Splitting a beamer animation

Question

I have a complex animation that evolves over a series of slides (a \foreach \frm in {1,...,12}\only<\frm>{ loop where certain parts are selected on individual slides with \visible).

So far I had all overlays inside a single frame, but I would like to distribute it to several frames so I can put other frames (with text and other things) in between, so I moved the entire tikzpicture where all overlays are created into a \newcommand{\makeAnimation}[1] such that I can say, e.g.

\begin{frame}{First part of the animation}
  \makeAnimation{1,...,4}
\end{frame}

\begin{frame}{Some explanations}
  Lorem ipsum
\end{frame}

\begin{frame}{Next part}
  \makeAnimation{4,...,7}
\end{frame}

(It is easier to create the whole thing in one go, because I would have to copy a lot of code otherwise)

However, the after the portion of the animation that is (correctly) generated by \makeAnimation{1,...,4} I get an extra empty slide, which I am unable to move away. Worse, if I change the \only after the \foreach statement, into a \visible (this keeps the centered tikzpicture from jumping around) it generates all slides from 1 ... 12 but leaves 5 ... 12 blank.

How can I fix this and get rid of the extra slides? At this point I would accept any hack as long as I don't have to split up/redo the animation.

---

Small example:

\documentclass[
  presentation,
  english,
]{beamer}
%\usepackage[utf8]{inputenc}
\usepackage{babel}
\usepackage{tikz}

\newcommand{\makeAnimation}[1]{
  \begin{center}
  \scalebox{0.7} {
    \begin{tikzpicture}
    \foreach \frm in {#1} {
      \only<\frm>{
        \node {FRAME \frm};
        \visible<3>{\node at (4,0) {three hippos};}
        \visible<4-5>{\node (a) at (2,0) {a couple giraffes};}
        \visible<5-6>{\node (b) at (4,2) {hello};}
        \visible<5>{\path[-] (a) edge (b);}
        \visible<6>{\node at (4,0) {six elephants};}
      }   
    }   
    \end{tikzpicture}
  }
  \end{center}
}

\begin{document}
\begin{frame}{foo}
  \makeAnimation{1,...,4}
\end{frame}
\begin{frame}{bar}
  next slide
\end{frame}
\end{document}

Produces

foo FRAME 1
foo FRAME 2
foo FRAME 3 three hippos
foo FRAME 4 a couple giraffes
foo                  % evil blank slide
bar next slide

Answer 1

If I may be so bold, I want to say that you are doing it wrong. beamer has a built-in way of "cutting" frames: the \againframe command, combined with overlay specifications on the frame to be cut. It's described in the manual, section 11.2 ("Repeating a frame at a later point"), and works like this:

\documentclass[
  presentation,
  english,
]{beamer}
%\usepackage[utf8]{inputenc}
\usepackage{babel}
\usepackage{tikz}

\begin{document}

\begin{frame}<1-4>[label = foo]{foo}
 \begin{tikzpicture}
 \onslide<1>
  \node {Frame 1};
 \onslide<2>
  \node {Frame 2};
 \onslide<3>
  \node at (4,0) {three hippos};
 \onslide<4-5>
  \node (a) at (2,0) {a couple giraffes};
 \onslide<5-6>
  \node (b) at (4,2) {hello};
 \onslide<5>
  \path[-] (a) edge (b);
 \onslide<6>
  \node at (4,0) {six elephants};
 \end{tikzpicture}
\end{frame}

\begin{frame}{bar}
  next slide
\end{frame}

\againframe<4-6>{foo}
\end{document}

Basically, the first frame is the "template" for the animation and replaces your \makeAnimation. The overlay specification says only to run it for four slides, and then cut to the explanation. Then \againframe comes in and produces slides 4 through 7 to finish the animation.

I am a little annoyed that it is necessary to write Frame 1 and Frame 2 separately, as there does exist a counter framenumber that should print the correct value in a single \theframenumber command. However, when I tried

\onslide<1-2>
 \node {Frame \theframenumber};

it just printed "Frame 1" twice. Using \only<1-2> did the same, and in fact, using Frame \theframenumber with two \onslide commands also printed "Frame 1" even after \onslide<2>! Perhaps this is a bug; certainly, this counter is not mentioned in the manual, so perhaps it is not supported.

Answer 2

Avoid unnecessary \visible<...> lines, let it depend on your argument range. After I removed the \visible<5> and \visible<6> lines in your example, the empty frame was gone.

The code (before the question has been updated) could even be reduced, doing the same but no empty slide:

\begin{tikzpicture}
\foreach \frm in {#1} {
  \only<\frm>{
    \node {FRAME \frm};
    \visible<\frm>{\node at (4,0) {its \frm};}
  }   
}   
\end{tikzpicture}

Even with the update, you could use \if or similar or a \foreach loop for \visible and a variable for the node content.

If it's just for preventing jumping, I would use \overlayarea.

Answer 3

I found that it is best to play dirty with tex. This does the trick, as it wraps the entire thing.

% parameter documentation
%  #1: current frame
%  #2: displayed range minimum
%  #3: displayed range maximum
%  #4: object range minimum
%  #5: object range maximum
%  #6: object
\newcommand{\showPart}[6]{\ifnum#1<#4\relax\else\ifnum#5<#1\relax\else\visible<\ifnum#2<#4 #4\else#2\fi-\ifnum#3<#5 #3\else#5\fi>{#6}\fi\fi}

Which allows the animation command to look like that:

\begin{tikzpicture}
    \foreach \frm in {#1,...,#2} {
      %\only<\frm>{  %%%%% IMPORTANT: remove this!
        \node at (-4,\frm) {FRAME \frm}; %% all four shown all the time!
        \showPart{\frm}{#1}{#2}{2}{4}{\node at (4,0) {it's 2-4};}
        \showPart{\frm}{#1}{#2}{4}{5}{\node at (4,2) {it's 4-5};}
        \showPart{\frm}{#1}{#2}{5}{5}{\node at (4,4) {it's 5};}
      %}
    }
    \end{tikzpicture}