Doubts grow over voice, image differences in Yukawa death video

by and

Staff Writers

The government says Saturday’s video containing what appears to be the corpse of hostage Haruna Yukawa is probably authentic, but less certain is whether the same can be said of the voice on the recording claiming in English to be Kenji Goto.

The government is fairly confident it’s him, but voice experts are mixed.

Experts who analyzed the recording and compared the speaker’s vocal characteristics with verified recordings of Goto speaking English found reasons to believe and disbelieve.

“There’s a high likelihood that the voice is Goto’s,” said Matsumi Suzuki, an independent voiceprint specialist.

Suzuki compared the recording from Saturday’s video with a video that Goto recorded in October to explain he would accept complete responsibility for whatever befell him in the area.

He compared the brief segments where Goto utters his name by performing frequency analysis to determine the spectrum of sounds used.

“Ten points of Goto’s vocal characteristics in the first video match the second one,” Suzuki said.

This correlation, he said, is sufficiently strong enough to conclude the same person is speaking in both recordings.

Others disagree.

Hajime Suzuki, president of Japan Acoustic Lab, doubts that the real Goto is speaking in Saturday’s video.

In his October video, Goto said in English: “It’s my responsibility if something happens. . . . Please don’t have a bad impression (of) Syrian people.”

The recording was considered poor for analysis because it contains a lot of background noise. As a result, the Japan Acoustic Lab chief could isolate only three words for comparison that crop up in both recordings: “don’t,” “my” and “please.”

He said analyzing the three words produced a very different voiceprint that suggests the speaker in Saturday’s video has a mouth of a “different shape and size” than Goto’s.

Although not as unique as a fingerprint, a voiceprint is nevertheless considered sufficiently reliable to justify arrest, he said.

Some who know Goto have said the latest video does not sound like him.

Toshi Maeda, a Japanese freelance journalist and friend of Goto’s, said the man who identified himself as “Kenji Goto Jogo” was someone else.

“I’ve heard him speak English before, including in his reports for NHK World,” Maeda said. “He has a bit of a Japanese accent. But the voice in the video sounded like that of a native English speaker.”

Some native Anglophones, however, describe the language used by “Goto Jogo” as accented English.

Meanwhile, the still image itself has raised suspicions as well.

Seijiro Kuroda, a professor of computing processing at Kinki University’s Junior College Division, said he believes the picture was manipulated — possibly by a group other than the Islamic State.

Kuroda pointed to the widely noted absence of the logo that was superimposed on the Jan. 20 video.

“The group normally includes a logo as some sort of propaganda. It wants to show off its authority, so it’s unlikely that they would forget to insert it,” he said.