j_hartling/paper_2025
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Wrote field data methods. Added some very general at the beginning of the methods section.

Browse Source
This commit is contained in:
j-hartling
2026-05-14 14:37:02 +02:00
parent 688f153bef
commit cbd0af7a5f
9 changed files with 172 additions and 164 deletions
Download Patch File Download Diff File Expand all files Collapse all files

19
main.aux
View File

@@ -173,7 +173,6 @@
\abx@aux@page{49}{4}
\abx@aux@page{50}{4}
\@writefile{toc}{\contentsline {section}{\numberline {2}Methods}{4}{}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {2.1}Functional model of the grasshopper song recognition pathway}{4}{}\protected@file@percent }
\abx@aux@cite{0}{windmill2008time}
\abx@aux@segm{0}{0}{windmill2008time}
\abx@aux@cite{0}{malkin2014energy}
@@ -182,6 +181,7 @@
\abx@aux@segm{0}{0}{machens2001discrimination}
\abx@aux@cite{0}{machens2001representation}
\abx@aux@segm{0}{0}{machens2001representation}
\@writefile{toc}{\contentsline {subsection}{\numberline {2.1}Functional model of the grasshopper song recognition pathway}{5}{}\protected@file@percent }
\abx@aux@page{51}{5}
\abx@aux@page{52}{5}
\abx@aux@page{53}{5}
@@ -194,7 +194,6 @@
\abx@aux@page{59}{5}
\abx@aux@page{60}{5}
\newlabel{eq:bandpass}{{1}{5}{}{}{}}
\abx@aux@page{61}{5}
\@writefile{lof}{\contentsline {figure}{\numberline {1}{\ignorespaces \textbf {Schematic organisation of the grasshopper song recognition pathway and structure of the functional model pathway.} \textbf {a}:~Simplified course of the pathway in the grasshopper, from the tympanal membrane over receptor neurons, local interneurons, and ascending neurons further towards the supraesophageal ganglion. \textbf {b}:~Schematic of synaptic connections between the three neuronal populations within the metathoracic ganglion. \textbf {c}:~Network representation of neuronal connectivity. \textbf {d}:~Flow diagram of consecutive signal representations~(boxes) and transformations~(arrows) along the model pathway. All representations are time-varying. 1st half: Preprocessing stage~(one-dimensional representation). 2nd half: Feature extraction stage~(high-dimensional representation). }}{6}{}\protected@file@percent }
\providecommand*\caption@xref[2]{\@setref\relax\@undefined{#1}}
\newlabel{fig:pathway}{{1}{6}{}{}{}}
@@ -212,6 +211,7 @@
\abx@aux@segm{0}{0}{hildebrandt2009origin}
\abx@aux@cite{0}{clemens2010intensity}
\abx@aux@segm{0}{0}{clemens2010intensity}
\abx@aux@page{61}{7}
\abx@aux@page{62}{7}
\newlabel{eq:env}{{2}{7}{}{}{}}
\abx@aux@page{63}{7}
@@ -247,17 +247,18 @@
\@writefile{toc}{\contentsline {subsubsection}{\numberline {2.2.1}Data sourcing}{10}{}\protected@file@percent }
\@writefile{lot}{\contentsline {table}{\numberline {2}{\ignorespaces Overview of the six grasshopper species from the \textit {Gomphocerinae} sub-family, the number of sources per species, the number of available recordings across sources, and the number of isolated songs across recordings.}}{11}{}\protected@file@percent }
\newlabel{tab:species_list}{{2}{11}{}{}{}}
\@writefile{toc}{\contentsline {subsubsection}{\numberline {2.2.2}Generating synthetic input signals}{11}{}\protected@file@percent }
\@writefile{toc}{\contentsline {subsubsection}{\numberline {2.2.2}Generation of synthetic input signals}{11}{}\protected@file@percent }
\newlabel{eq:noiseless}{{11}{11}{}{}{}}
\newlabel{eq:noisy}{{12}{12}{}{}{}}
\newlabel{eq:input_snr}{{13}{12}{}{}{}}
\@writefile{toc}{\contentsline {subsubsection}{\numberline {2.2.3}Quantifying signal intensity across representations}{12}{}\protected@file@percent }
\newlabel{sec:intensity_measures}{{2.2.3}{12}{}{}{}}
