Non-experts' perceptual dimensions of voice assessed by using direct comparisons

Three data sets of 13 speakers each are analysed using the elicitation phase of the Repertory Grid Technique in order to identify vocal perceptual dimensions of non-expert listeners. Sentences read by female and male speakers of German (Phondat 1 corpus) and by male Australian English speakers (AusTalk corpus) have been rated on (dis)similarity by same-sex listeners using triples. Applying a balanced incomplete design proposed for the Repertory Grid Technique, frequencies of dissimilar pairs are transformed into distance measures using non-metrical multidimensional scaling. For both German data sets, three dimensions describing the speaker differences are found, whereas for the Australian data, we found four dimensions. By inspecting the individual labels given to extreme stimuli on each dimension, the authors subjectively named the respective dimension. Identical names for similar dimensions were avoided on purpose, aiming at reflecting each connotation best without inferring identity. The names are “calmness”, “factual”, and “naturalness” for German men; “tension”, “positive timbre”, and “maturity” for women; and “pitch”, “remarkable timbre&voice”, “emotion”, and a fourth unnamed one for Australian men. Although an exact match cannot be supposed, there are strong similarities between dimensions from the different data sets (factual/tension/emotion, naturalness/maturity, timbre/timbre&voice). A smaller pretest supports these findings. Overall, the dimensions observed are found to be more complex and person-related than typically described in literature. These results add to the current state of research in perceptual dimensions of non-experts, and represent a foundation for developing a questionnaire to assess listeners’ impressions.