming from the sky? It's not a selective polariser is it? And it all exits the retarder the same? So shooting at right angles to the sun isn't going to produce a darker sky like it would with a normal polariser. Am I understanding this correctly..?

Also, doesn't tilting the retarder (as has been talked about here to cure reflections) effectively alter the thickness relative to the light path thereby influencing it's characteristics? Like, err, isn't the desired effect all down to the final relative phase of the light components, which is related to the birefringence characteristics and the thickness of the refracting material? This thickness would be increased further by the fact that as the retarder gets tilted the angle the light travels at inside the retarder is greater than the incident angle which effectively makes the filter even thicker (and this would be exhibited as a vertical shift in camera) thereby perhaps compromising the effect sought..?

And just to be even more awkward, the way I understand these things, they work by using an anisotropic material that splits the light into two perpendicular components travelling at different speeds and exhibiting different refractive indeces. At right angles to the filter all the light will go straight through so it's no big deal, but if they go through at an angle (such as when the retarder is tilted) these differences would mean that they will travel different paths and exit in slightly different places, thus potentially vertically smearing the image..? An extreme example can be seen with calcite crystals - http://massey.dur.ac.uk/NSEW/calcite.jpg - in which the anisotropic material refracts light, splitting the incoming rays into two components that take different paths during their journey through the crystal and emerge as separate light rays to yield double images.

Just asking...

Karel Bata
Director
London


>> Is it going to have any effect on polarised light coming from the sky?

Yes, but the intended effect (for the sky, at least) is to equalize the values between the transmitted and reflected images when the beamsplitter rig is pointed in a direction where the naturally polarized sky would
create rivalries.

>> It's not a selective polariser is it?

Again, it isn't a polarizer.

>> Also, doesn't tilting the retarder (as has been talked about here to cure reflections)effectively alter the

>> thickness relative to the light path thereby influencing it's characteristics?

Sin/cos calculations would say the difference in thickness is minimal, and also, the entire glass isn't the retarder, just the laminated filter. The glass is just optical glass.

>> At right angles to the filter all the light will go straight through so it's no big deal, but if they go through

>> at an angle (such as when the retarder is tilted) these differences would mean that

Like I said, educated hypothesis can be posted all day, but they won't mean anything until the parameters have been fully tested by someone. If you haven't played with a quarter wave retarder yet I would recommend
getting one and doing some testing before posting anything but questions. Until then, this thread might as well be about politics or who's got the cutest pictures of their kids... except it won't have words like
"birefringence".

Eric Deren
Dzignlight Studios
VFX & Animation Design
www.dzignlight.com
Atlanta, Los Angeles


Karel Bata wrote:

>>The 1/4 wave retarder sounds fantastic, and the demo I saw of OnSight's seemed to work a treat.

-------

For clarity the 1/4 Wave retarder used at The Advanced 3D Training programme at Twickenham Studios was kindly loaned by Eric Deren of Dzignlight. On Sight were the major equipment provider (Alexas, ET rugs etc) for the event but were impressed with the retarder too and no doubt investigating further as I write.

www.advanced3D.co.uk


Phil Streather
Stereo 3D Producer
PLF
The Old Vicarage
Leigh on Mendip
Somerset BA3 5QG

Tel: +44 1373 813301
Mob: +44 7771 783603


As it was explained to me by Stereographer Jason Goodman is the 1/4 wave retarder takes linear polarized light and converts it to circular polarized light diminishing any artifacts with the use of a mirror. No light loss
either.

Seems pretty simple, why haven't I seen this over the last 8 3D jobs I've done?

Dane Brehm
DIT: Data Tech
Oakland, CA


Dane Brehm wrote:

>> Seems pretty simple, why haven't I seen this over the last 8 3D jobs I've done?

Because 3D craftsmanship is not yet fully turned into common knowledge and experienced stereographers try to keep a few trump cards up their sleeves.

Remember the rolling shutter ?
Now it's the 1/4 wave trick.
There are a few more, but shhhh, gentlemen's NDA.

One in the open, just for the fun; On Hugo Cabret's steadiman shots, Larry McConkey had to change his heavy hiking boots for dancing shoes with thinner soles. He needed a better balance feeling than in 2D. Yes.  3D even requires different shoes for a steadicam operator !

All the Best.