\newlabel{eq:input_ratio}{{14}{13}{}{}{}}
\@writefile{toc}{\contentsline {subsection}{\numberline {2.3}Field data-based analysis of the model pathway}{14}{}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {2.3}Field data-based analysis of the model pathway}{13}{}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {3}Results}{14}{}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {3.1}Mechanisms driving the emergence of intensity invariance}{14}{}\protected@file@percent }
\@writefile{toc}{\contentsline {subsubsection}{\numberline {3.1.1}Full-wave rectification \& lowpass filtering}{14}{}\protected@file@percent }
\@writefile{lof}{\contentsline {figure}{\numberline {4}{\ignorespaces \textbf {Rectification and lowpass filtering improves SNR but does not contribute to intensity invariance.} Input $x_{\text {raw}}(t)$ consists of song component $s(t)$ scaled by $\alpha $ with optional noise component $\eta (t)$ and is successively transformed into tympanal signal $x_{\text {filt}}(t)$ and envelope $x_{\text {env}}(t)$. Different line styles indicate different cutoff frequencies $f_{\text {cut}}$ of the lowpass filter extracting $x_{\text {env}}(t)$. \textbf {Top}:~Example representations of $x_{\text {filt}}(t)$ and $x_{\text {env}}(t)$ for different $\alpha $. \textbf {a}:~Noiseless case. \textbf {b}:~Noisy case. \textbf {Bottom}:~Intensity metrics over a range of $\alpha $. \textbf {c}:~Noiseless case: Standard deviations $\sigma _x$ of $x_{\text {filt}}(t)$ and $x_{\text {env}}(t)$. \textbf {d}:~Noisy case: Ratios of $\sigma _x$ of $x_{\text {filt}}(t)$ and $x_{\text {env}}(t)$ to the respective reference standard deviation $\sigma _{\eta }$ for input $x_{\text {raw}}(t)=\eta (t)$. \textbf {e}:~Ratios of $\sigma _x$ to $\sigma _{\eta }$ of $x_{\text {env}}(t)$ as in \textbf {d} for different species (averaged over songs and recordings, see appendix Fig.\,\ref {fig:app_rect-lp}). }}{16}{}\protected@file@percent }
\@writefile{lof}{\contentsline {figure}{\numberline {4}{\ignorespaces \textbf {Rectification and lowpass filtering improves SNR but does not contribute to intensity invariance.} Input $x_{\text {raw}}(t)$ consists of song component $s(t)$ scaled by $\alpha $ with optional noise component $\eta (t)$ and is successively transformed into tympanal signal $x_{\text {filt}}(t)$ and envelope $x_{\text {env}}(t)$. Different line styles indicate different cutoff frequencies $f_{\text {cut}}$ of the lowpass filter extracting $x_{\text {env}}(t)$. \textbf {Top}:~Example representations of $x_{\text {filt}}(t)$ and $x_{\text {env}}(t)$ for different $\alpha $. \textbf {a}:~Noiseless case. \textbf {b}:~Noisy case. \textbf {Bottom}:~Intensity measures over a range of $\alpha $. \textbf {c}:~Noiseless case: Standard deviations $\sigma _x$ of $x_{\text {filt}}(t)$ and $x_{\text {env}}(t)$. \textbf {d}:~Noisy case: Ratios of $\sigma _x$ of $x_{\text {filt}}(t)$ and $x_{\text {env}}(t)$ to the respective reference standard deviation $\sigma _{\eta }$ for input $x_{\text {raw}}(t)=\eta (t)$. \textbf {e}:~Ratios of $\sigma _x$ to $\sigma _{\eta }$ of $x_{\text {env}}(t)$ as in \textbf {d} for different species (averaged over songs and recordings, see appendix Fig.\,\ref {fig:app_rect-lp}). }}{16}{}\protected@file@percent }
\newlabel{fig:rect-lp}{{4}{16}{}{}{}}
\@writefile{toc}{\contentsline {subsubsection}{\numberline {3.1.2}Logarithmic compression \& spike-frequency adaptation}{16}{}\protected@file@percent }
\newlabel{eq:toy_env_pure}{{15}{17}{}{}{}}
@@ -266,7 +267,7 @@
\newlabel{eq:toy_env_noise}{{18}{17}{}{}{}}
\newlabel{eq:toy_log_noise}{{19}{17}{}{}{}}
