| Some
little-known or little-appreciated details are given about video and
photographic imaging systems. These can cause false image artifacts and
consequently mislead investigators.
The Image On Your TV Set Some people are amazed to hear this, but here goes. There is actually no picture on the face of the picture tube on your TV receiver! At any one instant, all there is is a tiny, brilliant spot of light with a comet tail, somewhere on the screen. This spot of light flies across the face of the picture tube from left to right, and from top to bottom, to paint the rectangular area. The response lag of your eye is what averages this zipping sunlight-brilliant spot of light into a complete normal-brightness picture. In order to present details, and not just a blank white "raster", the intensity (and on color sets, the color) of the spot is varied, or "modulated", up to millions of times per second. The most obvious way to scan the face of the tube is to go from left to right and from top to bottom in sequence, that is line 1, 2, 3, 4, 5, 6 and so on to the bottom, exactly like your eye reads a printed page (unless you are Chinese!) and then repeat. This method is known to video engineers as "progressive scan" and to computer people as "non-interlaced scan". If you are peering closely at a computer screen, this is the way to go. Painting the entire screen 60 or more times per second prevents flicker and "line crawl" (described below.) This is also the best way to display video, and it is hoped that the future HDTV (High Definition Television) system will allow for this type of scan. In the adoption of today's television standards, practical considerations reared their ugly heads. The intent of NTSC (National Television Systems Committee) was to standardize for the United States a broadcast system that had 60 frames per second (to be above the eyes flicker threshold) and that displayed the same amount of picture detail as did a professionally produced 16mm black & white film print of the era. It was quickly established by a few calculations that this exceeded the available space in the VHF (Very High Frequency) broadcast spectrum if more than one or two channels were to be provided in each metropolitan area. This, of course, was decades before even transistors, let alone todays digital compression methods that will be used with HDTV. A compromise was reached involving "vestigial sideband AM" (way beyond the scope of this article) and "interlaced scan". For adequate picture detail, it was essential that there be 480 active lines, 525 total in the picture. Since it was only feasible to transmit half that many at a time, a technique known as interlace was adopted. This means that half of the lines (the odd-numbered ones; 1. 3. 5 etc.) are transmitted in the first "field" of 1/60 second, and the other half of the lines (the even-numbered ones; 2, 4, 6 etc.) are transmitted in the second field of 1/60 second, for a total of 30 frames per second.[1] This is a reasonable compromise, assuming the TV viewer is far enough away from the screen that he cannot quite make out the individual lines so the even and odd ones seem to be in almost the same place, so he does not see flicker. However, if you go up close to the screen you can see that the lines of light are alternating up and down; your brain interprets this as if the lines are actually drifting upwards and downwards. This effect is known as "line crawl".[4] The fact of interlaced scan has important consequences pertaining to VCRs as discussed in the next section, potentially causing 50-100% time measurement errors for the unwary. It might be pointed out that many incompatible standards exist worldwide, depending on local preferences (and pride?) Unless you have special multistandard or standards-converting equipment, you cannot play a video tape recorded with another video standard. It can have a different number of frames per second, a different number of scanning lines, the color recorded differently, a different tape speed, or all of the above. The Video Recorder First off, lets mention a few abbreviations that are often misused. "VCR" stands for "Video Cassette Recorder" and does not imply any particular size or format. "VHS" stands for "Video Home System" and describes a particular set of JVC patents and standards using 1/2" wide tape. Sony's Beta is from a Japanese word meaning the whole thing because that format, a predecessor of VHS and also 1/2" in width, was the first to record without guard bands and thus utilized the whole tape surface. It was not named after the Greek letter beta despite the fact that the latter is part of the trademark. "8mm" is named after the 8 millimetre (about 5/16") tape width. By way of comparison, an audio tape recorder only has to deal with frequencies up to about 20 kHz. ("kHz" meaning "kilohertz" or in English "thousands of cycles per second.") The video signal has frequencies up to 3 or 4 MHz ("megahertz" or "millions of cycles per second.") The upper response limit of a tape recorder is a function of the magnetic head's dimensions, the tape velocity, and the size of the ground-up bits of magnetizable rust that are glued to the tape surface and recorded on by the head. As it is not feasible to pull the tape past the head 200 times