A Multi-Stage Automatic Evaluation System for Sight-Singing
A Multi-Stage Automatic Evaluation System for Sight-Singing
复制标题
DOI：
10.1109/tmm.2022.3168132
复制
发表时间：
2023
影响因子：
7.3
通讯作者：
Weiming Yang;Xianke Wang;Bowen Tian;Wei Xu;W. Cheng
中科院分区：
文献类型：
--
作者：
Weiming Yang;Xianke Wang;Bowen Tian;Wei Xu;W. Cheng
视唱练习是音乐教育的基本组成部分。本文提出了一个客观、完整的视唱自动评价系统，该系统分为两个关键阶段：音符转写和音符对齐。在第一阶段，我们使用基于卷积递归神经网络（CRNN）的起始检测器进行音符分割，并使用（Kim et al. 2018）中描述的音高提取器进行音符标记。在第二阶段，提出了一种基于相对基音模型的对准算法。由于缺乏视唱音符对齐和整体系统评估的数据集，我们构建了视唱声乐数据集（SSVD）。系统的每个模块和整个系统都在这个数据集上进行了测试。起始检测器实现了90.61%的F-测量，并且音符转录和音符对齐阶段分别实现了88.42%和94.79%的F-测量。此外，我们还提出了一个客观标准的视唱评价系统。基于这个标准，我们的自动视唱系统在SSVD数据集上实现了77.95%的F测量。
Sight-singing exercises are a fundamental part of music education. In this paper, we present an objective and complete automatic evaluation system for sight-singing, which has two critical stages: note transcription and note alignment. In the first stage, we use an onset detector based on the convolutional recurrent neural network (CRNN) for note segmentation and the pitch extractor described in (Kim et al. 2018) for note labeling. In the second stage, an alignment algorithm based on relative pitch modeling is proposed. Due to the lack of datasets for sight-singing note alignment and the overall system evaluation, we construct the sight-singing vocal dataset (SSVD). Each module of the system and the entire system are tested on this dataset. The onset detector achieves an F-measure of 90.61%, and the stages of note transcription and note alignment achieve an F-measure of 88.42% and 94.79%, respectively. In addition, we propose an objective criterion for the sight-singing evaluation system. Based on this criterion, our automatic sight-singing system achieves an F-measure of 77.95% on the SSVD dataset.