I do deal with it every day... I live in Argentina, no problem up to there, but in the region I live Ss are not pronuciated at all when they are at the end of the world, so most plurals and singulars sound the same for example (not everybody but a lot of people talk like this and many sing like this to)
I have a couple of options always at hand, the first one is the obvious, if he sings this way, it sounds this way. Sometimes the S isn't there at all so you can't do anything... when it is but not so presence you could let it out in many ways...
1- a dynamic eq working like a reverse de-esser, but for this you need to have the S... so not the more choosed.
2- a transient processor like http://www.waves.com/Content.aspx?id=286 where you can boost hi frec transients.
3- a comp, 1176 is the correct one for this task, most short release and not so short attack, hi ratio, parallel compression with eq or filtering maybe.
4- waves C1 sidechain, play with hi frec dynamics to get something out of them when the Ss are there, then boost high, so you will be atten highs with the comp most of the time and boosting with the EQ, when not high signals there the comp release (fast, again) and let the eq works boosting this Ss... As a teacher said, C1 is when there is some trouble, in other case a lot of better comp plugins for playing with...
5- an exciter, like aphex waves plugin, but in this case you will be boosting all the highs from the track that I assume you don't want.
I rarely use a de-esser with any singer here around, but being short in Ss allow you to boost a lot highs when needed and most of the time high freq when no Ss around make no harm and bright up te vox, and when you are short of Ss is also a good thing (not always, ofcourse)
Parallel compression with 1176 as said before enhances transients and intelligibility... sometimes I use it in choirs so I can keep it well down the voice but still get what they sing...
I hope this would be usefull, anything else or clarification just ask.
JS