基于FM合成的多算子反馈模型的乐音重塑

通过FM合成基本原理,建立乐音重塑算子模型。提出基于FM合成的多算子反馈模型算法。对算法应用快速傅里叶变换理论,将算法生成的数据映射到频域上,并进行频谱分析以及和标准A4音比对,从而验证重塑乐音的优劣性。实验讨论了有反馈和无反馈算法频谱的区别,通过贝塞尔函数建立反馈后音色频谱表达式,分析反馈算法在不同时隙上频谱变化对乐音合成的影响。通过调查人们对合成乐音的感受,从客观和主观上验证多算子模型的品质。最后提出一种最优多算子反馈模型算法,该算法具有很好的频谱特性和音乐和谐性,可通过该算子对乐音进行良好的重塑。

Musical Tone Remodeling Based on the Multi-Operator Feedback Model of FM Synthesis

In this paper, based on the FM synthesis principle, the operator model is established to remold the musical tone. The multi-operator feedback model based on the FM synthesis is proposed. By using the Fourier transformation, the al- gorithm production data is mapped to frequency domain, whose spectral construction is analyzed and compared with the standard musical tone A4, thus it confirms the advantage and disadvantage quality of the remolded musical tone. The trials verify the differences in the spectral constructions between feedback and the non-feedback models. The expressions of musi- cal tone spectrum after feedback are built by Bessel function. With the change of time, spectral construction of feedback model has an influence on musical tone synthesis. Through the investigation, in which the people have different feelings of synthesize the musical tone, the quality of the musical tone is confirmed from objective and subjective views. Finally one of the most superior multi-operator feedback models is proposed, which has ascendant spectral construction characteristic and music concordance and remolds the musical tone in high quality.