Topics

The Making of the CML NFA Tests 2020 - part 2


Adam Wilt
 

Monday morning we showed up at the NFA and got stuck in. We laid out the lighting and camera idents so we could cycle through them easily: 

The gap in the camera list is the BMD Ursa’s ident, which was being used at the moment.

This worked as a status board or dynamic checklist. Cameras and lights moved up on the table as they were used; lights were arranged in priority order and those that hadn’t yet arrived were inset to the right to mark them off. We could easily tell at a glance where we were in the tests, what was to come next, and what kit we were still waiting for.

And yes: when we started, not all the lights had arrived. Geoff was busy behind the scenes talking to suppliers, and gear keep coming in throughout the week. In the end only one light never showed up, and we had a few others not initially planned for — hence the handwritten idents in some of the shots.

Right off the bat we deviated from plan: we mounted the trace frame (a wooden frame with Lee 216 diffusion) horizontally instead of vertically:

 

View from behind the models’ seats towards the camera

I had missed the word “vertical” in the prep document, an error I attribute to severe jet-lag.

The diffusion was about 2.6 meters from the models, not 1.2 meters as called for in Geoff’s instructions; we worked visually to maintain the angle of the key and the overall look of the lighting rather than focusing on hitting the numbers precisely.

That aside, we rigged the set much as Geoff prescribed, with a bounce card opposite the diffusion and with two sets of BB&S Pipeline LEDs — one 3200 K, the other 5600 K — rigged as backlights. 

 


The Pipelines not only separate the models from the background, they provide consistent color across all tests, serving as a subtle color reference.

The DSC Labs ChromaMatch chart was mounted in the same plane as at the models, while the other charts and test idents were attached to a swing-away flag so they could be positioned in the models’ laps, to keep all subjects of interest close to the same focus distance:

  



Both DSC charts were angled away from the key to eliminate flare. The other charts had matte surfaces so shadow-lifting flare was just a fact of life.

We spent the rest of our first day shooting just the tungsten 3200 K reference clips (DeSisti 5ks provided the light and one layer of 1/8 CTB cooled it off to precisely 3208 K). Fortunately we got faster on the subsequent days, developing a production-line “flow” to cycle through the cameras and their color settings. Once we moved away from tungsten lamps, the LEDs were switchable between 3200 K and 5600 K at the spin of a dial, and Erik rearranged our camera list by mounting-plate type in order to spend the least time switching tripod mounting adapters.

It soon became clear that we were not going to reach 100 fc (~1076 lux) with any of the LEDs, so we opted instead to light to 25 fc and open our aperture two stops. Even so, we fell short with a few fixtures, and compensated on those tests by moving the diffusion a meter closer to the subjects. At that, we sometimes had to move the lamp very close to the diffusion to get enough light, so overall we didn’t always have the consistency of lighting that Geoff had asked for.

 

More concentrated illumination, due to a lower-output lamp closer to the 216.

We also managed to shoot both the film stocks and the Sony FX9 one stop more exposed than intended. The film exposure was my mistake; I misread Geoff’s instructions and set the 500T stock’s aperture one stop wider than the digital cameras, and the 250T stock’s iris two stops wider. 

The FX9 was a more interesting mixup: its electronic shutter was set OFF, and Erik assumed that meant a normal, film-style 180º setting with a 1/50 sec. exposure. It was only when he checked clip metadata in Sony Catalyst that he saw it was 360º. For video people, “no shutter” means a 360º shutter with 1/25 sec. exposure, and Sony’s engineers, coming from a video-oriented background, made that the no-shutter default. 

 


Erik had a checklist on his iPhone and meticulously reviewed camera settings before every take, but even that wasn’t enough to bridge the film/video cultural divide!

Fortunately neither mistake was critical: both film stocks and the FX9 had more than enough headroom to fully capture chart colors and skintones without a problem. 



Adam Wilt
technical services
Vancouver WA USA (no, not that Vancouver, the other one)


Geoff Boyle
 

Can’t resist 😊

 

1.2 mtrs to 2.6mtrs

100FC to 25FC

 

A perfect demonstration of the inverse square law. Something that is too technical to be taught in most film schools…

 

cheers
Geoff Boyle NSC FBKS
EU based cinematographer
+31 637155076

www.gboyle.nl

www.cinematography.net

 

 


Geoff Boyle
 

“Geoff was busy behind the scenes”

 

Yeah, from my hospital bed with a nose feeding tube and 3 different pain killers.

