George Zimmerman's second-degree murder trial for the death of Trayvon Martin is scheduled to get underway in just three days, but first, the judge still needs to make a crucial decision in court Friday.

Zimmerman is back in court Friday as the judge continues to hear testimony regarding voice analysis of a 911 call made the night of the shooting, in which someone -- either Zimmerman or Martin -- can be heard screaming for help in the background. Tom Owen, one of the state's audio experts, says the audio does not match Zimmerman's voice.

———————————————

Latest updates from Friday's hearing

  • 4:45 p.m. -- Frye hearing will continue Saturday at 9:30 a.m. Defense has three witnesses that still need to be called. Jury selection is scheduled for Monday morning at 9 a.m. NEWS 13 will carry both! Also, a motion to extend George Zimmerman's curfew has been granted.
  • 4:38 p.m. - West asks if Reich he considered the fact that Zimmerman may have been punched in the face during the scuffle, which could have affected his speech. Reich said he could not detect any nasal quality in the screams. Reich also says he identified four or five different voices in the 911 call. Zimmerman, Martin, the caller and the dispatcher would be four.
     
    West has no further questions. Reich is dismissed. Judge Nelson asks West if he wants to call another witness now, or resume Saturday morning. Attorneys meet with the judge in sidebar.
     
  • 4:30 p.m. - The state objects to West's questioning in a Frye hearing. Judge Nelson agrees West is straying from the point, which is to determine if the methodologies were "new and novel" or generally accepted.
     
  • 4:25 p.m. - Reich said he received about $2,000 from the Washington Post to analyze the 911 calls before the state attorney's office contacted him for his services. He bills $100 per hour. He says it will take him a couple more weeks to finish his transcription of the audio.
     
  • 4:15 p.m. - Reich says he does not use speaker recognition software, calling it unreliable. He says his process includes slowing down the speech, and comparing different samples. He says he worked about 700 hours on the report given to the Washington Post, who reported his opinion that the screams on the 911 call were from Trayvon Martin.
     
  • 4 p.m. - Reich says the background speech he analyzed "very, very low in level." He specifically says he hears the phrases "I'm begging you," "Get off of me," and "Stop."
     
  • 3:45 p.m. - West asks what standards Reich follows. Reich says he does not adopt the FBI's standards, and doesn't appear to indicate that he follows a specific standard regarding the minimum amount of audio required for a sample. He says it's difficult to establish a standard that would work for every case.
     
  • 3:30 p.m. - Court resumes as West continues cross-examination of Dr. Reich, who says while the FBI's standard is 20 words, there's another standard that only needs 10.
     
  • 3:15 p.m. - Reich says he has a partial trasncript of the 911 call, but not with him during his testimony. He says it is possible to make out some of what's being said/yelled in the background without having to enhance the audio.  Reich also says there's research that says even a 1-second-long voice sample is long enough to conduct a proper analysis.
     
    West asks the court for a brief recess. Judge Nelson allows it.
     
  • 3 p.m. - Don West cross-examines Reich, questioning his "tentative" conclusion. Reich clarifies, saying he believes there is a "high likelihood" that his conclusion that the screams were by Trayvon Martin is correct.
     
    Reich explains he amplified the audio and heard the word "Stop." He says it was the last scream before the gunshot.
     
  • 2:45 p.m. - Reich said he used computer software that was not "new and novel" to analyze the 911 call. He said he initially did not have a sample of Trayvon Martin's voice, but later got one to compare with the screams. Reich notably called that conclusion "tentative."
     
    He adds the "very high-pitched" screams in the audio were not George Zimmerman, but Zimmerman's voice could also be heard on the 911 call. He says according to his findings, all the audio at the same level of the scream was Trayvon Martin's voice.
     
    Judge Nelson says the court reporter has been having difficulty understanding Reich, and asks the witnesses to slow down and speak more clearly.
     
  • 2:35 p.m. - Reich says he did a variety of spectral analyses on the 911 calls. Before that, only the foreground audio in the 911 calls had been accurately transcribed. He said he had to separate the screams from the background noise when he analyzed the audio.
     
  • 2:25 p.m. - Judge Nelson rules the defense's notes are work product and do not need to be turned over to the state.
     
    Prosecutor Rich Mantei calls Dr. Alan Reich, a speech expert, to testify over the phone. Reich has notably said before that he believes it was Trayvon Martin screaming "Stop," and "I beg you," in the 911 audio.
     
  • 2:15 p.m. - Judge Nelson reviews some emails from defense experts to Zimmerman's lawyers. The defense said it didn't turn over the emails to the state, because they considered the "work product" and not reports.
     
  • 2 p.m. - Court resumes. Judge Nelson asks if lawyers are prepared to continue Saturday if they don't finish up by 6 p.m. tonight.
     
  • 12:57 p.m. - Court is in recess for lunch until 2 p.m.
     
  • 12:55 p.m. - Owen says many 911 calls in many other cases contain screaming in the background, so it is common in his field to try and analyze the screams. He also admits audio from 911 calls are not the best quality, even under the best circumstances. Owen compares a short audio clip to a partial fingerprint, saying it can still rule out a non-match.
     
    West asks Owen if standards "don't mean anything." Owen laughs at that question, calling it ridiculous. West then claims Owen deviated from his own standards in looping the audio. Owen sayd he didn't deviate from his standards; he just didn't have 20 words to feed to the software.
     
    Owen says he has been paid $6,000 in fees so far for his testimony in the George Zimmerman case. Mantei points out that Owen's opinion on this case was made more than a year ago, shortly after the night of the shooting, and he has not seen a sharp increase in sales of the software.
     
  • 12:45 p.m. - Owen says spectrograph analysis has an error rate less than 1 percent, according to the FBI. He admits the 911 audio is "not really good evidence," and one probably could not make a "positive" conclusion that the voice was George Zimmerman's.
     
