Table 2.
Observed percentages of accurate automatic alignments (i.e., alignment intervals that cover the midpoint of the manual alignment).
| Subset | Aligner | Accuracy (%) |
|---|---|---|
| All sounds (33,545) | MFA-SAT | 87 |
| MFA-No-SAT | 79 | |
| Kaldi | 77 | |
| P2FA | 69 | |
| Prosodylab | 63 | |
| Plosives (9,149) | MFA-SAT | 85 |
| MFA-No-SAT | 75 | |
| Kaldi | 75 | |
| P2FA | 62 | |
| Prosodylab | 51 | |
| Vowels (13,010) | MFA-SAT | 90 |
| MFA-No-SAT | 87 | |
| Kaldi | 81 | |
| P2FA | 76 | |
| Prosodylab | 77 | |
| Fricatives (6,485) | MFA-SAT | 86 |
| MFA-No-SAT | 75 | |
| Kaldi | 75 | |
| P2FA | 62 | |
| Prosodylab | 60 | |
| Others (4,901) | MFA-SAT | 80 |
| MFA-No-SAT | 70 | |
| Kaldi | 72 | |
| P2FA | 68 | |
| Prosodylab | 55 |
Note. MFA = Montreal Forced Aligner; SAT = speaker adaptive training; P2FA = Penn Phonetics Lab Forced Aligner.