I'm using XeLateX, Unicode math, and the Latin Modern fonts (Latin Modern and Latin Modern Math). To me, it looks like subscripts in mathematical expressions are being properly (optically) scaled -- they are not merely small-size versions of regular characters, they are wider and have fatter strokes.
Here's some code
\documentclass{article}
\usepackage{unicode-math}
\setmainfont{Latin Modern Roman}
\setmathfont{Latin Modern Math}
\begin{document}
Here is some math: $x_2=2$.
\end{document}
and the output
In traditional TeX, the subscript scaling would be done by using a completely different font, like cmmi7. But the Latin Modern Math font only comes in one design size (10 pt, I suppose). So how is the subscript scaling achieved?