# Best subfigure placement wrt length of subcaptions

My Problem is twofold:

First i'm not sure what the established/best way is to place the captions in question. Secondly i don't know how to do implement my ideas with Latex.

My supervisor suggested, that i should place the graphics on the same height (as indicated by the red line). I'm not sure if this is the right approach, since it would generate whitespace as you can see in the following image (MS Paint ftw).

So my Questions are:

1. Is he right? Should i just force them somehow on the same height? (Maybe using a tabular approach?)
2. Is there maybe a automated scaling approach?
• That could mean with respect to both image size and text length.
• Or just with respect to image height like in the following picture (which might bring problems when the text is long and the width is small). How can something like that be accomplished? (i can't even do it in paint)
3. Or should i just put all the text in the main caption and write only one-liner in the subcaptions?

A MWE (i don't know how i can include graphics that everybody has in his installation):

\documentclass[twoside, 12pt, a4paper, parskip, draft]{scrbook}

\usepackage{hyperref}
\usepackage{graphicx}
\usepackage{caption}
\usepackage{subcaption}

\begin{document}
This is text. This is text. This is text. This is text. This is text. This is text. This is text. This is text.

\begin{figure}[ht]
\centering
\begin{subfigure}[b]{0.40\textwidth}
\includegraphics[width=\textwidth]{Bilder/related/guwip.pdf}
\caption{Erfassung der Fußposition mit einer der beiden Varianten.}
\end{subfigure}%
~
\begin{subfigure}[b]{0.54\textwidth}
\centering
\begin{minipage}{\linewidth}
\includegraphics[width=0.32\textwidth]{Bilder/related/swappdishsmall.pdf}
\includegraphics[width=0.32\textwidth]{Bilder/related/swappshoesmall.pdf}
\includegraphics[width=0.32\textwidth]{Bilder/related/swappcompsmall.pdf}
\end{minipage}
\caption{Erfassung mithilfe der anderen Variante und der mit magischen Bällen versehenen Schuhe. Der Nutzer wischt mit seinen Füßen innerhalb der Schale und erreicht damit etwas.}
\end{subfigure}
\caption{Bilder der auf Motion Capturing basierenden Ansätzen die Fußposition zu verfolgen.}
\end{figure}

This is text. This is text. This is text. This is text. This is text. This is text. This is text. This is text.
\end{document}

• Rule number 1: Your supervisor is always right ;-)
– user11232
Aug 30 '14 at 14:45
• Fortunately my supervisor is pretty cool and open to suggestions, as long as you have something to back up your argument :) Aug 30 '14 at 15:14

You may want to ask yourself if the captions really need to be so long. Captions usually work best if they're short and snappy. Might your readers be better off if you placed much of the text that's currently in the caption either in the body of the (sub)figure or in the main text itself? If you keep all captions nice and short, you won't need to agonize over the placement of overly long captions either, right? :-)

If you want to (or need to) keep the long captions, I suggest you do the following:

• Not having access to the actual graphics files, I can't be certain about this recommendation, but try [t] instead of [b] as the alignment specifier for both subfigures.

• No need to have a minipage environment inside the second subfigure environment. A subfigure is a minipage.

• Try to maximize the space available to the images by not centering the images within their respective spaces. Instead, spread out the images maximally so that they take up the full width of the text block.

\documentclass[twoside, 12pt, a4paper, parskip, draft]{scrbook}
\usepackage[latin1]{inputenc} % because of 'ß' character
\usepackage[T1]{fontenc}
\usepackage{graphicx}
\usepackage{caption,subcaption}
\usepackage{hyperref} % when in doubt, load hyperref last

\begin{document}
This is text. This is text. This is text. This is text. This is text. This is text. This is text. This is text.

\begin{figure}[ht]
\begin{subfigure}[t]{0.40\textwidth}
\includegraphics[width=\linewidth]{Bilder/related/guwip.pdf}
\caption{Erfassung der Fußposition mit einer der beiden Varianten.}
\end{subfigure}%
\hspace{\fill}
\begin{subfigure}[t]{0.56\textwidth}
\includegraphics[width=0.31\linewidth]{Bilder/related/swappdishsmall.pdf} \hspace{\fill}
\includegraphics[width=0.31\linewidth]{Bilder/related/swappshoesmall.pdf} \hspace{\fill}
\includegraphics[width=0.31\linewidth]{Bilder/related/swappcompsmall.pdf}
\caption{Erfassung mithilfe der anderen Variante und der mit magischen Bällen versehenen Schuhe. Der Nutzer wischt mit seinen Füßen innerhalb der Schale und erreicht damit etwas.}
\end{subfigure}
\caption{Bilder der auf Motion Capturing basierenden Ansätzen die Fußposition zu verfolgen.}
\end{figure}

This is text. This is text. This is text. This is text. This is text. This is text. This is text. This is text.
\end{document}

• Changing it to [t] worked: Now the bottom of all pictures is aligned. Guess i try to keep the captions approximately on the same length by hand. Aug 30 '14 at 13:50
• Is it important to use [width=\linewidth] instead of [width=\textwidth] as argument in includegraphics? Aug 30 '14 at 14:23
• @SebastianSchmitz - The parameters \textwidth, \linewidth, and \columnwidth conceptually have different meanings. The usage given the answer above is (I hope!!) correct. In practice, though, there tends to be so much confusion and sloppiness about which parameter means what that LaTeX packages (incl. graphicx!) are frequently written in a way to treat these terms essentially as synonyms.
– Mico
Aug 30 '14 at 16:14