audio
audioduration (s) 2.11
6.91
|
|---|
Description
155 Malaysian local speakers participated in the recording and the recoring environment is quiet. The recordings contain various categories like in-car scene, home, speech assistant. 50 sentences for each speaker. The valid time is 7 hours. All texts are manually transcribed with high accuracy.
For more details, please refer to the link: https://www.nexdata.ai/dataset/172?source=Huggingface
Format
16kHz, 16bit, uncompressed wav, mono channel
Recording environment
quiet indoor environment, without echo
Recording content (read speech)
smart car; smart home; speech assistant
Speaker
155 Malays, 66% of which are female
Device
Android mobile phone, iPhone
Language
Malay
Transcription content
text, 4 special identifiers
Accuracy rate
95% (the accuracy rate of identifiers is not included)
Application scenarios
speech recognition, voiceprint recognition
Licensing Information
Commercial License
- Downloads last month
- 0