You guys watch Joe Don Baker movies?
August 5, 2013 4:12 PM Subscribe
A six-minute documentary snippet discusses Kubrick's camera modifications for special, low-light f/0.7 Zeiss lenses used to film candlelit scenes in Barry Lyndon, now available to rent by aspiring filmmakers.
They sound like a pain in the ass to use (a depth of field you can measure in microns!), but boy would it be fun to try them out. Unfortunately I'm not made of solid gold and so cannot sell any of my body parts for enough money to rent them.
posted by chrominance at 4:26 PM on August 5, 2013 [1 favorite]
posted by chrominance at 4:26 PM on August 5, 2013 [1 favorite]
Can someone explain why this is technically difficult to make? A wide aperture…why is this so advanced? It seems to be a fairly simple...just make the aperture bigger....or?
posted by zardoz at 4:43 PM on August 5, 2013
posted by zardoz at 4:43 PM on August 5, 2013
zardoz: I believe it's due to the limitations of the refractive index of glass. The theoretical largest aperture possible for a glass lens is f/0.5.
posted by zsazsa at 4:51 PM on August 5, 2013 [4 favorites]
posted by zsazsa at 4:51 PM on August 5, 2013 [4 favorites]
Mitchell!
As for why it's difficult to make, it pretty much boils down to "lenses are hard to make as it is, and the bigger they are, the more difficult it becomes."
posted by ShutterBun at 4:53 PM on August 5, 2013
As for why it's difficult to make, it pretty much boils down to "lenses are hard to make as it is, and the bigger they are, the more difficult it becomes."
posted by ShutterBun at 4:53 PM on August 5, 2013
so you can rent the lens the Kubrick got for filming the fake moon landings? awesome!
posted by Dr. Twist at 4:56 PM on August 5, 2013
posted by Dr. Twist at 4:56 PM on August 5, 2013
It turns out that lighting a candle or cursing the darkness is a false dichotomy.
posted by GuyZero at 4:57 PM on August 5, 2013 [1 favorite]
posted by GuyZero at 4:57 PM on August 5, 2013 [1 favorite]
Holy shit, I watched Barry Lyndon for the first time just last night, and would love to see more camera footage with natural light in the same style!
posted by oceanjesse at 4:57 PM on August 5, 2013
posted by oceanjesse at 4:57 PM on August 5, 2013
God, I love Barry Lyndon.
The more I see it, the more I'm convinced that was Kubrick's true masterwork.
posted by Thorzdad at 4:57 PM on August 5, 2013 [3 favorites]
The more I see it, the more I'm convinced that was Kubrick's true masterwork.
posted by Thorzdad at 4:57 PM on August 5, 2013 [3 favorites]
Good grief...imagine being the focus puller on that movie. A 1" depth of field, combined with Kubrick's penchant for multiple takes. Oof!
posted by ShutterBun at 5:04 PM on August 5, 2013 [3 favorites]
posted by ShutterBun at 5:04 PM on August 5, 2013 [3 favorites]
I really wanna see what Shane Carruth could do with these lenses.
posted by cthuljew at 5:36 PM on August 5, 2013 [2 favorites]
posted by cthuljew at 5:36 PM on August 5, 2013 [2 favorites]
The thing I love about BL is simply how entertaining it is. You can take away the spectacle and gorgeous cinematography, and at root it is a good story, entertainingly told. I definitely love the spectacle and gorgeousity, but there really is something there beneath it, Pauline Kael notwithstanding.
posted by hwestiii at 5:44 PM on August 5, 2013 [2 favorites]
posted by hwestiii at 5:44 PM on August 5, 2013 [2 favorites]
Can someone explain why this is technically difficult to make? A wide aperture…why is this so advanced? It seems to be a fairly simple...just make the aperture bigger....or?
It is simple, if you want a big aperture you just make the lens bigger. Telescopes have some giant lenses in them, and while it costs a lot, it's not crazy.
