Audio experts rush to judgment on analysis of Mel Gibson tapes

 

[caption id="attachment_10795" align="alignleft" width="241" caption="Mel Gibson - wikimedia commons"][/caption]

Carol Forsloff - Mel Gibson has been in the news over recordings said occurred during arguments with his girlfriend, Oksana Grigorieva, in which the listener hears epithets and threats allegedly those of Gibson.  How do experts verify audio evidence and is it conclusive?

In language assessment, and forensic work of any kind,  the manner of collecting the evidence and how that evidence is examined can make a serious difference in whether it is determined to be authentic or not.

Most forensic experts would make no definitive conclusions based upon the evidence that has come in snippets.  They might explain how tapes are analyzed and what type of evidence is used for authentication, but making flat-out statements of absolutes can create a rush-to-judgment in any direction if people are not well-informed about what the process entails.  Professional Codes of Conduct specify what needs to happen in discussions of this sort.

In terms of how a tentative conclusion can be made, an expert would want to avoid using the word "all" in making any statement anyway, because the word allows for no exceptions.  Most scientific statements are given with qualifiers based upon the scientific principles used, the methods done, and the conditions under which the information was obtained.

In this case the Mel Gibson tapes were downloaded audio material from digital recordings.  Some experts quoted on television news never heard the recordings in any form, whereas others heard them over broadcasts.  But a laboratory examination is a whole different matter, and the process of determining authenticity is a process that has specific protocols.

So before Mel Gibson is hung by the press or folks rush to embrace his innocence in his defense, it is good to know that this process can be divisive under circumstances where evidence is copied in some manner or not found in original form, as occurs when sounds are recorded then downloaded into digitally.

ENF analysis describes a process of examination that may be applied to a digital audio/video recording in order to assist in determining its authenticity, according to the best practices outlines produced by The Working Group for Speech and Audio Analysis.

The process of authenticating audio material  requires determination if the recording is original, the first media on which the recording was made, not a copy, clone or the result of a file transfer.  An assessment is made of whether the recording's production came from a battery-operated recorder or one from an electrical outlet.  The examiner assesses whether the tape, disc, memory card, or other device has been used before and how the long the recorder has been used.  The recorder's settings are also important details.  Additionally it is determined whether the recorder has been switched on and off during the recording.  These are just several of the steps in the process of making a judgment about the authenticity or lack of authenticity of a tape or recording.

Other details examined include whether the recording has been played back, transcribed or copied and under what conditions.  The date and time of the recording, the person's responsible for them and the sequence and timing of events in the questioned recording are also important to examine.  Then the evaluator looks at any modifications or repairs that have been carried out before or after the recording has been undertaken.

Each forensic field has its own strategies based upon its best practices procedures, but they are similar in using a scientific method for determination.  It is easy, under the glare of a television light and a celebrity case, to make assumptions; but in a courtroom judges and juries want scientific process to be used.

Most cases have to fall under what is called the Daubert Rule, which is the establishment first that the expert is indeed an expert, so the qualifications, experience and education are examined.  Then the expert must attest to his/her field as a science and establish why it is by the procedures used ordinarily in the field and that he/she has used those procedures. 

Mel Gibson's reputation as an actor and his honor as well as the potential of a criminal charge and financial damages make the case critical enough for the proper assessment of evidence done in proper ways.  His girlfriend, who may have suffered verbal and physical abuse, most would agree, also deserves a fair treatment under the law with the evidence provided.

Before the rush to judgment and a decision is made about Mel Gibson's guilt or innocence is determined, the process of data collection and the examination of the evidence will take time and may indeed prove inconclusive based upon the difficulty in getting the auditory details that will hold up in court. 

"Does it sound to you like this was edited at all?" NBC's Jeff Rossen asked forensic voice analyst, Paul Ginsberg, in a taped segment on "Today."

"No, his voice is continuous in tone and her voice is continuous in tone, level and amount of distortion." Ginsberg said. "There's no evidence of any editing or erasures. I have examined this in minute details, down to the thousandth of a cycle."

Ginsberg addressed why the voice, widely acknowledged to be Grigorieva's, sounds so clear, while the voice said to be Gibson's sounds further away.

"He is the remote party, that is, he is far away coming in over a little speakerphone, where she's talking into this, equivalent of broadcast microphone," Ginsberg explained.

However, Arlo West, CEO of Creative Forensic Services, said on "GMA" on Thursday morning that he believed the recordings had been altered.

"The audio files were most certainly edited intentionally and the forensic terminology for this, is tampering," West said. "The edits that I found would have been done post-recording and are clearly done to redact dialog -- which means to intentionally remove words. I can tell there are gaps and edits in several ways by using waveform analysis and zero crossing analysis. These are most certainly digital recordings so there is no tape involved.

"I believe these are professionally done. I think she had help," West continued. "She clearly was speaking into what we call a large diaphragm microphone. Her voice is very well engineered. She sounds great." 

We can see how experts differ when all the facts aren't in.  That difference of opinion might remain when more is known because forensic science is science but not absolute because of variations that can occur in interpretation of certain details.

But media make assumptions on the glimpses of the experts.  The public then decides as well on the guilt or innocence of either Gibson or girlfriend Grigorieva before the facts are known.  It is entertaining certainly, but it is also important to understand the process itself as that is what our justice system maintains.

No one but the parties involved right now knows who is guilty of what and when, but it is certainly true that Gibson is in serious trouble, if only in proving his innocence in the face of public opinion where minds may be made up so the facts, when they are learned, make no difference.

-----------------------------------------------

The author/reporter of this article has been court qualified in the State of Hawaii as a forensics examiner in questioned document examination, rehabilitation counseling and life care planning and has been qualified under the Daubert rule.