Why I believe "open patent protection" might work.




ÒTrue TallyÓ

Providing Real-Time Data For and About Subjects Within a CameraÕs View

 

Michael Naimark

University of Southern California

Los Angeles, CA

 

 

Abstract

 

This invention is a specific enhancement of the Òtally lampÓ or Òcamera indicator lamp,Ó the red light visible on the front on virtually all video cameras when they are active. The specific enhancement is to make the light only visible to subjects when they are actually within the view of the camera, which may be called a Òtrue tallyÓ lamp. It can be achieved by placing the tally lamp behind the lens, on-axis with and the same size as the camerasÕ sensors, and pulsing on out-of-phase with the sensorsÕ activity. A preferred embodiment is described, based on creating through known means a uniform light source that closely resembles the camera sensors such that it can easily mount like one. Additional features include modulating the frequency of the true tally lamp to appear flashing as well as continuous and modulating the color of the true tally lamp, both enabling additional information to be conveyed to the subjects within the frame. Finally, the invention described here is a step toward an integrated on-axis projector/camera, or ÒprocamÓ system, a lively area today in the computer vision field as a means for enhancing both cameras and projectors.

 

Background

 

A Òtally lampÓ is a light on the front of a video camera that lets subjects know when the camera is active. Its original use was in multi-camera video studios such as newsroom sets, so the commentators could know at which camera to look as the director in the control room switches between cameras. Tally lamps on studio video cameras are typically big and red.

 

The tally lamp of the professional studio camera migrated into todayÕs consumer video camcorder as a standard item, where it is less prominent (e.g., a small red LED) and can be disabled via the cameraÕs settings menu. Many consumer camcorder companies now refer to them as the Òcamera recording lamp.Ó They are unobtrusive but visible to anyone looking for them.

 

ItÕs noteworthy that motion picture film cameras historically do not have tally lamps. In professional movie production, they are not important because everything is under the control of the director.

 

But in todayÕs world, where video cameras are cheap and ubiquitous, issues of control are not so simple. How important is it that subjects know when theyÕre being recorded or broadcast? Is it useful to let your daughter know when youÕre recording during her birthday party? Is it a simple courtesy to let your classmates know when youÕre recording a seminar? And are there more serious, sometimes legal, implications to not letting strangers know when youÕre recording them in a public place? In our new world, in which surveillance and privacy have become flashpoints for critical debate, cameras and their technologies are under scrutiny, offering both challenge and opportunity. For example, one well-supported area of current research is around technologies that automatically detect and hide faces from public surveillance cameras. The face data is actually stored under digital Òlock and keyÓ and only accessible by subpoena if required.

 

Another possible opportunity lies in providing more feedback from the camera to its subjects, like tally lamps. But tally lamps have a problem: they only indicate when the camera is active but give no indication where the boundaries of the cameraÕs view may be. Potential subjects in the general region in front of a camera cannot know if theyÕre actually in the view or not. And with zoom lenses, even if the camera is pointing near the subject, itÕs impossible to know the size of the field of view.

 

Disclosure

 

The solution is to place the tally lamp behind the lens, on-axis with and the same size as the camerasÕ sensors, and pulsing on out-of-phase with the sensorsÕ activity. If the tally lamp is a uniform light source whose surface dimensions match those of the sensors, the light refracts through the lens and is visible in the exact field of view seen by the sensors. It would appear to the subject as a light coming out of the lens. One might call this a Òtrue tallyÓ lamp.

 

The true tally lamp light source would need to be on-axis with the sensors. Many video cameras today have three camera sensors, for red, green, and blue (sometimes even a fourth) all on axis through the use of beam-splitting prisms or half-silvered mirrors. The true tally lamp light source would simply be an additional element using these known means for positioning multiple on-axis elements.

 

In order for the true tally lamp source to not interfere with the sensors, it would have to be pulsed on out-of-phase with the sensorsÕ activity. When it is on, it needs to come on quick enough to have no affect on the functional duty cycle of the sensors. When it is off, it needs to be completely off such that the sensors pick up nothing from it. Video cameras have two ÒfreeÓ opportunities to insert a pulse of light with no affect on the sensorÕs duty cycle: during the horizontal blanking and during the vertical blanking. For American video standards, horizontal blanking is about 1.3 milliseconds every 16 milliseconds and vertical blanking is about 10 microseconds every 60 microseconds. Given our eyesÕ retention ability, such pulsing, if of an appropriately strong intensity, would be more than adequate to see and could even appear continuous. Additionally, the camerasÕ sensors could have shorter duty cycles, for example, itÕs common for Òprogrammable shuttersÓ to be as fast as 1/1,000 second, leaving the remaining 99.9% of the duty cycle to pulse the tally light source.

 

Preferred Embodiments

 

The preferred embodiment would be based on making, through known means, a thin true tally light source of uniform light distribution out of a light-emitting semiconductor such as LEDs or OLEDs. It would be fast enough to pulse adequately, be sufficiently bright to be easily visible through the lens in daylight, be energy efficient, and produce no heat or other energy leakage that may affect the sensorsÕ performance. It could be made, through known means, to resemble camera sensor elements, such as the red, green, or blue CCDs. The more it resembles these sensor elements, the easier it is to mount like one.

