\documentstyle{speech}

\begin{document}

\thisischapter{1}

\author[A.U. Thor]{Author}

\address{Elsevier Science B.V., P.O. Box 521, 1000 AM Amsterdam,
         The Netherlands}

\chapter[Document style]{\LaTeX\ Manuscripts:\\
                         Speech Coding and Synthesis}

\section{Introduction}

Worldwide it is increasingly customary to compose scientific texts by
using some kind of electronic text processing system. For
type-setting scientific texts the systems \TeX\ and/or \LaTeX\ are
very popular (especially for highly mathematical texts).

Books or articles typed in \TeX\ or \LaTeX\ can also be very convenient 
in use for publishers, especially when authors adhere to some stylistic
elements. Reasons for specific style elements to be required are, for
instance, they are of common use in the field of interest, commonly
used in a certain periodical or book series, or a specific publisher
requires them. 

Adherence to the prescribed style will keep necessary editorial
changes to a minimum, and will therefore considerably speed up the
production process.

The following is a very concise manual for the style to be used 
as well as a description of the \LaTeX\ style files.

\section{Aspects of style}

\subsection{Variables, constants and labels}

In general, a scalar variable is type-set in an italic font. The way to
achieve this using \LaTeX\ is by typing variables in math mode only.
This means that scalar variables in running text should always be surrounded
by dollar signs. The same applies to one-letter identifiers having a
numerical value (barring some exceptions to be listed below). Labels
that are abbreviations for names or words, however, should always be
type-set upright, i.e., in a roman typeface (e.g., `E' for `Euler' in
$h_{\rm E}$, `m' for `minimum' in $\varepsilon_{\rm m}$ and `RK' for
`Runge--Kutta' in the step size $h_{\rm RK}$). 

Since these labels often occur in a mathematical environment, and
because \LaTeX\ type-sets all letters in math mode in an italic font,
the change to \verb|\rm| has to be made explicitly. To illustrate the
above, consider the following piece of text\footnote{In the examples
below a coding consistent with the old font selection scheme has been
used. If your local installation has the new font selection scheme
installed the coding is somewhat different.},

\begin{quotation}
We remarked earlier that a single dense row in $A$ will lead to a
full matrix $A^{\rm T}A$ and therefore, invoking the no-cancellation
assumption, a full Cholesky factor $R_{\rm C}$. Problems where the
matrix $A$ is sparse except for a few dense rows can be treated by
first solving the problems with the dense rows deleted. (\dots)

Consider the problem
\begin{equation}
\min \left\vert \left({A_{\rm s}\atop A_{\rm d}}\right)x-
\left({b_{\rm s}\atop b_{\rm d}}\right)\right\vert_2,
\end{equation}
where $A_{\rm s}\in {\Bbb R}^{m_1\times n}$ is sparse and 
$A_{\rm d}\in {\Bbb R}^{m_2\times n}$, $m_2 \ll n$ contains 
the dense rows.\ \dots
\end{quotation}
The \LaTeX\ code corresponding to the text above is

\begin{verbatim}
We remarked earlier that a single dense row in $A$ 
will lead to a full matrix $A^{\rm T}A$ and 
therefore, invoking the no-cancellation assumption, 
a full Cholesky factor $R_{\rm C}$. Problems where 
the matrix $A$ is sparse except for a few dense 
rows can be treated by first solving the problems 
with the dense rows deleted. (\dots)

Consider the problem
\begin{equation}
\min \left\vert \left({A_{\rm s}\atop A_{\rm d}}
\right)x-\left({b_{\rm s}\atop b_{\rm d}}\right)
\right\vert_2,
\end{equation}
where $A_{\rm s}\in {\Bbb R}^{m_1\times n}$ is 
sparse and $A_{\rm d}\in {\Bbb R}^{m_2\times n}$, 
$m_2 \ll n$ contains the dense rows.\ \dots
\end{verbatim}

Note that in the example the \lq d' for dense, \lq s' for sparse, \lq
T' for transpose and \lq C' for Cholesky all go upright.

\subsection{Review of the usage of different typefaces}

Normally speaking, all letters in a mathematical formula are
variables and therefore type-set in an italic typeface. Therefore, as
was mentioned above, \LaTeX\ has an italic typeface as default once
the program is in math mode. There are, however, a few exceptions and
these letters are to be type-set upright (i.e., using a roman
typeface). (Since italic is \LaTeX's default, there should be an
explicit font change to obtain upright letters in math mode.) The
most common symbols that are conventionally type-set upright are 
listed below.