The difference with this lens is not the size of the aperture, but that it has a very low focal ratio (the ratio of the aperture diameter to the focal length). Say I have a fixed size of film, and Kubrick wants a certain field of view, which together corresponds to a fixed focal length. At that specific focal length, the size of the aperture sets the amount of light that gets collected, so if I want to film a relatively dark room I need a big lens. Now, as that lens gets bigger, the rays of light that hit the edge have to undergo a lot more bending to get to the piece of film than the rays of of a smaller aperture lens. More bending requires greater curvature on the lens and requires higher index of refraction glass. Hard, but not impossible. Where this really starts to cause problems is that these highly curved surfaces and high index glasses introduce all sorts of optical aberrations, which cause the image to loose sharpness. E.g., spherical lenses (which are easiest to manufacture) introduce spherical aberration (because they're not parabolas), high index glass is often very dispersive, so different colors of light have different indices of refraction, and thus get refracted differently, creating chromatic aberration. Parabolic lenses are not quite good enough, because off axis even a parabola doesn't focus light perfectly, leading to coma off-axis. These are the low-order aberrations, and they have solutions that require adding more lens elements or more complex surface shapes, but these solutions always have their own higher order aberrations. E.g., if you change the surface shape you can fix coma at one wavelength, but it doesn't necessarily fix it at other wavelengths. Everything gets worse as you go to very fast (low focal ratio) lenses, so it ends up being a giant optimization problem to get a lens which is satisfactory in performance and cost. This is the general nature of lens design. It's a lot of trade-offs, and it's interesting to see one case where they made the trade-offs necessary to produce a really fast lens.
I could go on for days about optical aberrations, glass, optics in general, etc. It's a weird combination of mind-numbing and fascinating.
posted by kiltedtaco at 5:52 PM on August 5, 2013 [18 favorites]
It is simple, if you want a big aperture you just make the lens bigger. Telescopes have some giant lenses in them, and while it costs a lot, it's not crazy.
The difference with this lens is not the size of the aperture, but that it has a very low focal ratio (the ratio of the aperture diameter to the focal length). Say I have a fixed size of film, and Kubrick wants a certain field of view, which together corresponds to a fixed focal length. At that specific focal length, the size of the aperture sets the amount of light that gets collected, so if I want to film a relatively dark room I need a big lens. Now, as that lens gets bigger, the rays of light that hit the edge have to undergo a lot more bending to get to the piece of film than the rays of of a smaller aperture lens. More bending requires greater curvature on the lens and requires higher index of refraction glass. Hard, but not impossible. Where this really starts to cause problems is that these highly curved surfaces and high index glasses introduce all sorts of optical aberrations, which cause the image to loose sharpness. E.g., spherical lenses (which are easiest to manufacture) introduce spherical aberration (because they're not parabolas), high index glass is often very dispersive, so different colors of light have different indices of refraction, and thus get refracted differently, creating chromatic aberration. Parabolic lenses are not quite good enough, because off axis even a parabola doesn't focus light perfectly, leading to coma off-axis. These are the low-order aberrations, and they have solutions that require adding more lens elements or more complex surface shapes, but these solutions always have their own higher order aberrations. E.g., if you change the surface shape you can fix coma at one wavelength, but it doesn't necessarily fix it at other wavelengths. Everything gets worse as you go to very fast (low focal ratio) lenses, so it ends up being a giant optimization problem to get a lens which is satisfactory in performance and cost. This is the general nature of lens design. It's a lot of trade-offs, and it's interesting to see one case where they made the trade-offs necessary to produce a really fast lens.
I could go on for days about optical aberrations, glass, optics in general, etc. It's a weird combination of mind-numbing and fascinating.
posted by kiltedtaco at 5:52 PM on August 5, 2013 [18 favorites]
I've never understand the criticisms of Barry Lyndon, I get totally transfixed every time I watch it. I love how the story is on rails, there's no doubt that fate is going to swat Barry right back to Ireland and even though he deserves every bad thing that happens, you still feel a little bad for the poor schmuck.
posted by octothorpe at 5:58 PM on August 5, 2013
posted by octothorpe at 5:58 PM on August 5, 2013
Check out this history of aspect ratio that explains Panavision and other technical issues with movie cameras. The reasons are more complex for film due to installed projectors.
I think Barry Lyndon is way way underrated. I like costume drama and BL is one of my favorites.
The Candle lit scenes are really showstoppers.
posted by Ad hominem at 6:00 PM on August 5, 2013 [1 favorite]
I think Barry Lyndon is way way underrated. I like costume drama and BL is one of my favorites.
