Wednesday, 10 July 2013

Not hard to find if 'Thaksin's clip is real : Forensic Cop

THE NATION July 10, 2013 1:00 am The sophisticated technology of today could quickly and easily determine if the two men in an audio clip conspiring to bring former PM Thaksin Shinawatra back to Thailand were actually Thaksin and Deputy Defence Minister Yuthasak Sasiprapha. "If a clip is fake, you can see that the graph will rise and fall," Pol Lt Colonel Watcharat Chalermsuksan, a director at the Central Institute of Forensic Science (CIFS), said yesterday. Software like "Audacity" offered an easy and straightforward technique to dissect audio material by showing a sound graph, he said. Besides the voices, listening to background noise can be a giveaway. Fake clips do not have the same background noise, he said. "It is also easier to determine whether a clip is doctored if we have the original," he said. The clip that has gone viral has become another hot potato for the government. While many believe it is genuine, Pheu Thai MPs claim it was made to sabotage the Yingluck government. Audio frequency analysers have been widely used by various industries for some time, while the government can turn to the National Police's Scientific Crime Detection Institute and the Justice Ministry's CIFS for forensic investigation. However, it is harder to detect a voice modified to sound like a specific person. The National Electronics and Computer Technology Centre is developing a program to check voice authenticity that is 100 per cent correct. The centre's equipment is 80 per cent accurate in terms of voice manipulation analysis. Assoc Prof Sathon Vijarnwannaluk from Chulalongkorn University said it was more |difficult to determine if a voice is genuine or edited than to check if an audio clip is doctored. "It may be easier to detect a manipulated voice in a spoken word but it is harder to detect it in a sentence,'' he said. Examples of people whose voices have been imitated are popular Buddhist monk Phayom Kanlayano and even Thaksin, he said.

No comments:

Post a Comment