\newlabel{eq:toy_highpass_noise}{{20}{17}{}{}{}}
\@writefile{lof}{\contentsline {figure}{\numberline {5}{\ignorespaces \textbf {Intensity invariance through logarithmic compression and adaptation is restricted by the noise floor and decreases SNR.} Input $x_{\text {filt}}(t)$ consists of song component $s(t)$ scaled by $\alpha $ with optional noise component $\eta (t)$ and is successively transformed into envelope $x_{\text {env}}(t)$, logarithmically compressed envelope $x_{\text {log}}(t)$, and intensity-adapted envelope $x_{\text {adapt}}(t)$. \textbf {Top}:~Example representations of $x_{\text {env}}(t)$, $x_{\text {log}}(t)$, and $x_{\text {adapt}}(t)$ for different $\alpha $. \textbf {a}:~Noiseless case. \textbf {b}:~Noisy case. \textbf {Bottom}:~Intensity metrics over a range of $\alpha $. \textbf {c}:~Noiseless case: Standard deviations $\sigma _x$ of $x_{\text {env}}(t)$, $x_{\text {log}}(t)$, and $x_{\text {adapt}}(t)$. \textbf {d}:~Noisy case: Ratios of $\sigma _x$ of $x_{\text {env}}(t)$, $x_{\text {log}}(t)$, and $x_{\text {adapt}}(t)$ to the respective reference standard deviation $\sigma _{\eta }$ for input $x_{\text {filt}}(t)=\eta (t)$. Shaded areas indicate $5\,\%$ (dark grey) and $95\,\%$ (light grey) curve span for $x_{\text {adapt}}(t)$. \textbf {e}:~Ratios of $\sigma _x$ to $\sigma _{\eta }$ of $x_{\text {adapt}}(t)$ as in \textbf {d} for different species (averaged over songs and recordings, see appendix Fig\,\ref {fig:app_log-hp_curves}). Dots indicate $95\,\%$ curve span per species. }}{19}{}\protected@file@percent }
\@writefile{lof}{\contentsline {figure}{\numberline {5}{\ignorespaces \textbf {Intensity invariance through logarithmic compression and adaptation is restricted by the noise floor and decreases SNR.} Input $x_{\text {filt}}(t)$ consists of song component $s(t)$ scaled by $\alpha $ with optional noise component $\eta (t)$ and is successively transformed into envelope $x_{\text {env}}(t)$, logarithmically compressed envelope $x_{\text {log}}(t)$, and intensity-adapted envelope $x_{\text {adapt}}(t)$. \textbf {Top}:~Example representations of $x_{\text {env}}(t)$, $x_{\text {log}}(t)$, and $x_{\text {adapt}}(t)$ for different $\alpha $. \textbf {a}:~Noiseless case. \textbf {b}:~Noisy case. \textbf {Bottom}:~Intensity measures over a range of $\alpha $. \textbf {c}:~Noiseless case: Standard deviations $\sigma _x$ of $x_{\text {env}}(t)$, $x_{\text {log}}(t)$, and $x_{\text {adapt}}(t)$. \textbf {d}:~Noisy case: Ratios of $\sigma _x$ of $x_{\text {env}}(t)$, $x_{\text {log}}(t)$, and $x_{\text {adapt}}(t)$ to the respective reference standard deviation $\sigma _{\eta }$ for input $x_{\text {filt}}(t)=\eta (t)$. Shaded areas indicate $5\,\%$ (dark grey) and $95\,\%$ (light grey) curve span for $x_{\text {adapt}}(t)$. \textbf {e}:~Ratios of $\sigma _x$ to $\sigma _{\eta }$ of $x_{\text {adapt}}(t)$ as in \textbf {d} for different species (averaged over songs and recordings, see appendix Fig\,\ref {fig:app_log-hp_curves}). Dots indicate $95\,\%$ curve span per species. }}{19}{}\protected@file@percent }
\newlabel{fig:log-hp}{{5}{19}{}{}{}}
\@writefile{toc}{\contentsline {subsubsection}{\numberline {3.1.3}Thresholding nonlinearity \& temporal averaging}{20}{}\protected@file@percent }
\newlabel{eq:pdf_split}{{21}{20}{}{}{}}
@@ -279,13 +280,13 @@
\newlabel{fig:thresh-lp_species}{{7}{25}{}{}{}}
\@writefile{toc}{\contentsline {subsection}{\numberline {3.3}Intensity invariance along the full model pathway}{26}{}\protected@file@percent }
\@writefile{toc}{\contentsline {subsubsection}{\numberline {3.3.1}Including logarithmic compression}{26}{}\protected@file@percent }
