I've been reading up a bit about depth of field (dof) and the math has fried my brain. I think Jeff Donald's article on this site is helpful, but if you're pretty basic like me, you need it be presented as apple sauce.

I think mishandling the depth of field is probably one of the reasons why newbie/amateur cinematographers may not be able to get their projects to feel more like film, even after tackling other obstacles, like lighting, acting, sound, etc. Ever notice that in some of our dv projects, everything seems to be in sharp focus, all the time, even during close-ups?

Instead of throwing a wall of text at you about how i did it, I just turned on my DVX100A and made a quicktime clip of it in action.

I'd been watching X-files and Law & Order and noticed how much they present a short dof in their shots.

Take a look:

http://host212.ipowerweb.com/~chicagoh/scratch/

On the DVX100A, I am shooting in 24P advanced mode, shutter is off, fstop is 1.7. I'm manually controlling the focus and zoom rings (new experience).

Hmm, what else about the shot. That's a 60w bulb in the lamp. I don't think the brightness is fully up on the powerbook, but I can't remember.

Hope this helps someone.

Chris,
Of course you will get less depth of field on small CCDs if you zoom in as much as you did - coupled with a wide aperture as well. You will get some really nice effects like that, especially if you throw the focus from one person in the foreground to the other in the background and so on.
However, what you would get with 35mm film or the P+S Mini35 adapter is a narrow depth of field even on a wider shot.
Still, I appreciate that it is a voyage of discovery for you, so thank you for sharing it with us.


Robin

Ah, obviously this is all new to me in practice.

I'll play around with cam some more to see how wide I can go and still achieve this effect on the dvx100a.

I'm afraid that at $6,000, the P+S Mini35 adapter is out of my present financial range for now, but I've been drooling over one already.

http://www.dvxuser.com/articles/mini35/

We live in interesting times.

Hate to say it, but the Mini35 is really $10,000 out the door, set up for a given camera. And then you have to contend with a set of lenses, or budget to rent them.

<<I think mishandling the depth of field is probably one of the reasons why newbie/amateur cinematographers may not be able to get their projects to feel more like film, even after tackling other obstacles, like lighting, acting, sound, etc. Ever notice that in some of our dv projects, everything seems to be in sharp focus, all the time, even during close-ups?>>

Chris, that's an interesting statement. I would put it to you this way: if newbie or amateur cinematographers have spent enough time experimenting and shooting to the point where one can say that they have REALLY improved their lighting (and composition, and shot design) skills, they would then no longer be considered newbies or amateurs.

And I don't think that "mishandle" is perhaps the right term. A wide shot on DV is a wide shot on DV, and unless you are macro-focusing to an object that is inches from the lens, you ARE going to get an extended depth-of-field no matter what you do (barring the use of a device such as the Mini35). No degree of experience can change the physics of this situation.

Can DV look like film? With today's 24p cameras and image processing options, you have more tools at your disposal than ever to approach this Holy Grail; but still it won't look just like film. A better objective might be, can it FEEL like film; can it PLAY like film. Will the audience be able to register the movie without being distracted by a "cheesiness" to the visuals.

The fact is, if all of the other factors that make for good filmmaking are present, it JUST DOESN'T MATTER if the background is sharp. I've seen some amazing looking material shot on DV, and I never stopped to think about how much better it would look if the background had fallen out of focus. That said, I am a Mini35 advocate (and owner) and prefer that shooting method; but I also feel comfortable that I am able to shoot without it and make really nice pictures if necessary.

Understanding depth of field is important, and I commend you for putting in the time and effort to explore the concept. Shot design, composition, blocking and lighting; these are all MUCH deeper subjects and one can spend a lifetime as a student of these disciplines, and I think you'll find that your work will become much more "film-like" as a result, with or without out-of-focus backgrounds.

