My comment last month nagged at me, until I remembered an answer I gave at Rounded box around placeholder text that supports line breaking (in which I was trying to eliminate this gapping effect), and decided it was worth trying to adapt to enhance it. What I do is paint the natural spaces of alternating letters in alternating colors (here cyan!80 and cyan!60). I paint spaces their natural width in cyan and periods in cyan!95. I found it got too confusing if I chose vastly different colors.
Horizontally stretched lines can be quickly observed with white gaps between words. I chose a \vgap between lines of 2.5pt, since that value set the cyan color rules at a similar height to the letters themselves (if the gap is too small, the cyan rules themselves affect the line spacing). Nonetheless, vertical gap differences can be noted.
I should point out that this approach will inhibit hyphenation, and so it may not be a direct comparison to the unmarked text, where line breaks are involved. By the same token, while I have applied it to the complete document text in this MWE, one can selectively include partial text (it is of a macro form, embracing the text in a {} group) and possibly acquire the desired data by highlighting only a small fraction of the overall text.
\documentclass{article}
\usepackage{censor}
\usepackage{xcolor}
\newlength\vgap
\vgap=2.5pt% GAP BETWEEN ADJACENT LINES (THIS VALUE CONFORMS TO LETTER SIZES)
\textheight=3.03in
\makeatletter
\def\mystrut{\rule%
[\dimexpr-\dp\strutbox+.5\vgap]{0pt}{\dimexpr\ht\strutbox+\dp\strutbox-\vgap}}
% SET ALL LAPS TO 0pt...
\periodrlap=0pt\relax
\afterperiodlap=0pt\relax
\lletterlap=0pt\relax
\rletterlap=0pt\relax
% EXCEPT THE SPACE LAP, WHICH IS SET TO NATURAL SPACE WIDTH
\setbox0=\hbox{ }
\afterspacelap=\wd0\relax% THE NATURAL SPACE WIDTH
\renewcommand\censorrule[1]{%
\protect\colorbox{cyan}{\mystrut\rule[\censorruledepth]{#1}{0pt}}}
\newcounter{colindex}
\edef\cencolora{cyan!80}
\edef\cencolorb{cyan!60}
\def\chcolor{\stepcounter{colindex}\ifnum\thecolindex=2%
\def\cencolor{\cencolorb}\setcounter{colindex}{0}\else%
\def\cencolor{\cencolora}\fi}
\chcolor
\renewcommand\@cenword[1]{\colorbox{\cencolor}{\mystrut#1}\chcolor}
\def\censordot{\colorbox{cyan!95}{\mystrut.}}
\let\marktext\xblackout
\makeatother
\fboxsep=0pt
\parindent 0in
\flushbottom
\begin{document}
\marktext{Note however, that auto-hypenation will NOT work with this approach.
Can this procedure go across paragraphs boundaries?
Why yes it can.
Gaps can arise if glue is stretched too far.
NEXT LINE STARTS A PARBOX:}
\parbox{.8\textwidth}{\marktext{%
This tests marking a multiline block of text. This tests marking a multiline block of text.
This tests marking a multiline block of text. This tests marking a multiline block of text.
This tests marking a multiline block of text.}}
\marktext{%
NOT IN A PARBOX:
This tests marking a multiline block of text. This tests marking a multiline block of text.
This tests marking a multiline block of text. This tests marking a multiline block of text.
This tests marking a multiline block of text.}
\end{document}
In the global view, we can see that the vertical gap above and below the parbox has been squeezed out significantly.

In this zoom, we observe the horizontal stretching differences on different lines. The "f" on the last line is not squeezed... it naturally overhangs its bounding box.

If one looks closely, compressed text also exhibits visible cramping, as in this comparison of the letter "e", where the kern on the right has been abolished by an inter-word space.
uncompressed vs compressed
. . .
. . . . . . .
lua-visual-debug? – Torbjørn T. Mar 19 '14 at 22:20