【 FPGA 】FIR 滤波器之 Interpolated Filter 的系数数据（Filter Coefficient Data）

Interpolated Filter 

A previous section explained that an IFIR filter is similar to a conventional FIR, but with the unit delay operator replaced by k-1 units of delay. k is referred to as the zero-packing factor. One way to realize this substitution is by the insertion of k-1 zeros between the coefficient values of a prototype filter. When specifying an IFIR architecture, the full set of prototype coefficients is supplied in the coefficient file, without the zeros implied by the zero-packing factor. The zero-packing factor is defined through the filter user interface. For example, consider the filter coefficient data in the .coe file shown in Figure 3-48.

前一节解释了IFIR滤波器类似于传统的FIR，但单位延迟算子被k-1个延迟单位取代。 k被称为零填充因子。 实现这种替换的一种方法是在原型滤波器的系数值之间插入k-1个零。 在指定IFIR架构时，系数文件中会提供完整的原型系数集，零填充因子不会隐含零。 零填充因子通过过滤器用户界面定义。 例如，考虑.coe文件中的滤波器系数数据，如图3-48所示。

If a zero-packing factor of k=2 is specified, the equivalent filter impulse response is shown in Figure 3-49.

如果指定零填充因子k = 2，则等效滤波器脉冲响应如图3-49所示。

If the zero-packing factor is changed to k=3, the impulse response is as shown in Figure 3-50.

如果零填充因子变为k = 3，则脉冲响应如图3-50所示。

These examples use a symmetrical prototype impulse response; this is not a restriction of the filter core. The prototype filter coefficient set can be symmetrical, non-symmetrical, or negative-symmetric.

这些例子使用对称的原型脉冲响应; 这不是滤波器IP核的限制。 原型滤波器系数组可以是对称的，非对称的或负对称的。