Predicting single-tone responses from multitone responses. A Automatic gain control model. The input passes a variable-gain filter that reproduces the measured multitone responses. The overall magnitude of the output, \documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$ \left\langle {{{D}^{2}}} \right\rangle $$\end{document}, is fed back to the filter as a gain control, setting the shape of the filter. B Predicting a single-tone response. Passing a 50-dB near-CF single tone to all possible filters produces \documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$ \left\langle {{D_{\text{S}}}^{{2}}} \right\rangle $$\end{document} as a function of the multitone level I
M associated with the respective filters. For the filter setting to be self-consistent, \documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$ \left\langle {{D_{\text{S}}}^{{2}}} \right\rangle $$\end{document} must match the \documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$ \left\langle {{D_{\text{M}}}^{{2}}} \right\rangle $$\end{document} associated with the candidate filter. C, E Model predictions of the data in Figure 3 for normalized amplitude and phase, respectively. D, F Difference between predictions and data plotted versus frequency for amplitude and phase, respectively. Each curve represents a different SPL. Inset in D scatter plot of model predictions versus data for amplitude, with correlation R and variance accounted for (vaf) indicated. Variance accounted for (vaf) in phase indicated in F.
