top of page

Lesson 1: Speaker Profiling

Definition: the creation of speaker profiles is a process in which an unknown speaker is profiled based on their speech

Key Concepts

  • Speaker Profiling

  • Phonetic Features

  • Linguistic Features

  • Dialects

UNIT 1: Introduction to Speaker Profiling

Imagine you hear someone speaking in a crowded room, but you cannot see who it is. How can you often tell whether it's a friend, a family member, or a stranger? Subconsciously, our minds are sorting through speaker profiles and patterns of speech, trying to find a match to the voice. So in a way, we all create speaker profiles of those around us. And, from a scientific point of view, speaker profiling involves creating a demographic profile of a person solely based on their voice. Intriguing, right? Our voices are unique and carry a lot of information about us, such as demographic information like our age, gender or even where we grew up. 

 

It may sound like something from a spy novel, but speaker profiling is used in forensic science for investigation. Why? Because as mentioned before, our voice is one of the most personal and unique aspects of ourselves, providing insight into our identity and enabling the detection of crimes. Our voices provide even more insight into ourselves than a fingerprint would.

30.png

Activity 1: Mystery of

the Stolen Lindt Recipes

Answers Person 1 is American. Person 2 is Swiss. Person 3 is Swiss/Swedish. Person 4 is Egyptian. Person 5 is American.

Exercise 1: 
 

In the renowned Lindt chocolate factory, valuable chocolate recipes were stolen. The only clue: an audio recording containing voices of several individuals during the theft. Detective X, known for their expertise in speaker profiling, is tasked with creating a map of the thieves based on their voices to determine their origins. 

​

Listen to the recordings below and try to see if you can identify the country of origin that these speakers come from. Not all of them are native English speakers.

Person 1
00:00 / 00:09
Person 4
00:00 / 00:12
Person 2
00:00 / 00:21
Person 5
00:00 / 00:15
Person 3
00:00 / 00:11

UNIT 2: Diving Deeper into Features of Speech

Now, in order to create a concrete speaker profile of the suspect, we need to distinguish between phonetic and linguistic features.

​

By making this distinction between these two types of features, a more precise and detailed analysis of the speaker profile can be conducted. This allows for the identification of specific characteristics that are unique to a speaker.

​

We look at both phonetic and linguistic features as they represent different levels or dimensions of language. For example: 

​

Phonetic features pertain to the physical and acoustic properties of speech. They can be categorized into segmental and suprasegmental features. Segmental features, for instance, involve analyzing the fundamental frequency (think about the pitch of a voice, which for example tends to be lower in men due to longer and thicker vocal cords) and resonance or formants that determine the tonal characteristics of a voice. Suprasegmental features include aspects such as intonation, stress patterns, and rhythm, all of which contribute to the recognizable pattern of an accent.

​

On the other hand, linguistic features concern the structure and use of language itself, encompassing properties related to content and specific expressions in language (e.g., vocabulary, sentence structures, errors, or the use of idioms such as letting the cat out of the bag).

Exercise 2: 

​

Now as we heard in the audio recordings from the first activity, one member of the group of thieves is a person who studied mathematics and could be considered highly dangerous as they resorted to using the darknet to find the secret code to the safe. Not only is this person extremely clever, but they were also a bit nervous or may have been under the influence of some substances. As forensic linguists, it’s our job to help Detective X build a speaker profile to help us track down this gang of thieves. 

​

Listen to the following audio clip and circle the statements that apply to the speaker to help Detective X build a speaker profile.

Introduction.jpg

Activity 2: Creating a Speaker Profile

Mathematic Speaker
00:00 / 00:26

Let's Start with Phonetic Features

Phonetic Features.png

And Continue with Linguistic Features

Linguistic Features.png

Final thought for this lesson:

  • We have explored the concepts of linguistic and phonetic features that allow us to create a detailed speaker profile.

  • We have observed that each voice is unique, enabling speaker identification in fields such as forensics and technology.

​

In today’s digital world, it is crucial to consider the protection and ethical aspects of speech data. Is recording and analyzing someone’s speech without consent acceptable?

​​

Sources


Content​

Whaley, Lindsay J. (1997). Introduction to typology: the unity and diversity of language. London, UK: SAGE.

 

Videos

ariinbeijing (2017, June 23). What Languages Sound Like to Foreigners [Video File]. Retrieved from https://youtu.be/QxrDNRhYFyI?si=G_fBMn9Wxi0tISqH

Tom Scott (2015, June 11). Long and Short Words: Language Typology [Video File]. Retrieved from https://youtu.be/bxARj07jFp0?si=KBoqPYTsrnSStlA1

bottom of page