Translated and adapted by Paolo Attivissimo with the author's permission.
The original Italian article is available in the author's 11-settembre blog.
People who approach the analysis of the 9/11 Pentagon attack for the first time often ask this question. How come the released security camera recordings fail to show the aircraft clearly? I'll try to provide an answer.
Chapter 3 of the book Debunking 9/11 Myths, by Popular Mechanics editors David Dunbar and Brad Reagan, deals with the Pentagon camera recordings and with the analysis of the videos released by the US Department of Defense after the end of the Moussaoui trial.
Since their release, these videos have been the subject of countless analyses and reviews, because the general public expected that they would show an aircraft smashing into the Pentagon, in a manner similar to the many video recordings, taken from many angles, of the impact of United Airlines 175 into the South Tower of the World Trade Center. That expectation was let down.
The released video shows an impact, followed by a deflagration (often inaccurately described as an explosion) and the forming of a huge fireball, but there's no trace of an aircraft in any form you might expect to see.
But are we really sure that we should see a plane? In other words, is it right to expect to see an aircraft skimming the ground and then striking the Pentagon?
Debunking 9/11 Myths has this to say on page 61:
A Pentagon spokesperson tells Popular Mechanics that the video was taken with a Philips LTC 1261 security camera and recorded at one frame per second. Jerry Housenga is a technical product specialist with Bosch Security Systems, which bought the Philips camera division in 2002. According to Housenga, it was unrealistic to think that the low-quality security camera footage would reveal the crystal-clear image of a Boeing 757 traveling at 780 feet per second. While most advanced security and surveillance cameras can be set to capture real-time video, the attached recording systems are almost always set at significantly slower frame rates in order to conserve storage space. As a result, it is unlikely that the recording system of any nearby security camera would be set at a rate high enough to capture the speeding plane with decent resolution.As reported by Popular Mechanics, therefore, the aircraft was flying at approximately 780 feet per second, which is 237.9 meters per second or 858 kilometers per hour.
The Pentagon security cameras that took the videos used the NTSC standard and accordingly generated 29.97 frames per second in 60i mode, i.e., by capturing 60 fields per second (each frame of a television picture consists of two fields). Based on what data is publicly available and unless evidence to the contrary emerges, these cameras were therefore stationary analog interlaced models which used the NTSC standard.
For the technical analysis that follows, I'd like to clarify that I consulted many technical sources and discussed the matter with technicians who have worked in the field of audiovisual production for years and with CCTV and video production/transmission specialists. However, I would advise anyone not to trust my assessments blindly and to check my arguments personally.
Technical background: interlacingI refer initially to a source which is easily available to any Internet user: the Wikipedia. This source describes the NTSC standard by noting that it "consists of 29.97 interlaced frames of video per second." To understand the meaning of the term "interlaced", we can again look at the relevant Wikipedia entry:
With progressive scan, an image is captured, transmitted and displayed in a path similar to text on a page: line by line, from top to bottom. The interlaced scan pattern in a CRT (cathode ray tube) display completes such a scan too, but only for every second line. This is carried out from the top left corner to the bottom right corner of a CRT display. This process is repeated again, only this time starting at the second row, in order to fill in those particular gaps left behind while performing the first progressive scan on alternate rows only. Such scan of every second line is called a field. The afterglow of the phosphor of CRT tubes, in combination with the persistence of vision, results in two fields being perceived as a continuous image, which allows the viewing of full horizontal detail with half the bandwidth which would be required for a full progressive scan while maintaining the necessary CRT refresh rate to prevent flicker. Only CRTs can display interlaced video directly — other display technologies require some form of deinterlacing.The concept of interlacing is illustrated below: in NTSC, first the even lines, then the odd lines are captured and displayed (the PAL system instead starts with the odd lines).
In other words, a television field is a scan which only acquires every other line of an image: it's a sort of half-picture. Two fields, one with the odd lines and one with the even lines, are used to acquire and display the full picture, which is termed frame. The eye is fooled into seeing these two separate fields, these complementary half-pictures, as a single frame thanks to persistence of vision and to the persistence of the glow of the elements of a cathode-ray tube display (the kind of screen used in old non-flat TVs).
TV cameras don't work like still film camerasGoing back to the Pentagon videos, the available information suggests that the security cameras worked by interlacing at 60 Hz, i.e., they acquired 60 fields (not frames) per second, which yielded 29.97 frames per second.
