F-(<)1.0

[Jay Gannon]

In this case I'm trying to avoid props in order to use a real world environment as a 'miniature' set.



Macro levels of DOF.

It's a jib shot in a forest so tilt shift would be obvious and post DOF would require a lot of roto.

Use something like a 55mm F2.8 Macro Nikkor... incredibly thin DOF.

 

[Jeff Kilgroe]

And the 105mm f2.8 Macro Nikkor is even thinner than that. I'll take a shot later today to show.

I feel like we're stabbing in the dark though as we don't know what's required for focal length or FOV, etc..

One thing with most of the Macro lenses, they do have super-thin DOF, but that typically only occurs within the close-focus or macro focus region.

 

[Gavin Greenwalt]

Right. Just think normal lens ~50mm, wide framing but with macro DOF. I don't even know if that's optically possible but I figured maybe there was some bizzaro custom lens out there that would shoot that.

Something like this:

 

[C.H.Haskell]

Cheap trick could be using a lens baby. Its a fixed focal length of about a 50mm if I remember correctly and you can remove the aperture ring internally and your left with a very selective focusing method. I used it on a horror film in Dublin, (portrait of a zombie) as my zombie vision, or pov if you will. Not sure if this is at all what your looking for but an idea. good luck.

 

[Charles Angus]

Right. Just think normal lens ~50mm, wide framing but with macro DOF. I don't even know if that's optically possible but I figured maybe there was some bizzaro custom lens out there that would shoot that.

Something like this:

I don't think you're going to be able to make that look like a macro shot with any existing lens...

UPDATE:
Just poked my head out the door and looked through my 7D at a similarly framed shot with a Rokkor 58mm f/1.2 - background blur is not huge. Certainly does not look like a macro shot.

 

[Charles Angus]

What about a very long lens - would obviously have a very narrow field of view, but you would get a more obscured background.

 

[Jeff Kilgroe]

Lensbaby gives roughly the same effect as a tilt-shift system. It's DOF is not really thin, but rather focus is selective due to shifting the front element off-axis. It might be something to investigate though. It's not the same effect as your typical tilt/shift systems as it effects the focus region and not really the perspective.

Can the shot be done with a MoCo system? Then it could be repeated with varying focus settings and composited in post. Or measure depth cues to create a rough depth map or reference within your shot and artificially defocus in post.

I took some shots with the 105mm Nikkor. Its DOF doesn't get super shallow until you get down into the close focus range and then it becomes instantly razor-thin. Tried to get a usable clip of a focus rack, but it's really shaky from turning the lens by hand. Might be better to go back and shoot stills at different focus distances, but there's not much to see here. It behaves just like a 105mm prime until the object of focus is within about 6 feet of the lens, then it changes. For wider framing, I can actually get more shallow DOF with the Nikkor 50mm f1.2 than the 105 Micro at f2.8.

As Charles points out, you may be able to use a longer lens. But then you have to move back to retain the same framing. I don't know if a 300mm would help you here... Have you seen shots from The Exploding Girl? It was shot mostly with a 300mm prime, but none of it looks like macro shots.

 

[Andrew Wilding]

What about a very long lens - would obviously have a very narrow field of view, but you would get a more obscured background.

Yeah I would think that if you can far enough back to get the framing that you want that a super long lens would be the way to go here....

 

[Gavin Greenwalt]

300mm would probably require a helicopter since it's a 45* overhead shot.

It's ok. I presumed that a post DOF solution was the most likely requirement. But if I figured I would double double check that an optical solution wasn't really feasible.

Alternately I might be able too shoot stereo. And extract a Z-Depth from the disparity map.

 

[Jorge Díaz-Amador]

The Zeiss Planar 0,7/50mm that Kubrick used is very rare. I saw the two that were on display with the Kubrick Archives exhibition, which I saw in Ghent, Belgium. One was mounted on one of the modified Mitchell BNC's.

I believe there is another one with a German museum, which is the one shown at IBC.

I was at Carl Zeiss recently for training. I inquired with someone who worked in the prototype department. He could not find a single 0,7/50, but he did find some other prototype Hasselblad lenses for the Apollo program.

This lens would be completely impractical on a digital cine camera due to very short back focus. The closest thing would be the Leitz Noctilux.

Needless to say, if anyone has an 0,7/50mm Planar and would be willing to bring it to LA, I would be very interested in testing it.

 

[paulherrin]

gavin, how bout shooting motion control (or at the very least getting really close moves) and shooting a plate of the background, and then shooting your figure in the same location with a greenscreen cloth. then in post, add to the DOF.

at its simplest, you could add some foreground blur, and either blur the plate in post, or shoot a blurred version of the plate to begin with.

else, you can roto/extract a depth map, and/or use projections on proxy geometry to simulate a depth map, and tweak your DOF.

that's also an interesting idea about shooting with a 3d rig, that could be a possible solution... it just might be a little less clean.