\@writefile{lof}{\contentsline {figure}{\numberline {8}{\ignorespaces \textbf {Step-wise emergence of intensity-invariant song representations along the model pathway.} Input $x_{\text {raw}}(t)$ consists of song component $s(t)$ scaled by $\alpha $ with added noise component $\eta (t)$ and is processed up to the feature set $f_i(t)$. Different color shades indicate different types of Gabor kernels with specific lobe number $n$ and either $+$ or $-$ sign, sorted (dark to light) first by increasing $n$ and then by sign~($1\,\leq \,n\,\leq \,4$; first $+$, then $-$ for each $n$; five kernel widths $\sigma $ of 1, 2, 4, 8, and $16\,$ms per type; 8 types, 40 kernels in total). \textbf {a}:~Example representations of $x_{\text {filt}}(t)$, $x_{\text {env}}(t)$, $x_{\text {log}}(t)$, $x_{\text {adapt}}(t)$, $c_i(t)$, and $f_i(t)$ for different $\alpha $. \textbf {b}:~Intensity metrics over $\alpha $. For $c_i(t)$ and $f_i(t)$, the median over kernels is shown. Dots indicate $95\,\%$ curve span for $x_{\text {log}}(t)$, $x_{\text {adapt}}(t)$, $c_i(t)$, and $f_i(t)$. \textbf {c}:~Average value $\mu _{f_i}$ of each feature $f_i(t)$ over $\alpha $. \textbf {d}:~Ratios of intensity metrics to the respective reference value for input $x_{\text {raw}}(t)=\eta (t)$. For $c_i(t)$ and $f_i(t)$, the median over kernel-specific ratios is shown. \textbf {e}:~Ratios of standard deviation $\sigma _{c_i}$ of each $c_i(t)$. \textbf {f}:~Ratios of $\mu _{f_i}$. \textbf {g}:~Distributions of kernel-specific $\alpha $ that correspond to $95\,\%$ curve span for $c_i(t)$ and $f_i(t)$. Dots indicate the values from \textbf {b}. }}{28}{}\protected@file@percent }
\@writefile{lof}{\contentsline {figure}{\numberline {8}{\ignorespaces \textbf {Step-wise emergence of intensity-invariant song representations along the model pathway.} Input $x_{\text {raw}}(t)$ consists of song component $s(t)$ scaled by $\alpha $ with added noise component $\eta (t)$ and is processed up to the feature set $f_i(t)$. Different color shades indicate different types of Gabor kernels with specific lobe number $n$ and either $+$ or $-$ sign, sorted (dark to light) first by increasing $n$ and then by sign~($1\,\leq \,n\,\leq \,4$; first $+$, then $-$ for each $n$; five kernel widths $\sigma $ of 1, 2, 4, 8, and $16\,$ms per type; 8 types, 40 kernels in total). \textbf {a}:~Example representations of $x_{\text {filt}}(t)$, $x_{\text {env}}(t)$, $x_{\text {log}}(t)$, $x_{\text {adapt}}(t)$, $c_i(t)$, and $f_i(t)$ for different $\alpha $. \textbf {b}:~Intensity measures over $\alpha $. For $c_i(t)$ and $f_i(t)$, the median over kernels is shown. Dots indicate $95\,\%$ curve span for $x_{\text {log}}(t)$, $x_{\text {adapt}}(t)$, $c_i(t)$, and $f_i(t)$. \textbf {c}:~Average value $\mu _{f_i}$ of each feature $f_i(t)$ over $\alpha $. \textbf {d}:~Ratios of intensity measures to the respective reference value for input $x_{\text {raw}}(t)=\eta (t)$. For $c_i(t)$ and $f_i(t)$, the median over kernel-specific ratios is shown. \textbf {e}:~Ratios of standard deviation $\sigma _{c_i}$ of each $c_i(t)$. \textbf {f}:~Ratios of $\mu _{f_i}$. \textbf {g}:~Distributions of kernel-specific $\alpha $ that correspond to $95\,\%$ curve span for $c_i(t)$ and $f_i(t)$. Dots indicate the values from \textbf {b}. }}{28}{}\protected@file@percent }
\newlabel{fig:pipeline_full}{{8}{28}{}{}{}}
\@writefile{toc}{\contentsline {subsubsection}{\numberline {3.3.2}Excluding logarithmic compression}{29}{}\protected@file@percent }
