Figure 6.
The three stacked bar chart sections present Attack M O S, Release M O S and Overall M O S. Scores range from 1 to 5. Each section includes speakers m 01 to m 05, speaker f 01 and the whole group. T F, Ours and S A 2 are compared using score distributions and model means plus or minus 95 per cent confidence intervals. For Attack M O S, whole-group means are 2.73 plus or minus 0.05 for T F, 2.92 plus or minus 0.05 for Ours and 2.94 plus or minus 0.05 for S A 2. For Release M O S, whole-group means are 2.72 plus or minus 0.05 for T F, 2.97 plus or minus 0.05 for Ours and 2.85 plus or minus 0.05 for S A 2. For Overall M O S, whole-group means are 2.64 plus or minus 0.05 for T F, 2.88 plus or minus 0.05 for Ours and 2.80 plus or minus 0.05 for S A 2.Score distributions in subjective evaluations of nuance reflection for attack, release and overall impression (from top to bottom). For each speaker, the frequency distribution of nuance reflection scores for the synthesized sounds is shown as a stacked bar chart. The corresponding MOS (mean ± 95% confidence interval) is displayed next to each bar