Return to Transcripts main page


Continuing Coverage of the George Zimmerman Trial

Aired July 1, 2013 - 09:00   ET



CAROL COSTELLO, CNN ANCHOR (voice-over): Breaking right now, Zimmerman's side of the story.

GEORGE ZIMMERMAN, DEFENDANT: I fell down. He pushed me down. Somehow he got on top of me.

COSTELLO: This critical video evidence about to be front and center in a Sanford courtroom.

ZIMMERMAN: I felt his arms going down to my side and I grabbed it and I just grabbed my firearm and I shot him.

COSTELLO: Will the judge admit the video?

Plus, pivotal witness.

CHRIS SERINO, SANFORD INVESTIGATOR: Prior to you shooting him, he was on you, correct?

COSTELLO: Sanford investigator Chris Serino who found that Zimmerman acted in self-defense on the stand.

You're live in the CNN NEWSROOM. .


COSTELLO (on camera): This morning we'll bring you live testimony from the George Zimmerman murder trial. Among those that could take the stand, Chris Serino, the homicide detective who led the investigation in its early stages. In a recorded police interview that you're about to hear, Serino was clearly skeptical of Zimmerman's version of events.


SERINO: What happened this evening is that you essentially saw somebody who you in good faith thought was doing something wrong?


SERINO: ever heard of Murphy's Law? OK, that's what happened. This person was not doing anything bad. You know the name of that person this died?


SERINO: Trayvon.

ZIMMERMAN: Trayvon Martin.

SERINO: Trayvon Benjamin Martin. He was born in 1995, February 5th. He was 17 years old. An athlete, probably somewhere, somebody's gonna be in aeronautics, a kid with a future. A kid with folks that care. In his possession we found a can of iced tea and a bag of Skittles and about $40 in cash. Not the goon.


COSTELLO: Detective Serino wanted to charge Zimmerman with manslaughter. Of course, as you know, there was a bunch of controversy about that. We'll get into that a little later.

We'll take you live to the courtroom when things begin to heat up again. We have a team of reporters and analysts breaking down all of today's testimony. Let's begin our coverage, though, in Sanford, with CNN's George Howell. He's outside the courthouse. Good morning, George.

GEORGE HOWELL, CNN CORRESPONDENT: Carol, good morning. So there was supposed to be hearing at 8:30 this morning to decide about the statements. And we're talking about that video reenactment, two audio statements, and a statement, a written statement, that Zimmerman gave police after that shooting -- to decide whether those statements could be admitted as evidence.

Now we understand that attorneys have reached some sort of a decision about their concerns with that statement. Still unclear whether this evidence will be part of the trial. But if it is, this will be the first opportunity really to hear George Zimmerman in his own words explain to the jury and explain to attorneys exactly what he did.


ZIMMERMAN: All I could think about was when he was hitting my head against it (ph), it felt like my head was going to explode.

HOWELL (voice-over): It's George Zimmerman in his own words walking investigators through the incident less than 24 hours after shooting and killing Trayvon Martin. And it's this video, along with two audio statements, and a written statement he gave police, that could become critical new evidence in the case.

If Judge Deborah Nelson allows it in court, the jury could examine Zimmerman's demeanor and possible inconsistencies in his statements. Prosecutors are also expected to call on more pivotal witnesses this week, including the lead investigators from the Sanford police department, Chris Serino and Doris Singleton, who found that George Zimmerman acted in self-defense.

Jurors ended the week on Day 5 hearing from the only eyewitness to the struggle between Zimmerman and Martin. John Good told police he saw two people on the ground. The person on top was wearing darker clothes and the man underneath, he says, seemed to have a lighter complexion, like George Zimmerman.

MARK O'MARA, ZIMMERMAN'S DEFENSE ATTORNEY: What you saw was the person on top in an MMA-style straddle position, correct?


HOWELL: Good' testimony is important to the defense because it contradicts the testimony of two key witnesses for the state. Selma Mora told the jury she heard yelling, stepped out of her home, and saw a man who appear to be George Zimmerman on top.

O'MARA: The person on top did not respond to the first two times that you called to him, correct?

MORA: Correct.

O'MARA: And then the third time he said just call the police?

MORA: Correct.

HOWELL: Another important witness for prosecutors, Rachel Jenteal, who was on the phone with Martin moments before he was killed. Through more than four hours of questioning, she maintained it was George Zimmerman who was the aggressor.

DON WEST, ZIMMERMAN'S DEFENSE ATTORNEY: I thought in fact that you said that it could have been, for all you know, Trayvon Martin smashing George Zimmerman in the face -- is what you actually heard?