You might be interested in this creaky old thread (http://www.dvinfo.net/conf/showthread.php?s=&threadid=6281): this was a conversation about a short I shot a few years back, when the XL1 didn't have an S yet, and the Mini35 didn't exist. I'm afraid that the compression on the short as linked (at iFilm) is pretty terrible to behold by today's standards, but hopefully enough can be gleaned to illustrate the text. I talked a little bit about using lighting to create the depth in wider shots instead of selective focus. I hope you find it informative!

Best of luck with your endeavors,

I, too, thank you Chris for sharing your discovery journey with others here. We actually don't get too many such narratives.

In my opinion, depth-of-field is leaned upon far too heavily as a technique for guiding the eye of the viewer. Luminance, chrominance, size, position, contrast and movement are the characteristics that are manipulated to guide the eye in motion pictures. Focus manipulation deals largely with only contrast. It is a cheap trick for turning segments of the frame on and off, thus avoiding a much more complex undertaking of trying to make the background work for the scene or creating another shots setup. Rack-focus can be very effective occasionally, such as in "surprise" shots, but used very, very sparingly.

It's interesting to note that many of the great filmmakers of the last century often went to extreme measures to achieve deep depths of focus for shots. Faced with relatively slow films and lenses they would pump tens of thousands of watts of light into a scene just so they could squeeze their iris to pinholes to get reasonably good focus on backgrounds and background subjects. For an excellent example of this subject, buy or rent the late John Frankenheimer's "Seven Days in May" and listen to his commentary through the film. Watch the film carefully (actually any of his films) and you'll see that he was a master of enriching his scenes by capturing background performances that would have otherwise been lost from focus.

Have a ball with your education and your filming endeavors, Chris!

Thanks for all of your feedback and suggestions. I'm a sponge absorbing it all! :]

As most (but maybe not all) of you know, one major example of deep focus photography achievements was "Citizen Kane". Even in some of the darker scenes, they managed amazing results. They'd probably go nuts for our little rinky-dink video cams, and now we're all trying to get a shallow depth of field.

We are all (myself included) too busy thinking about using these techniques and looks, not to further the artisitic vision of our films, but to make them look like the medium of film, so as to "trick" the general public into thinking our smaller productions are more professional. In my opinion, through the plethora of smaller cable stations relying on video productions, some DV films ("Open Water"?), and recent popular documentaries (no matter where you stand politically, Moore's good for something!), the public seems more ready to accept video than 20 years ago when it was just for nightly news. Many people don't even realize that most of their favorite sitcoms were shot on video because the rest of said sitcoms' production values were so high.

If you want to be able to rack focus, get a large CCD cam with a 20X zoom, zoom all the way in, and you can still do some stuff.

<<<-- Originally posted by Ken Tanaka : In my opinion, depth-of-field is leaned upon far too heavily as a technique for guiding the eye of the viewer. -->>>

Very interesting perspective, Ken. The human retina seems to simulate DOF characteristics of about 1/3" chips- if you hold a finger up, focus on that, then "rack focus" to the wall, you notice your finger going out of focus a little bit, but certainly not to the wide open 35mm total blur level.

Another experiment: hold your hand at arm's length, then suddenly strike your forehead and return your hand to its starting position.

That last experiment doesn't have anything to do with DOF; just wanted to see how many people would do it. ;) Sorry.

Let me be the first to say, he's just full of these sorts of things. I'd like to start a sitcom called "That's Just Josh," let him write it; it'd be pretty funny. His beautiful wife Michelle could play it straight. We'd work the cats in too, somehow.

;-)

That's not the first time I've been said to be full of something. ;)

Just to bring this thread back on topic, I found this:

http://www.dvcreators.net/media/depthoffield.html

It is old and rather pedantic, and none of the links work, but it might serve as a good intro to DOF for those who would ignore Ken's advice, and go for the cheap shallow DOF effect!

I also want to add that increasing focal length doesn't actually effect DOF, but it does intensify the effect- when you enlarge something blurry, it appears blurrier (even though technically, it isn't).

Like walking closer to an impressionist painting- the paint does not change, yet the painting looks much more impressionist!

