Fig. 1
From: Listeners form average-based representations of individual voice identities
Illustration of the 2D voice space and the voice identities we created. VTL = vocal tract length, GPR = glottal pulse rate, ST = semitones. Shaded squares show the nominal within-person voice spaces for the 3 identities created from a sentence produced by one original talker (green). Empty dots arranged around the perimeter of the individual voice spaces show the locations of stimuli from the ring-shaped distribution used during the learning phase; filled dots show the locations of the stimuli forming the centre distribution introduced at test. Two sets of identities were created in this way, each based on a different original speaker’s voice. One set of identities was used during the learning phase; the second set was introduced to listeners as distractor identities at the old/new recognition task, which took place after the learning phase. Assignment of the two sets of identities as trained and distractor identities was counterbalanced across the participant sample. Please see the Methods section and Supplementary Note 1 for more details on the arrangement and assignment of training and distractor stimuli
