[OS X TeX] AMS-LaTeX question

[Michael Sharpe]

On May 28, 2012, at 11:28 AM, Martin Berggren wrote:

> How about
> 
> \begin{align*}
> 	&= \frac{A} { \left( k - m \omega_{0}^{2} \right)^{2} + c^{2} \omega_{0}^{2} } \sqrt{ \left[ \left( k - m \omega_{0}^{2} \right) \left( m \omega_{0}^{2} + k - m \omega_{0}^{2} \right) + c^{2} \omega_{0}^{2} \right]^{2} + \left( m c \omega_{0}^{3} \right)^{2} }
> 	\\
> 	&= \frac{A} { \left( k - m \omega_{0}^{2} \right)^{2} + c^{2} \omega_{0}^{2} } \sqrt{ \smash[b]{\underbrace{ \left[ k \left( k - m \omega_{0}^{2} \right) + c^{2} \omega_{0}^{2} \right]^{2} + \left( m c \omega_{0}^{3} \right)^{2} }_{Q}} }.
> \end{align*}
> 
> On May 28, 2012, at 19:09 , Gary Gray wrote:
> 
>> \begin{align*}
>> &= \frac{A} { \left( k - m \omega_{0}^{2} \right)^{2} + c^{2} \omega_{0}^{2} } \sqrt{ \left[ \left( k - m \omega_{0}^{2} \right) \left( m \omega_{0}^{2} + k - m \omega_{0}^{2} \right) + c^{2} \omega_{0}^{2} \right]^{2} + \left( m c \omega_{0}^{3} \right)^{2} }
>> \\
>> &= \frac{A} { \left( k - m \omega_{0}^{2} \right)^{2} + c^{2} \omega_{0}^{2} } \sqrt{ \underbrace{ \left[ k \left( k - m \omega_{0}^{2} \right) + c^{2} \omega_{0}^{2} \right]^{2} + \left( m c \omega_{0}^{3} \right)^{2} }_{Q} }.
>> \end{align*}
> 

I think this can be improved a bit by adding an invisible \vrule before \smash[b]:

\begin{align*}
&= \frac{A} { \left( k - m \omega_{0}^{2} \right)^{2} + c^{2} \omega_{0}^{2} } \sqrt{ \left[ \left( k - m \omega_{0}^{2} \right) \left( m \omega_{0}^{2} + k - m \omega_{0}^{2} \right) + c^{2} \omega_{0}^{2} \right]^{2} + \left( m c \omega_{0}^{3} \right)^{2} }
\\
&= \frac{A} { \left( k - m \omega_{0}^{2} \right)^{2} + c^{2} \omega_{0}^{2} }\sqrt{\vrule width 0pt depth 5pt\smash[b]{\underbrace{ \left[ k \left( k - m \omega_{0}^{2} \right) + c^{2} \omega_{0}^{2} \right]^{2} + \left( m c \omega_{0}^{3} \right)^{2} }_Q}} .
\end{align*}

The difference in Computer Modern math at 11pt is small, but can be quite noticeable at some other sizes and with other math fonts. Like other extensible math objects, there is a discrete collection of available sizes that will not produce the same results in all cases.

Michael