license: mit
Data used in PercePiano dataset.
Full credit goes to the original authors: Jisoo Park, Jongho Kim, Jeong Mi Park, Ahyeon Choi, Wen-Syan Li, Jonghwa Park & Seung-won Hwang
Data includes the mean and standard deviation for each of the 19 features. Table 2 from the paper:
| Category | Label | Option 1 | Option 2 | D.960 mv2 | D.960 mv3 | D.935 | WoO.80 | No. of I don't know | ICC(1,1) | ICC(1,k) |
|---|---|---|---|---|---|---|---|---|---|---|
| Timing | 1 | Stable beat | Unstable beat | 2.96(1.82) | 3.35(1.67) | 3.94(1.71) | 3.31(1.62) | 5 | 0.22 | 0.95 |
| Articulation | 2 | Short | Long | 3.33(1.56) | 4.48(1.38) | 3.90(1.49) | 3.63(1.56) | 64 | 0.22 | 0.95 |
| 3 | Cushioned | Solid | 3.74(1.72) | 4.13(1.46) | 3.68(1.49) | 3.92(1.54) | 67 | 0.2 | 0.94 | |
| Pedal | 4 | Sparse/dry | Saturated/wet | 2.88(1.59) | 3.99(1.71) | 2.22(1.50) | 3.87(1.46) | 76 | 0.42 | 0.98 |
| 5 | Clean | Blurred | 4.26(1.84) | 3.59(1.65) | 3.42(1.52) | 3.09(1.37) | 115 | 0.33 | 0.97 | |
| Timbre | 6 | Even | Colorful | 3.66(1.69) | 3.58(1.44) | 3.84(1.54) | 3.59(1.49) | 23 | 0.18 | 0.93 |
| 7 | Shallow | Rich | 3.41(1.65) | 4.10(1.44) | 4.03(1.62) | 3.91(1.57) | 20 | 0.22 | 0.95 | |
| 8 | Bright | Dark | 3.82(1.50) | 3.13(1.21) | 3.13(1.31) | 3.99(1.24) | 38 | 0.25 | 0.96 | |
| Dynamic | 9 | Soft | Loud | 3.75(1.61) | 4.48(1.21) | 3.50(1.48) | 3.73(1.52) | 89 | 0.22 | 0.95 |
| 10 | Sophisticated | Raw | 3.48(1.60) | 4.02(1.36) | 3.58(1.65) | 3.54(1.59) | 46 | 0.24 | 0.95 | |
| 11 | Little range | Large range | 4.13(1.56) | 4.35(1.33) | 3.83(1.48) | 3.63(1.43) | 32 | 0.16 | 0.92 | |
| Music making | 12 | Fast paced | Slow paced | 4.23(1.34) | 3.24(0.99) | 3.40(1.17) | 3.62(1.30) | 30 | 0.39 | 0.98 |
| 13 | Flat | Spacious | 4.20(1.65) | 3.68(1.30) | 3.90(1.50) | 3.87(1.48) | 51 | 0.2 | 0.94 | |
| 14 | Balanced | Unbalanced | 3.36(1.55) | 3.92(1.30) | 3.98(1.53) | 4.16(1.49) | 28 | 0.18 | 0.93 | |
| 15 | Pure | Dramatic | 3.98(1.74) | 3.90(1.41) | 4.32(1.52) | 4.05(1.58) | 54 | 0.17 | 0.93 | |
| Emotion | 16 | Pleasant | Sad | 3.99(1.48) | 3.06(1.21) | 3.13(1.26) | 4.10(1.18) | 46 | 0.29 | 0.96 |
| 17 | Low energy | High Energy | 3.96(1.54) | 3.66(1.25) | 4.25(1.21) | 4.02(1.30) | 36 | 0.2 | 0.94 | |
| 18 | Honest | Imaginative | 3.81(1.62) | 4.44(1.20) | 3.84(1.64) | 3.69(1.59) | 69 | 0.2 | 0.94 | |
| Interpretation | 19 | Unsatisfactory | Convincing | 3.69(1.63) | 4.22(1.37) | 3.64(1.64) | 3.77(1.61) | 32 | 0.18 | 0.93 |
Table 2. Statistical analysis of piano performance evaluation. Each row represents a perceptual feature with 2 opposite options, divided into categories. Musical piece-named columns show mean (M) and standard deviation (SD) of labeled value (M(SD)). ICC(1,1) and ICC(1,k) show intraclass correlation coefficients for single and average ratings.
Each feature has a corresponding two item list with the float for the mean of all responses, followed by the float for the standard deviation.