Skip to main content

Table 3 Occurrences and distributions of Mongolian letters in the 30-min and 12-h target language datasets and IPA phonetic symbols

From: Text-to-speech system for low-resource language using cross-lingual transfer learning and data augmentation

# Letter Phoneme #Occurrences and distribution # Letter Phoneme #Occurrences and distribution
    30 minutes 12 hours     30 minutes 12 hours
1. a a 1870 11.00% 54403 11.33% 18. m 372 2.19% 10349 2.15%
2. n, 1378 8.11% 41565 8.65% 19. v 234 1.38% 8175 1.70%
3. e 1625 9.56% 38739 8.07% 20. 190 1.12% 6744 1.40%
4. g 1030 6.06% 28254 5.88% 21. z 218 1.28% 6691 1.39%
5. p r 1060 6.24% 26879 5.60% 22. i 148 0.87% 5109 1.06%
6. i 686 4.04% 25673 5.34% 23. 259 1.52% 4890 1.02%
7. d 827 4.86% 23810 4.96% 24. i 176 1.04% 4641 0.97%
8. x h 856 5.04% 22237 4.63% 25. e j 132 0.78% 4390 0.91%
9. l 742 4.36% 21224 4.42% 26. 114 0.67% 3516 0.73%
10. Y u 684 4.02% 19969 4.16% 27. c 72 0.42% 2351 0.49%
11. t 647 3.81% 19362 4.03% 28. ja 27 0.16% 1765 0.37%
12. o o 805 4.74% 19019 3.96% 29. j 17 0.10% 843 0.18%
13. i 613 3.61% 18724 3.90% 30. ë jo 36 0.21% 727 0.15%
14. y 604 3.55% 17868 3.72% 31. k 24 0.14% 591 0.12%
15. ö 532 3.13% 14145 2.94% 32. f 5 0.03% 471 0.10%
16. c s 495 2.91% 13438 2.80% 33. p 1 0.006% 401 0.08%
17. b 519 3.05% 13330 2.78% 34. i 1 0.006% 26 0.005%