 

[Matt W.]

On House, to get the shallowest focus possible, they used a 5D with a 50mm f1 Canon lens (if you want a 50mm Super35 equivalent field of view, then the more common 85mm f1.2 would be your pick). The larger sensor size grants about a stop shallower depth of field so these would "feel" like 35mm f0.7 and 50mm f0.8, respectively.

There's a super farron available online for about $600 if it hasn't sold yet, so that plus a 5D would be really shallow, but I'm not convinced it would work. My guess is there would be so much spherical aberration it would look "soft focus" and not "macro." Not sure if that's what you want, but if it is it could work... It's physically impossible to construct a lens faster than f0.5 and exceedingly difficult to get decent performance past f1, so if you want a "normal" perspective and remotely acceptable sharpness, the 5D set up is probably the best bet.

There's some apocryphal story of an old movie where the filmmakers constructed a 35mm adapter type thing with a big ground glass and large format lens. The shallowest focus available from off-the-shelf optics would then come from the legendary 210mm f2.8 Schneider Xenotar, so legendary that Schneider has no record of it on their website. Assuming it has a similar image circle to the other Xenotars, it would project a 7.9''*4.3'' image with a normal/wide focal length, and so correspond with something like 40mm <f0.5 on the 5D or 28mm <f0.35 on Super35, field of view/depth of field-wise. I think this would end up being more of a science project than a functional system, though, even if you could get your hands on one of the four of these that are rumored to exist. (Forgive me if my math is wrong, I am SO not a math person.)

The moral of this story is: use motion control and then roto. A 3d rig won't derive meaningful enough depth information for a clean matte, don't even bother, and an optical solution will either be a fuzzy, expensive mess or no more shallow than what they did on House.

Edit: The 210mm is a real object of myth; the story goes that the US Government special ordered five of them fifty years ago for missile testing, and most were lost or destroyed in the process. The lenses were so optically great that their resolution tests were classified. Those who have used them claim they are the sharpest lenses ever designed in their focal length. Which is hard to disprove given that only two are confirmed to exist (one even showed up on eBay and went for way too cheap).

 

[Gavin Greenwalt]

By the way. I just finished last night.

Ended up using CG and photos projected onto cards.

 

[Andy Jarosz]

Something I've always wanted to try is fly a huge sheet of visqueen behind a subject, effectively blurring out the background. Similar to a play that utilizes a curtain of scrim material.

 

[Erik Bien]

IIRC, one of the 'reality-TV' series uses a large frame of Hampshire Frost behind the subject for talking-head interviews in small spaces: maybe Miami Ink?

 

[Gavin Greenwalt]

Something I've always wanted to try is fly a huge sheet of visqueen behind a subject, effectively blurring out the background. Similar to a play that utilizes a curtain of scrim material.

I've seen it done and done small scale tests but it looks pretty crappy. It's blurred instead of defocused so you don't get that nice 'crisp' bokeh.

 

[Charles Angus]

IIRC, one of the 'reality-TV' series uses a large frame of Hampshire Frost behind the subject for talking-head interviews in small spaces: maybe Miami Ink?

This was done with interesting effects in the film "Eternal Sunshine of the Spotless Mind", iirc.

 

[Florian Stadler]

You need a bigger sensor. You could mount a large format still camera and shoot it's ground glass. Kind of like a Depth of Field adapter on a different scale...

 

[Roberto Lequeux]

You need a bigger sensor. You could mount a large format still camera and shoot it's ground glass. Kind of like a Depth of Field adapter on a different scale...

This could have done it, though it would have been tricky. It sounds like he would have needed something like a large format screen and find a way to vibrate it (just guessing required size here), or a ground glass disc about 2' in diameter and turn that into a massive spinning DoF adapter. If you didn't move it you would have had stationary "grain" from the GG. Not practical, but it would give great results. Right?

 

[Matt W.]

You need a bigger sensor. You could mount a large format still camera and shoot it's ground glass. Kind of like a Depth of Field adapter on a different scale...

Read a few posts up, I suggested just this and the exact lens for the job, the coverage, and approximate focal length/f-stop equivalent it would result in. (210mm Schneider Xenotar, 7.9''*4.3'' image plane for 1.85:1, wide/normal, about f0.35 in terms of equivalent depth of field.)

For a narrower field of view, the Zoomatar 240mm f1.2 would be another good (and even more expensive) option for an approximately IMAX-sized ground glass, which would offer a 75mm f0.37 equivalent for Super35 on a 71.2mm*38.5mm ground glass.