ggfox00000/stt-vibravox-fr-test

VibraVox FR测试分割数据集是一个用于法语自动语音识别（ASR）和多传感器音频处理的数据集。它包含两种配置：speech_clean（在安静环境中录制的语音）和speech_noisy（在嘈杂环境中录制的语音）。每个样本包含6个同步的音频通道，分别来自不同类型的麦克风和传感器，如头戴式麦克风、额头加速度计、软/硬耳内麦克风、太阳穴振动传感器和喉部麦克风。数据集还提供了说话者的性别、ID、句子ID、持续时间、原始文本、标准化文本和音素化文本等信息。该数据集由CNAM-LMSSC实验室创建，遵循CC-BY-4.0许可。

The VibraVox FR test split dataset is designed for French automatic speech recognition (ASR) and multi-sensor audio processing. It includes two configurations: speech_clean (recorded in a quiet studio environment) and speech_noisy (recorded in a noisy environment). Each sample contains six synchronized audio channels from different types of microphones and sensors, such as a headset microphone, forehead accelerometer, soft/rigid in-ear microphones, temple vibration pickup, and throat microphone. The dataset also provides information on the speakers gender, ID, sentence ID, duration, raw text, normalized text, and phonemized text. Created by the CNAM-LMSSC lab, the dataset is licensed under CC-BY-4.0.