# Symbol with overline and subscript leads to wrong kerning

I stumbled over a problem, which I assume should be quite common, but I couldn't find a satisfying answer (e.g. this and this are not what I am looking for).

I'd like to get an overline over a symbol, but not over its subscript. Though this is no problem, the grouping destroys the kerning of the subscript. Consider the example:

How can I get the overline just over f like in the second line, but the kerning of the subscript of the first line?

Of course I could adjust it manually with negative spacing, but I wonder if there is a more generic solution.

## MWE

\documentclass{article}

\usepackage{amsmath}
\usepackage{helvet}
\usepackage{sansmathfonts}

\begin{document}

without overline:
$f_t$
with overline and subscript outside:
$\overline{f}_t$
with overline and subscript inside:
$\overline{f_t}$

\end{document}


You can use \bar or adjust the space after \overline

\documentclass{article}

\usepackage{amsmath}
\usepackage{helvet}
\usepackage{sansmathfonts}

\begin{document}

without overline:
$f_t$
with overline and subscript outside:
$\overline{f}_t$
with overline and subscript inside:
$\overline{f_t}$
with bar and subscript outside:
$\bar{f}_t$
\setbox0\hbox{$f_t$}
\setbox2\hbox{$\overline{f}_t$}
$\overline{f}\kern\dimexpr\wd0-\wd2\relax_t$
\end{document}


If you need this more than once, something like

\newcommand\mean[2][]{%
\overline{#2}%
\ifx\relax#1\relax\else
\sbox0{$#2_{#1}$}%
\sbox2{$\overline{#2}_{#1}$}%
\kern\dimexpr\wd0-\wd2\relax_{#1}%
\fi}

with macro
$\mean[t]{f}$

• @thewaywewalk obviously if doing this more than once you would wrap the measuring in a two argument macro use (although personally I think I'd use \bar) – David Carlisle Jan 3 '17 at 11:04
• @thewaywewalk no that is wrong, \bar (unlike \overline) is a fixed font glyph the length does not depend on the width of the base. But I added a macro version for \overline to the answer. – David Carlisle Jan 3 '17 at 11:12
• You're right, I got confused. However a fixed length is also not desirable for my case. Thanks for the macro! – thewaywewalk Jan 3 '17 at 11:24