请教一个FIR滤波器的设计问题
时间:12-12
整理:3721RD
点击:
问题描述如下:
Please provide a design that achieves maximum throughput and implements the
following FIR filter:
y(n) = 5*x(n) + 0.3*x(n-1) + 0.25*x(n-2), where x is a 8 bit unsigned signal
[0,255] and y is the output. Please provide the wordlength of the buses and for
the coefficients, as well as a justification for your design choices.
由于本人没有数字滤波器设计经验,抓不住此问题的关键处,求教相关人士这个问题如何
回答比较理想?
多谢
Please provide a design that achieves maximum throughput and implements the
following FIR filter:
y(n) = 5*x(n) + 0.3*x(n-1) + 0.25*x(n-2), where x is a 8 bit unsigned signal
[0,255] and y is the output. Please provide the wordlength of the buses and for
the coefficients, as well as a justification for your design choices.
由于本人没有数字滤波器设计经验,抓不住此问题的关键处,求教相关人士这个问题如何
回答比较理想?
多谢
第一项用 {x(n), 2'b00} + x(n)算
中间项的结果用最接近(0.3 * 2^k)的整数来乘x(n-1),得到结果舍去低k位,k越大误差越小
最后一项右移2位
乘法根据需求可以有单拍或多拍实现 //FPGA的话可以用自带的乘法单元?
我也没做过滤波器,也不知道我这个思路可不可行。。呼唤大牛拍砖
最后一项是右移2位吧
另外最后输出y的位数应该怎么定啊,如果按照题目中max output through来定的话,y的位宽应该是8位吧。
我也不太懂,瞎猜的