# libertine-newtxmath some symbols

I'm using libertine with newtxmath. Actually it is very good but some symbols don't seem good to me as in the figure.

Can I use different font from these symbols like given below without changing the other ones? (these are from mathpazo.)

\documentclass{article}
\usepackage[english]{babel}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage[lining]{libertine}
\usepackage[libertine,cmintegrals]{newtxmath}
\usepackage{ucs}
\usepackage{amsfonts,amsmath,amssymb}
\usepackage{graphicx}
%------------TIKZ-PGF----------------
\usepackage{tikz,pgf,pgfplots}
\pgfplotsset{compat=1.10}
\usepgfplotslibrary{fillbetween}
\usetikzlibrary{shapes.multipart}
\usetikzlibrary{patterns}
%---------------------------------

\begin{document}

$f\left( \mathbf{x}|\theta\right) =\underset{h\left( \mathbf{x}\right) }{\underbrace{1}}\cdot\underset{g_{1}\left( T\left( \mathbf{X}_{1}\right) |\mu\right) }{\underbrace{\left( 2\pi\sigma^{2}\right) ^{-n/2}\exp\left[ -\frac{n\left( \overline{x}-\mu\right) ^{2}}{2\sigma^{2}}\right] }% }\underset{g_{2}\left( T\left( \mathbf{X}_{2}\right) |\sigma\right) }{\underbrace{\exp\left[ \frac{-1}{2\sigma^{2}}\sum\left( x_{i}-\overline {x}\right) ^{2}\right] }}%$
$\sqrt{\exp\left[ \frac{-1}{2\sigma^{2}}\sum\left( x_{i}-\overline {x}\right) ^{2}\right]}$

$\forall\exists= \pm+-$
\end{document}

• Which font(s) do you want them to come from? – cfr Apr 21 '14 at 3:31
• for example computer modern or mathpazo. – mert Apr 21 '14 at 6:31
• newtxmath provides alternatives for \forall and \exists, namely \forallAlt and \existxAlt – Guido Apr 21 '14 at 8:48
• @Guido did you mean \existsAlt? – cfr Apr 21 '14 at 11:40
• @cfr yes I did :-( – Guido Apr 21 '14 at 12:17

Adding the following at the end of your preamble should do the trick:

\DeclareSymbolFont{mysymbols}     {OMS}{cmsy}{m}{n}
\SetSymbolFont{mysymbols}  {bold}{OMS}{cmsy}{b}{n}
\DeclareSymbolFont{myoperators}   {OT1}{cmr} {m}{n}
\SetSymbolFont{myoperators}{bold}{OT1}{cmr} {bx}{n}
\DeclareMathSymbol{\forall}{\mathord}{mysymbols}{"38}
\DeclareMathSymbol{\exists}{\mathord}{mysymbols}{"39}
\DeclareMathSymbol{\pm}{\mathbin}{mysymbols}{"06}
\DeclareMathSymbol{+}{\mathbin}{myoperators}{"2B}
\DeclareMathSymbol{-}{\mathbin}{mysymbols}{"00}
\DeclareMathSymbol{=}{\mathrel}{myoperators}{"3D}


Here's the full MWE:

\documentclass{article}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage[lining]{libertine}
\usepackage[libertine,cmintegrals]{newtxmath}
% \usepackage{ucs}% not compatible with utf8 inputenc - see output!
\usepackage{amsfonts,amsmath,amssymb}
\DeclareSymbolFont{mysymbols}     {OMS}{cmsy}{m}{n}
\SetSymbolFont{mysymbols}  {bold}{OMS}{cmsy}{b}{n}
\DeclareSymbolFont{myoperators}   {OT1}{cmr} {m}{n}
\SetSymbolFont{myoperators}{bold}{OT1}{cmr} {bx}{n}
\DeclareMathSymbol{\forall}{\mathord}{mysymbols}{"38}
\DeclareMathSymbol{\exists}{\mathord}{mysymbols}{"39}
\DeclareMathSymbol{\pm}{\mathbin}{mysymbols}{"06}
\DeclareMathSymbol{+}{\mathbin}{myoperators}{"2B}
\DeclareMathSymbol{-}{\mathbin}{mysymbols}{"00}
\DeclareMathSymbol{=}{\mathrel}{myoperators}{"3D}

\begin{document}

$f\left( \mathbf{x}|\theta\right) =\underset{h\left( \mathbf{x}\right) }{\underbrace{1}}\cdot\underset{g_{1}\left( T\left( \mathbf{X}_{1}\right) |\mu\right) }{\underbrace{\left( 2\pi\sigma^{2}\right) ^{-n/2}\exp\left[ -\frac{n\left( \overline{x}-\mu\right) ^{2}}{2\sigma^{2}}\right] }% }\underset{g_{2}\left( T\left( \mathbf{X}_{2}\right) |\sigma\right) }{\underbrace{\exp\left[ \frac{-1}{2\sigma^{2}}\sum\left( x_{i}-\overline {x}\right) ^{2}\right] }}%$
$\sqrt{\exp\left[ \frac{-1}{2\sigma^{2}}\sum\left( x_{i}-\overline {x}\right) ^{2}\right]}$

$\forall \exists \pm = + -$
\end{document}


Note that combining ucs with [utf8]{inputenc} produces warnings in the output. I've commented out ucs here but see the compilation output for other options.

The result:

• is it also possible to use "=" symbol from computer modern? – mert Apr 21 '14 at 12:35
• @mert See my edit. Sorry, I didn't have time to find all the bits I needed to complete this earlier as I had to go out. This code should, I think, meet all the desiderata. – cfr Apr 21 '14 at 14:56
• @mert No problem ;). – cfr Apr 21 '14 at 18:11