When we listen to a voice the perception of the voice, the language and the speaker is perceived as a unity. But it is possible to distinguish between them. The following categories were created for my study of continuity announcers in the Danish broadcast radio. About 100 different voices were compared through listening.

The sound of the voice can be divided into sub categories such as

• Tone of voice
• Proximity of the voice
• Voice range (soprano – mezzo –alto)
• Dynamics (modulation/monotonous)

The language use can be divided into subcategories such as:

• Dialect
• Perceived speech tempo (fast/slow)
• Spontaneous speak vs. reading aloud

The speaker can be discribed by subcategories such as

• Estimated age ( e.g. 0-19, 20-39, 40-59 ect.)
• Gender

The voice of somebody evokes a range of images of the speaker. Shortly after the introduction of radio technology T.H. Pear wrote 'Voice and Personality' (1931). He claims that the voice express the personality and in 1934 Allport and Cantrill followed up with the study 'Judging Personality from Voice. Based on this more emotional questions of the perception of the speaker might be added such as

• Is the speaker likeable?
• What do you think the speaker looks like.

What affects both the use of voice, the language and our perception of the speaker is the performative situation or the context. There is a huge different in the use of the voice if the recording has been done in a studio or in the street in the middle of car noise.

The research question is defining what your schema should look like. In order to be able to categorize the voices it is effective to oppose two elements on a scale such as monotonous/modulation. When you are starting categorising your voices invite somebody else to listen to the same voice sample with your scheme and your definition of the categories in their mind. If you and your test group disagree on what the voice is like you might want to change your category.