几个基本公式
基本信号的傅里叶变换
以下是冲击信号、直流信号、虚指数信号的傅里叶变换
\[\mathcal{F}(\delta(t)) = 1 \\ \mathcal{F}(1) = 2\pi\delta(\omega) \\ \mathcal{F}(\delta(t-T)) = exp(-j\omega T) \\ \mathcal{F}(exp(jw_0t)) = 2\pi\delta(w-w_0) \]冲击信号作用
相乘:
\[F(t) = f(t) \delta(t) = f(0) \delta(t)\\ \int_{-\infty}^{\infty}F(t)dt = f(0) \\ f(t)\delta(t-T) = f(T) \]卷积:
\[F(t) = f(t)*\delta(t) = \int_{-\infty}^{\infty}f(t-\tau)\delta(\tau)d{\tau} = \\ f(t)\int_{-\infty}^{\infty}\delta(\tau)d\tau = f(t) \\ f(t)*\delta(t-T) = f(t-T) \]信号采样
\[S(t) = \sum_{-\infty}^{\infty}\delta(t-nT_s) \\ x_s(t) = x(t)S(t) = \sum_{-\infty}^{\infty}[x_s(nT_s)\delta(t-nT_s)] \]信号周期延拓
\[P(t) = \sum_{-\infty}^{\infty}\delta(t-nT_0) \\ \widetilde{x}(t) = x(t)*P(t) = \sum_{-\infty}^{\infty}[x(t)*\delta(t-nT_0)] = \\ \sum_{-\infty}^{\infty}[x(t-nT_0)] \]信号采样的频域情况
由傅里叶变换性质 时域乘积等于频域卷积,我们考虑采样信号的傅里叶变换:
\[\mathcal{F}(S(t)) = \mathcal{F}(F_s[\widetilde{S}(t)]) \]其中\(F_s\)表示周期信号的傅里叶级数,对于\(\widetilde{S}(t)\) 而言,其傅里叶级数的各项系数是:
\[C_n=\frac{1}{T_s}\int_{-\frac{T_s}{2}}^{\frac{T_s}{2}}[\widetilde{S}(t)exp(-jn\omega_st)]dt =\\ \frac{1}{T_s}\int_{-\frac{T_s}{2}}^{\frac{T_s}{2}}[\sum\delta(t-kT_s)exp(-jn\omega_st)]dt \]因为积分周期\([-\frac{T_s}{2}, \frac{T_s}{2}]\) 内只有一个冲击信号\(\delta(t)\),所以上公式可以写为:
\[C_n = \frac{1}{T_s}\int_{-\frac{T_s}{2}}^{\frac{T_s}{2}}[\delta(t)exp(-jn\omega_st)]dt = \frac{1}{T_s} \]因此信号\(S(t)\)就可以写作:
\[S(t) = \sum_{n=-\infty}^{\infty}[C_nexp(jnw_st)] = \frac{1}{T_s}\sum_{n=-\infty}^{\infty}exp(jnw_st) \]现在对\(S(t)\)做傅里叶变换:
\[\mathcal{F}(\omega) = \frac{1}{T_s}\int_{-\infty}^{\infty}[\sum_{n=-\infty}^{\infty}exp(jnw_st)]exp(-jwt)dt = \\ \frac{1}{T_s}\sum_{n=-\infty}^{\infty}[\int_{-\infty}^{\infty} e^{-j[w-nw_s]t}dt] = \frac{1}{T_s}\sum\delta(w-nw_s) \]或者,根据\(\mathcal{F}\)的线性性质,也可以这么来看:
\[\mathcal{F}(S(t))= \frac{1}{T_s}\sum\mathcal{F}[exp(jnw_st)] = \frac{2\pi}{T_s}\sum\delta(w-nw_s) \]设原信号的傅里叶变换是\(x(w)\);现在,再根据傅里叶变换的性质,可以得出:
\[x_s(t) = x(t)S(t) \\ \mathcal{F}[x_s(t)] = \mathcal{F}[x(t)]*\mathcal{F}[S(t)] =\\ K\mathcal{F}[x(t)] * \sum\delta(w-w_s) = K\sum x(w-nw_s) \]上式中,K表示缩放系数。
这个结果表明了,时域的采样等于频域的周期化。这里有两个重要的结论:
- 对于非带限信号,其频域周期化后一定存在混叠,故而不可能恢复
- 对于带限信号,要避免混叠, 必须要满足\(-w_c + w_s > w_c\) 即 \(w_s>2w_c\),采样频率必须是信号带宽的2倍