Wireless Auto-focus System

sean90291

Veteran
Pulling focus in run-and-gun situations is too hard with a sensitive HD camera. And auto-focus systems usually jump around all over the place, as they're unable to decide what to focus on. So...my brilliant idea is a wireless system that maintains perfect focus with a subject at all times by way of a "transmitter" that transmits from the subject to the camera, telling the camera precisely how far the subject is from the camera at all times.

For example, you have Anderson Cooper running around post-hurricane damaged New Orleans. You have your Red camera in hand. And behind Anderson's ear, or on his lapel is the transmitter. When Anderson steps away from you, the camera automatically re-focuses for a further distance in precise lockstep with the speed at which Anderson is moving. He steps closer, and again the camera is remotely re-focused. If he even bends forward, the transmitter will automatically send a signal to tell the camera to adjust focus for those few inches.

The transmitter should be small and placeable at a point that approximates the person's "eyeballs"--that is how far the subject's eyes are from the camera, since the eyes are the focal point the camera should be concentrating on.
 
If this means two points of transmission -- one on camera or even into an accessory as a follow focus remote and the other with the subject, I think you already won my vote.

It would be "the" great add feature and it will sell this camera as popcorn in any Stars Wars release.

A film camera with the better glass available better than the Auto Focus of any video camera?

Imagine how many ENG shooters would like to have a feature like that to run & gun work...

Excellent idea! And I even bet it will not be difficult nor expensive to get it.

Are you listen Jim? Or Ted Schilowitz, the leader of the rebellion? Even as an accessory to any camera available in the market... Are you already thinking to have RED accessories to other cameras? This would be truly revolutionary... Even more that already is!


EDIT -- And I hope the same opinion from the jury itself, specially, from the RED team in order to really implement that. If not why would we have this contest?
 
Last edited:
I had thought of something similar and Sean90291 describes it perfectly. Can you imagine the area you could use this in......Cars, Live action Sports etc.
If you could get the data recorded as well you would have a perfect compositting tool too! especially if you could have a primary(which Red focused on) and then subsidiary transmitters which could be used in post.

Good luck Sean90291!!! Great Idea.
 
This is an excellent idea! It should be done 5 years ago! I bet that if RED team ignores it, someone soon or later will make it a general accessory but a direct camera support will be a really great advantage.
 
as i posted some hours ago, Emanuel you are an institution.. good ideas to jim.. are you a RED consultant?.. eheh

sean90291, if the jury doesn't think the same than ourselves, they're wrong.. your idea is the best.. kudos
 
this is a great idea! id put the transmitter out of sight under ther hair right above the hairline on top of their head. there is a similar device in existance that uses triangulation to hol focus. i think it was used in some sort of track and field film. i like you idea much better with a singular point for tracking. ..a sort of localized GPS. make it manually overideable and its a great addition to any set
 
You could have different transmitters on different frequencies. Change the # on the Red camera and suddenly it rack focuses to the other transmitter. Say you have a reporter and two subjects in the field (or 3 actors on the set). If each were wearing a transmitter emitting a different frequency, you could toggle between your subjects' focal point.
 
But what if I want to put the transmitter, say, behind the person's head (like clipped to their colar) out of sight? ...

How about alow the user to program an offset from the avtive focal plane to the transmitter. Let me tell the camera that it should focus 4 inches closer than the transmitter (this beeing the eyes if the TX is on the back colar)

- Mikko
 
This idea would only work with Red lenses instilled with a smart focusing system, not a cooke or zeiss etc. etc.

Good idea for Red only lens users but pointless for others....

Doesn`t get my vote.
 
mikkowilson said:
But what if I want to put the transmitter, say, behind the person's head (like clipped to their colar) out of sight? ...

How about alow the user to program an offset from the avtive focal plane to the transmitter. Let me tell the camera that it should focus 4 inches closer than the transmitter (this beeing the eyes if the TX is on the back colar)

- Mikko

Good point. The details could be customizable for such situations (offsetting by X inches as the case may be).

This idea would only work with Red lenses instilled with a smart focusing system, not a cooke or zeiss etc. etc.

Good idea for Red only lens users but pointless for others....

All the more reason for Red to have specialized lenses for certain circumstances, like in-the-filed run and gun. True, it wouldn't work for other lenses, but there could be an automated follow focus mechanism that could be calibrated to other lenses), which would probably be too much...but for any situation where you had time to pull focus with other lenses, you probably wouldn't want the "transmitter" anyhow. You'd have time to set up and control your focus. In run and gun, you might have no choice but to use a lens system with an accurate automated focus system.


Oh damn. I just noticed that I'm not supposed to post more than one entry myself in a thread (only comment on other posts). Hope I'm not disqualified. I won't post anymore here.
 
Last edited:
I also had this though and wondered what barriers existed from keeping this from the market place. Then I thought about critical focus situations with shallow DOF and how a leaning actor or one that had turned their head might miss the XCU on the eyes. Crazy thoughts like a transmitter that was sensitive to angle of the body and could be given offset settings as well crossed my mind but ultimately the system wouldn't work perfectly all of the time. Never the less it would be very useful in most situations. This reminds me of a system that video conferencing uses where the speaker carries a tiny transmitter that the camera tracks by motorized pan and tilt. The set up was pretty inexpensive too. Seems like the technology could be barrowed from that. A real world option Sean. Hope we get it in Red
 
Maybe an ultrasonic transmitter on the subject synced to the camera beforehand may be the trick...the camera would know how far away the subject is because it knows what time it is, and therefore how far away the sound of the ultrasonic transmitter is.

This might be a very simple solution to the problem.
 
dailyrushes said:
Maybe an ultrasonic transmitter on the subject synced to the camera beforehand may be the trick...the camera would know how far away the subject is because it knows what time it is, and therefore how far away the sound of the ultrasonic transmitter is.

