The Channel 4 video on Sri Lanka is an elaborate hoax that was propagated two years ago with a few minutes of video. For the record Ch4 said it cannot vouch for the authenticity of the video, but somehow its message was that the whole scenario did happen and the footage was filmed using a mobile phone. When I analysed the video, I said it never came from a mobile phone and the quality of motion vectors was high. Also, the video has cinema drop frame data, indicating it was filmed with a ‘cinema film camera’ and never came from a mobile phone or a normal video camera.
When the United Nations got hold of the video, somehow the video’s technical metadata was changed; it indicates that someone tried to change the metadata to show that it is from a mobile camera. Doing so created more metadata anomalies and the UN technical analyst said it has 17 frames of anomaly that he cannot explain, but the video is authentic and not edited. Later I found out that he is not a video expert at all and has a security background and his only experience is subjective analysis which is not acceptable in a Court of Law.
Then UN Rapporteur Hens appointed another video specialist named Grant Frederick to analyse the video and his findings are quite the opposite to that of the first video specialist. His findings clearly indicate that the first video did not come from a mobile phone at all and the video was edited and created on July 15, 2009, two months after the war against terrorism ended.
The real story is that when the first video was discredited by me, somehow Ch4 created three more videos to back up its claims and in one of the new videos a person using a mobile camera is shown in the frames, which is quite hilarious since this video tries to back up Channel 4’s incompetence with a new video with ‘new information’ which is discredited in the first video.
The whole Ch4 issue is a forgery that had gone wrong (two years ago) and to fix it, three more videos have been created this year to add weight to the first. I think Ch4 is being taken for a ride and it is unfortunate that we have a UK TV station not doing its due diligence properly.
How did the UN get in to a corner, considering this video as authentic when it is not?
The issue arises from the small clip (of a few minutes duration) that was shown on Ch4 two years ago, claiming to have come from a mobile phone… It was an amateurish video filmed on a cinema type camera (which is rare since average cameras are not cinema type), then transcoded to a mobile camera format to show that it came from a mobile camera. In doing so, the transcoded video leaves all the cinema video metadata intact and this is the section which indicates that the video never came from a mobile camera.
UN specialist Grant Frederick came to the conclusion that this video (Video 1) is edited and used an optical zoom. Those were my observations as well.
An edited video is not acceptable in a Court of Law.
The second issue is to date, there are no optical zooms in a mobile phone including iPhones. The third issue is, the date on the video metadata is 15.07.2009, two months after the final battle ended.
Although Frederick mentions all these data, he sidesteps the conclusion without saying it is a forgery or authentic… Instead, he confuses everyone by saying the second video is authentic according to “subjective analysis” and this is only an opinion… How dishonest can one get, not stating the obvious, that the video cannot be authenticated by subjective analysis without 100 percent technical authenticity.
The three other videos are new creations and two of them are Sri Lanka Army public videos and stills.
These were shown inter-mixed with others to get public sympathy, but Ch4 avoided saying they are public material already broadcast in Sri Lanka. The other one is a major forgery.
The bottomline is TV stations have sophisticated equipment to verify their videos, but for some reason they decided not to use their editing hardware to verify this particular video.
Matter of concern
This report is from Grant Frederick:
“The appearance of an ‘A’ in the last 17 frames of the first video was also a matter of concern.
Encoded date: UTC 2009-07-15 13:17:23
Duration: 5 minutes 25 seconds
The encoded date is well after the alleged date of the incident and, if accurate, would tend to discredit the recording’s authenticity without a logical explanation for the discrepancy, as with the first file analysed.
It is now readily apparent that both videos originally broadcast on Channel 4 and the newly released extended versions are the product of a basic editing software.
Comparison of audio spectrographs from the two weapon discharges reveals a significant variation between the two, indicating that these are two separate and distinct audio events. This disparity between the two audio events would preclude the possibility that a single gunshot sound effect was added in an editing process.
The graphic or logo present for the final 17 frames of VideoDJ.3GP is absent from this segment.
It is clear that the camera used to record this segment has an ability to jump to a 2X zoom. The zoom appears to be an optical zoom, since digital zoom artifacts are not present. (Mobile cameras do not have optical zoom)
The existence of five separate and distinct recorded segments within the file, one of which contains 542 frames of video identical to the previously released recording (VideoDJ.3GP) provides conclusive evidence that these files are the product of editing.
The fact that editing in any form has been applied to the recordings submitted for authentication will quite understandably generate skepticism or even suspicion.
While there is no way to know what may have been deleted before or after the video segments, the remaining intact recorded video is accurate based on all available evidence. In the absence of the actual device, authenticity of the recordings cannot be determined to an absolute certainty.”
Image authentication must not be confused with the requirement to authenticate evidence as a precondition to admissibility in a Court of Law.
Likewise, authenticity differs significantly from integrity. Integrity ensures that the information presented is complete and unaltered from the time of acquisition until its final disposition. For example, the use of a hash function can verify that a copy of a digital image file is identical to the file from which it was copied, but it cannot demonstrate the veracity of the scene depicted in the image.
(The writer is Director IPTV and Media – CTO Office Sydney Area, Australia Broadcast Media. He was the former head of Cisco’s global broadcast and digital video practice, at present an Executive Director at IPTV Systems. He currently serves as the chief architect of Optus’ Network Systems Design Broadcast and Satellite TV operations.)