The Candle lit scenes are really showstoppers.
posted by Ad hominem at 6:00 PM on August 5, 2013 [1 favorite]
The one and only time I saw Barry Lyndon, I was suffering from a fever and out sick from school. I have vague recollections of the film and should probably go back and see it sometime.
(Amusing celebrity encounter: Marisa Berenson at the Bowery Ballroom, in the bar area. We talked about knitting and angora rabbits. Only in retrospect did I realize who I was talking to.)
posted by pxe2000 at 6:08 PM on August 5, 2013
(Amusing celebrity encounter: Marisa Berenson at the Bowery Ballroom, in the bar area. We talked about knitting and angora rabbits. Only in retrospect did I realize who I was talking to.)
posted by pxe2000 at 6:08 PM on August 5, 2013
Good grief...imagine being the focus puller on that movie. A 1" depth of field, combined with Kubrick's penchant for multiple takes. Oof!
And the Mitchell cameras appear not to have had a reflex viewfinder, either.
posted by junco at 6:39 PM on August 5, 2013
And the Mitchell cameras appear not to have had a reflex viewfinder, either.
posted by junco at 6:39 PM on August 5, 2013
Not by default, but I think they installed some after the custom milling was done. That's really neither here nor there fr a focus puller, though. Apparently what they did was mount/graft a video camera & monitor on the film camera overlayed with a grid system of points of known distance (i.e. Actor #1's eyes, Actor #2's eyes, etc.) so that when focus was required to change to a certain area, the markings on the grid would indicate the correct distance. They didn't do many moving shots (and probably none with the Zeiss lens) so that would at least make it less impossible.
posted by ShutterBun at 6:56 PM on August 5, 2013
posted by ShutterBun at 6:56 PM on August 5, 2013
Blazecock, you and I should be best friends. "You guys watch Joe Don Baker movies?" has been one of my catchphrases since freshman year of high school. Got me to watch an interesting little video on something in which I previously had no interest though!
posted by Partario at 7:43 PM on August 5, 2013 [1 favorite]
posted by Partario at 7:43 PM on August 5, 2013 [1 favorite]
It's amazing how hard to conceive nightlife by candlelight is in this day and age.
posted by sp160n at 8:43 PM on August 5, 2013 [4 favorites]
posted by sp160n at 8:43 PM on August 5, 2013 [4 favorites]
That's fantastic, too bad they can only be used with a camera that looks like it's at least three generations behind the current crop of digital cameras. 11 T-stops is less than the Red One, and up to full HD resolution with "2k planned" is ridiculous today.
In fact, given the dynamic range of that sensor, I'm not convinced you'll get any better low light performance on this camera with those lenses than you get on, say, a Red Epic or Sony F65 with Leica Summilux-C lenses. They're T1.4, but the cameras are so much more sensitive that I think it might all in all work better. Not to mention the just now rolling out Red Dragon sensor, which has even less noise and 16 stops or so of range.
posted by Joakim Ziegler at 9:45 PM on August 5, 2013 [1 favorite]
In fact, given the dynamic range of that sensor, I'm not convinced you'll get any better low light performance on this camera with those lenses than you get on, say, a Red Epic or Sony F65 with Leica Summilux-C lenses. They're T1.4, but the cameras are so much more sensitive that I think it might all in all work better. Not to mention the just now rolling out Red Dragon sensor, which has even less noise and 16 stops or so of range.
posted by Joakim Ziegler at 9:45 PM on August 5, 2013 [1 favorite]
@cthuljew - Well Shane Carruth has got a Panasonic GH2 and you can get Voightlander lenses for that that are f0.95 for the price of a can of coke compared to the Kubrick lenses. Not the same thing, but it would get you a good portion of the way there in terms of low light performance (kiltedtaco would know exactly how much most likely!)...
posted by 6am at 12:41 AM on August 6, 2013 [1 favorite]
posted by 6am at 12:41 AM on August 6, 2013 [1 favorite]
Yeah, the performance improvement from CCDs over film will completely dominate over the gains from this lens. Film has a quantum efficiency of a few percent, maybe up to 10% if you do crazy things like bake it in hydrogen gas for an hour (astronomers had some weird methods). Backside-illuminated CCDs have QE's easily in the >80% range, plus the improvements in optical coatings since then, etc. So a factor of 10 improvement in QE is a factor of 10 in aperture area that you don't need. 35mm f/0.7 is a 5cm aperture diameter, ~20cm^2 area, if I go to only needing 2cm^2 then that's 16mm aperture diameter, f/2.2. That's a much easier lens, though you do loose the depth of field effects you got with the low f number.
posted by kiltedtaco at 8:42 AM on August 6, 2013 [1 favorite]
posted by kiltedtaco at 8:42 AM on August 6, 2013 [1 favorite]
ShutterBun:Not quite. Bigger is harder, yes, but large f-stop has nothing to do with lens size. Microscope objectives can have sub-1.0 f-stops; they're only a few millimeters wide.