\@writefile{lof}{\contentsline {figure}{\numberline {9}{\ignorespaces \textbf {Effects of disabling logarithmic compression on intensity invariance along the model pathway.} Input $x_{\text {raw}}(t)$ consists of song component $s(t)$ scaled by $\alpha $ with added noise component $\eta (t)$ and is processed up to the feature set $f_i(t)$, skipping $x_{\text {log}}(t)$. Different color shades indicate different types of Gabor kernels with specific lobe number $n$ and either $+$ or $-$ sign, sorted (dark to light) first by increasing $n$ and then by sign~($1\,\leq \,n\,\leq \,4$; first $+$, then $-$ for each $n$; five kernel widths $\sigma $ of 1, 2, 4, 8, and $16\,$ms per type; 8 types, 40 kernels in total). \textbf {a}:~Example representations of $x_{\text {filt}}(t)$, $x_{\text {env}}(t)$, $x_{\text {adapt}}(t)$, $c_i(t)$, and $f_i(t)$ for different $\alpha $. \textbf {b}:~Intensity metrics over $\alpha $. For $c_i(t)$ and $f_i(t)$, the median over kernels is shown. Dots indicate $95\,\%$ curve span for $f_i(t)$. \textbf {c}:~Average value $\mu _{f_i}$ of each feature $f_i(t)$ over $\alpha $. \textbf {d}:~Ratios of intensity metrics to the respective reference value for input $x_{\text {raw}}(t)=\eta (t)$. For $c_i(t)$ and $f_i(t)$, the median over kernel-specific ratios is shown. \textbf {e}:~Ratios of $\mu _{f_i}$. \textbf {f}:~Distribution of kernel-specific $\alpha $ that correspond to $95\,\%$ curve span for $f_i(t)$. Dots indicate the value from \textbf {b}. }}{30}{}\protected@file@percent }
\@writefile{lof}{\contentsline {figure}{\numberline {9}{\ignorespaces \textbf {Effects of disabling logarithmic compression on intensity invariance along the model pathway.} Input $x_{\text {raw}}(t)$ consists of song component $s(t)$ scaled by $\alpha $ with added noise component $\eta (t)$ and is processed up to the feature set $f_i(t)$, skipping $x_{\text {log}}(t)$. Different color shades indicate different types of Gabor kernels with specific lobe number $n$ and either $+$ or $-$ sign, sorted (dark to light) first by increasing $n$ and then by sign~($1\,\leq \,n\,\leq \,4$; first $+$, then $-$ for each $n$; five kernel widths $\sigma $ of 1, 2, 4, 8, and $16\,$ms per type; 8 types, 40 kernels in total). \textbf {a}:~Example representations of $x_{\text {filt}}(t)$, $x_{\text {env}}(t)$, $x_{\text {adapt}}(t)$, $c_i(t)$, and $f_i(t)$ for different $\alpha $. \textbf {b}:~Intensity measures over $\alpha $. For $c_i(t)$ and $f_i(t)$, the median over kernels is shown. Dots indicate $95\,\%$ curve span for $f_i(t)$. \textbf {c}:~Average value $\mu _{f_i}$ of each feature $f_i(t)$ over $\alpha $. \textbf {d}:~Ratios of intensity measures to the respective reference value for input $x_{\text {raw}}(t)=\eta (t)$. For $c_i(t)$ and $f_i(t)$, the median over kernel-specific ratios is shown. \textbf {e}:~Ratios of $\mu _{f_i}$. \textbf {f}:~Distribution of kernel-specific $\alpha $ that correspond to $95\,\%$ curve span for $f_i(t)$. Dots indicate the value from \textbf {b}. }}{30}{}\protected@file@percent }
\newlabel{fig:pipeline_short}{{9}{30}{}{}{}}
\@writefile{toc}{\contentsline {subsubsection}{\numberline {3.3.3}Intensity invariance in a naturalistic setting}{31}{}\protected@file@percent }
