Research On Timbre Modeling Of Virtual Musical Instrument Based On Emotion

The timbre modeling technology of emotion-based virtual musical instrument in this article belong to the scope of computer music technique, its aim is using the computer to construct a series of digital virtual musical instruments which is convenient to"write"emotional expression of music.The current generation of music is mainly by the way of recording, it requires live performance by musicians,post-processing by software engineers and other procedures to achieve it, this need consume enormous human,material and time cost, while the music generated in this way has a strong uncertainty, and it is difficult to make precise, fine adjustment. Through timbre model, using computer to"write"the music is very convenient, and it is able to make a precise control to what is"written", further more,modify it is also very convenient, fast, this will make a significant savings in resource cost recorded and time cost recorded.Music is an art of emotional expression, to express the emotion that composers want to show clearly is often not an easy thing. But through constructing a series of special timbre model for emotional expression, the difficulty of emotional expression will be lower, the needs of people's emotional expression can be meeted.Division from different angles, the emotional manifested in music can be divided into a number of different types. on the basis of two-dimensional model proposed by Thayer and Juslin's classification theory, the emotion of music is divided into sadness, happiness, anger and tranquil by this article. In the analysis of the existing reality musical instruments and according to the characteristics of emotional correspondence with the musical musical instruments'timbre, the musical instruments which are good at expressing various emotion are found and regarded as sampling of the main study object and the auxiliary, the timbre model of emotion-based computer musical instrument are provided the base.Because the sound of musical instruments is produced by vibration, its timbre depends on the amplitude envelope and spectral characteristics of features. So, the timbre model of emotion-based computer musical instrument are based on these two aspects. Construct amplitude envelope sub-model and spectral characteristics of features sub-model, then synthesize to create the finally model. After construct the model, using the C# language. in .NET framework to built an experimental platform to test effect of the model. Enter parameters of the model in experimental platform, write the music data in memory , then save it as WAV audio files, validate the accuracy of the model by playing test.This group of models not only reduce the costs in music's generation, but also improve its production efficiency, reduce the difficulty of emotional expression. This research has a certain theoretical guidance and practical value to the study of algorithmic composition and emotion-based retrieval of music.