    Prosecutor Rich Mantei redirects, criticizing West's line of questioning that Owen isn't a programmer by pointing out that Owen may know how to use a microscope, but not how to build one.
     
  • 12:30 p.m. - Owen says there are not enough words for a "positive" elimination of George Zimmerman as the person screaming on the 911 call. "I can't be positive, but it's probable." He explains he would need a 20-word sample to make a truly positive ID. So, he could only make a probable ID in this case.
     
    Owen has had to ask Don West several times to let him finish speaking before continuing to question him.
     
  • 12:15 p.m. - Owen says he believes the person screaming on the 911 call cried for help multiple times, all in different ways, and all before the gunshot goes off. He says the screamer called for help at least eight times in seven seconds.
     
    Owen also says he was recently told the NCIS bought his software, but he doesn't know in what capacity it's being used. He says is not a programmer, and is not trained to know if the program is working as designed.
     
  • 12 Noon - Owen again explains how he raised the pitch of Zimmerman's voice in another recording to match the pitch of the screams, and they did not match. West questions the methodology behind that, claiming there's no research or studies that support it.
     
    Owen goes on to say he doubled the length of the screams by editing them together, so it would be a long enough sample to analyze. The software requires a 16-second voice sample. West questions whether looping the audio is scientifically acceptable. Owen says it's common to loop audio, but West argues the methodology is "new and novel."
     
  • 11:45 a.m. - Owen explains he receives a small percentage from sales of the software from one of the Tracer technologies. West has Owen break down his financial relationship with the company. Owen confirms he has promoted the software he receives profits from on national TV, because that's what he used to analyze the 911 call.
     
  • 11:30 a.m. - Judge Nelson said she agreed with the state's objection that today's hearing should not discuss the results of the analysis of the 911 scream. This hearing should only focus on whether the experts' methodology is generally accepted in the field. Defense attorney Don West objected with Nelson's ruling on the focus of the hearing. Nelson says she will hear the defense's objections after Owen's testimony is over.
     
  • 11:12 a.m. - Court is about to resume. Prosecutors are working to set up audio again to play another recording.
     
  • 10:40 a.m. - Judge Nelson asks Owen to email the transcript of the 911 screams to attorneys now, before questioning can continue. Court is in a brief recess.
     
  • 10:30 a.m. - West asks Owen about his education. Owen has a degree in history and no advanced degrees. He also makes no claim to understand the mathematics behind the software he uses.
     
    Owen says he identified certain words in the screams on the 911 call, and will send a transcript to attorneys.
     
  • 10:15 a.m. - Mantei has no further questions. Defense attorney Don West is now cross-examination, beginning with questioning Owen about his history in working with audio.
     
    West is trying to undermine Owen's credibility by pointint out that part of his 50 years of experience was recording weddings. Owen explained he started working in audio at age 15, with wedding recordings. He published his first paper on voice ID in 1981.
     
  • 10:12 a.m. - Owen says he raised Zimmerman's voice in his re-enactment to match the pitch of the screams to better compare the two. "The screams don't match at all."
     
    Owen says experts in the same field can often disagree. He compares it to asking if lawyers disagree.
     
  • 10:10 a.m. - Owen says he also analyzed the 911 call George Zimmerman made moments before confront Martin to see if he made a racial slur after getting out of his car. Owen determined Zimmerman did not.
     
  • 10 a.m. - Owen is explaining how the different software he uses for voice analysis works. The technology he uses measures the rate at which someone speaks. He explains if there is enough voice frequency, then it's a usable sample.
     
  • 9:55 a.m. - Owen: "Let's put it out there that we don't have the perfect tape" of the scream. He adds it's easier to eliminate someone's voice than to identify who is speaking.
     
  • 9:45 a.m. - Owen testifies that he has 50 years' experience in the audio field. As prosecutor Rich Mantei questions him, Owen explains the different schools of thought on voice recognition, saying the more you do, the better you get at it.
     
    Audio from the 911 call of the scream in the background is played in court. Owen said he compared that with a "scream" test recorded of George Zimmerman, but he initially had no audio of Trayvon Martin to use for testing, so he tried to either match or rule out Zimmerman using audio from his re-enactment of the events.
     
  • 9:37 a.m. - The state calls its first witness of the day: Tom Owen, a forensic audio analyst based in New Jersey. Owen is sworn in to testify via video conference. He is submitted as an expert in his field.
     
  • 9:27 a.m. - George Zimmerman enters the courtroom with defense attorneys Mark O'Mara and Don Wesat. A loud, repeated buzzing can be heard in the courtroom.
     
  • 9:15 a.m. - Today's George Zimmerman hearing at the Seminole County Courthouse is running late. Attorneys were meeting with Judge Debra Nelson in her chambers. There was no immediate indication of what the delay is.

———————————————

Recap of Thursday's hearing

During a Frye hearing Thursday, the judge heard testimony from Dr. Hirotaka Nakasone, an audio analyst with the FBI who listened to the entire 911 call, and said the audio quality was too poor to scientifically determine who, in fact, was screaming.

The defense called Nakasone to the stand, because Zimmerman's lawyers want Judge Debra Nelson to decide if the state's own voice experts are generally accepted by others in their field.

Three expert state witnesses who analyzed the 911 audio all reached the same conclusion about whom they thought was screaming: Trayvon Martin. Those analysts are expected to testify Friday.

The trial is set to begin Monday, June 10, with jury selection. Lawyers need to pick six jurors and four alternates.

———————————————

Last-minute motions

Before Thursday's hearing, the defense filed two more motions, hoping to:

If the judge does not find time rule on those motions Friday, she could possibly take up the issues before bringing in the first jury pool on Monday.