This might be a very simple solution to the problem.

(to amplify my own post here)

The ultrasonic transmitter, having been synced to the a device at the camera, would continue emitting an ultrasonic sound which, as the subject got further/closer to the camera, would have an easily-measured "difference" as sound only travels at 660 fps. Therefore, the camera focus device (a simple attachment) would know exactly how far away the subject was, barring errors from reflections, etc.

The foucus aid would then measure the difference (delay in arrival of the sound) and know exactly how far away the subject is. This error voltage could then be converted to a signal to focus the camera through an appropriate interface.

You could even use two or three of these and rack focus between subjects in the scene with the flick of a switch or something.

With available current technology, the transmitter could probably be as small as a button on a shirt, or taped to the actor's body beneath the shirt.
 
I was in the mood to post a few ideas about this idea...

I was in the mood to post a few ideas about this idea...

About ultrasonic method:
The sound speed is affected by many factors like the air density, temperature, humidity, air speed and by the subject and/or camera speed. So you can’t just “sync” the ultrasonic transducer once and rely on it.

For example, in a situation where you have a medium-strong wind of 30miles/h (~50km/h) this will affect the sound speed by as much as 4% (wind speed per second /sound speed per second = 50000/3600/340) so if the subject is 33feet away (10meters) there will be 1.32 feet of inaccuracy (40cm) and this will be variable depending on the wind direction.
Of course if the subject is running to or from the camera with that speed, even if there is no wind the effect will be the same.

There is a way to minimize these errors if the camera will send a short ultrasonic signal to the unit on the subject and the unit will respond by sending immediately another short ultrasonic signal back to the camera. The camera then, will measure the delay between the sent signal and the received signal (divided by 2) to find the distance.
Assuming that the wind conditions and all other factors doesn’t change instantly (unfortunately wind *does* change instantly many times) and assuming that the same conditions remain while the ultrasonic signal travels from and to the subject, the sound will travel faster in one direction and slower at the opposite direction at the same amount, so the errors will be subtracted. Also, the camera should measure the temperature to compensate for this important factor.

Unfortunately the ultrasonic sound is directional but maybe this can be solved with a point sound source, but what if the subject turns his back to the camera? Two units will complicate things further.

Cons:
-Visible contact with the transducer on the subject required
-Two units required
-Low precision in some cases.
-Requires temperature sensor.
-Rare but possible interference from inaudible sounds (machines, automobile’s breaks, pressured gas, etc)
-Cannot pass through glass
-Only for sort distances.
-Prone to reflections

Pros:
-Each subject can be distinguished
-Lower cost

Infrared idea:
I thought of this idea: It is possible to use a multi-angle reflector on the subject to send an infrared beam instantly back to the camera, having the camera measure the light delay -just like the professional hand-held distance-measurement instruments using lasers ! Of course these does not use reflectors, the laser beam is strong enough to be reflected on the subject’s surface.
But a reflector instead of a receiver-transmitter unit is a huge advantage! You can put two of these if required, it will be low-cost, not harmful as is the laser and not as directional. If the emitter is strong enough, a well-designed multi-angle reflector will do the job perfectly!

Knowing that the light speed is 300.000.000 meters/second it takes 3.3 nanoseconds per meter but because of the light reflection, the delay will be 6.6 nanoseconds per meter (1 meter=3.32 feet)

With this method the measurement won’t be affected by any environmental factor except the circuit’s measurement accuracy.
(It might not work in fog, but there is no point to focus on someone you don’t see!)
Both the infrared transmitter and the receiver on the camera should have a lens to be able to work in a relatively long distance and maybe optionally a wide one to track the subject moving in the frame in a short distance from the camera. Also, for short distances a smaller reflector could be used.

Cons:
-Visible contact with the reflector on the subject required
-It might require dual emitters and receivers on the camera with different lenses for short and long distances.
-Different subjects cannot be distinguished
-a reflective object passing, might give a false reflection, but most false reflections can be screened.
-Higher cost
-Maybe too bulky to integrate on the camera...or maybe not.

Pros:
-Very high precision
-A passive low-cost multi-angle reflector can be used on the subject
-Not affected by environmental factors
-Lower cost for more than one subject
-Light weight (for the subject)
-No interference issues
 
Last edited:
I see your points on the ultrasonic system. Now that you've solved that, let's throw that one out!

On the infrared system, that looks somewhat promising to me. I vote for further research and responses!

On a different tack, it seems to me I've seen someone mention he used goggles like they use for gaming applications displaying the scene tapped from the camera's video output. Now that might prove useful...providing you could see in the resolution that they get on a large screen like a focus puller with a big screen can do nowadays.

You could just use one eye, and the other is on the scene and the camera. That might be interesting! Boy, if we're not careful, pretty soon, with technology we're going to eliminate every job but the director and the actors!
 
dailyrushes said:
Boy, if we're not careful, pretty soon, with technology we're going to eliminate every job but the director and the actors!

It might sound... scary, but I think in 50 years the actors might have a very limited role. There is a heavy ongoing research and development (and some projects are in many ways secret) about computer autonomous characters. I believe that *much* sooner we won't be able to distinguish a human from a computer character, not only by the looks but also by his behaviour...

I think it's going to happen what Benny Hill (the comedian) had envisioned in one of his episodes (a long time ago) where it was supposed that while viewing a movie on TV he could use the buttons on his remote to decrease or increase the violence, sex, etc in real-time, just like the brightness and colour saturation! It was about a normal rather boring scene in a hotel room, with the room service (a girl) coming inside to serve the breakfast or something. That scene converted to total chaos when he turned sex and violence to the max. I remember I was laughing to tears!

Uhmm… sorry, that was totally off topic!
 
Back
Top