Forensic speechreading


Forensic speechreading is the use of speechreading for information or evidential purposes. Forensic speechreading can be considered a branch of forensic linguistics. In contrast to speaker recognition which is often the focus of voice analysis from an audio record, forensic speechreading usually aims to establish the content of speech, since the identity of the talker is usually apparent. Often, it involves the production of a transcript of lipread video-records of talk that lack a usable audiotrack, for example CCTV material. Occasionally, 'live' lipreading is involved, for example in the Casey Anthony case. Forensic speechreaders are usually deaf or from deaf families, and use speechreading in their daily lives to a greater extent than people with normal hearing outwith a deaf community. Some speechreading tests suggest deaf people can be better lipreaders than most hearing people.

Speechreading expertise

No tests of speechreading have yet been developed in a forensic context: that is, to benchmark individual skills in speechreading from a video record, including production of a reliable transcript. For many years, UK agencies made extensive use of one particular speechreader, whose reports are now not to be used for evidential purposes. Several speechreaders and deaf professionals currently offer these services. In the UK, these include Terry Ruane, , among others. In the US, is a leading lipreading expert. Expert speechreaders may be able to advise on various issues, including whether a videorecord is or is not speechreadable, and the accent and language used by a talker. Commissioning agents need to be aware of issues inherent in the unreliability of speechreading, and be prepared to treat such advice with caution.

The law

In the UK, a landmark case and appeal established the admissibility of lipreading evidence. However, the appeal court also required that the judge should issue a special warning as to its risks and limitations.
While lipread speech can carry useful speech information, it is inherently less accurate than heard speech because many distinctive features of speech are produced by actions of the tongue within the oral cavity and are not visible. This is a limitation imposed by speech itself, not the expertise of the speechreader. It is the main reason why the accuracy of a speechreader working on a purely visual record cannot be considered wholly reliable, however skilled they may be and irrespective of hearing status. The type of evidence and the utility of such evidence varies from case to case.
In the US, there is debate concerning the admissibility of speechreading evidence and its status, especially in relation to variations in state and federal evidential procedures, and with respect to the privacy implications of the Fourth Amendment to the US constitution.

Three UK cases involving speechreading evidence