<<<-- Originally posted by Ken Tanaka : It's interesting to note that many of the great filmmakers of the last century often went to extreme measures to achieve deep depths of focus for shots. -->>>

Funny that you mention this. I am just now reading Hitchcock (http://www.amazon.com/exec/obidos/ASIN/0671604295/qid=1094331399/sr=ka-1/ref=pd_ka_1/002-6222585-1100838) by Francois Truffaut, which is a fascinating biography consisting of interviews during the 1960's.

At the end of Spellbound there is a shot viewed from the perspective of Leo G. Caroll as he points his gun at Ingrid Bergman, then turns it towards himself to commit suicide. Hitchcock wanted the hand with the gun and Bergman to both be in sharp focus. His solution was to build a giant hand and gun! Evidently he used similar tricks in other films.

He even comments that later he learned you could achieve greater depth of field by using lots of light and shooting at small apertures, but his cameramen discouraged it because it wasn't flattering to the actors faces.

Boyd,
Just going slightly OT, but you mentioning Hitchcock/Truffaut reminds me that a couple of weeks ago, while on holiday in France, I was driving back from the local boulangerie with the morning baguette and croissants and listening to France Culture on the car radio, when I heard an English voice that I thought I recognised. It turned out to be a recording of Hitchcock being interviewed by Truffaut (with a translator in the middle doing stirling work.)
It was fascinating and I thought - where else but France with their love of Cinema would I get a treat like this?.
The family's breakfast was a little delayed that day...

Robin.

<<I also want to add that increasing focal length doesn't actually effect DOF, but it does intensify the effect- when you enlarge something blurry, it appears blurrier (even though technically, it isn't).>>

Hmm. I think I have to call on this one, Josh...focal length does affect DoF; the increased blurrines is a result of the given plane falling further outside of the circle of confusion. It's as bona fide an "out of focus" look as it would be if you were to rack the focus further away from that focal point.

Actually, that may not be a solid technical argument. I was always terrible at geometry, when we were asked to "prove" something by quoting theorems; the teacher would say "well, that doesn't prove why the angle was square" after my explanation, and I would respond "but it does to me!".

In this instance, I know that a long focal length will definitely result in softer backgrounds, and the softness is indeed greater than if I had taken a frame blowup of the center of a wider-angle version of the scene (which I think is what you are suggesting) due to the decreased DoF.

Charles,

This debate has raged on for over a century... the bottom line is, zooming in does make the background look softer. So that's what we do when we want the background blurrier. And that's what we teach!

In one of my early classes was an optical physicist that explained that although what I was teaching was practically true, it was technically false.

The diameter of the CoC is determined only by the size of the aperature, not the focal length. If you do as you say, enlarge a section of background of a wider shot, it will appear just as blurry as the zoomed in shot. (Actually, it will look blurrier, due to the unrelated factor of there being less pixels, or sparser density of silver halide crystals.)

This is all academic, since it doesn't affect how you shoot. Perception is everything. There is no truth in art, only in optics!

Another related disparity between what our eyes see and what is true, and a widely misunderstood optical phenomenon is that a lot of people attribute focal length to changing the whole look of a shot. That is, a "telephoto look" shows distances compressed, a larger background, and so on, while a "wide angle look" produces distortion of lines, and increased disparity of relative size, etc.

Technically, focal length has no effect on any of these factors. It's the distance from the lens that determines all these things. If you blow up a small section of a shot with a wide angle lens (with a camera with infinite resolution- wouldn't that be nice!) you will see an identical picture as one taken from that same spot with a telephoto lens or setting.

Hey, I am looking for a DP for a one day shoot in LA week after next! What's your schedule like?

Josh, find a formula for DOF that does not include focal length. It will be next to impossible, I know of only one formula for DOF that excludes focal length and it use magnification instead. But since focal length changes magnification, it is indirectly a component of that formula as well.

The confusion comes about because Coc was originally determined by making contact prints. As we all know contact prints are hard to make from video files and small format photography, such as 35mm. Other factors have been incorporated into modern practices of determining and utilizing Coc. Among these factors are resolution and MTF.

On your second point, your absolutely correct. Changing lenses (focal length) only changes magnification and does not change perspective. Changing the camera position in relationship to the subject will change perspective.

Hi Jeff,

True, focal length is part of the formula but makes very little difference, provided subject size in frame (and f-stop and image size) stay consistent.

Look at the photos on this page:

http://www.luminous-landscape.com/tutorials/dof2.shtml


The pictures of Gromit on this page tell the whole story- because the car is equally blurry in pic A and pic B, because the DoF is the same regardless of focal length. Of course, the background looks blurrier in A because it's bigger!

http://www.vanwalree.com/optics/dof.html

A little further down the page is what Charles was talking about- blowing up the car to the same size and comparing the two pictures- you can see the "blurriness" is consistent.

(Interestingly, this page criticizes the first link I posted!)


From this page:

http://www.normankoren.com/Tutorials/MTF6.html#DOF_focal_length

"DOF expressed in distance is nearly independent of focal length."

You can see by the chart that DoF remains almost constant (within 2mm) from 50mm to 1000mm focal length. With the short focal range of 20mm DoF only increases to 217 mm.

From this one:

http://www.cinematography.net/Pages%20DW/DepthOfFieldImageSize.htm

"Image size and f-stop are the main determinants of depth of field. Focal length of the lens does not matter, although on the extreme ends of focal length, image diffraction and interference play some part. If one moves an 18mm shot close enough to make the subject exactly the same size as it was with the 250mm lens, then the DoF is identical assuming the f-stop and circle of confusion choice are exactly the same."

Our whole teaching philosophy at DVcreators.net is "all practical, real-world techniques, no fluff, no theory"! So it's ironic that I am the one that brought up such an impractical, though perhaps interesting diversion.

For anyone besides an obsessed technogeek reading this: "If you want your background fuzzy, zoom in!"

The key to all this is that we are assuming the same subject size in frame. My example of blowing up the center of the image was not making this assumption, rather the same subject photographed from the same distance. In this case, the focal length changes and the DoF changes with it. A wide-angle version of the scene (in which the subject is perhaps 1/10th the height of the frame) and an extreme telephoto (in which the subject's head is seen in close-up) will deliver different depth of field characteristics.

An example: given 35mm 1:33 as the shooting format:

A 17mm lens focused at 6 feet will deliver a depth of field from 4'2" to 10'10".

A 100mm lens at 6 feet will deliver from 5'11" to 6'1".

So, if we had another subject positioned at 10', it will be in focus on the 17mm, but not on the 100mm. And given a perfect recording medium where resolution is not an issue, by blowing up the center of the 17mm shot, that subject at 10 ft should be sharper on the 17mm than on the 100mm.

However, if we were to match the image size of the first subject as shot on the 100mm using the 17mm, we would have to march in to 1'. And our subsequent depth of field would be from 11.4" to 1'1"--in other words, very similar to what we saw with the 100mm (roughly one inch on either side).

Or am I missing something?

(p.s. thanks for the inquiry Josh--I'm about to ship out to Montreal to work on a feature until the beginning of October).

This is a confusing part of DOF for many photographers. The arguments makes assumptions about the position of the camera in relationship to the subject and subject size. In my experience motion picture, video and portrait photographers think of DOF in terms of a constant or fixed subject size. Landscape, sports and wildlife photographers think of DOF not in terms of a constant subject size, but rather how to minimize or maximize DOF for the scene in front of them. The way each of these groups thinks, in terms of their own photography, effects their perception of how DOF works. Josh, you may want to review the article we have posted (http://www.dvinfo.net/articles/optics/dofskinny.php) for our community here at DV Info.net.

I am posting in this thread because of the vast skill and experience of those who are posting in it.

My question is this.

I am shooting out the front window of a vehicle moving at freeway speeds. When I zoom in fully, the lane stripes move as though the vehicle is travelling very slow (looks like 5 to 10 mph). When I pull back to full wide, the vehicle then appears to be travelling at the same speed that my eyes see.