 

The mounting would be through the known means of using half-silvered mirrors, or more commonly, prisms, to split the beam of light (which is the same in both directions) in such a way to keep all elements on axis. For simple Òone-chipÓ (CCD sensor) video cameras, such a beam splitter would allow a second chip to be added and for Òthree-chipÓ video cameras, such a beam splitter would allow a fourth chip to be added.

 

The simplest preferred embodiment for pulsing the true tally lamp is to pulse it on during the vertical interval, since it is a thousand times longer than the horizontal interval, as well as being a larger percentage of the duty cycle. It would pulse at 60 hertz and be on for one-sixth of the duty cycle, resulting in, worst case, a partially noticeable flicker (depending on what region of oneÕs field of view itÕs in).

 

In a preferred embodiment, the true tally light would complement rather than replace a regular tally lamp, since they provide different but complementary functions: Òcamera activeÓ and Òcamera pointing at me.Ó If the tally light is red, the true tally light may be a different color, but red is probably still preferred.

 

Optional Features and Future Applications

 

The simplest additional feature for a true tally light is to modulate its frequency, which requires no change in hardware. Hidden during the vertical interval, it will appear almost continuous, allowing it to flash at human-detectable speeds, like once per second. At the very least, information could be conveyed to the subjects by having the light appear continuous or flashing, and additional information if the light flashes slowly or quickly. For example, this information could be to distinguish when the camera is active or in standby mode.

 

A more ambitious additional feature would result if the true tally light could emit more than one uniform color, for example, by constructing it of red, green, and blue LEDs. Even more additional information could be conveyed to subjects if the true tally light could change color.

 

Finally, a very ambitious additional feature would result if the true tally light could emit more than one uniform color and be modulated on a pixel-by-pixel basis. For one thing, it would allow specific colors to be conveyed anywhere in the cameraÕs view on a pixel-accurate basis. It could, for example, Òsend a messageÓ to only one particular subject (identified through known computer vision techniques).

 

But even more ambitious, such a light source could double as a projector, perfectly on-axis with the camera. Such projector/camera configurations, when used in phase as a closed-loop system, have numerous applications for expanding both camera capabilities and projector capabilities. In computer vision, these systems even have a name, Òprocams,Ó and are currently a lively source of activity (e.g., see www.procams.org). Procam projection can be both low-energy, invisible to the naked eye but useful to project detectable structure, for example, to help determine depth, or high-energy and visible to the naked eye, useful where the camera provides feedback for the projection itself, for example, to keep the projected frame aligned when the projection screen is moving in real time.

Some additional in-phase applications are:

- as a light source for retro-reflective ÒmarkersÓ used in motion capture;

- if the light source is strong enough, as the means to actually illuminate the scene and subjects, either as a flash or continuous light. On-axis illumination of subjects results in absolutely no shadow effects.

 

No fully integrated procam system currently exists.

Additional Features

 

1. Single integrated sensor + light source chip

The original disclosure assumed that sensor chips and light source chips must be different chips, placed on axis through the use of known means such as prisms and partially silvered mirrors. There is no reason why a single chip cannot have both sensors and light source integrated together, for example, through micro-optical engineering. Such engineering is common for integrating different color sensors on the same chip. The advantage of an integrated sensor + light source chip is size: just like single-chip color cameras, no additional prism or mirror optics is required.

 

2. Integrated focusing lens for the light source chip

It would be most efficient for the light source to ÒaimÓ all of its light out through the camera lens with minimum scattering in other directions. This can be accomplished several ways, such as external focusing optics in front of the light source chip or integrated in the chip itself.

 

It may, however, be important to minimize the effect of stopping down the camera lensÕ aperture on the view-ability of the tally light, for example in bright sunlight. This can be accomplished by further focusing the light source chip primarily through the center of the camera lens.


3. Use of Class IIIa lasers

ÒClass IIIaÓ lasers emit no more than 5 milliwatts of power and are the class of lasers used in common laser pointers. These laser pointers have proven to be safe to both cameras and human eyes, and can be found for sale on the Internet for one or two dollars apiece. The core of these lasers is a very tiny, very inexpensive, solid-state laser emitting diode.

 

Using laser-emitting diodes as the light source may simplify the design for this invention due to the coherent nature of laser light. For example, the need for the tally light source to be the same size as the camera sensors may not be necessary. Clever placement of the tiny emitter may also allow for smaller alternatives to sensor-size prisms or half-silvered mirrors.

 

Class IIIa lasers work very well when used in the Òopposite direction:Ó when a laser pointer is aimed directly into the lens of a digital camera, the entire field of view is temporarily neutralized, or Òzapped,Ó at distances as far as several hundred meters. [See Naimark, M. (2002). How to Zap a Camera. (self published on www.naimark.net).] ItÕs reasonable to assume that this process may be generally reversed.


Back to Projects Pending 2010