TLT-school

TLT-school是一个由意大利特伦托布鲁诺凯斯勒基金会创建的语音数据集，旨在评估学习英语和德语的非本土儿童学生的表现。该数据集包含2017年和2018年从9至16岁学生收集的语音记录，涵盖小学、中学和高中。所有语音记录均由人类专家根据预定义的熟练度指标进行评分。此外，2017年收集的大部分语音记录已仔细手动转录。数据集的部分内容将免费提供给对非本土语音识别和第二语言熟练度自动评估特别感兴趣的科学社区。该数据集适用于研究非本土儿童语音识别、语音评分自动化以及第二语言学习行为分析等领域。

TLT-school is a speech dataset developed by the Bruno Kessler Foundation of Trento, Italy, aiming to evaluate the performance of non-native child students learning English and German. This dataset includes speech recordings collected from students aged 9 to 16 across primary, secondary, and high school levels in 2017 and 2018. All speech recordings were scored by human experts based on predefined proficiency metrics. Furthermore, most of the speech recordings collected in 2017 have been meticulously manually transcribed. Portions of this dataset will be freely made available to the scientific community with a particular interest in non-native speech recognition and automatic assessment of second language proficiency. This dataset is suitable for research areas such as non-native child speech recognition, automated speech scoring, and behavioral analysis of second language learning.