as fast (you'd have awfully large reels!) it was decided instead to make video heads move rapidly past the tape. This is done by attaching two video heads, 180 degrees apart, to the outside of a drum. The video tape is extracted from the cassette and wrapped halfway around the drum at a slight angle. The drum spins 30 times per second so one head records and plays field 1, and the other head records and plays field 2. The recorded tracks are only about .001" or .002" wide, each recorded diagonally on about 5" of tape. Now, when the tape is stopped in the "still" or "frame advance" mode, both heads follow the same path across the stationary tape and therefore both display the same field. If you are stopped on field 1, you will see lines 1, 1, 3, 3, 5, 5 and so on; if you are stopped on field 2, you will see lines 2, 2, 4, 4, 6, 6 and so on; in either case your vertical resolution is cut in half. When you advance a conventional VCR to the next recognizable picture, it will not be the next frame 1/30 of a second later, but the next field of 1/60 of a second later. However, if you have a fancy "jog/shuttle" VCR, it is likely that the designer has included a true frame advance, by sensing what field you are on and if it is the wrong one (an arbitrary choice by the designer) forcing an advance of the tape a few extra thousandths of an inch to the next correct field. In this case, you would advance in 1/30 second steps although you are still looking at a reduced-definition video field. You could determine whether you are advancing in 1/60 or 1/30 second steps by recording an event of known duration and advancing your VCR slowly while counting how many pictures you go through for 1 second of real time. In either class of recorder, it is very unlikely that you are actually viewing a full video frame unless your machine is equipped with some variation of "dynamic tracking" that physically moves one or both spinning heads up and down in the still mode. One way to tell is as follows: if you are looking at a frozen field, the moving object will just be blurred equal to its travel distance in 1/60 second; if you are looking at a frozen frame, the moving object will be blurred and will also seem to be jiggling back and forth in the direction of travel, relative to stationary objects. Video Cameras Todays modern video camera is an impressive technological achievement. However, to make it affordable a number of compromises have been made in its design, which can introduce image artifacts. A professional camera is of the "3-chip" design, where one image sensing array, usually a light-sensitive CCD (Charge Coupled Device,) is used for each of the three additive primary colors, to wit: Red, Green and Blue. A dichroic beam-splitter is used with a single lens, to direct the appropriate part of the light spectrum to each. An array of 480 by 640 (or more) light-sensitive cells in each CCD gives full NTSC resolution video, excellent color rendition, and no false colors in small objects or sharp details. These cameras are expensive, consequently most consumer cameras are different in that they employ a single CCD. You may well ask how just a single CCD chip can respond to color information. It is done by making a few major (and not totally valid) assumptions about the image being made, and by some tricky engineering. The first major assumption is that most objects aimed at will occupy a large number of CCD cells. Another is that the objects being shot don't move around much. A repeating predictable array of over 300,000 tiny color filters is applied over the face of the CCD; a number of schemes have been used such as alternating red, green and blue filters; currently the most popular seems to be green, yellow (red plus green), magenta (red plus blue) and cyan (green plus blue). By delaying the video and matrixing the signals from cells of adjoining lines or rows, the objects color can be derived.[5] This works well for large stationary objects. For small or rapidly moving objects, the system breaks down and false colors are generated even where the original object is not colored at all. This effect is minimized by inserting a "crystal filter" in the image path to make the image slightly fuzzy so it is forced to land on more than one cell, by increasing the number of cells so there are a larger number of color filters, and by the color encoding scheme that prevents small objects from having much color saturation in the video signal. The net result, however, is that one-chip video cameras should be regarded as highly suspect for recording objects or details that are very small on the screen. Another possible problem is that not all of the CCD chip is light-sensitive. There are insensitive bands between the approximately square light-sensitive cells, so a very small object can appear to change in brightness and color as it moves around between cells. CCD chips often also have a number of cells that are defective and do not respond at all; their existence is concealed by a custom memory chip that is automatically programmed for each defect-carrying CCD, that arranges for the information in an adjacent cell to be used to replace the missing information. Again, this works well for large areas but could misinform the investigator that a small