/Bernard Mendiburu/
VP Innovation, Volfoni
Paris, Los Angeles


>> "Seems pretty simple, why haven't I seen this over the last 8 3D jobs I've done?"

Because as usual it is the obvious that we fail to recognize.

Argyris Theos
DoP
Athens Greece
+306944725315


I've been shooting nights and sleeping days on the 3D project John Schwartzman is lensing and I posted the other night in a state of slight delirium. I should clarify one thing about my post. 21st Century 3D is not officially involved in this project, only me (the CEO) as 2nd unit stereographer. I am working closely with the 3ality team, the synergy is awesome all around.

Jason Goodman
CEO
21st Century 3D New York
505 8th Avenue #1006
New York, NY 10018
212-244-8585 ext 109

http://www.21c3d.com


Sorry for the delay replying but I've been having some software problems that I think I'll be writing about somewhere....

Anyway, the 1/4 wave worked really well in front of the mirror box but the angle was critical, a very small angle where it was perfect.

This wasn't due to the scene which has no effect that I could see on the issue but is I suspect simply due to the nature of a 1/4 wave retarder and as such I suspect that the "correct" angle will vary from mirror to mirror.
This seems pretty logical and straightforward to me.

Dane asks why he hasn't used one yet, maybe because we still have people in 3D who don't like to share and keep their "secret sauce" to themselves, if so they are going to get steamrollered out of the way soon.

Using a huge pola in front of the mirror only seemed to have one disadvantage as it worked well in matching at nearly every angle, cost!

Individual polas on the lenses worked extremely well but are a bit fiddly, that was a masterpiece of understatement!

If/when I get another 3D movie I will certainly try and get 1/4 wave retarders for the front of the mirror box, in fact I'll do it for a commercial with a big enough budget!

I'll be putting up a webpage or 10 of the results in a week or so.

The rig was setup with zero IA to help us with seeing the differences so I'll post split screens.

Cheers

Geoff Boyle FBKS
Cinematographer
EU Based
Skype geoff.boyle
mobile: +44 (0)7920 143848
www.gboyle.co.uk


>> maybe because we still have people in 3D who don't like to share and keep their "secret sauce" to

>> themselves

There are arguments for both sides, certainly. I personally enjoy sharing. But someone would have to be an idiot to not concede that there are two very rational sides to an issue like this. How many people here
are currently working with gear that is under an NDA? Why do you think that is? Why would someone put an NDA on something? Answer this and you'll have your answer as to why this wasn't talked about much.

The reality is that the quarter wave retarder was mentioned or referenced at least 10 times on CML-3D in the past year... folks just aren't paying attention.

Eric Deren
Dzignlight Studios
VFX & Animation Design
www.dzignlight.com
Atlanta, Los Angeles


Eric, what's an NDA?

Argyris Theos
DoP
Athens Greece
+306944725315


Non Disclosure agreement. A legal document that binds a party from disclosing sensitive or confidential information.


In this case a vendor may require an NDA to be signed by the DP or the production in exchange for testing our their rigs or in order to get their rigs early or test out prototype versions of their rigs.

Daniel Colmenares
(818) 552-4590 Work


A few estimates of cost have come in for 1/4 wave retarders. Not as expensive as we first thought for such a big piece of glass but there is a big price range for the same size filter from vendors. Assuming that these filters are just a 1/4 wave retarder and there’s nothing else incorporated in to it apart from AR both sides. I’m sure there must be differences in quality between manufacturers but I’m not sure what that might be and how to define a good or bad filter. My assumptions are given that its good quality glass, and the orientation is correct what else might affect the performance of the filter ?

Dave Blackham,
Esprit Film and Television,
UK


Things to consider:

How flat is the glass, how clear is the glass, how thick is the glass. How well is the de-polariser applied (often a film applied with a roller). How scratch resistant is it and any AR coatings, is it affected by moisture. How is the edge of the glass finished (any small chips and rough edges will weaken the glass). If it is a plastic film retarder applied to a glass substrate what are its static electricity properties i.e. will it be a dust magnet?

Many brands of Cling Film (saran wrap) act as a 1/4 wave retarder, but I wouldn't recommend it for big budget productions!!

Alister Chapman
DoP - Stereographer - UK based
Office +44 1344 483748
Mobile +44 7711 152226
US Mobile 405-822-1764

www.hurricane-rig.com