Because of this, it is incorrect to compare a TV camera with a still film camera, because for an equal shutter speed (1/60th of a second) a film camera captures all the picture on film, whereas an interlacing analog TV camera actually captures only a field, i.e., only the odd or even lines of the picture, not both. This is determined by the electronic systems that control the camera.
Knowing this, we can now determine what to expect when we view on a computer video player a video of a scene which is crossed at very high speed by an object, if the video is taken by a fixed analog TV camera, saved on videotape with a time-lapse recording rate of one frame per second, and then converted to the MPEG1 digital video format.
This is not a trivial task, since it involves many steps:
- A fixed interlaced analog TV camera which acquires the pictures;
- A time-lapse system which records to videotape (as a best-case scenario, it is assumed that a digital recording system was used at the Pentagon; if the recording system was analog, the recording and playback quality would be even worse);
- Possible conversion of the video to another medium, e.g. CD-ROM or DVD;
- Playback of the digital video on a computer video player.
1. What would the TV camera acquire?Let's start by looking at what the fixed analog interlaced TV camera would acquire. Based on the information currently available, the Pentagon surveillance cameras that took the well-known videos of the impact were NTSC interlaced analog color cameras with an automatic iris and a shutter speed of 1/60th of a second. The camera acquired a static scene, in which the background does not move, some objects are stationary and some objects (cars, people, clouds, and others) instead move. It acquired fields at a shutter speed of 1/60th of a second.
Let's now consider a simplified model in order to understand the behavior of the Flight 77 aircraft. The plane enters the field of view at a speed which can be estimated, for the sake of simplicity, as equal to the impact speed, i.e., approximately 237.9 meters per second. Its path is, as a first approximation, roughly at 46° with respect to the imaginary line that connects the aircraft to the camera.
With this simplified assumption, the speed of the plane in each point of the linear path traced by the plane can be split into two components, which are radial and tangent to the center of an imaginary circle which is centered on the TV camera and passes through the current position of the aircraft.
The tangent component of the aircraft's speed tells us how far the plane would shift in each second on an imaginary line which is almost parallel to the plane of the CCD sensor of the camera. The radial component tells us the speed at which the plane moves toward or away from the camera. The first component causes linear displacements on the CCD of the camera, while the second component determines the changes in apparent size of the plane.
Two moments in time in the same frameIn other words, we are dealing with an object whose representation on the CCD sensor shifts and at the same time changes size and shape. Under the simplified assumption of an initial angle of 46°, the object moves by 2.85 meters parallel to the CCD and moves 2.75 meters closer during each individual field, i.e., every 1/60th of a second.
This means that each individual frame of the 30 acquired each second by the Pentagon TV cameras is composed of odd and even lines which refer to an object which has moved by 2.85 meters parallel to the CCD and has moved closer by 2.75 meters. This extreme simplification applies only to the initial point of the actual flight path, since the distance from the camera and the angle with which the speed vector is divided vary for each point of the path and therefore the displacements and size variations perceived and recorded by the CCD vary as well.
Accordingly, it is entirely incorrect to compare a complete picture taken in 1/60th of a second by a film camera with an interlaced frame captured by an interlaced analog TV camera, in which the even and odd lines of the same frame actually show the moving object in two different places, because the even and odd lines of a same frame actually originate from fields acquired 1/60th of a second apart.
As the aircraft approaches the Pentagon (assuming, for the sake of simplicity in this model, that its speed is constant, although the plane was actually physically accelerating), the radial component of its speed increases and therefore the tangent component also increases (according to the rule of vector parallelogram composition, along the new radial and tangent directions on concentrical circles which are centered on the camera and pass through the point where the plane is located).
Accordingly, the plane, merely due to geometrical considerations (not physical ones), would undergo a tangent acceleration and a radial deceleration with respect to the camera: the enlargement of the object seen by the CCD would decrease, but the displacement between the spatial positions projected onto a plane parallel to the CCD by the even and odd lines of the same frame would increase. The effect, in the real world, would be further increased by the actual acceleration of the plane.
This, in itself, is already enough to clarify that it is unlikely to see a sharp picture of the plane, as one might expect based on ordinary experience.
Motion blurThere is also another problem: what does the CCD sensor capture if, during the 1/60th of a second for which its shutter remains "open", the light that strikes the CCD belongs partly to the object and partly to the background, which is fixed?