As for why it's difficult to make, it pretty much boils down to "lenses are hard to make as it is, and the bigger they are, the more difficult it becomes."
The f-stop is an inverse measure of the amount light is bent/refracted as it travels through the system. Refracting light further draws in all sorts of problems: lower transmission at high angles, higher aberrations (many of which go up faster than linear with radius: r^2, r^3...), and high bending implies high index of refraction, which limits the materials one can use (material choice is key to limiting color aberrations).
So, while the nominal f/1.0 is not really a hard limit, it serves as an easy distinction between "we can do that for X hundred or thousand dollars" to "don't let our competitors know we're even attempting this!".
And, yes: modern digital cameras are capable of better performance with higher f-stops (which additionally means deeper focal ranges) than Stanley's f/0.7 masterpiece, thanks to detector technology.
kiltedtaco: Telescopes have some giant lenses in them, and while it costs a lot, it's not crazy.About $100k for a 1-meter-wide lens 10 years ago (to NASA), IIRC; the price goes up greater than a square of radius (a 2-meter would be closer to $1M). Not pocket change, but reasonable-ish.
posted by IAmBroom at 9:12 AM on August 6, 2013
sp160n: It's amazing how hard to conceive nightlife by candlelight is in this day and age.Having just returned from a week of camping by candlelight at Pennsic... well, it is pretty different, yes!
posted by IAmBroom at 9:17 AM on August 6, 2013
What a great post and fantastic, informed comments - this is why I keep coming back to metafilter.
posted by exogenous at 9:24 AM on August 6, 2013
posted by exogenous at 9:24 AM on August 6, 2013
Bigger is harder, yes, but large f-stop has nothing to do with lens size. Microscope objectives can have sub-1.0 f-stops; they're only a few millimeters wide.
Yes, but we're dealing with a few fixed constants here, first and foremost being "a large enough image to cover 35mm film" and "a focal length within the realm of what human eyeballs see." In other words: a focal length of about 50mm is optimal, and in order to provide adequate coverage of a piece of 35mm film and be less than 40mm from the focal plane, we're necessarily talking about a pretty hefty piece of glass.
In short, a large f-stop has an awful lot to do with lens size.
posted by ShutterBun at 7:48 PM on August 6, 2013
Yes, but we're dealing with a few fixed constants here, first and foremost being "a large enough image to cover 35mm film" and "a focal length within the realm of what human eyeballs see." In other words: a focal length of about 50mm is optimal, and in order to provide adequate coverage of a piece of 35mm film and be less than 40mm from the focal plane, we're necessarily talking about a pretty hefty piece of glass.
In short, a large f-stop has an awful lot to do with lens size.
posted by ShutterBun at 7:48 PM on August 6, 2013
ShutterBun: In short, a large f-stop has an awful lot to do with lens size.Meh, I used to make 1-meter telescope lenses. Those are dinky.
The f-stop is a ratio, plain & simple: a tangent slope. You're still confusing size with angles. And here's why that's true:
What's the f-stop of a lens composed of two, tiny, 1-mm-wide pieces of a lens: one piece 50mm in front of the optical centerline (the central core of the hole lens), and one piece from the very edge, 25mm above the first piece? The "lens" would be composed of a trivial amount of glass; the holder holding them both in position would surely weigh more than both put together.
f/1.0, of course. It really isn't about size.
posted by IAmBroom at 12:40 PM on August 8, 2013
f/1.0, of course.
Err, how on earth are you able to calculate the f-stop of this hypothetical lens without knowing its focal length? In order for a lens with an effective aperture of 1mm (which it sounds like what you're describing) to have an f-stop of 1.0, it would have to have a focal length of 1mm, which would obviously be unusable for 35mm photography. Granted, with digital cameras using smaller image sensors (such as the micro 4/3rds system) you can get away with a smaller pupil while maintaining a large f-stop more easily. But the bottom line is: with all else being equal, a lower f-stop number means larger glass, period.
