SANFORD, Fla. -

Hirotaka Nakasone is an expert on voice recognition systems, who works for the operational technology division of the Federal Bureau of Investigation (FBI).

During a pre-trial hearing, Nakasone said "no one can speak in the same way twice."

Special section: George Zimmerman trial

He was initially a witness for the defense. On July 1st, he was a State witness without objection from the defense.

Nakasone told prosecutor Richard Mantei that recognizing the voice of a teenager whose voice is changing is difficult.

"Guessing age is a little complicated," Nakasone said.

READ: Witnesses' profiles

He also added there wasn't enough clear sound on the 911 recording of a call made to police Feb. 26, 2012 to determine whose voice it was.

"I decided it was not possible to determine" on the 911 recording who the screamer was, Nakasone said.

Nakasone also told defense attorney Don West that there was only 2.53 seconds of screams in the 911 call in reference. He needs 30 seconds of natural speech for reliable results.

READ: 5 key moments from 1st week of trial