Subscribe to Audio Forensic Expert by Email!
Archive for April, 2012
Ever since I began my career as an audio engineer back in the 70’s, noise and sound quality have always been a focus mine as an audio expert. I remember when my interest first started back in the day when records—those twelve inch round black things—would play at parties and the background noise was never a point of focus. But today as vinyl is making a comeback, audio philes all over the world are starting to collect vinyl again. Last week, in fact, The Who released a 10 inch vinyl record to celebrate National Record Store day, but I digress…
There are several things that I do with almost every audio recording that needs to be clarified. Step one is to listen to a good portion of–if not the entire–audio recording and make observations about the noises that I feel can be removed. The second step is to determine what tools I have available that I can use to remove those unwanted sounds. Lastly, I begin the process of noise reduction in the steps of audio clarification.
Almost every software program that’s available has a noise reduction program in it. Once you learn how to use that noise reduction program, you first select the audio file you are working with and begin by removing unwanted noise. You’re only going to be able to remove one type of that unwanted noise, so my advice is to try different settings within the noise reduction program. Then, the next step is to use equalization to first remove additional unwanted sounds and then use a second equalization process to increase the desired or wanted sounds. When that’s complete, the last step of the restoration and clarification process is compression. Compression basically helps to make all of the remaining sound even so that if what you’ve done so far has, for instance, decreased the volume of the digital file to such an extent that the volume is very low, the output of the compressor plug in will help you raise that volume to a more desirable level.
Also, if it is an audio recording that you are trying to hear a conversation of, the playback of what you have done with all of these tools is almost as equally important as the entire clarification process is. I have found that listening to a very hard to hear audio recording on different devices helps my ability to create a transcript or to hear all of the sounds more clearly. For example, if I am restoring an audio that has bad or hard to hear conversation on it even after I have finished the restoration process, I will then burn that file onto a CD and listen to it, for instance, on a boom box or on a clock radio. Because both devices will play that recording back differently, my physical sound perception will pick up those sounds differently and aid in the transcription process.
This whole process is all about the science of what the ear perceives. Because the perception of sound is different from person to person, it is important to listen back to the hard to hear conversations, music in the background, any other sound that you’re trying to identify (gunshots, screams, etc.) on more than one playback device. The main reason is because headphones and the various types of playback devices that have external speakers will reproduce that sound differently based on each device that you’re listening to the sound on.
So in conclusion, follow the steps that I’ve outlined above as you are using your software program and do the best you can with the tools that you have to work with when restoring your audio file. It is also almost as equally as important to listen to it played back on multiple devices, which will help your clarification process out.
The human voice is like a symphony: a follow up on the 911 calls
By Ed Primeau
There are always two sides to any litigation. One side believes and defends while the other disagrees and prosecutes. Yesterday, April 2, 2012 was a day full of requests, demands and criticism. It wasn’t much different from when I testify. It’s the opposing council’s position to discount my findings and any of the forensic cases I have successfully completed.
As forensic experts we use our available tools to examine and bring truth into the courtroom. Regarding the Trayvon Martin case and voice identification, as forensic experts, Tom Owen and I may have the tools and experience available to us to conduct voice identification, but we do not have the proper voices recorded to arrive at a positive ID and conclude the voice yelling on the 911 recording is Trayvon Martin yelling for help. By process of elimination, both Tom Owen and I agree the voice yelling in the 911 recording is not George Zimmerman. That is our opinion, and there is no money involved.
Tom used a computer software program to analyze his findings and I used my ear: critical listening skills. Lawrence O’Donnell summed it up well (last night on MSNBC) when he compared my skill to that of a piano tuner.
The human voice is like a symphony; each voice’s uniqueness is based on the physical and psychological characteristics of the individual. A symphony uniqueness is based on the type of instruments played and skill of the musicians playing the instruments. Similarly, human voice is created using many physical components, like the lungs, larynx and wind pipe. A person’s voice exits the body through the mouth, involving the tongue, teeth and lips.
When the human voice is electronically recorded, we add other electronic characteristics to the voice through the electronic equipment used to make the recording.
During playback of the recording, forensic experts have the opportunity to listen to these components in harmony and examine them forensically. As a voice identification expert, I have the skill to look for symphonic characteristics of each human voice and make note of my observations to arrive at a conclusion. Does it match the person in question or not?
I believe like a fingerprint, no two human beings have the exact vocal components, which is why no two voices are exactly alike. When conducting voice identification, I listen for voice tone, articulation, how words are pronounced, vowels etc. I also view the sound spectrum by electronically measuring the sound frequencies with a spectrum analyzer which reveals the frequency spectrum range in the voice being examined.
In the background of the 911 call, I hear a male voice yelling for help which cracks like teen male’s does when going through puberty. This is one of the characteristics I observed during my initial investigation. The tone of Zimmerman’s voice is also different than the tone of the voice yelling for help. That’s about all I can conclude as absolute at this point.
It would help us continue our investigation if we could convince George Zimmerman to agree to let a forensic examiner record a sample of his voice screaming in the exact same location as the voice in the 911 call (we would have the same acoustics on the sample which would serve as an exemplar), and use as many of the same electronic components that were used in the recording of the woman’s 911 call where we hear the screaming and the gun shot.
However, in this case all we have is Zimmerman’s voice speaking, not yelling. At this point Tom and I agree based on our skills and what we have to work with that the voice is not George Zimmerman yelling for help. Until we get the right tools, we cannot proceed to determine if the voice is Trayvon.
Right now I believe only one person knows the answer: George Zimmerman.