Differential Motion Smear in the Mexico City Video
The Mexico City video of Aug. 6, 1997 has been studied frame by frame. Initial reports were favorable. However, "fingerprints of a hoax" were discovered when the motion smear or edge blur or "diffuseness" of the UFO image was compared with the smear/blur/diffuseness of the images of the buildings. Jeffrey Sainio published his overall analysis of the differential image smear in the October 1998 issue of the MUFON Journal. This showed that on the average the motion smear of the building was greater than any measurable motion smear of the UFO image. Also discovered were two frames in which the building motion smear was great enough as to make the horizontal top edge of the building image very diffuse. At the same time in these frames the images of the wind sock on top of the building were so blurred as to be difficult to see against the sky background. By way of comparison, in the unsmeared frames (when the camera was basically steady) the horizontal top edge of the building image is very sharp and the wind sock makes an obvious dark image against the sky. The two smeared frames are about 1 second apart in the video. Along with each smeared frame in the pictorial analysis below is an adjacent unsmeared frame for comparison. One should realize that these highly smeared frames are "surrounded" by frames in which the building image is not smeared or smeared only slightly by camera motion. In other words, the camera motion occurred suddenly, with the maximum amount of image motion from frame to frame taking place between the smeared frame and the frame immediately preceding or following. Since the frames occur at a rate of 30/second, the maximum camera motion occurred over a time of 1/30 to 2/30 (two frames) of a second. A more explicit discussion follows. The camera did pan with the UFO, following it quite closely, so that, overall, there was not much image smear apparent to the naked eye. However, that was because the shutter was faster than normal. Jeff Sainio and I independently concluded that the shutter time was probably about 1/200 sec rather than the more normal 1/60 (the time of a single field or 'half' a frame of the video). The faster shutter reduced the image smear. I believe (I could be wrong) that most cameras, when you turn them on, set a 'default' shutter speed of 1/30 sec. Then you have to manually change the shutter speed if you want it faster (shorter shutter time). That would mean that in the "heat of the moment" the witness had the presence of mind to increase the shutter speed before beginning the video. I know that on my camera it takes a little time to set the faster shutter. Point: the faster shutter would be more likely to reduce image blur of the distant _real_ buildings thus making the edges of the building _more_like_ the unblurred edges of the UFO image... If it was a hoax, the hoaxer would WANT the edges of the UFO image and the edges of the building images to have comparable smear. The hoaxer would choose a faster-than-normal shutter speed for just that reason (to make the motion blur of the UFO and building images similar.) In this case the presumed hoaxer was reasonably successful except in two frames where the camera motion was very rapid. The "discovery" of a suspicious difference in blur was made first on a statistical basis: Jeff Sainio's computer-aided measurements, made frame-by-frame throughout the video, of the blur of the edges of the UFO and the buildings showed that on the average the UFO edge blur was less than that of the buildings. But then we discovered a couple of frames where the camera moved too fast for the fast shutter. Little wind socks on the buildings were so blurred as to be almost invisible. Yet the edges of the UFO image were not changed. Now, if this had happened over a few seconds we might argue that it was a result of fast panning of the camera... But it did not occur over several seconds. These are individual frames, _separated_ by several seconds... Thus, the faster motion of the camera occurred primarily during one frame time (1/30 sec) and then, after each of these frames of fast motion, the camera reverted to its more normal rate of motion (random hand vibration) without much blur of the building edges. More specifically, in these two frames there is a rapid downward motion of the image of the horizontal roof edge, a motion that smears the image of the horizontal, top edge of the roof of the building next to the UFO. (The downward motion of the building image means the camera suddenly tilted upward.) That motion lasted only 1/30 sec. During that time the camera tilted by an angle which, if projected to the distance beyond the building, corresponds to something like 10 feet. Since the UFO image didn't smear that would mean that, if the UFO were a real object "out there," it moved UPWARD (along with the camera upward rotation) by that number of feet in 1/30 sec. After the smear frame, however the UFO image was in the same position relative to the building as before. Thus, if a real object, it must have moved downward by that number of feet immediately (1/30 sec) after moving upward, In other words, a jump up and down lasting no more than 1/15 sec, and timed perfectly with the camera motion (the camera tilted up and then down in the same time period due to hand vibration). There were two frames with smear of the building images that was large enough to be easily seen. As mentioned above, these frames were separated by some seconds. Hence the 'jump' occurred twice, perfectly timed to match the accidental (I presume) camera vibrations. If it was a real object 'out there'. For the expert: Jeff's computer program measured the magnitude of the (brightness) gradient of the edges of the UFO and and of the building in each frame of the video. The magnitude of the edge gradient is large when the image is well focused and not smeared by motion. The edge gradient magnitude decreases with increasing defocus and with image blur due to motion. He discovered that the magnitude of the edge gradient of the UFO was about constant throughout the video whereas the magnitude of the edge gradient of the building changed with the image motion from frame to frame. (The image motion from frame to frame is a measure of the rate of rotation of the camera about some axis... said motion being caused by hand vibration or mechanical vibration... and said motion being the cause of the image smear.) ................................... The following composite picture contains the only two frames in the video in which the edge blur of the buildings is easy to see with the naked eye. It also contains, for comparison, frames for which there is negligible blur of the building edges. The UFO images in the four frames have been copied and placed near one another for easy comparison. (Note that the vertical left edges of the building are always rough and grainy because of the horizontal scanning of the video readout system.) Compare the amount of blur or diffuseness of the bottom (horizontal) edge of the UFO image in the four frames with the amount of blur or diffuseness of the top (horizontal) edge of building in the four frames.Note: this video and the visual sightings were discussed in messages posted on UfoUpdates in the fall of 2000. See messages starting Saturday, 18 Nov 2000.