A 50-hour Children Bilingual Speech Recognition Database
A 50-hour Children Bilingual Corpus for Recognition has been released. It covered 140 children with the age from 5 to 12 with 1:1 gender distribution. The corpus contains 25-hour mandarin and 25-hour English (including word’s pronunciation and alphablocks ), which covered the texts and words in elementary school textbooks.
It is recorded in real environments using 16 kHz sampling frequency and 16-bit quantization accuracy mono PCM format.
The corpus is collected using various Android smart phone models and primarily in indoor (quiet and anechoic ) environments.
During the corpus design and collection process, we have formed an efficient team with high standard capabilities. This leads to the great success for our corpus to have below 2% sentence-wise error, which is dominating in the current market.
This corpus can be used for training and testing the children speech recognition system, as well as speech analysis. It has been well-acknowledged by industry as a corpus with high speech quality and recognition accuracy.