It really isn't about size.
It is, when you're dealing with other fixed variables, such as the size of the film, and the desired focal length. It is ALL about (relative) size.
I used to make 1-meter telescope lenses. Those are dinky.
Really? Look up a few 1000mm lenses with an f-stop of less than 8.0, and let me know if you find one that could be described as "dinky." For that matter, find any f/1.0 lens for a 35mm (equivalent) camera that doesn't resemble a big chunk of glass the size of a soup can.
posted by ShutterBun at 10:41 PM on August 9, 2013
Err, how on earth are you able to calculate the f-stop of this hypothetical lens without knowing its focal length? In order for a lens with an effective aperture of 1mm (which it sounds like what you're describing) to have an f-stop of 1.0, it would have to have a focal length of 1mm, which would obviously be unusable for 35mm photography. Granted, with digital cameras using smaller image sensors (such as the micro 4/3rds system) you can get away with a smaller pupil while maintaining a large f-stop more easily. But the bottom line is: with all else being equal, a lower f-stop number means larger glass, period.
It really isn't about size.
It is, when you're dealing with other fixed variables, such as the size of the film, and the desired focal length. It is ALL about (relative) size.
I used to make 1-meter telescope lenses. Those are dinky.
Really? Look up a few 1000mm lenses with an f-stop of less than 8.0, and let me know if you find one that could be described as "dinky." For that matter, find any f/1.0 lens for a 35mm (equivalent) camera that doesn't resemble a big chunk of glass the size of a soup can.
posted by ShutterBun at 10:41 PM on August 9, 2013
ShutterBun: f/1.0, of course.I told you the focal length:
Err, how on earth are you able to calculate the f-stop of this hypothetical lens without knowing its focal length?
IAmBroom: one piece 50mm in front of the optical centerlineIt's a 50-mm lens.
ShutterBun: In order for a lens with an effective aperture of 1mm (which it sounds like what you're describing) to have an f-stop of 1.0, it would have to have a focal length of 1mm, which would obviously be unusable for 35mm photography.Wrong. The effective aperture (for f-stop calculations) would be 50 mm, the diameter over which light rays are being gathered. (To be fair, the diffraction limit wouldn't be achievable without another 1-mm lens at -25mm from the optical axis, but I never claimed it was operating at the diffaction limit.)
The light-gathering ability would be 0.5pi mm^2, or essentially the same as a 1.43-mm diameter lens.
posted by IAmBroom at 10:44 AM on August 13, 2013
I guess we're not exactly seeing eye to eye on a lot of key definitions, then, because none of what you just described makes any sense to me from a photographic lens standpoint.
posted by ShutterBun at 8:27 PM on August 13, 2013
posted by ShutterBun at 8:27 PM on August 13, 2013
On re-reading, I guess it sounds like you're saying that two tiny fragments of a lens, placed at the appropriate distances, could be described as having an f-stop of 1.0
I'm sure you have your reasons for suggesting that, but to me it sounds like a pinhole camera with 2 holes in the front, which obviously isn't going to behave anything like an f1.0 lens.
posted by ShutterBun at 8:35 PM on August 13, 2013
I'm sure you have your reasons for suggesting that, but to me it sounds like a pinhole camera with 2 holes in the front, which obviously isn't going to behave anything like an f1.0 lens.
posted by ShutterBun at 8:35 PM on August 13, 2013
"Behaving like an f1.0 lens" is not what I claimed.
I claimed it would have an f-stop of 1.0. It would, in classical optics.
You appear to be a user, familiar only with camera terms. I'm not talking about what they print on the packages; I'm talking optical science.
posted by IAmBroom at 10:22 AM on August 14, 2013
I claimed it would have an f-stop of 1.0. It would, in classical optics.
You appear to be a user, familiar only with camera terms. I'm not talking about what they print on the packages; I'm talking optical science.
posted by IAmBroom at 10:22 AM on August 14, 2013
« Older Kiese Laymon may be the best writer and curator in... | 18 Days Newer »
This thread has been archived and is closed to new comments
posted by Blazecock Pileon at 4:20 PM on August 5, 2013 [1 favorite]