What is Enhanced Dialogue as an Ear Witness?
Enhanced Dialogue as an Ear Witness is one of the most important forms of digital media evidence and the focus of this blog post. With the number of audio recordings used in court today, it is important that the audio recording be reliable and understandable. However, this is not always the case. Forensic enhancement of dialogue and preservation of the recording integrity allows an audio recording the opportunity to be an ear witness and reliable evidence for the Trier of Fact.
Audio Forensic Experts have a plethora of tools to choose from when forensically enhancing recorded dialogue. This post will feature Dialogue Isolate which is an application in iZotope RX and can also be used as a plugin for other digital work stations.
Why is it Important to use Audio Forensics?
The Scientific Working Group of Digital Evidence (SWGDE) publishes best practice guides for digital forensics practitioners to follow.
SWGDE Best Practices for Forensic Audio Version 2.2 (October 8, 2016) Section 7 describes:
“A forensic audio examiner should be specifically trained in the procedures performed. Refer to  and  for recommendations on forensic audio training. Do not attempt an examination of evidence without training, experience, and meeting laboratory qualifications”.
Background Noise is an Enemy of Dialogue
SWGDE Best Practices for Forensic Audio Version 2.2 (October 8, 2016) Section 7.4 Enhancement describes:
The goal of audio enhancement also referred to as clarification, is to increase the intelligibility of voice information or to improve the signal to noise ratio of a target signal by reducing the effects that mask it. This process typically begins with critical listening ,  and analysis ,  to determine the signal/noise ratio and the nature of distortions or interferences masking the target signal. This will aid in identifying a processing strategy to mitigate unwanted sounds or distortions. This strategy may include:
- Analysis tools to identify the limiting characteristics;
- Selection of filters that address the limiting characteristics;
- An appropriate signal path;
- The sequence of filters to be applied;
- Selection of playback and processing equipment or applications;
- Identifying whether regions require different filtering approaches and minimizing any boundary effects of regional processing.
For speech enhancement, improved intelligibility means more words can be understood. To determine if intelligibility has been increased, compare the original recording against the processed recording, repeating as frequently as necessary. This continued comparison will help avoid over-processing. For lengthy listening sessions, take breaks to avoid listener fatigue.
In less technical terms, the goal is to hear the dialogue and any sounds that get in the way of understanding the dialogue are unwanted noise. Some examples of this unwanted noise are car engines, HVAC and static and other electronic interference. Enhanced dialogue as an ear witness is most reliable when forensically enhanced.
Tools for Dialogue Enhancement used by Audio Forensic Experts
Unwanted Sounds: Non-Continuous vs. Continuous Noise
Unwanted sounds can be separated into two broad categories. The continuous noise is steady and has a similar tone and volume level throughout its entire duration, like the fan of your computer. Non-continuous noise changes over time like the sounds of lawnmowers, traffic on a busy street, and music. Between the two categories, non-continuous noise is harder to remove and requires particular tools and training in order to do so.
Dialogue Isolate is a relatively new tool that is available in iZotope RX. iZotope best describes Dialogue Isolate as follows:
“Dialogue Isolate in iZotope RX 7 Advanced is designed to separate spoken dialogue from non-stationary background noise such as crowds, traffic, footsteps, weather, or other noise with highly variable characteristics. It can be particularly effective at increasing the level of dialogue in challenging low signal-to-noise ratio conditions. In RX 7, Dialogue Isolate has been made available directly in Pro Tools as an AudioSuite plug-in—now this powerful technology is usable without needing to even open RX!”
How does Dialogue Isolate work?
Dialogue Isolate has recently had its functionality improved with version 7 of iZotope RX. This relatively new tool is extremely useful for improving the intelligibility of dialogue during the forensic enhancement process. It specifically avoids attenuating the frequencies where the plugin detects dialogue.
iZotope best describes the operational features of the Dialogue Isolate as follows:
“Dialogue Isolate uses a machine-learning algorithm (deep neural network) that was trained on a large library of speech and noise data to automatically recognize speech content in an input signal. It leverages this training data to detect and separate dialogue and noise into two separate signal components. The levels of the separated dialogue and noise components can then be independently adjusted using the Dialogue and Noise gain controls.”
Dialogue Isolate works best when the program is able to properly perceive the dialogue information. If the voice is buried too deeply under background noise or if the microphone is not picking up the dialogue, the program may consider some of that dialogue as noise and would attenuate that dialogue. However, if the microphone is able to pick up the dialogue well enough, Dialogue Isolate is very useful for filtering the unwanted noise.
Dialogue Isolate Module from iZotope RX7:
This is the first in a series of blog posts about enhanced dialogue as an ear witness in court. Check out the rest of our website for more information and to find out how to contact us.