object is disappearing or changing shape when in fact it is passing over a dead cell. Fortunately the CCD yields are improving and many cameras may have only a few dead cells out of 300,000+. One more thing: the greatest sensitivity of a silicon CCD chip itself is in the invisible near infrared (at about 800 nanometres for you number freaks.) A cyan-appearing IR (infrared) rejecting filter is also installed in the light path to reduce the amount of IR that gets through, but it is not 100% effective. A strong source of IR will get through the filter and create a video image of something that is not visible to your eye. In experiments by the author, infrared is seen by his Hitachi VM-E10A 1-CCD camcorder as being white in color and not red; this is probably because the dyes in the color matrix filters are transparent to IR and therefore generate no color difference signals, as is the case with white light which also gives equal response from all cells. The response of a 3-CCD camera to infrared was not tested. The high sensitivity of the CCD to IR is used to advantage in surveillance cameras; when the IR-blocking filter is omitted in a B&W camera, fully exposed video can be generated with just a relatively small amount of infrared flooding the area, which is totally invisible to the human eye. We might mention here that despite a few shortcomings the solid-state camera is a vast improvement over the 1-tube cameras that once were used. The chip camera has virtually unlimited life, far better and more uniform color and sensitivity, and no lag from one video frame to the next, much like a motion picture camera which offers an unspoiled fresh area of film for each movie frame. The CCD soaks up light like a sponge for generally 1/60 second, then it is read-out and erased to start fresh for the next 1/60 second. With very bright pinpoint subjects a vertical smear may occur. Some cameras have a "high-speed shutter" which electrically makes the CCD insensitive to light for a selectable portion of the usual light integrating time. A shorter shutter speed gives a less blurred rendition of moving objects but gives less sensitivity in the same proportion, so more light is needed. Motion Picture Cameras A movie camera is basically like a still camera, except that it continuously takes one photograph after another, generally 16, 18 or 24 of them per second, and usually in a smaller film format to save money. As with a still camera, the shutter is closed while advancing from one frame to the next. Unlike a still camera (where the exposure time is defined by a dial and is independent of how often you take a picture,) the cine-camera shutter is geared to the mechanism and rotates once per frame. If the camera is running slowly, the exposure per frame will be increased. This is an important point as will become evident below. In some famous contactee movie footage, the number or position of
claimed UFO's would frequently change between frames, itself a
remarkable coincidence, accompanied by an increase in exposure in the
first frame of each new arrangement. Apologists have supposed that the
exposure change is caused by some burst of unknown energy that caused
the film to come out lighter at that With a still camera, if you set the exposure time to "bulb" or "time" you can integrate light over a period of seconds, minutes or hours causing point light sources to leave trails behind them if either the camera or light source is moved during this time. Possibly unknown to many people is the fact that the same result can be obtained with a movie camera in two main ways: camera malfunction, or use of a camera that is not designed to stop with the shutter closed. If the cameraman is excited he might not notice that in the middle of filming a sequence the camera has stopped running because of momentary partial letting go of the release button, because the battery contacts are corroded, or otherwise. Most amateur cameras have a mechanism that normally prevents the camera from stopping with the shutter open, to prevent objectionable "flash frames" [6] that would spoil the hobbyist's movies forever because most amateurs did not even splice their little reels together into bigger reels, let alone actually edit their films. If the camera stops because of an electrical malfunction however, the normal electromechanical interlock would be bypassed and there would be about a 50-50 chance of the shutter staying open, thereby recording street lights as trails across the sky unless the camera is secured to a tripod, in which case they would merely be overexposed. In a sprocketless camera, including all super-8 cartridge cameras, it is fairly common for the film to stick in the gate under damp conditions, or if the camera is defective, also causing a similar effect and with little risk of tearing the film. The difference in this case is that the shutter would still be interrupting the light, and the light trails would show as dashed or dotted lines across the sky. Actually, a dotted trail can also be left by fluorescent lights, or by discharge type street lights (mercury vapor, sodium vapor and metal halide,) since they actually flash 120 times per second on 60-Hz line current. Most professional cameras are made for generating film that is edited, consequently