\vspace{12pt plus 3pt minus 3pt}
{\it To be set upright are:}
\begin{itemize}
\item i when used as imaginary unit, e.g.,
$ a+{\rm i}b$,  $\rm e^{\rm i\phi}$, etc.
\item Geometric functions, e.g., sin, cos, tan, etc. It is easy to use 
macros \verb|\sin|, \verb|\cos|, and \verb|\tan| for this, since these 
also give proper spacing in formulas.
\item The base of the natural logarithm, as in ${\rm e} ^x$.
\item Differential operators, e.g., d$x$.  (Note: ${\mit\Delta}$ as
common variable, as in ${\mit\Delta} =3.2$ is to be set
slanting.\footnote{Slanting capital Greek letters are obtained by 
explicitly changing to the mathematical italic font {${\tt
\backslash}$mit}.})
\item Groups of more than one character like SU(2) and SU(3), etc.
\item Subscripts and superscripts which are used as an abbreviation,
e.g., $h_{\rm E}$ for Euler step.
\item The identifiers Im and Re for the imaginary
and real parts of complex numbers, respectively.
\end{itemize}
Simple scalar variables and (usually) vectors are set in an italic 
typeface. However, vectors can also be set in a bold italic typeface 
(like this $\vec{r}$) in order to distinguish them from variables. 
To obtain this, please use \verb|\vec{r}|. In case the need is felt 
to distinguish matrices or tensors from ordinary variables, or vectors, 
a boldface sans-serif font can be used (use \verb|\tens|), e.g., 
$\tens{T}$. Since tensor components like $T_{\mu\nu}$ or matrix 
components like $a_{ij}$ are in themselves scalars, they are type-set 
simply using an italic font.

Note. Use font changes only if it is useful: too rich a use of 
different fonts will lead to an ugly text and is prone to lead 
to many errors.

\subsection{Abbreviations and contractions}

As is well-known, certain frequently recurring words in a scientific
text are, in general, not written out in full on every occurrence.
For the Handbook series the following usage is recommended:
\begin{itemize}
\item eq. for `equation' and eqs. for `equations' (when followed
by a number, e.g., eq. (4.2));
\item fig. for `figure' and figs. for `figures' (when followed
by a number, e.g., fig. 1);
\item table for `table';
\item section for `section';
\item chapter for `chapter';
\item e.g. for `for instance' or `for example';
\item i.e. for `that is'.
\end{itemize}
If these words occur at the beginning of a sentence, they should not
be abbreviated.

For reasons of clarity, referring to, e.g., an equation by simply
giving the equation number should be avoided. The readability of the
text can be much improved by writing `eq.' or `eqs.' explicitly. (For
instance, `As can be seen from eqs.~(4.2)--(4.5), this leads to a
contradiction, whence the proposition is not true.' is much clearer
than `As can be seen from (4.2)--(4.5), this leads to a
contradiction, whence the proposition is not true', since readers
might think that lemmas~4.2--4.5 are meant.)

\section{Displayed equations}

\subsection{One-line displayed equations}

In \LaTeX{} there are essentially two ways to indicate a one-line
displayed equation. The first is the environment \verb|equation|, the
second is the environment \verb|displaymath|
(\verb|\begin{displaymath}|\dots \verb|\end{displaymath}| can be
abbreviated to \verb|\[|\dots \verb|\]|). The first produces a
numbered equation and steps the equation counter and the latter
produces an unnumbered equation. Note that plain-\TeX's \verb|$$|
construction is {\em not} used. Using this will have two undesirable 
effects: first it may interfere with the automatic numbering, and
second it will produce centered equations, even if this is not in
accordance with the document style.

\subsection{Multi-line displayed equations}

This type of equation can be divided into two main categories:

\begin{itemize}
\item displayed material that is simply too long to fit on a line, and
\item systems of equations.
\end{itemize}

The first category is (normally) best handled by using the \verb|eqnarray|
environment. This is a three-column array environment, which are aligned 
right, centered, and left, respectively. \LaTeX{} adds an equation number 
to every line automatically, unless the line ends with the 
\verb|\nonumber| command. (If no equation numbers are needed at all, it is 
best to use the \verb|eqnarray*| environment.)

The second category can be handled best by using the construction:
\begin{verbatim}
\begin{equation}
    \begin{array}{x...y}
       ....
    \end{array} 
\end{equation} 
\end{verbatim} 
The main difference with the \verb|eqnarray| construction is that
only one number is added, and that this number is centered behind the
material. If no number is needed at all, use the same construction as
above, but change the \verb|equation| environment to the
\verb|displaymath| environment.

\section{Theorem-like environments}

For the convenience of the reader, a few theorem-like environments have 
been defined. These are
\begin{itemize}
\item \verb|question|
\item \verb|example|
\item \verb|proof|
\item \verb|remark|
\end{itemize}

The fonts used are in accordance with the commonly accepted ones for
these kinds of constructions. The use matches the use of the
\verb|theorem| environment as described in the \LaTeX{} manual.


\section{Getting started}

Secondary document styles speech11 and speech12 will not become available.

Please do not change the typesetting area. One enourmous advantage 
of the fact that authors have document styles available is that they 
can anticipate whether equations will fit. 

The start of the document should look like 

\begin{verbatim}
\documentstyle {speech}

\begin{document}
\thisischapter{1.2}
\author[short author]{<author_name>}
\address{<author_address>}
\support{<Grants, etc., if any>}

\section[Short title]{Long title}
\end{verbatim}

\verb|\thisischapter| is a chapter identifier. If you do not know the
number of your contribution in the book yet, use any number you like.

\verb|\author| has the author's name as it will appear on the title
page as an argument. If the name appearing in the running-head line
is to be different, then the latter should be entered as an optional
argument.

\verb|\support| is similar to the \verb|\thanks| macro in \verb|\LaTeX|.

\verb|\chapter| has an argument that contains the title as it will appear
on the title page. If the title to appear in the running-headline is
to be different, the latter should be entered as an optional argument.

Since it is the \verb|\chapter| command that generates the title page,
\verb|\maketitle| should not be used.

\end{document}