文章摘要

文章信息/Info

Title:: An iterative function system-based synthesis method of multi-track personalized tinnitus rehabilitation sound

Author(s):: DENG Mingyue; HE Peiyu; FANG Ancheng; College of Electronics and Information Engineering, Sichuan University, Chengdu 610065, China

Keywords:: Tinnitus sound therapy; Iterative function system; Musical instrument digital interface; Natural sound; Frequency matching; Sound synthesis

摘要:: 针对经典疗法在耳鸣声治疗有效性或康复音可听性上存在的不足，本研究基于迭代函数系统（iterative function system, IFS）和乐器数字接口技术（musical instrument digital interface, MIDI），提出一种多音轨个性化耳鸣康复音的合成方法，使合成康复音在保持频率特性的同时提升心理感知特性。从个性化匹配的MIDI音乐中提取音调和节奏组合，在IFS算法映射下合成时长可控分形音乐。以丰富的频率集中自然声、流畅的类白噪声自然声和分形音乐合成多音轨个性化耳鸣康复音，其中频率集中自然声匹配患者耳鸣频率，代替传统掩蔽中听感不佳的窄带噪声。仿真结果表明，客观上本研究方法所合成康复音中的旋律与原曲相似，符合1/f波动和分维特性，康复音输入耳鸣模型能有效抑制对应频段耳鸣；试听结果表明，主观上整体试听效果舒缓自然。本研究方法对耳鸣治疗具有一定参考价值。

Abstract:: Aiming at the deficiencies of classical therapy in the effectiveness of tinnitus sound therapy or euphony of rehabilitation sound，we proposed an IFS-based synthesis method of multi-track personalized tinnitus rehabilitation sound (MPTRS).MPTRS had both the frequency characteristics and the psychological perception characteristics.Synthesizing time-controlled fractal music with the tones and rhythm groups were extracted from the personalized MIDI music based on IFS.MPTRS was constituted by various frequency-concentrated natural sound, fluent white-noise-like natural sound and fractal music, and frequency-concentrated natural sound sounded better than narrowband noise,and matched the frequency of patient tinnitus.The results of simulation showed that the melody of MPTRS was similar to the original one, and it had no sudden change of tone and conformed to 1/f fluctuation and fractal characteristics, that MPTRS could inhibit the tinnitus of corresponding frequency in the tinnitus model objectively, and that MPTRS sounded soothing in audition experiment subjectively. Thus, it has certain referential significance for tinnitus therapy.