it is not important if several flash frames occur at the beginning or end of each shot since they will be cut out. In fact, most editors prefer that there be a few overexposed frames in order to quickly locate the start of the new shot. These cameras include, in the 16mm size, the Arriflex 16-S, 16-BL, and 16-M; the Eclair NPR and ACL; the Bolex EBM; all of the newsfilm sound cameras including the Auricon conversions, Frezzolini, General Camera, and Cinema Products; and others. Again, each time the cameraman stops the camera there is about a 50-50 chance the shutter will be open until the next shot is begun. In the daytime this shows as overexposed frames; at night against a dark sky long trails will be generated by light sources if the light or camera is moved. Even a spring-wound camera that normally stops with the shutter closed can stop with the shutter open at the end of the useful run, as in this case a separate mechanism does the stopping. It may be well to point out here that some super-8 cameras are designed to eliminate the hot frame at the start of each shot; also most cameras don't generate a hot frame if operated at low filming rates, say 8 FPS (frames per second.) Movie cameras have what might be called a sampling error that is not shared by video cameras unless the electronic shutter is engaged. With the camera filming at 18 FPS, generally for each frame the shutter is open for about 1/40 second and closed for 1/30 second for each 1/18 of a second. When filming something that is flashing or moving at about the same rate, distortion will result. If you'll think back to many Western movies, maybe you'll remember the wagon wheels that seemed to be standing still, or even turning backwards. That is because the spokes were passing by at about the same rate that the film was being exposed, generally 24 FPS unless they were "under-cranking" or running the camera a bit slowly to speed up the projected action and make it more exciting. Still Cameras With Focal Plane Shutter Aha! You thought still cameras were perfect, didn't you? Actually they are not immune to recording problems, particularly if you have a high-grade camera with a "focal-plane shutter." This type of shutter is strictly speaking not in the focal plane (that's where the film is located) but is maybe 1/8" in front of it. It is found in most cameras with interchangeable lenses, including the 35mm single lens reflex. This is in contrast to the behind-the-lens or between-the-lens shutter commonly used in point and shoot cameras without interchangeable lenses. The quirk of the F.P. shutter is that all of the film frame is not being exposed simultaneously. At long shutter speeds it is very close to being all exposed all once. At the shortest shutter speed the amount of time skewing (for lack of a better word) increases to a value much greater than the actual exposure time itself. This comes about because the F.P. shutter is made of two opaque black curtains or blades that move across the frame either vertically or horizontally, depending on the camera. When you take a picture, the first curtain is released and begins its travel across the frame, exposing the film to the light. Depending on the shutter speed that you or the camera have selected, after the right interval the closing curtain is released to terminate the exposure of that particular frame. The curtains cannot open and close instantly, so consequently one part of the frame begins to be exposed before the rest, and that same part begins to be blocked again before the rest. The cameras electronic flash "X-Sync" speed is a good indication of how fast the curtains travel. At the X-Sync speed, this is about the shortest exposure time at which the film frame is capable of all being exposed at the same time by a short flash. The flash is fired the instant the first curtain is fully open and the second curtain has not yet begun to close. On older cameras this was often 1/30 second; newer ones are commonly 1/125 or even 1/250 second. At slow shutter speeds, say 1 to 1/30 second, the curtain speed will have little effect on the correct rendering of the subject. How, you may well ask, is a shutter speed of 1/1000 second obtained if the curtains take about 1/125 second to travel across the film? In essence, it is by triggering the curtains very closely together in time so they pass across the film only open by about 1/8 of the picture height or width. This can be verified easily by taking an electronic flash picture with the camera set to 1/1000 second; the only area illuminated by the flash will be a strip of about this proportion, increased somewhat by the duration of the flash, which could be 1/500 to 1/50,000 second. The consequences of this are twofold. If the subject is emitting or is illuminated by a brief flash, like an electronic flash, and your shutter speed is short, only part of the subject will be visible. If the subject is moving, its apparent shape will be distorted by the "slit scan" [7] effect of the small opening in the curtains. If it is moving in the same direction as the curtains, it will be stretched. If it is moving in the opposite direction as the curtains, it will be squashed. If it is moving perpendicular to the curtains, it will be distorted, i.e., a rectangle will become