DV DOF Trick for No Budget

the only problem is that sometimes you just don't have the room.

I have shots where I did this in my demo reel and my teaser for my feature I'm finishing up, Average Joe. The problem I've found with this is that it really flattens the image alot. You can also move the subject painfully close to the camera, you can keep the zoom wide. Or put a silk between the subject and the background (tight sheer fabric stretched over a wooden or pvc frame.

Actually, the flattening of the image turns out to be a bonus for stuntwork. If you ever need someone to jump out of the way of a moving vehicle, you can add those precious inches by zooming in from afar and shortening the perceivable distance.

Will Vincent said:

the only problem is that sometimes you just don't have the room.

Click to expand...

Aint that the truth. We had to basically plan out all our shots where we needed shallow DOF and schedule them for places where we had room to do this.

Technically, DOF doesn't change by zooming in the lense, but the perception of depth is compressed, so it appears that the DOF has been decreased. The root of all evil is small sensors. I'm keeping my fingers crossed on the Red camera (http://www.red.com/), because that will have a nearly full-frame 35mm CMOS sensor and interchangeable lenses. I think it will blow every other digital movie camera out of the water ... if it actually materializes.

In the mean time, stay in close and keep the aperature wide open. The magic word is "neutral density filter" ... sorry, that was 3 words.

actually, if you abbreviate it, NDF can be considered a single word. I like to throw in a polarizer on most shots too to cancel out off axis light since the chips are really bad at sorting that out (hazy footage is ugly).

Yes, when you zoom in, the background is already out of focus, but to such a small degree that you can't tell until you zoom in and emphasize the blurriness...you then have to move the camera back to keep your framing. So it's a trick, and it's directly, specifically related to the fact that we're shooting on small format cameras. DoF is mathematical though, and you can graph it for your camera.

I often advocate to folks that they take the time to use a tape measure and graph out the DoF for your camera. The DoF will be the same at any focal length (zoom setting) so long as the iris is at the same setting (wider is better for shorter DoF). As you zoom, the same range of distances from the CCD will be in focus, but the relationship of the elements in the frame changes.

at widest zoom:
a----------b----------c

at 2x zoom, the apparent distance between the camera and each element halves:
a-----b-----c

The b (the subject) and c (the background object) appear closer to one another. Since mathematically, c has moved farther than b, it appears to have increased in size more dramatically than b.

In 3-D space, our eyes give us information about how far out their focal vectors cross. This allows our brains to calculate the distance to an object. If you close one eye (try fencing with an eye patch sometime), this information leaves and only 2 pieces remain. Parallax and comparative size.

Parallax is why the corn near your car as you drive seems to move faster than the sun. This is the effect we get with a dolly move across a scene...stuff farther from the camera slides more slowly than stuff closer...if we pan with the foreground items, the background seems to shift behind the subject.

As far as comparative size (the actual topic of conversation for this point), I know that you are taller than your car (unless you have a truck), so if the car looks taller than you, you must be very far away. As we zoom in and move back on a scene (which looks pretty cool when done at the same time), the apparent distance between your nose and your ear change ever so slightly and flatten the scene.

We will get better depth if we keep the zoom wide and move the camera in to frame..very tightly on the actor so we have to focus closer to the lens, thereby decreasing the Actual depth of field without distorting the apparent distance between the elements of the frame.

so there! that's a long post...I'll shut up now.

Wow, Knightly; I couldn't have said it better myself. My DV camcorder has 1/3 inch CCDs. The only way I can get the selective focus I really want to see (that which I learned to expect with my 2x2 and 35mm film cameras) is to be about 3 feet from my subject. The reason I haven't jumped on the HDV bandwagon is that I'm waiting for that larger CCD/sensor camera to move into my price range. I am so tired of tiny focal planes, I don't care if it has 1080 lines of resolution!

pulls cord - "The Google Says, .33 inches = 8.38200 millimeters"

So you're basically shooting 8mm with half the effective resolution and less grain.

you mean something like this? Shot with a shoulder mount so forgive movement.. was just a test.

http://www.scarlet-films.com/ant.mov

that's the stuff Pretty flora!

Ncje, did you shoot that on HDV? Great example of selective focus. You might also notice that when the bicyclist is a few meters away, and in sharp focus, the distant background is nearly in focus. Even with moderately wide aperature settings, I've found that selective focus gets less effective as the subject moves away from the lens and becomes almost useless around 3-4 meters out (using my 1/3" DV camcorder).