WEST: Just earlier today.

JEANTEL: By who?

WEST: By you.

JEANTEL: You ain't get that from me.


HOWELL (on camera): Week 2 picks up. Live pictures here in Sanford, Florida, inside the courtroom. You see these attorneys having a hearing there in front of the judge. We can't exactly hear what they'res saying, but we do expect to have audio very soon and to hear from the first witness today.

We do expect, Carol, that we could hear from several important witnesses. Obviously Chris Serino and Doris Singleton, those two lead investigators who were the first to decide that George Zimmerman acted in self-defense, decided not to hold him, to arrest him. But we could also hear from a ballistics expert, a person who can tell us - and prosecutors, we expect, will make this point -- that the gun was loaded and ready to be fired. Also from the medical examiner to explain the bullet wound to Trayvon Martin's chest and possibly the trajectory of that bullet, which could give some insight, some information, about that struggle that ensued before the fatal shot, Carol.

COSTELLO: All right. George Howell reporting live from Sanford. Joining me now with some perspective, CNN legal analyst Sunny Hostin. She's in Sanford, Florida. And here in Atlanta, defense attorney Page Pate and HLN contributor Jason Johnon. Welcome to all of you. Thanks for coming back in.

Let's head back to Sanford for a second and check in with Sunny about this video reenactment done by George Zimmerman Prosecutors, of course, want to keep that evidence out. Defense attorneys want it in. Do you know exactly what went down in court this morning in regards to that reenactment?

SUNNY HOSTIN, CNN LEGAL ANALYST: You know, I'm not so sure that the prosecutors want to keep it out. We have limited information about what was going to happen this morning. There was supposed to be a mini hearing at about 8:30 this morning, but the hearing was canceled. The judge let everyone know that the lawyers had come to some sort of an agreement and there was no longer an issue about any of the videotape statements.

And so I suspect that we will be hearing testimony and perhaps also seeing those videotape statements. Of course, we've all seen them, but the jury has not.

COSTELLO: OK, as you can see, the first witness is about to take the stand. This is a doctor. He has testified as a defense voice analyst expert for the FBI during a separate hearing. Don't know exactly what he's going to testify about but we suspect we do. Let's listen.


HIROTAKA NAKASONE, VOICE ANALYSIS EXPERT: My name is Hirotaka Nagasone, H-I-R-O-T-A-K-A. Last name is N-A-K-A-S-O-N-E. Currently I'm assigned to the operational technology division of the FBI and performs a variety of duties as a senior scientist.

I conduct research and development activities for the development of automated speaker recognition systems, and also as a correlational (ph) duties, I perform on a routine basis voice related forensic examinations including speaker recognition, speech analysis, and acoustic analysis of signals we encounter, including gunshot analysis.

Lastly, as a senior scientist, I provide advice, guidance, recommendations for the operational technology divisions in other FBI field office.

RICHARD MANTEI, PROSECUTOR: Dr. Nakasone, how long have you been in your current position?

NAKASONE: Since 2009 as a senior scientist for four years. And prior to that I was hired in the seminal (ph) position since 1996. MANTEI: Tell the ladies and gentlemen of the jury essentially what it is that you have done, how long you've been in this field and what has made you sort of qualified to hold your current position.

NAKASONE: Well, I initially, when I was doing my Ph.D. thesis at Michigan State University, I got into this forensic speaker recognition as part of my thesis under Professor Oskatoshi (ph). And when I graduate with the Ph.D. degree, I taught at Michigan State as an assistant professor for 18 months or so.

And after that, I moved to Los Angeles County Sheriff Department in Los Angeles, California, to take up a federal research project for the development of computer assisted voice identification system. I stayed at Los Angeles County Sheriff Department for seven years, since 1985 to 1992. My title with the sheriff department was voice specialist.

And then consequently, I joined FBI in August of 1992 for the same responsibilities, type of work that I did. I think I was qualified by FBI to conduct forensic examinations on the audio evidence in 1994. Ever since, I've been doing this kind of work until today.

MANTEI: Doctor, have you -- you mentioned a Ph.D. thesis. Have you done work independently of just your work?

NAKASONE: Yes. My first independent scientific work would be my own Ph.D. thesis. This was titled as a "Automatic Speaker Recognition by Using Deviation Spectra." And when I moved to Los Angeles, the research grants was provided to continue my work as I wrote in the thesis under the project name, titled "CAVIS: Computer-Assisted Voice Identification System."

When I joined the Bureau, I started specific modernization effort of the FBI's technique to do voice identification. Until 1997, since early part of 1960s, FBI was using so-called spectrograph (ph) voice comparisons, or commonly known as voiceprints. And it was getting to be very ineffective to establish the scientific validity because of the subjectivity in the examination involved.

