A forensic voice expert yesterday gave evidence in the trial of a man accused of a string of arsons and making hoax calls to the police and fire service.

Professor John French was asked by Leicestershire police to compare recordings of Omar Mohammed Suleman with those of the hoax caller.

Suleman is on trial at Leicester Crown Court having denied being responsible for starting 26 fires, including at his uncle's house and his family's firm, between 2007 and 2009.

Prof French, whose previous work includes giving evidence in the Soham murder trial and the Who Wants to Be a Millionaire "coughing" scandal, analysed seven recorded telephone calls made allegedly by the defendant to the emergency services and his police interview recordings.

He told the court the maker of the hoax calls and Suleman, 29, of The Fairway, Oadby, shared "distinctive" characteristics.

"They have all the hallmarks one might expect to have been made by Mr Suleman and there is nothing to indicate they were not made by him," he said.

Prof French said that to indicate how unusual a voice was, a one-to-five scale had been developed.

This starts at one, where the speech is not at all distinctive and where you would find a lot of people talking like that, to five, which is exceptionally distinctive," he said. "As far as the voice on both the calls and the interview goes, the speaker is level three – distinctive. There are features shared by the calls and interview."

Prof French told the court that, after editing out the operator's voice he was left with an endless stream of speech ranging from 16 to 17 seconds for one call to 54 seconds for another.

He identified 11 separate speech characteristics, some of which were detectable only by using computerised "sound spectrography" and concluded that all seven disputed calls and the recorded interview were consistent to have been made by the same person.

Prof French said some of the hoax calls contained non-native English speech and that the caller had tried to disguise his voice. "There is a lot of grammatically-broken English," he said.

"The pronunciation is markedly more Asian than in the first batch and I had the view that the speaker was trying to disguise his voice.

"I had to exclude one of the calls where the speaker sounded as though he was deliberately attempting to adopt regional English pronunciation. He was covering over his natural voice and there was very little I could get from it."

He said that voice analysis could not be compared with, and be as accurate as, fingerprint evidence.

"It should not be used as stand-alone evidence," he said.

"People watch CSI and get a lot of misleading information about voice-prints."

The trial continues.