jimbozhang commited on
Commit
ed16922
1 Parent(s): 1b06aad

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +54 -119
README.md CHANGED
@@ -12,14 +12,26 @@ size_categories:
12
  ---
13
  # speechocean762: A non-native English corpus for pronunciation scoring task
14
 
15
- ## Example
 
 
 
 
 
 
 
 
 
 
 
 
 
16
  ```bash
17
  pip install datasets
18
  pip install librosa
19
  pip install soundfile
20
  ```
21
 
22
-
23
  ```python
24
  >>> from datasets import load_dataset
25
 
@@ -44,69 +56,9 @@ pip install soundfile
44
  'total': 9}
45
  ```
46
 
47
- ## Introduction
48
- Pronunciation scoring is a crucial technology in computer-assisted language learning (CALL) systems. The pronunciation quality scores might be given at phoneme-level, word-level, and sentence-level for a typical pronunciation scoring task.
49
-
50
- This corpus aims to provide a free public dataset for the pronunciation scoring task.
51
- Key features:
52
- * It is available for free download for both commercial and non-commercial purposes.
53
- * The speaker variety encompasses young children and adults.
54
- * The manual annotations are in multiple aspects at sentence-level, word-level and phoneme-level.
55
-
56
- This corpus consists of 5000 English sentences. All the speakers are non-native, and their mother tongue is Mandarin. Half of the speakers are Children, and the others are adults. The information of age and gender are provided.
57
-
58
- Five experts made the scores. To avoid subjective bias, each expert scores independently under the same metric.
59
-
60
-
61
  ## The scoring metric
62
  The experts score at three levels: phoneme-level, word-level, and sentence-level.
63
 
64
- ### Phoneme level
65
- Score the pronunciation goodness of each phoneme within the words.
66
-
67
- Score range: 0-2
68
- * 2: pronunciation is correct
69
- * 1: pronunciation is right but has a heavy accent
70
- * 0: pronunciation is incorrect or missed
71
-
72
- For the phones with an accuracy score lower than 0.5, an extra "mispronunciations" indicates which is the most likely phoneme that the current phone was actually pronounced.
73
- An example:
74
-
75
- ```json
76
- {
77
- "text": "LISA",
78
- "accuracy": 5,
79
- "phones": ["L", "IY1", "S", "AH0"],
80
- "phones-accuracy": [0.4, 2, 2, 1.2],
81
- "mispronunciations": [
82
- {
83
- "canonical-phone": "L",
84
- "index": 0,
85
- "pronounced-phone": "D"
86
- }
87
- ],
88
- "stress": 10,
89
- "total": 6
90
- }
91
- ```
92
-
93
- ### Word level
94
- Score the accuracy and stress of each word's pronunciation.
95
-
96
- #### Accuracy
97
- Score range: 0 - 10
98
- * 10: The pronunciation of the word is perfect
99
- * 7-9: Most phones in this word are pronounced correctly but have accents
100
- * 4-6: Less than 30% of phones in this word are wrongly pronounced
101
- * 2-3: More than 30% of phones in this word are wrongly pronounced. In another case, the word is mispronounced as some other word. For example, the student mispronounced the word "bag" as "bike"
102
- * 1: The pronunciation is hard to distinguish
103
- * 0: no voice
104
-
105
- #### Stress
106
- Score range: {5, 10}
107
- * 10: The stress is correct, or this is a mono-syllable word
108
- * 5: The stress is wrong
109
-
110
  ### Sentence level
111
  Score the accuracy, fluency, completeness and prosodic at the sentence level.
112
 
@@ -137,68 +89,51 @@ Score range: 0 - 10
137
  * 3-4: Unstable speech speed, speak too fast or too slow, without the sense of rhythm
138
  * 0-2: Poor intonation and lots of stammering and pauses, unable to read a complete sentence
139
 
140
- ## The original scores
 
141
 
142
- The all five experts' original scores could be found in `resource/raw-scores.json`. An example item:
143
- ```
144
- {
145
- "000010011": {
146
-
147
- "text": "WE CALL IT BEAR",
148
- "accuracy": [7.0, 9.0, 8.0, 8.0, 9.0],
149
- "completeness": [1.0, 1.0, 1.0, 1.0, 1.0],
150
- "fluency": [10.0, 9.0, 8.0, 8.0, 10.0],
151
- "prosodic": [10.0, 9.0, 7.0, 8.0, 9.0],
152
- "total": [7.6, 9.0, 7.9, 8.0, 9.1],
153
- "words": [
154
- {
155
- "accuracy": [10.0, 10.0, 10.0, 10.0, 10.0],
156
- "stress": [10.0, 10.0, 10.0, 10.0, 10.0],
157
- "total": [10.0, 10.0, 10.0, 10.0, 10.0],
158
- "text": "WE",
159
- "ref-phones": "W IY0",
160
- "phones": ["W IY0", "W IY0", "W IY0", "W IY0", "W IY0"]
161
- },
162
- {
163
- "accuracy": [10.0, 8.0, 10.0, 10.0, 8.0],
164
- "stress": [10.0, 10.0, 10.0, 10.0, 10.0],
165
- "total": [10.0, 8.4, 10.0, 10.0, 8.4],
166
- "text": "CALL",
167
- "ref-phones": "K AO0 L",
168
- "phones": ["K AO0 L", "K {AO0} L", "K AO0 L", "K AO0 L", "K AO0 {L}"],
169
- },
170
- {
171
- "accuracy": [10.0, 10.0, 10.0, 10.0, 10.0],
172
- "stress": [10.0, 10.0, 10.0, 10.0, 10.0],
173
- "total": [10.0, 10.0, 10.0, 10.0, 10.0],
174
- "text": "IT",
175
- "ref-phones": "IH0 T",
176
- "phones": ["IH0 T", "IH0 T", "IH0 T", "IH0 T", "IH0 T"]
177
- },
178
- {
179
- "accuracy": [3.0, 7.0, 10.0, 2.0, 6.0],
180
- "stress": [10.0, 10.0, 10.0, 10.0, 10.0],
181
- "phones": ["B (EH0) (R)", "B {EH0} {R}", "B EH0 R", "B (EH0) (R)", "B EH0 [L] R"],
182
- "total": [4.4, 7.6, 10.0, 3.6, 6.8],
183
- "text": "BEAR",
184
- "ref-phones": "B EH0 R"
185
- }
186
- ],
187
- },
188
- ...
189
- }
190
- ```
191
 
