I have analyzed parts of
this YouTube video clip.
The minute-long video seems to show a flying object manouvering in the proximity
of the observer and fly over him/her. At this point there is a sound with a clear
Doppler effect audible.
The aim was to determine if the sound is consistent with the scenario seen in the video or whether it shows evidence of being synthesized or otherwise manipulated. Approximating audio frequenciesThe audio track was extracted from the YouTube video with the mplayer program to a WAV file. The right and left channel turned out to be identical, so the file was monoized. A program called snd was used to produce a sonogram of the audio track and specifically of the time window 26..36s which supposedly contains sound of the passing flying object. This WAV file contains the sound of the passing flying object.
Sound of the object seems to consist of two main frequencies and show a Doppler effect. In addition there is some "echo" effect spanning over several (even five) seconds. This seems to be inconsistent to a scenario of single moving sound source. The echo sound seems to repeat after every 1,7 seconds (rough approximate from the sonogram). Each repetition is in a lower volume than the previous. There is no apparent physical reason for this effect. The apparent Doppler effect of the two main frequency curves were examined. The following frequencies were approximated from the upper curve: when the sound source is approaching the observer, the frequency is about 500Hz. After passing the camera, the frequency decreases to about 319Hz. The base frequency (average) is 409,5Hz. The frequency decreased by 181Hz. The lower frequency curve has the values 223Hz, 142Hz, and base 182,5Hz respectively. The lower frequency decreased by 81Hz. Calculating speed of the sound sourceThe following equation was used to calculate the speed: fd = 2*f0*v*vs/(v2-vs2) Here fd is the frequency difference between approaching and receding sound source; f0 is the base frequency of the sound source; v is the speed of sound; vs is the speed of the sound source. The resulting speed for the object is vs = 76m/s ~ 274km/h ~ 170mph. The calculation with the lower frequency curve gave approximately the same result. The frequency calculation can be easily checked on this page (calculate receiver velocity, source frequency=409,5Hz, wave velocity=334m/s, new frequency=500Hz). Size approximation
Using a cloud in the background we can approximate the time it takes for the object to advance its own diameter length. This happens during frames 906 to 911. The video has a frame rate of 30fps, so 0,167s passes during this advance. Knowing the speed at that time we can calculate an approximate size for the object, which is 76m/s * 0,167s = 12,7m. This calculation gives a conservative, minimum value for the speed (and thus size of the object) because the calculation result is actually a component of the speed vector which points directly towards the camera. The following frames may indicate that the object passes in front of some branches of a tree. GIF animation: There is an enhanced video clip of this part here. It was produced with the "unsharp" operator of the ImageMagick software package. Some individual frames:
This would require either the tree to be of gigantic size, or the object diameter to be considerably smaller than 12-13 meters. Aerial perspectiveTo find out whether any aerial perspective effect is measurable in the object, these four frames were examined: 865 (object very close to the camera), 900, 930, and 950 (object in the distance but the three lights still visible). Assuming the velocity stays approximately constant, the object moves over 200 meters during these frames. Objects in the distance should seem more blue than objects close to the viewer, so the proportion of the blue component should increase with the distance. The bottom of the craft including the lights was separated from each frame to image files. Average of red, green, and blue components of the pixels were calculated. Pixels having any component of value 0.0 or 1.0 were discarded in this calculation to prevent saturated light sensors from distorting the component ratio. Resulting R, G, and B values were normalized to find out the proportion of red, green and blue component in each frame.
The measurement did not find any shifting towards blue. In fact proportion of the blue component has decreased slightly. This could mean that the object is computer generated, or that the aerial perspecive effect is too small at this distance (~ 200m) to be measured. ConclusionAt least these two independent observations suggesting a fabrication can be made of the video: 1) Echo effect. Doppler effect of the last "echo repetition" starts about five seconds later than the "main" Doppler effect. At this point the object has already passed the camera and is in the distance. There seems to be no physical explanation to what Doppler effect would start when the object is already several hundreds meters in the distance. If there was some kind of natural echo, ie. sound reflection from an object in the distance, the echoing sound wouldn't show an identical Doppler effect. This suggests that the original sound with a rather realistic Doppler effect was repeated and faded in order to make it more dramatic. 2) Speed and size of the object. Speed calculated using the audio frequencies is 76m/s ~ 274km/h. Diameter of the object would be about 13 meters. The video gives, especially at the tree branch episode, the impression of a smaller object hovering at a lower speed. It should be noted that video compression artifacts might distort how the branches appear, so this point may need further investigation. Based on these observations it is a reasonable conclusion that the video clip is a fabrication. 2008/11/29 pouko(a)touko.cjb.net |
