一种基于IFS分形算法和分解和弦的耳鸣康复音合成新方法

摘要

耳鸣是在无外界声源刺激下的一种主观感知,在临床上十分常见.习服治疗是目前公认的一种有效的耳鸣治疗方法.研究表明,具有相似却又不重复的分形音乐可以作为耳鸣习服治疗的治疗声源.本文提出了一种基于IFS(Iterated Function System)分形算法和分解和弦的耳鸣康复音合成新方法,该方法将分形图形处理技术运用到耳鸣康复音的合成中,使得生成的音符整体在可控范围内随机出现,并且相邻音符在可调节的范围内波动,可有效的避免音调突变现象.此方法保留了图像信号中的分形特征,实现了可控范围内的确定性与可预测性的统一;在加入了分解和弦后,音乐更加连贯.试听结果表明,该耳鸣康复音具有相似不重复、流畅且突变少的特点,对耳鸣治疗有一定的参考价值.%Tinnitus is a sensation of sound without external stimulation and it is common on clinical.Tinnitus retraining therapy (TRT) is a recognized treatment of relieving tinnitus.Scientific researches show that some similar and unrepeatable music can be used as the sound of TRT.This paper proposed a new method of generating fractal tones based on IFS and broken chord.Applying this new method in the synthesizing could make all the notes within the scale and appeared randomly.Furthermore,the adjustability of neighboring notes avoided the tone mutation.By using this method,the tinnitus rehabilitation sound achieved a balance of predictability and randomness.Experimental results showed that it had obvious fractal characteristics and remained the fractal image features.By adding broken chords,the music turned into more coherent and mildly pleasant songs.So it had reference significance to the treatment of tinnitus.