192
- In this file, the phoneme-level scores are notated in the following convenient notation:
 
 
 
193
 
194
- * for score 2, do not use any symbol
195
- * for score 1, use "{}" symbol
196
- * for score 0, use "()" symbol
197
- * for the inserted phone, use the "[]" symbol
198
 
 
 
 
 
 
 
 
199
 
200
- For example, "B (EH) R" means the score of EH is 0 while the scores of B and R are both 2,
201
- "B EH [L] R" mean there is an unexpected phone "L" and the other phones are scored 2.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
202
 
203
  ## Citation
204
  Please cite our paper if you find this work useful:
 
12
  ---
13
  # speechocean762: A non-native English corpus for pronunciation scoring task
14
 
15
+ ## Introduction
16
+ Pronunciation scoring is a crucial technology in computer-assisted language learning (CALL) systems. The pronunciation quality scores might be given at phoneme-level, word-level, and sentence-level for a typical pronunciation scoring task.
17
+
18
+ This corpus aims to provide a free public dataset for the pronunciation scoring task.
19
+ Key features:
20
+ * It is available for free download for both commercial and non-commercial purposes.
21
+ * The speaker variety encompasses young children and adults.
22
+ * The manual annotations are in multiple aspects at sentence-level, word-level and phoneme-level.
23
+
24
+ This corpus consists of 5000 English sentences. All the speakers are non-native, and their mother tongue is Mandarin. Half of the speakers are Children, and the others are adults. The information of age and gender are provided.
25
+
26
+ Five experts made the scores. To avoid subjective bias, each expert scores independently under the same metric.
27
+
28
+ ## Uses
29
  ```bash
30
  pip install datasets
31
  pip install librosa
32
  pip install soundfile
33
  ```
34
 
 
35
  ```python
36
  >>> from datasets import load_dataset
37
 
 
56
  'total': 9}
57
  ```
58
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
59
  ## The scoring metric
60
  The experts score at three levels: phoneme-level, word-level, and sentence-level.
61
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
62
  ### Sentence level
63
  Score the accuracy, fluency, completeness and prosodic at the sentence level.
64
 
 
89
  * 3-4: Unstable speech speed, speak too fast or too slow, without the sense of rhythm
90
  * 0-2: Poor intonation and lots of stammering and pauses, unable to read a complete sentence
91
 
92
+ ### Word level
93
+ Score the accuracy and stress of each word's pronunciation.
94
 
95
+ #### Accuracy
96
+ Score range: 0 - 10
97
+ * 10: The pronunciation of the word is perfect
98
+ * 7-9: Most phones in this word are pronounced correctly but have accents
99
+ * 4-6: Less than 30% of phones in this word are wrongly pronounced
100
+ * 2-3: More than 30% of phones in this word are wrongly pronounced. In another case, the word is mispronounced as some other word. For example, the student mispronounced the word "bag" as "bike"
101
+ * 1: The pronunciation is hard to distinguish
102
+ * 0: no voice
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
103
 
104
+ #### Stress
105
+ Score range: {5, 10}
106
+ * 10: The stress is correct, or this is a mono-syllable word
107
+ * 5: The stress is wrong
108
 
109
+ ### Phoneme level
110
+ Score the pronunciation goodness of each phoneme within the words.
 
 
111
 
112
+ Score range: 0-2
113
+ * 2: pronunciation is correct
114
+ * 1: pronunciation is right but has a heavy accent
115
+ * 0: pronunciation is incorrect or missed
116
+
117
+ For the phones with an accuracy score lower than 0.5, an extra "mispronunciations" indicates which is the most likely phoneme that the current phone was actually pronounced.
118
+ An example:
119
 
120
+ ```json
121
+ {
122
+ "text": "LISA",
123
+ "accuracy": 5,
124
+ "phones": ["L", "IY1", "S", "AH0"],
125
+ "phones-accuracy": [0.4, 2, 2, 1.2],
126
+ "mispronunciations": [
127
+ {
128
+ "canonical-phone": "L",
129
+ "index": 0,
130
+ "pronounced-phone": "D"
131
+ }
132
+ ],
133
+ "stress": 10,
134
+ "total": 6
135
+ }
136
+ ```
137
 
138
  ## Citation
139
  Please cite our paper if you find this work useful: