Here at the Forensic Voice Centre, forensic voice comparison is the most common service sought by police and defence teams in the UK and internationally. But, what is forensic voice comparison, how is it useful, and what can you expect from an evidential report? In the first of our myth-busting series, the specialists at the Forensic Voice Centre uncover more about forensic voice comparison and how it can assist your case.
What is forensic voice comparison?
Forensic voice comparison, also known as forensic voice analysis, forensic speaker comparison, speaker recognition, or speaker identification, is the forensic science activity of comparing the voice and speech patterns of a known speaker (usually a suspect or defendant in a criminal case) with the voice and speech patterns of an unknown offender in an evidential recording.
We compare “known recordings”, usually police interviews, body-worn camera footage and custody CCTV recordings of the “known speaker” (the suspect or defendant) with the voice of the offender in an evidential recording. The aim is to determine the likelihood that the speech comes from the same person or from different people.
How does Forensic voice comparison work?
Our expert forensic practitioners carry out analytical listening and acoustic analysis of various aspects of the voice, such as accent, vowel and consonant pronunciation, pitch, vowel resonances, speech tempo, fluency, word usage, vocabulary and grammar. By examining phonetic, linguistic, and acoustic components, they assess the likelihood of their findings if the speech originated from the known speaker or from a different speaker. The nature and strength of the conclusion is judged by how similar the voices are, and how distinctive the voice features are compared to other speakers from a similar language background.
How are the results of a voice comparison presented?
At the end of the analysis, you will receive an evidential report that contains a conclusion (or set of conclusions) as well as further information about the analysis, the methods used, the expert analyst and any limitations with the recordings,
Like other forensic science activities, the conclusion is selected from a standard 13-point scale. There are six levels to indicate a “positive” conclusion (i.e., support for the view that the speakers are the same person) and six levels to indicate a “negative” conclusion (i.e., support for the view that the speakers are different people). In the middle of the scale is the level of “inconclusive” whereby the results of the analysis do not support either view.
We do not use terms such as “match” or “non-match” and we do not report numerical values (e.g., “99% match”).
Is forensic voice comparison the same strength evidence as DNA?
The results of voice comparison analysis can be tendered in evidence and are robust to cross-examination. However, it does not provide the same strength of evidence as a full DNA comparison typically does. We follow the recommendations put forward by the International Association for Forensic Phonetics and Acoustics (IAFPA) that this evidence is used in conjunction with other evidence.
What are the limitations of forensic voice comparison analysis?
There are limitations that affect the extent to which we can compare recordings. These limitations can affect the strength of the conclusions, or render the analysis unsuitable. Limitations include:
| BREVITY Recordings that contain a small amount of free-flowing speech. | SPEAKING LEVELS Extremely raised speech or whispered speech. |
| DISTANCE Speech spoken at a distance from the recording device that was not recorded clearly at the time of recording. | DISGUISE Speech where the speaker is disguising their voice either to sound unlike themselves or to sound like someone else. |
| STYLE Speech where the speaker is reading aloud or adjusting how they talk to address children or animals. | BACKGROUND NOISE Loud background noise can affect how well the speech signal is captured. |
| QUALITY Poor technical quality. | MISMATCH Where the recordings to be compared are very different in terms of any of the factors listed above. |
After we have screened the recordings for suitability, we will inform you which limitations are likely to affect the conclusion(s) in your case. This can assist you with deciding whether to authorise the estimate.
How much does forensic voice comparison cost?
The cost of forensic voice comparison depends on the number of “known speakers” and the number of recordings that you would like compared. Please call our team to receive a quote over the phone.
What do I need to prepare to submit a forensic voice comparison enquiry?
You can read all about our submission process here.
We require the following information for analysis, so it’s best to prepare this as soon as possible.
- Confirmation that all recordings are in their best/most original format and that they have not been reformatted or compressed in any way from the original. (If you are not able to obtain the original versions, please share the information on how they have been reformatted or changed.)
- Confirmation that you have uploaded all available known material of the ‘known speaker(s)’ including interviews, BWC footage and/or custody footage.
- Name(s) of the ‘known speaker(s)’
- Operation name/Reference number
- Dates and times of all recordings
- Exhibit references for all recordings (if applicable)
- Critical dates (such as trial or hearing dates)
- Information about how the recordings have been recorded, downloaded and transferred from the original recording device (e.g., the device type and software used)
- Information about how the recordings have been enhanced (if applicable)
- MG21 form (if applicable)
How long does forensic voice comparison take?
We aim to complete most cases within 6 weeks from the point of financial authorisation and receipt of all original recordings.
We are able to complete some urgent cases on a case-by-case basis. Please contact the team if you require an urgent turnaround time.
Does Voice Comparison require accreditation?
At the moment, the Forensic Science Regulator Codes do not apply to voice comparison, and so there is no requirement for accreditation, and accreditation is not available. However, we still follow the guidance and standards from the Regulator where possible in order to guarantee the quality of our reports.
got a question?
Our team is available to discuss case strategy and quotations over the phone or via email. Contact us today to find more about forensic speaker comparison.
The Forensic Voice Centre 