 

Sometimes my “help” was a bit strange 😊

 

cheers
Geoff Boyle NSC FBKS
EU based cinematographer
+31 637155076

www.gboyle.nl

www.cinematography.net

 

 


Adam Wilt
 

1.2 mtrs to 2.6mtrs

100FC to 25FC

 A perfect demonstration of the inverse square law. 

Yes, however, the large area-lit diffusion was not a point source*, so falloff with distance (or increase with closeness) was not as dramatic. When we moved the diffusion in from 2.6m to 1.6m we went from around 15 fc to 25 fc; inverse-square would have seen us get to 40 fc.

Still, point taken: read the fine manual, follow the instructions FFS, and we’ll be more likely to get closer to the target levels!

*(Had we used an infinitely-long line source, like BB&S Pipelines end-to-end, falloff would have been inverse-linear: doubling the distance would halve the light. With an infinitely-wide planar source, like a sheet of Lee 216 bisecting the universe, falloff would not occur at all: illumination would not vary with distance. The closer the 8x4 diffusion got to the subject, the more it acted like a planar source; the farther it got, the more it behaved like a point source.)

Cheers,
Adam Wilt
technical services
Vancouver WA USA (no, not that Vancouver, the other one)


jmartinson@...
 

I have been looking for the math behind line and planar light sources Adam referred to:

 

Had we used an infinitely-long line source, like BB&S Pipelines end-to-end, falloff would have been inverse-linear: doubling the distance would halve the light. With an infinitely-wide planar source, like a sheet of Lee 216 bisecting the universe, falloff would not occur at all: illumination would not vary with distance. The closer the 8x4 diffusion got to the subject, the more it acted like a planar source; the farther it got, the more it behaved like a point source.

 

I haven’t seen it covered in any lighting textbook and want to change that.  I am interested in the math that predicts the falloff as the light transitions to acting as a point source as distance increases.

 

Joe Martinson

Filmmaker Institute

jmartinson@...

 

From: cml-lighting@... <cml-lighting@...> On Behalf Of Adam Wilt
Sent: Friday, March 6, 2020 10:28 AM
To: cml-lighting@...
Subject: Re: [cml-lighting] The Making of the CML NFA Tests 2020 - part 2

 

1.2 mtrs to 2.6mtrs

100FC to 25FC

 A perfect demonstration of the inverse square law. 

 

Yes, however, the large area-lit diffusion was not a point source*, so falloff with distance (or increase with closeness) was not as dramatic. When we moved the diffusion in from 2.6m to 1.6m we went from around 15 fc to 25 fc; inverse-square would have seen us get to 40 fc.

 

Still, point taken: read the fine manual, follow the instructions FFS, and we’ll be more likely to get closer to the target levels!

 

*(Had we used an infinitely-long line source, like BB&S Pipelines end-to-end, falloff would have been inverse-linear: doubling the distance would halve the light. With an infinitely-wide planar source, like a sheet of Lee 216 bisecting the universe, falloff would not occur at all: illumination would not vary with distance. The closer the 8x4 diffusion got to the subject, the more it acted like a planar source; the farther it got, the more it behaved like a point source.)

 

Cheers,

Adam Wilt
technical services
Vancouver WA USA (no, not that Vancouver, the other one)

 


Glenn Lee Dicus
 

I’m not sure how useful this would be, but here is a stab at it.


θ = angle of light.

θ=2*Tan-1(R/D)

Where R is the radius of the fixture lens, D is the distance.

As D approaches ∞ (infinity), θ approaches 0, a parallel line and is then considered an ideal point source.  As D approaches 0, θ approaches 180 degree’s.


I’m going to hate myself in the morning.


Glenn Lee Dicus
IATSE (ICG Local 600)
310.903.7069
glenndicus@...




On Mar 6, 2020, at 11:15 AM, jmartinson@... wrote:

I have been looking for the math behind line and planar light sources Adam referred to:
 
Had we used an infinitely-long line source, like BB&S Pipelines end-to-end, falloff would have been inverse-linear: doubling the distance would halve the light. With an infinitely-wide planar source, like a sheet of Lee 216 bisecting the universe, falloff would not occur at all: illumination would not vary with distance. The closer the 8x4 diffusion got to the subject, the more it acted like a planar source; the farther it got, the more it behaved like a point source.
 
