基于神经网络的声乐演奏评价系统研究及其在钢琴教学中的应用

发布时间：2018-01-14 11:43

本文关键词：基于神经网络的声乐演奏评价系统研究及其在钢琴教学中的应用　出处：《贵州大学》2008年硕士论文　论文类型：学位论文

【摘要】： 随着近年来钢琴教学的兴起,大量的人员加入到了学习钢弹奏的队伍。但昂贵的钢琴教学费用以及它特有的教师和学生一对一的教学模式造成了钢琴教育资源非常的紧缺,学习钢琴演奏成为了一项奢侈的活动。于是采用电脑多媒体软件进行钢琴教学就成为了缓解该矛盾的一种可行的方式。本文讨论了钢琴教学软件实现方法,对电脑钢琴教学中的难点(即电脑教学是单向的知识传授而缺少交互的环节)提出了采用神经网络模型对钢琴演奏进行评价的方法,并用来模拟教师指导学生进行弹奏练习。对于一首音乐的演奏,影响其效果的因素是多种多样的,对其评价的指标也有很多种,如节奏感、表现力、乐感、风格的把握等。采用电脑来模拟这个评价过程,实质上就是要找出影响音乐演奏效果的因素和评价指标之间的数理关系。神经网络是人工智能里通过模拟人类大脑思维方式提出来的一种数学模型,它具有对数据分布要求不严格、非线性的数据处理方法、强鲁棒性和动态性等优点,非常适合作为评价系统的数学模型。另外神经网络也有很强的理论基础,其在各行业中的应用也发展得基本成熟,本文尝试着把神经网络数学模型引入到钢琴演奏的评价系统中来。本论文所完成的主要工作: 1、论文首先说明了神经网络的原理,然后根据音乐理论找出了影响演奏效果的因素,包括音符、节奏、节拍、旋律、调性等。然后对每一种因素进行量化,并作为神经网络的输入参数。 2、根据音乐演奏的评价指标设计神经网络模型,在钢琴教师和学生的帮助下获得了钢琴演奏样本数据,完成了神经网络的训练。 3、介绍了现阶段钢琴教学软件的基本情况,设计了一个钢琴教学软件的框架。然后实现了《欢乐颂》钢琴弹奏练习的功能,并采用神经网络评价模块对教师和学生的弹奏进行了评价,检验模型的性能。钢琴学习的初期功能完备的MIDI键盘和真实的钢琴相比对于演奏者的发挥并无很大的影响,因而弹奏系统采用了MIDI键盘作为获取演奏特征的工具。弹奏软件采用了VC++编写实现,运行于Win32平台上面。在六盘水“音韵琴行”邀请了一位钢琴教师和两位学生进行了弹奏试验。在比对了他们的不同的弹奏结果后得出,系统设计的各项指标均达到了要求。
[Abstract]:With the rise of piano teaching in recent years. A large number of people have joined the team to learn how to play steel. But the high cost of piano teaching and its unique one-to-one teaching mode between teachers and students cause the shortage of piano education resources. Learning to play piano has become an extravagant activity. Therefore, the use of computer multimedia software for piano teaching has become a feasible way to alleviate this contradiction. This paper discusses the implementation method of piano teaching software. This paper puts forward a method of evaluating piano playing by using neural network model, which is difficult in computer piano teaching (that is, computer teaching is one-way knowledge imparting but lack of interactive link). And used to simulate the teacher to guide students to play. For the performance of a music, the factors that affect its effect are various, and there are many kinds of evaluation indicators, such as rhythm, expressiveness, musical sense. Use a computer to simulate the evaluation process. In essence, it is necessary to find out the mathematical relationship between the factors that affect the effect of music playing and the evaluation index. Neural network is a mathematical model proposed by simulating the way of thinking of human brain in artificial intelligence. It has the advantages of not strict data distribution, nonlinear data processing, strong robustness and dynamic, etc. It is very suitable for the mathematical model of evaluation system. In addition, the neural network also has a strong theoretical basis. Its application in various industries has also developed maturely. This paper attempts to introduce the neural network mathematical model into the evaluation system of piano performance. The main work of this thesis is as follows: 1. Firstly, the paper explains the principle of neural network, and then finds out the factors that affect the performance according to the music theory, including notes, rhythm, rhythm, melody, tonality and so on. Then it quantifies each factor. And as the input parameters of the neural network. 2. According to the evaluation index of music performance, the neural network model is designed. With the help of piano teachers and students, the sample data of piano performance are obtained, and the training of neural network is completed. 3. This paper introduces the basic situation of piano teaching software at present, designs a framework of piano teaching software, and then realizes the function of piano playing practice. The neural network evaluation module is used to evaluate the performance of the model. In the early stage of piano learning, the MIDI keyboard has no great influence on the player's performance compared with the real piano. Therefore, the MIDI keyboard is used as the tool to obtain the performance features in the playing system, and the software is implemented by VC programming. Run on the Win32 platform. In Liupanshui, a piano teacher and two students were invited to play. After comparing their different results. All the indexes of the system design have met the requirements.
【学位授予单位】：贵州大学
【学位级别】：硕士
【学位授予年份】：2008
【分类号】：J624.1

【参考文献】