\@writefile{lof}{\contentsline {figure}{\numberline {10}{\ignorespaces \textbf {Intensity invariance along the model pathway in a naturalistic setting.} Input $x_{\text {raw}}(t)$ consists of a song of \textit {P. parallelus} recorded in the field at eight different distances $d$ and is processed up to the feature set $f_i(t)$. Different color shades indicate different types of Gabor kernels with specific lobe number $n$ and either $+$ or $-$ sign, sorted (dark to light) first by increasing $n$ and then by sign~($1\,\leq \,n\,\leq \,4$; first $+$, then $-$ for each $n$; five kernel widths $\sigma $ of 1, 2, 4, 8, and $16\,$ms per type; 8 types, 40 kernels in total). \textbf {a}:~$x_{\text {filt}}(t)$, $x_{\text {env}}(t)$, $x_{\text {log}}(t)$, $x_{\text {adapt}}(t)$, $c_i(t)$, and $f_i(t)$ at each $d$. A noise segment from the same recording is shown for reference. \textbf {b}:~Intensity metrics over $d$. For $c_i(t)$ and $f_i(t)$, the median over kernels is shown. \textbf {c}:~Average value $\mu _{f_i}$ of each feature $f_i(t)$ over $d$. \textbf {d}:~Ratios of intensity metrics to the respective value obtained from the noise reference. For $c_i(t)$ and $f_i(t)$, the median over kernel-specific ratios is shown. \textbf {e}:~Ratios of standard deviation $\sigma _{c_i}$ of each $c_i(t)$. \textbf {f}:~Ratios of $\mu _{f_i}$. }}{32}{}\protected@file@percent }
\@writefile{lof}{\contentsline {figure}{\numberline {10}{\ignorespaces \textbf {Intensity invariance along the model pathway in a naturalistic setting.} Input $x_{\text {raw}}(t)$ consists of a song of \textit {P. parallelus} recorded in the field at eight different distances $d$ and is processed up to the feature set $f_i(t)$. Different color shades indicate different types of Gabor kernels with specific lobe number $n$ and either $+$ or $-$ sign, sorted (dark to light) first by increasing $n$ and then by sign~($1\,\leq \,n\,\leq \,4$; first $+$, then $-$ for each $n$; five kernel widths $\sigma $ of 1, 2, 4, 8, and $16\,$ms per type; 8 types, 40 kernels in total). \textbf {a}:~$x_{\text {filt}}(t)$, $x_{\text {env}}(t)$, $x_{\text {log}}(t)$, $x_{\text {adapt}}(t)$, $c_i(t)$, and $f_i(t)$ at each $d$. A noise segment from the same recording is shown for reference. \textbf {b}:~Intensity measures over $d$. For $c_i(t)$ and $f_i(t)$, the median over kernels is shown. \textbf {c}:~Average value $\mu _{f_i}$ of each feature $f_i(t)$ over $d$. \textbf {d}:~Ratios of intensity measures to the respective value obtained from the noise reference. For $c_i(t)$ and $f_i(t)$, the median over kernel-specific ratios is shown. \textbf {e}:~Ratios of standard deviation $\sigma _{c_i}$ of each $c_i(t)$. \textbf {f}:~Ratios of $\mu _{f_i}$. }}{32}{}\protected@file@percent }
\newlabel{fig:pipeline_field}{{10}{32}{}{}{}}
\@writefile{toc}{\contentsline {subsection}{\numberline {3.4}Interspecific and intraspecific feature variability}{33}{}\protected@file@percent }
\@writefile{lof}{\contentsline {figure}{\numberline {11}{\ignorespaces \textbf {Interspecific and intraspecific feature variability.} Average value $\mu _{f_i}$ of each feature $f_i(t)$ against its counterpart from a 2nd feature set based on a different input $x_{\text {raw}}(t)$. Each dot within a subplot represents a single feature $f_i(t)$. Different color shades indicate different types of Gabor kernels with specific lobe number $n$ and either $+$ or $-$ sign, sorted (dark to light) first by increasing $n$ and then by sign~($1\,\leq \,n\,\leq \,4$; first $+$, then $-$ for each $n$; five kernel widths $\sigma $ of 1, 2, 4, 8, and $16\,$ms per type; 8 types, 40 kernels in total). Data is based on the analysis underlying Fig\,\ref {fig:pipeline_full}. \textbf {Lower triangular}:~Interspecific comparisons between single songs of different species. \textbf {Upper triangular}:~Intraspecific comparisons between different songs of a single species (\textit {O. rufipes}). \textbf {Lower right}:~Distribution of correlation coefficients $\rho $ for each interspecific and intraspecific comparison. Dots indicate single $\rho $ values. }}{34}{}\protected@file@percent }