8.3: Spectral Density and Bandwidth

Fluctuations Near Equilibrium · LibreTexts

0%

Contents

    Text

    Size
    100%
    Theme
    Typeface
    Line length
    Spacing

    Read aloud

    How

    Speed: 1×
    Light up
    Highlight

    Symbols: Mulberry Symbols, © 2018–2026 Steve Lee, CC BY-SA 4.0 (licence, the set), the site's own copies.

    From LibreTexts, under CC BY-NC-SA 4.0, as the source states it; its equations are written in TeX, which the PDF shows typeset. This page at LibreTexts.

    8.3: Spectral Density and Bandwidth

    Fourier convention: We adopt the convention

    \[ V(\omega)=\int_{-\infty}^{\infty} d t V(t) e^{i \omega t} \tag{8.8} \label{8.8} \]

    with inverse

    \[ V(t)=\frac{1}{2 \pi} \int_{-\infty}^{\infty} d \omega V(\omega) e^{-i \omega t} . \tag{8.9} \label{8.9} \]

    For any voltage signal \(V(t)\), the voltage noise spectral density \(S_{V}(\omega)\) is defined by

    \[ \langle V(t) V(0)\rangle=\int_{-\infty}^{\infty} \frac{d \omega}{2 \pi} S_{V}(\omega) e^{-i \omega t} . \tag{8.10} \label{8.10} \]

    For real-valued signals, \(S_{V}(\omega)\) is an even function: \(S_{V}(-\omega)=S_{V}(\omega)\).

    Equation \ref{8.11}, below, is the Wiener-Khinchin theorem [18, 19]: the power spectral density \(S_{V}(\omega)\) is the Fourier transform of the autocorrelation function \(C(t)=\langle V(t) V(0)\rangle\), or equivalently

    \[ S_{V}(\omega)=\int_{-\infty}^{\infty}\langle V(t) V(0)\rangle e^{+i \omega t} d t \tag{8.11} \label{8.11} \]

    Setting \(t=0\) in Eq. (4) gives

    \[ \left\langle V^{2}\right\rangle=\int_{-\infty}^{\infty} \frac{d \omega}{2 \pi} S_{V}(\omega), \tag{8.12} \label{8.12} \]

    which is Parseval's theorem: the total mean-square equals the integral of the spectral density over all frequencies [18].

    A voltmeter with a bandpass filter that passes only \(|\omega| \in\left[\omega_{1}, \omega_{2}\right]\) contributes from both the positive-frequency band \(\left[\omega_{1}, \omega_{2}\right]\) and its mirror image \(\left[-\omega_{2},-\omega_{1}\right]\) :

    \[ \left\langle V^{2}\right\rangle_{\mathrm{meas}}=\int_{-\omega_{2}}^{-\omega_{1}} \frac{d \omega}{2 \pi} S_{V}(\omega)+\int_{\omega_{1}}^{\omega_{2}} \frac{d \omega}{2 \pi} S_{V}(\omega) . \tag{8.13} \label{8.13} \]

    Since \(S_{V}(-\omega)=S_{V}(\omega)\) for a real signal, the two integrals are equal, giving

    \[ \left\langle V^{2}\right\rangle_{\mathrm{meas}}=2 \int_{\omega_{1}}^{\omega_{2}} \frac{d \omega}{2 \pi} S_{V}(\omega) . \tag{8.14} \label{8.14} \]

    Finally, substituting \(\omega=2 \pi f, d \omega=2 \pi d f\) :

    \[ =2 \int_{f_{1}}^{f_{2}} S_{V}(2 \pi f) d f=\int_{f_{1}}^{f_{2}} G_{V}(f) d f \tag{8.15} \label{8.15} \]

    where \(G_{V}(f) \equiv 2 S_{V}(2 \pi f)\) is the one-sided spectral density. The factor of 2 in the definition of \(G_{V}\) absorbs both the factor from folding negative frequencies and the Jacobian of the \(\omega \rightarrow f\) substitution. Collecting these steps:

    A voltmeter sensitive to frequencies in the band \(\omega \in\left[\omega_{1}, \omega_{2}\right]\) measures the mean-square voltage

    \[ \left\langle V^{2}\right\rangle=2 \int_{\omega_{1}}^{\omega_{2}} \frac{d \omega}{2 \pi} S_{V}(\omega)=\int_{f_{1}}^{f_{2}} G_{V}(f) d f \tag{8.16} \label{8.16} \]

    where \(\Delta f=\left(f_{2}-f_{1}\right)\) is the measurement bandwidth, exactly as in Nyquist's original analysis. The one-sided spectral density \(G_{V}(f)=2 S_{V}(2 \pi f)\) absorbs both the folding of negative frequencies and the Jacobian of the \(\omega \rightarrow f\) substitution, which explains the factor of 4 commonly found in engineering books [20] versus the factor of 2 in statistical mechanics texts.


    This page titled 8.3: Spectral Density and Bandwidth was last modified on Mon, 02 Nov 2026 06:09:31 GMT and is shared under a CC BY-NC-SA 4.0 license and was authored, remixed, and/or curated by Pabitra N. Sen.