I haven’t seen it covered in any lighting textbook and want to change that.  I am interested in the math that predicts the falloff as the light transitions to acting as a point source as distance increases.
 
Joe Martinson
Filmmaker Institute
 
From: cml-lighting@... <cml-lighting@...> On Behalf Of Adam Wilt
Sent: Friday, March 6, 2020 10:28 AM
To: cml-lighting@...
Subject: Re: [cml-lighting] The Making of the CML NFA Tests 2020 - part 2
 
1.2 mtrs to 2.6mtrs
100FC to 25FC
 A perfect demonstration of the inverse square law. 
 
Yes, however, the large area-lit diffusion was not a point source*, so falloff with distance (or increase with closeness) was not as dramatic. When we moved the diffusion in from 2.6m to 1.6m we went from around 15 fc to 25 fc; inverse-square would have seen us get to 40 fc.
 
Still, point taken: read the fine manual, follow the instructions FFS, and we’ll be more likely to get closer to the target levels!
 
*(Had we used an infinitely-long line source, like BB&S Pipelines end-to-end, falloff would have been inverse-linear: doubling the distance would halve the light. With an infinitely-wide planar source, like a sheet of Lee 216 bisecting the universe, falloff would not occur at all: illumination would not vary with distance. The closer the 8x4 diffusion got to the subject, the more it acted like a planar source; the farther it got, the more it behaved like a point source.)
 
Cheers,
Adam Wilt
technical services
Vancouver WA USA (no, not that Vancouver, the other one)
 



Adam Wilt
 

I have been looking for the math behind line and planar light sources Adam referred to...

I remember it from first-year physics, but that textbook is long gone. I see the math for electric field computation for point, line, and planar sources in Tipler, chapter 22: https://books.google.com/books?id=AttDBYgLeZkC&pg=PA728&lpg=PA728&dq=tipler+physics+22-1&source=bl&ots=mFyOs3W_U4&sig=ACfU3U2EHbupo0i7M7acPFr138bwMfUXRw&hl=en&sa=X&ved=2ahUKEwj88JeCjInoAhXaJDQIHfQVC0kQ6AEwBnoECDoQAQ#v=onepage&q=tipler%20physics%2022-1&f=false

Analogous math works for light and gravity, too (and explains why there is no net gravitational force inside a hollow planet for Ross Rocklynne fans, but I digress…). But I’m having a hard time finding user-friendly references that apply to lighting. I did come across a couple, for those less intent on deriving Maxwell’s equations:

http://www.cs.cmu.edu/afs/cs/academic/class/15462-s12/www/lec_slides/lec13.pdf (statement of falloff for point, line, and planar sources)
https://blog.kasson.com/the-last-word/light-falloff-vs-distance-for-round-sources/ (experimental measurements close to and far from a large round source)

Cheers,
Adam Wilt
technical services
Vancouver WA USA (no, not that Vancouver, the other one)


Glenn Lee Dicus
 

θ=2*Tan-1(R/D)

I’ve been thinking and it occurred to me that there is an interesting application of an equation of this sort.  

For example, you had an 6k on a scissor lift shooting through a 4x4 of 216 on the day of production.  However, later in the edit, it is determined that some pickups are needed, some inserts and/or medium closeups.

Where do I put the M18 the reduced budget allows for and what size 216 frame should I choose?

As it turns out, equating the equations for each light source you want to match, and solving for distance, you get the following.

D2=D1*R2/R

The distance is directly proportional to the ratios of the two lenses. In this case, the two frames.  We know the size of the first frame, 4 feet.  And we know what frames are available, let’s say they are 4x4, 2x2 and 1x1 frames. 

All we have is an M18 for the day of pickups.  So, we’ll error on the side of getting as close as possible with the 1x1..  We can always scrim down.

So, now the equation becomes:

D2=D1*1/4

We place the 1x1 25% the distance.  And if you notice, it is the ratio of the two lenses which dictate the distance.  In other words, if a lense is twice as large as the one available, place it half the distance. 

And here is the fun one, if it is 5 times the size, like a 20x20 with the sun as it’s source, then it’s replacement should be a 4x4 at 1/5th the distance.

Correct me if I’m wrong.  Unless I’m grossly misguided, this seems to be too simple and must be common knowledge, no?



Glenn Lee Dicus
IATSE (ICG Local 600)
310.903.7069
glenndicus@...




On Mar 6, 2020, at 11:15 AM, jmartinson@... wrote:

falloff