Dataset Viewer
Auto-converted to Parquet Duplicate
Audio Sample ID
int64
LAB ID
int64
성별
int64
연령(세;개월)
string
자음정확도
string
집단
int64
stimuli types
string
audio
audio
1
201
1
6;3
100
1
APAC_word
2
271
2
5;4
100
1
APAC_word
3
272
2
4;9
48.6
2
APAC_word
4
275
1
3;4
70
2
APAC_word
5
278
1
4;10
87.1
1
APAC_word
6
308
1
7;9
100
1
APAC_word
7
309
1
9;0
98.6
1
APAC_word
8
310
2
8;5
97.1
1
APAC_word
9
314
2
9;4
100
1
APAC_word
10
315
1
8;5
98.6
1
APAC_word
11
317
2
8;4
92.9
2
APAC_word
12
318
2
8;7
98.6
1
APAC_word
13
321
2
8;3
97.1
1
APAC_word
14
322
2
9;1
100
1
APAC_word
15
323
1
7;3
97.1
1
APAC_word
16
324
2
5;10
100
1
APAC_word
17
327
2
9;11
98.6
1
APAC_word
18
328
2
7;11
90
2
APAC_word
19
329
1
9;10
91.4
2
APAC_word
20
357
2
7;3
95.7
1
APAC_word
21
389
2
3;0
74.3
1
APAC_word
22
391
1
2;8
70
1
APAC_word
23
392
1
7;3
72.9
1
APAC_word
24
396
2
6;10
87.1
2
APAC_word
25
400
1
6;9
90
1
APAC_word
26
406
1
6;2
94.3
2
APAC_word
27
413
2
5;7
91.4
2
APAC_word
28
416
1
7;2
88.6
2
APAC_word
29
427
1
7;9
100
1
APAC_word
30
435
2
4;9
87.1
1
APAC_word
31
440
1
2;11
92.9
1
APAC_word
32
454
1
7;2
94.3
1
APAC_word
33
467
1
6;9
100
1
APAC_word
34
524
1
7;5
100
1
APAC_word
35
571
1
8;9
100
1
APAC_word
36
594
2
6;8
98.6
1
APAC_word
37
621
1
3;8
91.4
1
APAC_word
38
623
1
3;6
82.9
1
APAC_word
39
624
1
3;9
82.9
1
APAC_word
40
625
2
3;9
94.3
1
APAC_word
41
626
2
3;6
100
1
APAC_word
42
627
2
3;6
94.3
1
APAC_word
43
629
2
3;11
100
1
APAC_word
44
645
1
5;0
65.7
2
APAC_word
45
651
1
4;6
70
2
APAC_word
46
654
1
4;2
67.1
2
APAC_word
47
657
2
6;8
55.7
2
APAC_word
48
659
1
4;1
78.6
2
APAC_word
49
660
1
4;8
51.4
2
APAC_word
50
317
2
8;4
92.9
2
KAPP_word1
51
317
2
8;4
92.9
2
KAPP_word2
52
317
2
8;4
92.9
2
KAPP_nonword
53
328
2
7;11
90
2
KAPP_word1
54
328
2
7;11
90
2
KAPP_word2
55
328
2
7;11
90
2
KAPP_nonword
56
372
2
3;7
71.4
2
KAPP_word1
57
372
2
3;7
71.4
2
KAPP_word2
58
372
2
3;7
71.4
2
KAPP_nonword
59
406
1
6;2
94.3
2
KAPP_word1
60
406
1
6;2
94.3
2
KAPP_word2
61
406
1
6;2
94.3
2
KAPP_nonword
62
411
1
6;4
91.4
2
KAPP_word1
63
411
1
6;4
91.4
2
KAPP_word2
64
412
2
6;8
92.9
2
KAPP_word1
65
413
2
5;7
91.4
2
KAPP_word1
66
413
2
5;7
91.4
2
KAPP_word2
67
416
1
7;2
88.6
2
KAPP_word1
68
418
1
7;2
64.3
2
KAPP_word2
69
418
1
5;10
64.3
2
KAPP_nonword
70
419
1
7;10
82.9
2
KAPP_word1
71
419
1
7;10
82.9
2
KAPP_word2
72
419
1
7;10
82.9
2
KAPP_nonword
73
421
1
5;4
80
2
KAPP_word1
74
421
1
5;4
80
2
KAPP_word2
75
421
1
5;4
80
2
KAPP_nonword
76
459
2
6;5
65.7
2
KAPP_word1
77
459
2
6;5
65.7
2
KAPP_word2
78
459
2
6;5
65.7
2
KAPP_nonword
79
505
1
7;10
87.1
2
KAPP_word1
80
505
1
7;10
87.1
2
KAPP_word2
81
505
1
7;10
87.1
2
KAPP_nonword
82
522
2
6;3
88.6
2
KAPP_word1
83
522
2
6;3
88.6
2
KAPP_nonword
84
576
2
9;4
92.9
2
KAPP_nonword
85
579
2
9;0
91.4
2
KAPP_nonword
86
635
1
5;1
87.1
2
KAPP_word1
87
635
1
5;1
87.1
2
KAPP_word2
88
635
1
5;1
87.1
2
KAPP_nonword
89
637
1
5;4
71.4
2
KAPP_word2
90
638
1
5;5
80
2
KAPP_word1
91
638
1
5;5
80
2
KAPP_word2
92
638
1
5;5
80
2
KAPP_nonword
93
639
1
7;1
77.1
2
KAPP_word1
94
641
1
5;6
77.1
2
KAPP_word1
95
641
1
5;6
77.1
2
KAPP_word2
96
641
1
5;6
77.1
2
KAPP_nonword
97
642
1
6;5
74.3
2
KAPP_word1
98
643
1
3;11
78.6
2
KAPP_word1
99
645
1
5;0
65.7
2
KAPP_word1
100
645
1
5;0
65.7
2
KAPP_word2
End of preview. Expand in Data Studio

Speech Dataset for AI-Based Language Assessment in Children

The "Speech Database of Typically Developing and Speech-Impaired Children" is an open speech dataset designed to support the development of AI-based language assessment systems. It contains speech samples from children aged 2 to 9 who are either typically developing or have reduced consonant articulation accuracy.

