What's so hard about a DIY Follow Focus?

[Dan Diaconu]

>>>>>>>>>>In other words; all the electronics out there can't replace the judgment and response of a human being who can pull focus. You need to be able to react instantly and intuitively<<<<<<<<<<<<<

For now.
But,
With all due respect, I am just about to further demonstrate the opposite (using the 35 image converter instead of an Arri or another film camera) and the focusing system I have made.
Hold your thoughts for a second and follow this rationale:
Scenario:

Camera to actor 6 ft.
Lens 135mm.
Aperture 2.8.
DOF a few inches.
Actor is on the mark never moving his feet.
Actor is "arguing" and leaning back and forth. Not much, but just enough to drive the first AC crazy. He can't say what he has in mind, for that would be his last day on the show...... (he got used to it anyway, so it don-hurt-no-mo...)
(BTW, this is a dirty OTS and he is leaning on a diagonal line not straight on the optical axes of the lens !!!!)
Panatape or digitape shows the first AC (at best and if aimed at his face and never reading the BG as he leans in and out)
the very precise distance of 5' 8" to say 6'3".
Marks on the lens:
5' , some 30deg later a 5.5 and some other 20 deg further the 6' and so on....
There is no indication of 5.8 and 6.3.
So... where is that 5.8 and 6.3???
But the first AC knows all that and he is taking marks (on the white donut eyeballing the viewfinder) during rehearsals.
Action is slightly different from rehearsals (faster/slower pace, more dynamic, etc)
Rolling focus in and out with action on a lens that has no marks whats-oh-ever for the amount of movement in the field has nothing to do with science. Is pure educated and polished by practice intuition or second nature.
There is no time to read distance and conform the lens. By that time, the action is long gone. INSTINCT IS ALL it takes (for now)! So, I second your respect for this skill.
Good? As good as many years of eyeballing distances can get a pro. Perfect? Nope.

Scenario 2:

Kid on a swing.
Same 135 (or let's be reasonable here and pick an 85)
Average swing movement? Let's say 5 ft.
Camera on dolly coming from 25 to 15 ft diagonal.
The take is 30 seconds long.
Infinite combinations of positions between camera and the kind's position on the swing lead to the ultimate nightmare for the first AC.
Mark zeat. Rehearse zeat. Panatape zeat.

I have just done "zeat" (not only in the clip I am on the swing) but actually dolly-ing in on 135mm from 12 to 5 ft while a matchbox was swinging from the kitchen light via dental floss thread. Clip is kinda long to post, but "zeat" (and other shots such as following a soap bubble and sprinkler water jets on 135mm at 2.8) might as well have been the reason behind the Canadian' Academy to award me the Gemini for tech achievement.

Now, obviously the question: if is soooooo good, why is it not in use?
Answer: a premature views discrepancy between my initial business partner and I, lack of promotional funds as well as FEAR of first AC's that ANY new-kid-on-the-block using the device could replace their polished skills and positions led to the current situation.
Film (for which it was designed) is going down and for 35mm image converter is way too expensive.
I will shoot further tests to demo the above but I will knot waste my time anymore to make a point. Not worth it. Just for the luve-of-it. nil homini naturae sine magno labore dat
(for all of us: der aint no free lunch) I am not playing S M R T here but is fun and that's all I remember from first year in HS of latin.
(they took it of the qurikulim after zeat.) so... don ask fo-mo.
No humans/animals were harmed during any of the above real and fictional takes.
Casualties: soap bubbles...,.... water on the ground (oh well...and the DV tape....;-)<
(Just entertaining my virtual friends here......hey?(a'la'Canada)

[Charles Papert]

Dan:

Also with due respect--I'm having a hard time following you. I re-read the thread to see if you already referenced this device--perhaps you have been discussing it in another thread that I haven't seen. And can you explain what you mean by "zeat"?

If you have developed a system that can accurately pull focus under the situations you describe, I applaud you and would still encourage you to actively pursue development, regardless of the issues you describe. I have several friends who have pioneered radical film equipment and faced the various battles you describe. Yes, AC's will tend to resist things that "undermine" their skill level but mostly if they don't work or are unreliable, or difficult to use. But that doesn't mean there isn't a place for it. I worked with a prototype system that required an actor to wear a sensor like a wireless mike, and transmitted distance back to the camera. It was in a rudimentary form, and it seemed to me to be delivering confusing information, but one day that might come to market. Preston Cinema's Light Ranger system is also occasionally in use for difficult shots like the swing scenario you described, but it is reportedly a bit cantankerous to work with.

<<Film (for which it was designed) is going down and for 35mm image converter is way too expensive.>>

It's pretty clear that high end HD has already embraced a 35mm sized image sensor (Genesis, D20, Dalsa). Certainly the Mini35 is not cheap compared to the cameras it is attached to (then again, those cameras are amazingly sophisticated for their price, so that is the wild card, not the adaptor), but perhaps one of the hard-working DIY crowd will eventually come out with a commercial version.