What confuses me even further is the fact that the illusion of faster speed is seen by the naked eye with a narrower field of vision. Exactly opposite of how the camera behaves.

Thoughts, comments?

Thanks in advance for your help.

regards,

-gb-

This is one of those tricks that the mind and eye play on each other. The mind needs points of reference to accurately perceive speed. When the lens is zoomed in on the pavement, the eye loses any stationary point of reference.

The mind and eye play the same tricks with the size of objects also. This is why the eye is never an accurate judge of size. Have you ever seen the moon rising on the horizon and it looks huge? Then later in the evening when the moon is higher, it has somehow shrunk or the the earth has adjusted it's orbit and the moon is much, much smaller. What is happening? The mind judges all things on the horizon as being a great distance away and if seen, they must be very large, such as mountains, skyscrapers, the moon etc. But things overhead are perceived differently by the mind because they lack the reference of the horizon. Next time you see a large moon on the horizon, get a quarter from you pocket and hold it next to the moon at arms length. You will actually see the moon shrink to the size of the quarter (it's actual size) right before your eyes.

I've been doing this trick with my photography students for years. It never ceases to amaze them.

The diameter of the CoC is determined only by the size of the aperature, not the focal length
That's a confusing statement... the diameter of the Circle of Confusion is actually a subjectively-determined value that is dependent on the size of final blowup. The more blown-up the final image will be, the tighter the circle of confusion needs to be in order to survive the magnification process. 16mm DOF charts use a tighter CoC because 16mm DOF charts are designed to tell you what the depth of field will be WHEN BLOWN UP TO 35mm.

Circles of Confusion are the most variable element in the DOF formula, because the CoC is entirely dependent on the final viewing format. Film cinematographers prefer to see their dailies projected on the big screen, rather than video dailies, because something may look perfectly in focus on the small monitor but be quite out of focus when projected.

The more your image will be magnified, the tighter the CoC needs to be, which is why 35mm has the largest CoC and 1/6" CCD's have the tightest. The CoC isn't dependent on aperture or focal length or anything else; it's a subjective value that is tied to how large you intend to blow up the final image.

Thanks for the info Jeff. I will do the quarter trick. The horizon reference is also possibly the explanation for watching a sunset, or sunrise. Ever notice how 'fast' the sun goes when it gets next to the horizon. It goes from full sphere to completely gone (sunset) in less than 5 minutes. Same statement applies. Does the earth suddenly turn faster when the sun is near the horrizon, of course not.

regards,

-gb-

Ok, enough of my questions. We now return this thread to its 'circle of confusion' regarding 'circle of confusion'. (big grin)

<<<-- Originally posted by Boyd Ostroff : At the end of Spellbound there is a shot viewed from the perspective of Leo G. Caroll as he points his gun at Ingrid Bergman, then turns it towards himself to commit suicide. Hitchcock wanted the hand with the gun and Bergman to both be in sharp focus. His solution was to build a giant hand and gun! -->>>

Just a followup on this. A couple minutes ago I turned on the TV just as Turner Classic Movies was showing the final minutes of this film. Watching that scene again I noticed that while the giant gun was quite convincing, the giant hand was actually very comical and the details of the knuckles and fingernails had an almost cartoon look to them! But even with the gimmick of the oversized props, the background is in focus but the hand and gun aren't quite. It also appears that they used some sort of silk effect to give Ingrid Bergman a soft look.

It's funny what lengths we'll go to in order to realize some crazy idea we've dreamed up... and Hitchcock was no different in this regard.

Ten years ago I shot a short that involved one character looking at his reflection in a spoon, with another character in the background, and both were to be in focus. I can't remember the specifics but I do remember that I had to use both a split diopter and light the set to something like an f11 (this was a 16mm project) to hold the deep focus. It was a real pain; the set was ridiculously bright, we were burning every light we had but the shot worked.

On DV, it would have taken hardly anything to achieve the same (or better) results.

Michael, your post got moved to Photon Management (lighting).