Table 3.

Perceptual ratings for similarity to other voices in the artificial language, similarity to the test voice, and similarity to native English speakers.

	Mean (SD) similarity to other voices	Mean (SD) similarity to test voice	Mean (SD) native-like rating
Speaker 1	3.15 (1.12)	3.70 (1.25)	2.80 (1.22)
Speaker 2	3.17 (1.22)	3.10 (1.45)	3.70 (1.06)
Speaker 3	3.09 (1.14)	3.10 (1.10)	2.20 (1.23)
Speaker 4	3.41 (0.99)	4.10 (0.88)	3.00 (0.82)
Speaker 5	3.24 (0.98)	3.10 (1.10)	2.60 (1.26)
Speaker 6	3.23 (1.29)	3.20 (1.48)	2.80 (1.03)
Speaker 7	3.38 (1.07)	3.70 (0.95)	2.40 (1.07)
Speaker 8	4.11 (0.89)	4.30 (0.82)	1.20 (0.42)
Overall	3.35 (1.13)	3.54 (1.12)	2.59 (1.01)

Note: The similarity ratings used a scale from 1 (very similar) to 5 (very different). The native English speaker judgments were based on asking how much each speaker sounded like a native English speaker from 1 (definitely a native speaker) to 5 (definitely not a native speaker).