a parallelogram. The direction of movement, as used here, is of the image at the film plane which is inverted and reversed (or rotated 180 degrees) from the subject as seen by your eye. Conclusion It is hoped that this article will be of benefit to investigators and perhaps even explain some puzzling evidence. For some added considerations, please refer to the authors previous article8 which looked at unique film types and some techniques that can be used in motion picture and still-picture trick photography. Footnotes 1. For the sake of not confusing the reader (and the author) we are taking some liberties with the phantom [2] EIA standard. What we call "line 1" is unofficially called "field 1 or 3, line 21." What we call "line 2" is unofficially "field 2 or 4, line 21." In reality, the first visible thing at the top of the video frame [3] is the second half of "field 2 or 4, line 20" which is actually above the first visible line of field one. Also a purist would correctly point out that there are really 59.94 fields and 29.97 video frames per second. 2. The latest revision of EIA (Electronic Industries Association) standard RS-170 has reportedly been stuck in committee for years (maybe 14 years by now), as one might perhaps expect from an organization based in Washington, D.C., and only an unofficial draft has ever been released. Despite its phantom status, it is unofficially referred to as RS-170A and is complied with by all makers of high-end equipment worldwide. Since it does not officially exist it cannot be obtained from EIA but can be had under the table from manufacturers. Don't tell EIA but it was also published in a booklet entitled "Establishing and Maintaining SC/H Phase" issued in 1980 by Grass Valley Group, Grass Valley, CA. 3. You can only see the actual top (and bottom and sides) of the full video frame if you are using an "underscanned monitor", otherwise some 5 to 10% of the picture is lost on all edges because of the "overscan" past the physical dimensions of the picture tube. Overscan protects the viewer from the presumably unspeakable horror of seeing an occasional black line at the edge of the picture. Many question whether this much overscan is necessary with today's more stable solid-state TVs and monitors. Also the public is becoming accustomed to seeing black edges around most computer monitors, which are underscanned to show everything. 4. Some large-screen TV displays include a digital memory circuit that enables the screen to show all of the lines at one time, repeating the process 60 times per second. Their common name of "line doubler" is a bit misleading as the number of lines is not actually changed, they are just all there for each scan so have to be written twice as fast. The picture is improved in apparent resolution and line crawl is eliminated, but it is not quite as good as with true progressive scan. 5. A description of how a 1-chip color camera works is usually found in the service manuals for same. One currently in print with a largely but not completely incomprehensible explanation, plus relevant graphs, illustrations and formulas is the "WV-D5100 Service Manual", part number AVS9003282C1 which at last report sold for $30 from Panasonic in Secaucus, N.J. 6. A flash frame should not be confused with a hot frame. A hot frame that is only about one F-stop overexposed, at the beginning of a shot, is not too noticeable on projection. A flash frame is completely fogged to water clear and will startle or irritate the viewer. 7. Slit scan was used to advantage in the film, 2001, A Space Odyssey for the Stargate sequence near the end of the film, where images appeared squeezed at an artificial horizon and became stretched as they approached the protagonist and the audience. Very narrow slits were used in the optical printer to preserve the sharpness of the images. 8. Some Notes on Photographic Trickery, by Clive Tobin, MUFON UFO Journal Number 244, August 1988, pp 12-14. While in 1988 the author was ga-ga over "Who Framed Roger Rabbit" the most impressive recent examples are probably "Terminator II" and "Jurassic Park" which employ a multiplicity of special visual effect techniques including ultrahigh resolution computer generated images. =============================================================
Above: Field 1 (solid) on your TV screen. Field 2 (dotted) is written after Field 1 is finished. Enlarged fro clarity; the actual picture has about 480 lines except for those marked off in the overscan.
Above: An array of light-sensitive CCD cells. This is enlarged for clarity, an actual CCD camera chip has about 480 x 640 rows or over 300,000 total, roughly 1/4 inch wide overall.
Left: UFO's sense the position of the camera shutter and move between frames, and must be emitting a huge burst of unknown energy that makes the film come out lighter. Or maybe this is just a hot frame that always results from stopping and starting the camera. Right: Ashcan Command here makes their mother ship zip all over the sky, warning Earthlings to sign the standard intergalactic abduction-consent form on the dotted line or suffer the terrible consequences. Or maybe the camera stopped with the shutter open, and the camera was moved around. Shown as a negative to avoid large black areas on the page.
|