If you really need selective focus, as Knightly explained, you just need to know the parameters of your camera and minimize your DOF by using NDF filters, Polarizers, and tight framing. The tighter you can frame your subject, the better ... just as your video demonstrates. I have had strange, strobe-like effects from increasing the camera's shutter speed to allow me to open the aperature further in bright light. Therefore I hesitate to recommend higher shutter speeds. Certain subjects will look strange without motion blur, and panning the camera is out; unless you're going for a special effect.

It's like Clive pointed out a while back. Knowledge and experience are more important than top of the line equipment. I've found that shooting experimental clips like Ncje's teach me what to expect from my equipment, and how to get the most out of it.

I'd still like to know what camera was used to shoot that clip, Ncje.

ncje said:

you mean something like this? Shot with a shoulder mount so forgive movement.. was just a test.

http://www.scarlet-films.com/ant.mov

Click to expand...

Sweet! What camera did you use for that????

Shallow depth of field is one thing thats used in an entire toolbag to tell a story. When you watch a movie depending on the the DP its used in varying degrees from very little to a lot. I do have a 35mm adapter for the cam this was shot on, but find few but very artistic shots for it. This was shot on a DVX100a. I am picking up the Canon A1 soon as it hits the shelves, for the price with a 20x zoom lens it offers my style a better range. Just hoping that the lens is good at this stage as Ill be one of the first of the shelf, there will be no way of telling before buying.

This clip was shot in my garden on Tokyo.

Interesting conversation, but just thought that I'd mention that you can buy cheap ND filters for the front of most camcorders.

The use of ND (Neutral Density) to decrease the amount of light, which then allows you to open the aperture and thus decrease depth of field, is the most effective way to control depth of field.

The other option is to move the lights further back -- or scrim them down by adding increased layers of diffusion. It's the same formula, less light equal increased aperture, equal less DOF.

for 1/3" chippers, I'm going to see if I can get a DoF chart for a super 8 camera (about the same capture size). I can't seem to find any laying around the web that are useful and easy to understand. Maybe someone around this board has one laying around they could link

Found one:

http://www.filmshooting.com/community/articles/dof.php

clive said:

Interesting conversation, but just thought that I'd mention that you can buy cheap ND filters for the front of most camcorders.

The use of ND (Neutral Density) to decrease the amount of light, which then allows you to open the aperture and thus decrease depth of field, is the most effective way to control depth of field.

Click to expand...

Bit late... but better late than never. The clip I shot was using a circular polariser which pretty much achieves the same thing.

knightly said:

pulls cord - "The Google Says, .33 inches = 8.38200 millimeters"

So you're basically shooting 8mm with half the effective resolution and less grain.

Click to expand...

That is not correct, because the given size of the CCD means the diagonal, not the width.

width = a
height = b
diagonal = c

for 4:3 ratio 1/3" CCD:

a / b = 4 / 3 -> a = 4 * b / 3
c = 8.4667 mm (1 inch = 25.4 mm)
c² = a² + b²

b = sqr(0.36 * c²) = 0.6 * c = 5,08 mm
a = 4 * b / 3 = 6,773 mm

for 4:3 ratio 1/4" CCD like Canon GL-2 / XM-2:

a = 5,08 mm

btw: 35 mm film has only a usable width of 24.9 x 18.7 mm (negative) or 21.3x17.8 mm (positive), anamorph 2x

Here is an DOF example picture I took with the Canon XM-2 (GL-2) using 1/50 f2.4 and half zoom (10x):

Just to throw another wrench into the gears, DOF is also somewhat subjective. DOF is based on circles of confusion, which quantify the acceptable amount of blurriness for a specific lens, medium, etc. There is really only one point of precise focus, with a range of acceptable focus. Therefore, in order to measure DOF, you must determine what is acceptable focus (minimum circle of confusion diameter). With mini-DV, which is rather low resolution, you might be able to use a relatively large circle of confusion (CoC), for an even wider DOF. Whereas, with HDV or other HD formats, the CoC might have to be smaller, which might enhance selective focus somewhat, even with the same lens and CCD size. (i.e. If the sharp parts of the image are sharper, the slightly out of focus areas will look noticably softer)