- Status: Planned Program
Background
The transcription of linguistic data and the identification of speakers have been the subject of some research activity in the analysis of audio data. However, the evaluation of meta-information that can be extracted from acoustic signals in addition to spoken language is still a largely unexplored field. In particular, it is not currently known which information can be derived about the environment in which audio data was recorded. This is highly relevant in time-critical application scenarios, such as emergency calls or missing person investigations, or in the context of so-called “secret recordings”. These are audio recordings that are created covertly and without the knowledge of other people. They can occur, for example, during kidnapping cases.
Aim
The “Audio Forensics” (AuFo) research program aims to analyze low-quality audio data regarding its recording location. The objective is to examine the extent to which partial room acoustic parameters allow the characterization of the recording environment, including but not limited to size, geometry, composition, and location (interior or exterior). This is necessary in order to ideally be able to narrow down the regional scope. Besides the reconstruction of the recording location, it would also be possible to analyze any information that can be derived about the recording device, e.g. the special characteristics of a smartphone microphone. Depending on the distinctness of the gained data, information could also be extracted to uniquely identify, recognize or limit possible sound sources.
Ideally, law enforcement authorities will be able to obtain precise location information, which will enable them to respond to time-critical applications, such as kidnapping or missing person investigations. Even initial investigative leads offer considerable value to law enforcement.
Disruptive Risk Research
We believe that this research objective has a tremendous potential for innovation. According to the current state of the art, it is unclear whether it is possible to reconstruct the recording location based on low-quality audio data. Currently, it is not feasible to predict whether the research results will be applicable in the context of criminal investigations. A significant challenge is ensuring that forensic methods can be applied to safeguard legal evidence. From a scientific and technical perspective, the program must therefore be considered to be high-risk research.