Preparation of Audio Evidence

Media Translation is the process of converting media from one format into another format.

There are many different sound file formats like wav, aiff, au, raw pcm, ape, flac, wv, tta, atrac lossless, m4a, wma lossless, shn, mp3, wma, opus, vorbis, atrac, aac, wma lossy.

Sound/Voice recording functions are available on most smart phones. The recording parameters need to be carefully chosen when the sound file is to be used as evidence in court.

There are many programs on computers and phones that can used to change the format of sound files.

For Court Evidence, only programs that retain the original quality and maintain the integrity of the original auditory scene should be used.

Listening Optimisation is the process of getting the evidence amplified and filtered so as to optimise evidence for hearing and understanding.

Amplification can be used as a listening optimisation technique that is either static or adjustable over Time, Intensity and/or Frequency.

Various speech band filters can be used to optimise listening. These filters can also be static or adjustable over Time, Intensity and/or Frequency.

Telephone Speech is commonly 300Hz to 3300Hz. The extended speech band is from 100Hz to 8000Hz. By applying a band pass filter to speech, out of band can be reduced or eliminated which can optimise the speech for Listening

Two different microphones that capture two different scenes or perspectives, can be placed on two channels of the sound file. If one channel is better quality than the other then channel separation can be used to optimise listening.

Region of Interest (ROI) Extraction involves separating out the important sections of the sound files as audio evidence. This ROI can be a section of an audio file from x to y minutes or seconds where y > x.

If microphone A whose signal is stored on the left channel of a sound file, contains the important audio evidence then channel separation can be used to extract that ROI.

Care must be taken in extraction of any ROI because of the implicated bias of removing a section of audio evidence from its overall context.

Either a sound editor or technical audio editing software can be used to do ROI extraction.

For Court Evidence, the ROI must be ethically extracted and only programs that retain the original quality and maintain the integrity of the original auditory scene should be used.

This means that the audio evidence has to shown represent the original auditory scene beyond reasonable doubt no matter what form it presented in, in court.

The key factor in determining if the audio evidence has sufficent credibility to be used in court, is the signal to noise ratio. To hear speech clearly the signal to noise ratio has to be significantly > 1 across all speech band frequencies.

If the signal to noise ratio is = 1, just > 1 or < 1 that means that the level of speech is not significantly greater than the level of noise and that some or most speech sounds are embedded in noise and therefore not intelligible.

In a listening test, where noisy speech has a low signal to noise ratio, different things will be heard when different people listen to the audio file and that audio file is not suitable as court evidence

Clear intelligible speech has range of intensity differences in the waveform, a falling slope in the spectrum, fine detail in the spectrogram and a broad range of energy in the intensity histogram.