What is Forensic Transcription?
Forensic transcription can be considered the transcription of anything of an evidentiary nature. However, it is often used to refer specifically to the transcription of marginally intelligible speech. Actually - not just the transcription of it but the scientific methods used in doing so. This second definition of forensic transcription is the one this article addresses.
The best forensic transcriptionists are going to be linguists with a mastery of phonetics. They don't just listen closely to the dialog and type what they here. They use various audio enhancement and noise reduction methods as well as more traditional means that transcriptionists use such as looping, slowing down the speed of the speech and using high quality headphones. In addition, these forensic transcriptionists will pay close attention to pragmatics, syntax, and other characteristics of a person's speech and finally, they may use a spectrogram to help in the decoding process. A spectrogram is a visual representation of a speech waveform.
Speech and Spectrograms
When we speak we produce different types of sounds and these sounds are created in different ways. Sounds are created by different parts of the tongue, lips, soft palate, teeth, and so-forth. How and when air movement occurs in different parts of the vocal tract such as the oral and nasal cavities is also important. The point is that there is a variety of very different types of sounds and these different types of sound show up differently on the spectrogram. Spectrograms show harmonics, plosives, fricatives, nasal stops and so-forth in distinct ways. A linguist can't read a spectrogram like a book, but it can give clues by showing the types of sounds that are made; even when the speech itself isn't all that intelligible. This in turn helps the linguist decode difficult-to-understand speech.
Finding the Right Company
If you have a speech recording that is for some reason difficult to understand, you will want to find a company that provides forensic transcription services. Many transcription companies will accept marginally intelligible recordings (and charge you more for it) but the only difference in service is that they will spend more time listening to it. They will use the tried and true methods that transcriptionists have used for years with recordings of either good or fair quality (looping, slowing the speech, and turning up the volume). This will be a waste of your money. It may be okay for slightly difficult or noisy recordings but not the really difficult stuff.
So in order to keep the "unintelligible" and "inaudible" entries to a minimum you will have to look beyond the typical transcription company (i.e., medical, legal, business, and general) to take a crack at your recording. You can take a look at forensic audio and "more advanced" transcription and detective related companies. These will often produce better results but you still have to be careful. Many of these companies will have an audio engineer use audio enhancement techniques at an attempt to improve the intelligibility - but then they assign the actual transcription to a regular transcriptionist. People need to understand that audio enhancement only goes so far. Oftentimes it doesn't make anything easier to understand. It'll make the recording less noisy or "more listenable" by eliminating buzzing, tones, hum, and the like, but the speech itself will often remain mostly unintelligible. Since the cost is much more at this level your recording will be given more attention and the additional time may yield greater results with the deciphering process but it isn't the same as having someone with a background in phonetics do the decoding.
In addition, a forensic linguist probably won't want an enhanced recording. He will know which parts of the recording should receive filtering and can do it himself. This way none of the recording will be over enhanced or enhanced unnecessarily. Why pay an audio engineer to work on it when the forensic transcriptionist can do it himself and only to the appropriate sections?
Not Just a Transcript
When your forensic transcription is done by a forensic linguist or phonetician (or sometimes a speech scientist), you not only get someone with the ability to provide the best possible transcript, but you also get a forensic analysis and report explaining the methods used along with how and why the linguist came to his conclusions. A curriculum vitae may also accompany the transcript and report which can be helpful convincing a judge to allow its use for trial. Having a qualified forensic linguist do the work and submit his analysis and CV may even be useful in preventing the case from going to trial at all.
Below are two real examples of forensic transcription by a linguist being used in an investigation.
1) A prisoner who had previously been on probation was also suspected of murder. The police decided to place a tape recorder in the visiting room of the prison where the he would meet with his girlfriend. Unfortunately the visiting room was crowded with noisy people. Thus, the recorded dialog of the prisoner and his girlfriend was mostly unintelligible. After the forensic examiner applied speech enhancement techniques to the recording and preformed a difficult forensic transcription, it was established that the suspect was telling his girlfriend about how he had committed the murder.
2) A group of men were traveling in a car and were discussing a crime they had recently committed. Unbeknownst to them their discussion was being recorded. However, the radio was on at the time as well as a noisy car engine making it difficult to understand. After using speech decoding and enhancement techniques the dialog was successfully transcribed. It was discovered that the men who were suspected of auto theft, had been actually talking about a murder.