Let me put it another way. If the shutter opens at a certain instant to capture a field (for example the even lines of the frame) and we consider, again by way of example, a point of the field of view that lies 2.5 meters ahead of the nose of the aircraft (which is moving at a tangent speed of 171.1 meters per second), that point remains lit by the light of the fixed background for most of the time and is reached by the nose of the plane (which moves tangentially by 2.85 meters) only when the shutter is about to close. So what would the CCD acquire?
Would it acquire the background, with a ghost produced by the plane which arrived as the shutter was about to close? Probably. The same effect would occur for a point just behind the plane: when the shutter opened, the tail of the aircraft would be there for a very short time and then would be replaced by the fixed background for most of the period of exposure.
Therefore, to summarize, we would have dynamic effects linked to the geometry between the actual trajectory of the aircraft and the fixed viewpoint of the TV camera; dynamic effects due to the physical forces which act on the plane; and "video-electronic" effects linked to the way in which the TV camera works and acquires images.
2. What would the video recorder record?As mentioned, the recording system records the pictures that arrive from the video camera. However, in order to record for suitably long periods, the system does not record all the 29.97 NTSC frames per second (each composed of two fields): it records a single frame per second. This is known as time-lapse recording.
But which single frame does it record, among the 30 acquired by the camera every second? That depends on how the system is set up, but undoubtedly whenever the second ends, the corresponding frame is recorded.
Knowing how the recorder is set up is important anyway, because those 29.97 frames per second differ greatly from each other in terms of image content. What is the likelihood of recording that exact frame in which the aircraft is present in full, if such a frame exists and with all the doubts raised above?
The selection caused by time-lapse recording is such that the aircraft, despite being acquired by the TV camera, might not have been recorded, since it appeared in the frames that were not considered by the recorder.
Another aspect that should be ascertained before making any conjecture is whether there was or not a device for deinterlacing the input signal ahead of recording and any other circuitry which might have altered the structure of the recorded fields with respect to the ones acquired by the TV camera.
3. What would be exported to CD-ROM or DVD?A further problem would then arise from the subsequent export to CD-ROM or DVD of the frames stored by the videocassette recorder. Since the recorder records one frame per second from the TV camera, if the recorded video were to be played back at the normal speed of 29.97 frames per second, the result would be a 30-fold speed-up effect. Accordingly, the videos to be exported may have been obtained with methods allowed by ordinary CCTV digital players, which allow to set at will the playback speed. Another, cruder technique would consist in repeating each individual frame 30 times before exporting the video.
Video recorded in NTSC format has a definition of at least 640x480 useful pixels in each frame, but the video release by the DoD is in the MPEG1 format, which has a resolution of 352x240 pixels in each frame, i.e., much lower than the native recorded format (assuming that the native format was full NTSC).
Accordingly, in addition to all the factors cited above, the export process necessarily entailed a further reduction in image resolution, in addition to a compression which uses a lossy format and may introduce a further complication in the way in which each individual pixel is shown with respect to the original.
One would need to know, therefore, the processing undergone by the videotape recording before being exported to MPEG-1. For example, it may have been deinterlaced. One crude form of deinterlacing and compression is the complete elimination of one field, leaving only the even or odd lines. Certainly something must have happened, since the number of lines of the released MPEG1 video is half that of the native NTSC format.
4. What would a player show?Clearly, the quality of the hardware and software playback system affects the usability and final level of the MPEG1 video for the viewer. Seeing a Video-CD with an MPEG1 file on a home player is one thing; playing back an MPEG1 file in a professional computer video editing program is another. But these doubts are beyond the scope of any technical remarks regarding the video in itself, although different players unquestionably yield different playback qualities of the same digital file.
ConclusionsIn view of these simple remarks, my answer to the question "Why don't the Pentagon videos clearly show a plane?" is that as long as we don't know:
- the level of detail of the original images on videotape
- the composition of the chain of acquisition, processing and storage of the frames (including the settings of the equipment)
- what processing was done to the video before MPEG1 compression
In view of all this, I believe that statistically speaking, a video which clearly showed an aircraft striking the Pentagon would have been more suspicious. The videos released by the DoD following a FOIA request after the end of the Moussaoui trial are not, for me, conclusive evidence of the presence or absence of an aircraft at the Pentagon: but if I had to choose, there are more clues in favor of the presence of a plane than of its absence.