This dataset is based on standardized Korean articulation tools:

  • APAC (Articulation and Phonology Assessment for Children)
  • K-APP (Korean Articulation and Phonology Profile)

Only words and sentences spoken by children were included.

The dataset is distributed under the CC BY NC ND 4.0 license and can be used by clinicians and researchers for non-commercial, public-interest purposes.

This dataset contains audio samples and their corresponding metadata.
Audio files are located in the audio/ directory.
Metadata is stored in metadata.csv, where the file_name column links to each audio file.

Dataset Overview

  • Total Data Duration: Over 300 minutes of speech recordings
  • Number of Speakers: 385 children
  • Sampling Rate: 16kHz

Dataset Composition

  • The dataset consists of meaningful and non-meaningful words and sentences spoken by children. It includes metadata on each child’s age, consonant articulation accuracy, and group classification (typical or speech-impaired).

Dataset Fields

  • Audio Sample ID: Unique identifier for each audio sample.
  • LAB ID: Unique identifier assigned during utterance collection.
  • 성별: Gender of the speaker (1 = Male, 2 = Female).
  • 연령: Age in months.
  • 자음정확도: Consonant articulation accuracy.
  • 집단: Group classification (1 = Typical, 2 = Speech Sound Disorder).
  • Stimuli types: Type of stimulus used in the assessment (e.g., APAC, K-APP) and form (word, nonword, sentence).
  • File name: Name of the audio file.

License

This dataset follows the CC BY NC ND 4.0 license. It can be shared and used in its original form for non-commercial purposes if the source is credited. Users must notify the creator about the intended use and scope before utilizing the dataset.

The uploader did not participate in the dataset's creation. This dataset is provided to help broaden access to the original research.

This dataset is a result of the Glocal University 30 Project, supported by the Ministry of Education and the National Research Foundation of Korea.
Project ID: GLOCAL-202407990001
Contact: Seung-Hee Ha (shha@hallym.ac.kr)


아동 음성 언어평가용 AI 학습 데이터셋

소개

「정상 또는 병리적인 아동 음성 데이터 베이스」셋트는 AI 기반 언어 평가 시스템 개발을 위해 제작된 공개 음성데이터셋입니다. 이 데이터는 만 2세부터 9세까지의 정상 또는 말소리 정확도가 떨어지는 아동을 대상으로 수집되었습니다.

데이터는 한국어 표준화 발음 평가 도구인

  • 아동용 발음평가(APAC)
  • 한국 조음음운프로파일(K-APP)
    에 기반한 낱말과 문장 발화로 구성되어 있습니다.

데이터셋은 CC BY NC ND 4.0 라이선스 하에 배포되며, 공익적인 목적을 위해 임상가 및 연구자들이 비영리적으로 사용할 수 있습니다.

이 데이터셋은 오디오 샘플과 이에 상응하는 메타데이터로 구성되어 있습니다.
오디오 파일은 audio/ 디렉토리에 있으며,
메타데이터는 metadata.csv에 저장되어 있고 file_name 컬럼을 통해 각 오디오 파일과 연결됩니다.

데이터 개요

  • 총 데이터량: 300분 이상의 발화
  • 화자 수: 385명
  • 샘플링 레이트: 16kHz

데이터 구성

  • 아동이 발화한 의미/무의미 낱말 및 문장으로 구성되며, 각 아동의 나이, 자음정확도, 말소리 장애 여부에 대한 메타데이터를 포함합니다.

컬럼 정보

  • Audio Sample ID: 각 오디오 샘플의 고유 식별자
  • LAB ID: 발화 수집 당시 부여된 고유 식별자
  • 성별: 1 = 남성, 2 = 여성
  • 연령: 나이 (개월 단위)
  • 자음정확도: 자음 정확도 (Consonant accuracy)
  • 집단: 1 = 일반, 2 = 말소리장애
  • Stimuli types: 검사어 도구(APAC, K-APP)_유형(word, nonword, sentence)
  • File name: 오디오 파일 이름

라이선스

이 데이터셋은 CC BY NC ND 4.0 라이선스를 따릅니다. 출처를 명시할 경우, 비영리 목적에 한해 원형 그대로 사용할 수 있습니다. 데이터셋 사용 전에는 반드시 사용 의도와 범위를 저작자에게 알려야 합니다.

이 데이터셋은 원 데이터셋의 접근성과 활용도를 높이기 위한 목적으로 제공되며, 업로더는 데이터 제작에 직접 참여하지 않았습니다.

본 데이터셋은 교육부와 한국연구재단의 재원으로 수행된 ‘글로컬대학 30’ 연구 결과입니다.
과제번호: GLOCAL-202407990001
연락처: 하승희 (shha@hallym.ac.kr)

Downloads last month
16