Table 7 Summarization of all systems tested

Spectrogram prediction models

1.

M-MN (baseline)

MN12h

-

-

-

-

-

2.

M _SJ-TL

JP

MN30

-

-

-

-

3.

M _SE10-TL

EN10

MN30

-

-

-

-

4.

M _SE24-TL

EN24

MN30

-

-

-

-

5.

M _SEJ-TL

EN24

JP

MN30

-

-

-

6.

M _S-DA

AD ₃₀ + MN30

-

-

-

-

-

7.

M _SEJ-TL-DA

EN24

JP

AD ₃₀

MN30

-

-

8.

M _SEJ-TL-DA _D

EN24

JP

AD ₃₀-set1

AD ₃₀-set2

AD ₃₀-set3

MN30

9.

M _MJ-TL

JP + MN30

MN30

-

-

-

-

10.

M _ME10-TL

EN10 + MN30

MN30

-

-

-

-

11.

M _ME24-TL

EN24 + MN30

MN30

-

-

-

-

12.

M _MEJ-TL

EN24 + JP + MN30

MN30

-

-

-

-

13.

M _M-DA

AD ₃₀ + MN30

-

-

-

-

-

14.

M _MEJ-TL-DA

EN24 + JP + AD ₃₀ + MN30

MN30

-

-

-

-

15.

M _MEJ-TL-DA _D

EN24 + JP + AD ₃₀ + MN30

AD ₃₀-set1

AD ₃₀-set2

AD ₃₀-set3

MN30

-

16.

M _MEJ-TL-DA _1hour

EN24 + JP + AD _1h + MN1h

MN1h

-

-

-

-

17.

M _MEJ-TL-DA _2hours

EN24 + JP + AD _2h + MN2h

MN2h

-

-

-

-

18.

M _MEJ-TL-DA _3hours

EN24 + JP + AD _3h + MN3h

MN3h

-

-

-

-

Neural vocoders

19.

NV-MN (baseline)

MN12h

-

-

-

-

-

20.

NV-DA

AD ₃₀ + MN30

-

-

-

-

-

Model type

M _SXXX = Single-speaker TTS model

M _MXXX = Multi-speaker TTS model

NV = neural vocoder

Method used for model training

TL = Cross-lingual transfer learning

DA = Data augmentation

TL-DA = Cross-lingual transfer learning and data augmentation

TL-DA _D = Cross-lingual transfer learning and data augmentation with additional fine-tuning

Databases used for training stages

EN10 = 10 hours of the English dataset

EN24 = 24 hours of the English dataset

JP = 10 hours of the Japanese dataset

MN12h = 12 hours of the target language dataset

MN30 = 30 minutes of the target language data

MN1h = 1 hour of the target language data

MN2h = 2 hours of the target language data

MN3h = 3 hours of the target language data

AD ₃₀ = augmented data generated from 30 minutes of the target language data

AD ₃₀-set1 = the first set of the augmented data generated from 30 minutes of the target language data

AD ₃₀-set2 = the second set of the augmented data generated from 30 minutes of the target language data

AD ₃₀-set3 = the third set of the augmented data generated from 30 minutes of the target language data

AD _1h = augmented data generated from 1 hour of the target language data

AD _2h = augmented data generated from 2 hours of the target language data

AD _3h = augmented data generated from 3 hours of the target language data