So in 1997, I was charged to look into modernizing speaker recognition technology. After that, I initiated a lot of research projects to identify any existing technology viable for forensic applications. I published a variety of papers related to speaker recognition technologies.

MANTEI: If you had to estimate for us how many I'll call them peer- reviewed articles or articles you published over the years, what would that number be?

NAKASONE: I really never counted but probably -- maybe 30 or 40.

MANTEI: OK. As far as your actual work in lab for both the FBI and the other agencies that you described, how many actual voice samples have you attempted to conduct identification on over the years?

NAKASONE: I can give you just a rough idea. Judging from the fact that I handled approx -- anywhere from about 40 to 80 cases per year, and if I worked for the Bureau, that would amount to 1400 cases. But since I've been working for more than ten years or so, my best guess would be probably actually handled and analyzed and written report somewhere in the neighborhood of 3,000 cases. And maybe involved maybe five times of the number in the voices.

MANTEI: Have you ever testified in courts before as an expert in the field of forensic speaker recognition?

NAKASONE: Yes, I have. However, most of the time when I testify in the courtroom I never testified to the results of the analysis and examination that I conducted. Rather, I was requested to testify in the court to talk about the reliability and the scientific foundations of speaker recognition technologies.

MANTEI: How many times would that be?


COSTELLO: All right, while the doctor goes through his background, we're going to take a quick break. We'll be back with much more from Sanford.


COSTELLO: All right. We have not gotten to the meat of the doctor's testimony. He's an expert voice analyst. If you're confused as I am and as Jason Johnson is, because, Jason, we thought a voice analyst wasn't allowed to be part of this trial.

JASON JOHNSON, HLN CONTRIBUTOR: Right. The Frye hearings, we spent weeks with both sides arguing as to whether or not you could have someone come in and talk about whether it was Trayvon or George Zimmerman yelling.

So, I don't know if this was a trick by the prosecution. I don't know why the defense didn't object. This should be very interesting. I wonder what he's going to talk about.

COSTELLO: Let's talk to Mr. Defense Attorney, Page Pate, because you suspect you know.

PAGE PATE, CRIMINAL DFENSE ATTORNEY: Right. You know, the judge said you can't have an expert testify to that question of whose voice is it we heard making the scream. Maybe there are other things about that tape an expert can shed light. What I thought was interesting is the defense said, yes, let's bring him up. They seem excited about this testimony.

COSTELLO: The defense testimony said I'm eager to hear what he has to say.

PATE: So am I.

COSTELLO: Very curious. OK. Let's go back to testimony now.


NAKASONE: To evaluate the suitability of the data and we check and listen very carefully and then want to make sure speech sample we have at hand naturally spoken. It would be devoid of disguise attempt. Any evidence of intoxication, there must be removed from analysis, and also we pay attention to emotional state of this speech.

If the person is uttering voice samples with the extreme happiness, their voice would be too joyous including lots of giggling and laughter, those must be removed. Sometimes we run into voice being uttered when the person is going through stress or undue depression. The voice must be also be removed and any other unnatural voicing say like yelling, screaming, moaning, this type of voice must be removed from the analysis and the standard operating procedures.

And when the speech is determined to be natural, we would check the other aspects of conditions. For example, the duration must be at least 30 seconds after blank was removed. Usually, we can do meaningful analysis if the voice sample is 30 seconds, one minute, maybe up to three minutes or so. After three minutes, the performance hits the ceiling.

But when the speech sample drops down to 10 second, 14 seconds or so, the examiner has to conclude that the voice is -- the duration is not good enough for analysis and then we will also check the amount of vibration included the speech. If there's a long amount of vibration in the recording, it will be removed. We will not proceed.

Another is that samples, we have two voices to compare. One, coming from the recorded, incriminating recording, and the other voices from the suspect's voice recorded in the -- you know, quiet room. Whereas, incriminating voice as recorded through the telephone line has, say, a bad quality -- these days most of the telephone, especially land line carries decent quality. It can be usable. But depending on the cell phone that distorts human voice.

When we have to confront this disparity the examiner is mandated to stay away from further analysis.

MANTEI: So, you've talked about sort of the -- some of the factors that affect your ability to conduct an analysis. I think you mentioned there's variations in what is going on with the speaker terms of mental state or their physical state or emotional state, the way the speech is expressed.

And you were just talking about sort of variables that could go on with the mechanism of the recording, the cell phone distortion and that sort of thing.

