数字信号处理过程中信号截位误差抑制方法研究  

Study on Restraining Signal truncation Error in DSP

作　　者：郭连平[1] 田书林[1] 王志刚[1] 

机构地区：[1]电子科技大学自动化工程学院,成都611731

出　　处：《信号处理》2013年第5期544-549,548-549,共6页Journal of Signal Processing

基　　金：国家自然科学基金(60827001);高等学校博士学科点专项科研基金资助项目(20120185130002)

摘　　要：在数字信号处理过程中,受一些数字算法的计算精度影响,需要对运算过程中的中间信号实现从高量化精度到低量化精度的转换,即数据位宽的截断。直接对高量化精度信号进行截位会导致截位得到的低量化精度信号无杂散动态范围(SFDR)的下降。在高量化精度信号中加入Dither后再截位可以有效降低直接截位误差中的谐波失真,明显改善低量化精度信号的动态范围。MATLAB仿真结果说明了加入Dither后截位可有效增加低量化精度信号的SFDR,增加量可达到16dB。In DSP, restricted by the calculating precision of some digital algorithms, the digital signal transition from high quantizing precision to low quantizing precision was necessary, i.e. the truncating of bit width of signal. The direct truncating operation on signal of high quantizing precision led to the decline of the spurious free dynamic range （SFDR） of truncated signal. Truncating after adding Dither signal to the signal of high quantizing precision could decline the harmonic distortion of the truncation error efficiently and extend the dynamic range of the truncated signal significantly. The simula- tion results of MATLAB shown that truncating after adding Dither signal could extend the SFDR of the truncated signal by about 16dB.

关 键 词：精度转换 截位误差 Dither信号 无杂散动态范围 

分 类 号：TN911.6[电子电信—通信与信息系统]