Don't get me wrong, I'm interested in new technologies myself. I think a device like you describe would be pretty handy--for certain situations. But since you have a pretty good understanding of the process, I'm sure you'll also agree that pulling is much more than just keeping a subject in focus; many shots will require pulling between two characters with specific timing and at specific rates that are entirely up to the assistant to determine based on his judgement, the dialogue, the energy of the scene, etc. Sometimes they will calculate splits to hold two characters in partial or full focus; sometimes they will opt to roll the focus a little bit deeper so that the foreground person is just inside acceptable focus so that the background person can be as sharp as is possible--or vice versa. Again, all of this is based on human judgement and cannot be duplicated by any device that says "show me the subject, and I'll keep it in focus".

[Dan Diaconu]

Thank you for you re Charles.
By "zeat" I mean "that".
The system does not pull focus by itself. It is a visual indicator for the first AC.
In very short:
two tiny CMOS/CCD (surveillance type) on the sides of the MB counter rotate in sink with the lens of the film camera. Wherever the images overlap, that is where the focus is. The first AC could not care less, if the actual distance is 6.2 or 6.3 or anything else. As long as the images overlap on his monitor (wireless) he can use his FIZ (as he normally does) to operate the focus ring on the lens and SEE what he is doing on the monitor without guessing the distance.
As for "split" focus, all he needs to do is keep the two images NOT perfectly overlapped by equal amounts (hence his actual focus being in the middle) carrying both actors within the DOF of a certain lens (more or less subject to distance, focal length and aperture setting)
An encoder on the focus ring reads the movement of the lens and “translates” this movement into a circular counter rotating movement for the two CMOS cameras.
Calibration to a lens takes one minute and is based on three points (plus infinity which is the same for all lens - the two CMOS are parallel at this starting point)

>>>>>>>Again, all of this is based on human judgment and cannot be duplicated by any device that says "show me the subject, and I'll keep it in focus".<<<<<<<<<<<

I agree with you on this. The device itself does not do the job for the focus puller. It is up to him to actually pull focus, BUT:
this time he can be right on (to the very inch) at any given time on anything that can be photographed. (including but not limited to FIRE, WATER, SMOKE, MIRROR REFLEXIONS, SMALL OBJECTS (relative to the frame size) and other "hard to get" the least to say situations.
You can see some pics and clips here:
http://dandiaconu.com/gallery/Oustanding-Technical-Achievement

Charles, my email is on the website if you need it.
Sorry to go off topic for all that are not interested.

[Charles Papert]

Very interesting Dan. So essentially, this is based on the rangefinder focusing principle--the assistant dials the lens until the two images of the desired object overlap, right?

Will this system work beyond a certain distance, such as a super-telephoto image like someone running at the camera on a 600mm or longer (seen in full figure or wider)?

In the scenario you described, where the camera is dollying in as the subject is on the swing, isn't it still pretty tricky to pull focus using this method--I would guess it takes a bit to guess where to stop and reverse each end of the "swing" since it changes each time, due to the camera getting closer. Certainly I can see the advantage when using a still camera lens with poor markings, or how an inexperienced assistant could achieve better results with this perhaps more intuitive system...like I said, very interesting.

I'll have my assistant check this out; he was actually one of the designers behind the Panatape so I'd be curious to see his reaction to something like this.

[Dan Diaconu]

>>>>....dials the lens until the two image....>>>>
Yes.
>>>>Will this system work beyond....<<<<
Yes again. Replacing the middle tube with a longer one (increasing thus the base of the triangle) and replacing the lens on the CMOS with longer focals (set of primes to match the Zeiss primes, yes, they are interchangeable and have their own tiny MB, no filters though) longer focals can be helped:
(I have just uploaded another picture with the lens kit) (same album, last picture)
Longest lens I had it calibrated was 250mm and gave focus to the first AC (that was checking it out) whatever he aimed the camera at.( way beyond measuring tape's range: trees, street signs and people on the sidewalk.)When that was OK, I got it up on the camera as you saw in the pics.
>>>>....it takes a bit to guess where to stop and reverse each end.....<<<<<<
Please watch the first clip 5 years... (on that page)
With each swing, the movement decreases. Not a lot, but IT DOES. It takes about 90 sec for it to settle.
I have been rehearsing my "moves" for over two hours. (practice) (I have pulled focus before a few times by eye and tape and marks on Zeiss and Panavision.
Without it, I failed miserably with all the marks that I could have even wanted under my nose and the "moves" practiced for a while....
Lens marks are not linear. The movement is accelerated towards close focus. A "rhythmic" back and forth will not do.
Have some fun (in a break) if you get the chance with your assistant and see. Have him see that clip first, please, so he knows the set up.
Is the easiest set up ever. A tennis ball and dental floss. Swing it two-three-five feet or whatever the ceiling allows. Lens 50mm wide open. Distance 5?-6? ft. On axis (not diagonal as I have done it). Free to use panatape/cinetape. Have him hold focus longer than 10 seconds. (done it for 2 minutes)Look through the viewfinder. You be the judge. If you could get it on tape to see how many times during one swing he will be soft (behind or ahead)….. but that would be too much trouble. (I have done it though)
Thank you for checking the site.