Are there other environmental factors, the distance between the recording and the mic microphone, for example? Are there other factors over and above with the speaker and the microphone that could have an effect?

NAKASONE: Yes. Thank you for opening that up. The effects of the distance to the voice is also regarded as a -- something would have to be careful. There are two terms to describe those. The first one is far field recording versus near field recording.

When we typically talk on the phone our lips are very close to the microphone, maybe an inch away. That's a typical near field recording. It captures our voice very clearly.

On the other hand when the person is say far end of the room and there are say 10 feet, 15 feet away, we call this voice when captured by recording system, we call it far-field recording or far-field voice.

And far-field voice is problem because not only weakened signal. It has a noise with it, mixed with noise created by the room and also vibrations and echoes created by (IANUDIBLE) bouncing walls. That sort of smears the voice.

Another thing is that far-field voice tends to have very poor signal to noise ratio. Signal to noise ratio is the ratio of signal divided by noise signal. Say like in the courtroom the signal level when we measure it by sound pressure meter it might give something like 40 dB decibels. I'm talking probably with the total of about 80 dB loudness and the signal to noise ratio is sort of you know, it's a logarithmic expression.

We just use subtraction. So, 80 minus 40 gives me the voice. My voice carries about a 40 dB signal to noise ratio and that's what you hear. And 40 -- 20 is 100. OK, it's more than.


COSTELLO: All right. This is Dr. Nakasone testifying. He's a voice analyst for the FBI. All of us is struggling to understand why he's testifying because we thought that any voice analysis that would be introduced in court would not be allowed. And we're talking about those 911 tapes and whose voice was on the tapes.

Sunny, what do you make of this?

SUNNY HOSTIN, CNN LEGAL ANALYST: I think I'm just as confused as all of you are, Carol. You know, you're right. During the Frye hearing, not only was the decision made that there would be no expert analysis in terms of whose voice was crying on the tape. The judge found that type of science was unreliable and just hadn't been accepted in the scientific community, but this particular witness testified for the defense at the Frye hearing, which is why I suspect they said we look forward to hearing them. They have no objection to his qualifications, no objection to his testimony.

So, I'm not quite sure what he is going to testify to by the prosecution. Let's remember that this prosecution team is really sort of a dream team. Between the three of them, Bernie Del La Rionda has been a prosecutor for 30 years, he's tried hundreds of murder cases.

So, I suspect they know what they're doing. There's a reason why this witness is on the stand. I just don't know why.

COSTELLO: Page, earlier you said you were surprised this was the first witness of the day because after last week which was kind of devastating to the prosecution, you'd think there would be this exciting witness on the stand so the jury could sit up and pay attention.

PATE: Certainly, or at least one that's very strong for the prosecution. Trials are a lot about momentum. I think Friday ended strong for the defense.

So, Monday morning, if you're the prosecution you want to catch them when they are away and interested and start rebuilding some credibility towards your theme.

COSTELLO: All right. He's asking more questions. Maybe they will get to the meat of the testimony soon.

So, Jason Johnson --

JOHNSON: I don't know. Maybe there's some huge twist ending to this and look like geniuses at the end of the day. But right now, this makes absolutely no sense to anybody, and probably the jury.

COSTELLO: Well, they are talking about the 911, so let's listen.


NAKASONE: Voice compressed in portion, and another examiner was assigned to do something else. But I'm going to concentrate on what I did.


NAKASONE: By following this standard operating procedure I listened to entire recording. I got it over four minutes. I want to check my note.

I correct, four minutes to 2:34. Out of that 2:34, the portion was somewhere around 45 seconds, starting from the beginning of the recording, until there was a loud noise, almost like gunshot like audio event.

And that duration was about 45 seconds. We listen to the entire conversation exchange between the female caller and female 911 operator. And when I was listening to those conversations, of course, I could hear the voices going in the background, altercation, somebody yelling for help. That was screaming. They caught screaming.

Screaming goes in and out when the dispatcher and the female caller were talking and sometimes the screaming comes up isolated from those. But since we cannot really analyze any voice which is a step over by something else or super-imposed with other people's voice, I discounted those area and it came up a little less than three seconds, 2.53 seconds of screaming was standing by itself. UNIDENTIFIED MALE: Let me make sure I got this clear, Doctor. There were, as you reference, in that 45-second portion, there was a series that you described as the screaming noise. How long total you call step on and unstepped on?

NAKASONE: Total, if I -- if I combine all of those segments, it reached a little less than 19 seconds, 18.80 seconds or so. Of those, 16 some seconds was overlapped by other voices, and less than three seconds to 45-second, was the area where nothing else but the screaming what was coming through.