Whistle: Data-Efficient Multilingual and Crosslingual Speech Recognition via Weakly Phonetic Supervision
Paper • 2406.02166 • Published
Datasets:
text stringlengths 3 182 |
|---|
a ə |
a'a ə ɑ |
aa ɑ |
aaa ɑ ə |
aaaa ɑ ə |
aaaaa ɑ ɑ ə |
aaaaaa ɑ ɑ ə |
aaaaaaa ɑ ɑ ɑ ə |
aaaaaaaa ɑ ɑ ɑ ə |
aaaaaaaaa ɑ ɑ ɑ ɑ ə |
aaaaaaaaaa ɑ ɑ ɑ ɑ ə |
aaaaaaaaaaa ɑ ɑ ɑ ɑ ɑ ə |
aaaaaaaaaaaa ɑ ɑ ɑ ɑ ɑ ə |
aaaaaaaaaaaaa ɑ ɑ ɑ ɑ ɑ ɑ ə |
aaaaaaaaaaaaaa ɑ ɑ ɑ ɑ ɑ ɑ ə |
aaaaaaaaaaaaaaa ɑ ɑ ɑ ɑ ɑ ɑ ɑ ə |
aaaaaaaaaaaaaah ɑ ɑ ɑ ɑ ɑ ɑ ɑ |
aaaaaaaaaaaaah ɑ ɑ ɑ ɑ ɑ ɑ ə |
aaaaaaaaaaaaand ɑ ɑ ɑ ɑ ɑ ɑ ɑ n d |
aaaaaaaaaahh ɑ ɑ ɑ ɑ ɑ |
aaaaaaaaaai ɑ ɑ ɑ ɑ ɑ i |
aaaaaaaaaau ɑ ɑ ɑ ɑ o ʊ |
aaaaaaaaah ɑ ɑ ɑ ɑ ə |
aaaaaaaah ɑ ɑ ɑ ɑ |
aaaaaaaahhh ɑ ɑ ɑ ɑ |
aaaaaaaand ɑ ɑ ɑ ɑ n d |
aaaaaaah ɑ ɑ ɑ ə |
aaaaaaahaaaahaa ɑ ɑ ɑ ɑ h ɑ ɑ ə |
aaaaaaand ɑ ɑ ɑ ɑ n d |
aaaaaagggh ɑ ɑ æ ɡ |
aaaaaah ɑ ɑ ɑ |
aaaaaahaahaa ɑ ɑ ɑ h ɑ ə |
aaaaaahhh ɑ ɑ ɑ |
aaaaaahhhhh ɑ ɑ ɑ |
aaaaaahhhhhh ɑ ɑ ɑ |
aaaaaand ɑ ɑ ɑ n d |
aaaaah ɑ ɑ ə |
aaaaahh ɑ ɑ ɑ |
aaaaahhh ɑ ɑ ɑ |
aaaaahhhh ɑ ɑ ɑ |
aaaaahhhhh ɑ ɑ ɑ |
aaaaahrrr ɑ ɑ e ɻ ɻ |
aaaaahs ɑ ɑ z |
aaaaand ɑ ɑ ɑ n d |
aaaaannnnndd ɑ ɑ æ n n d |
aaaaaow ɑ ɑ a ʊ |
aaaaargh ɑ ɑ ɻ ɡ |
aaaaauggggh ɑ ɑ o ɡ ɡ |
aaaaauggghh ɑ ɑ o ɡ |
aaaaayyyyyy ɑ ɑ a i j i ə i |
aaaagh ɑ ɑ ɡ |
aaaah ɑ ɑ |
aaaahh ɑ ɑ |
aaaahhh ɑ ɑ |
aaaailanta ɑ ɑ ə l æ n tʰ ə |
aaaallllll ɑ ɑ l l l |
aaaamooguuus ɑ ɑ m u ɡ u z |
aaaand ɑ ɑ n d |
aaaannndd ɑ æ n d |
aaaannnnnddddd ɑ ɑ n n m d d |
aaaarah ɑ ɑ ɻ ə |
aaaauahauah ɑ o ə h a ʊ ə |
aaaaugh ɑ o |
aaaauugghhh ɑ o ə ɡ |
aaaauuuuuuuhhhh ɑ o u u ə |
aaaay ɑ ɑ i |
aaab e i ɑ b |
aaabveyjvcmlnaw ɑ ə b v e i h e v kʰ m l n o |
aaad ɑ æ d |
aaadaaaaa ɑ ɑ d ɑ ɑ ə |
aaads ɑ æ d z |
aaaffair ɑ ə f e ɻ |
aa'ages ɑ i dʒ i z |
aaages ɑ i dʒ i z |
aaaggh ɑ æ ɡ |
aaah ɑ ə |
aaahahah ɑ ɑ h ɑ h ɑ |
aaahh ɑ ɑ |
aaahhh ɑ ɑ |
aaai ɑ a i |
aaaiat ɑ e æ tʰ |
aaaiatanee ɑ e ə tʰ e i n i |
aaaigned ɑ e i n d |
aaaiiimmhmmm ɑ e i m h m |
aaailable ɑ e i l ə b ə l |
aaainst ɑ e n s tʰ |
aaaiter ɑ e i tʰ ə ɻ |
a'aal ə ɑ l |
aaala ɑ ɑ l ə |
aaald ɑ o l d |
aaale ɑ e i l |
aaalk ɑ o kʰ |
aaall ɑ o l |
aaam ɑ ə m |
aaamage ɑ æ m i dʒ |
aaamiin ɑ ɑ m i æ n |
aaamma ɑ ɑ m ə |
aaand ɑ ɑ n d |
aaanhmmm ɑ æ h m |
aaanieil ɑ ɑ n i i l |
End of preview. Expand in Data Studio
English and Chinese IPA Lexicons and Phoneme Sets
This repository provides English and Chinese IPA pronunciation lexicons and phoneme inventories collected from the vocabularies of multiple ASR corpora. The resources can be used for ASR, grapheme-to-phoneme conversion, pronunciation modeling, TTS, and related speech research.
Repository Structure
.
├── en/
│ ├── lexicon.txt
│ └── phone_list
└── zh/
├── lexicon.txt
└── phone_list
Source Corpora
The vocabulary list covers transcribed words from the following ASR resources. Only the compiled pronunciation dictionary and phoneme set are published here; the audio, annotations, and metadata for each corpus are still provided by the original publishers.
Chinese
English
- AMI Meeting Corpus
- Earnings-22
- GigaSpeech
- SPGISpeech
- LibriSpeech
- VoxPopuli（English subset）
References
Saierdaer Yusuyin, Te Ma, Hao Huang, Wenbo Zhao, Zhijian Ou. "Whistle: Data-Efficient Multilingual and Crosslingual Speech Recognition via Weakly Phonetic Supervision." IEEE Transactions on Audio, Speech and Language Processing, 2025. https://arxiv.org/pdf/2406.02166
Te Ma, Min Bi, Saierdaer Yusuyin, Hao Huang, Zhijian Ou. "LLM-based phoneme-to-grapheme for phoneme-based speech recognition." INTERSPEECH 2025. https://www.isca-archive.org/interspeech_2025/ma25_interspeech.pdf
- Downloads last month
- 141