[Dan Diaconu]

BTW,
If the above is successful, next is having two balls, 8-16 inches apart from each other (diagonal on the ceiling) and also 8-16" apart at different heights (one edging the bottom frame and the other one the top of the frame. Camera is locked. No pan, no tilt. Shift focus from one to the other at will or by request.
At different heights, they will have different oscillation (frequency). When this one is mastered, dolly in (out) while they swing. (I did not do the last one myself, but I did shift from the "real thing" to its reflection in a mirror (some 4 ft away from the real thing) while dolly-in). Next is the same ball describing a circle....smaller and smaller (I used a candle for a "dramatic effect")

Oh well, the list is long, but the fun is worth it.

[Steven Fokkinga]

Hi,

Just a stupid question in between all this die-hard technotalk :)

I'm planning to make my own FF together with the micro35 (when is that guide coming?! :), and would like to know from one of you pros: how far do you normally turn a FF knob from one focus end to the other? Is it one turn, multiple turns, or isn't there really a standard?

I was looking for lenses (nikons) today for my future micro35 and already figured out I need MF since there focus ring turns much farther from end tot end and much smoother than AF.

Thanks, Steven

[Dan Diaconu]

Some FF units have two speeds, most only one.
You should be able to cover infinity to close focus on your Nikon lens in one hand turn (or a bit less)
MF have that range spread over 200 deg (more or less) . Find a larger dia. gear to mount on the focus ring of the lens, so when engaged to your FF, the hand knob will rotate about 300 deg. Do not get stuck with the above figures. Any alternative ratio that you can find works for you is BEST. I am glad you pursue it.

[Oscar Spierenburg]

I have a separate viewfinder on my (double DV) 35mm adapter, with DOF and nice focusing circle in the middle. It has almost the same lens as the adapter itself. I wounder if it is possible to make a follow focus that simultaneously focus both lenses. Maybe a wire between them or something. (they are a bit apart and parts of the adapter between them)

If someone has a suggestion...(putting all my bets on Dan)

[Dan Diaconu]

"Almost the same lens" spells trouble.
If they do not rotate in sink (via gear) you do not see what you record! Infinity (as well as close focus) should be identical for both, same as here:
http://images.google.ca/images?q=rolleiflex&btnG=Search&hl=en&lr=&client=firefox-a&rls=org.mozilla%3Aen-US%3Aofficial_s

any size gear will do:

http://www.care2.com/c2c/photos/view...DER%20LENS.jpg

(make sure there is no backlash and not too tight if you use gears)
Otherwise, O-rings on bearings will do just as fine. Cross the O-ring (as figure eight) so it does not slip and add a thin plastic sheet in between to cut the friction. Good luck.

[Oscar Spierenburg]

Dan, the second link is to a restricted area.

Thanks for the info. I'll look for an identical lens.
Just one more thing about my camera (adapter), it is to big to put the second lens next to the camera's lens. This image explains the difficult setup:

http://doublecam.250free.com/viewfinder.jpg

That's why I figured a wire (flexible axis) or something would help to extend the follow focus .

[Dan Diaconu]

You will get quite a bit of parallaxe there....
argh........fo-get-a-bout-it.... a lot of headache and you will be stuck in gears for months without consistent results....

For the set up you have, I would slide the mirror back about 17-18mm and mount a MC filter (that will act as a mirror) at 45 deg (closest to the GG) and a tiny CMOS survailance type on top to capture the image of the GG (reflected in the filter)and display it on a monitor (of your choice) You can mount the CMOS so you see it up right. They (CMOS) do not need A LOT OF LIGHT to give you an image (remember, it "steals" the image of the reflection in the filter) they are cheap (less than $20) and your camcorders will not suffer any amountable light loss or image degradation.
(this is kind of a tiny prompter or videoassist for film cameras)
You could have it up and running in no time!!!! (I guess)

[Oscar Spierenburg]

Hmmm...too bad for me. But the idea of the mirror and surveillance CMOS is quite good. Thanks.

[Dan Diaconu]

why is too bad? Are you stuck with the previous idea?Let's "grease" the mind:
Steadicam, crane, dolly. How do you frame (with what you wanted to do?) CMOS can go in a tranmitter and you can have the pic wireless......
I'll leave it with you.....

[Oscar Spierenburg]

Too bad because I already made a separate viewfinder mounted on the camera. Well, I can always use it for a directors finder. I like the idea of the CMOS.