Edit: Spoke too soon. Did find one paper[1] (though I can't read it) which uses it to compare the production of 's' and 'z' sounds. Would love to know if there are any more papers though.
We tried to image sound waves, but the density gradient for sound is much less than that produced by changes in temperature. We attempted to make a resonant chamber and use a high-intensity ultrasonic source, and make a standing wave that we might capture photographically, and while we saw something once we could not produce it. It's likely that the ultrasonic source we were using was gradually degrading.
The research paper you link to almost surely uses the heat differences to see the jets of air coming out of the mouth, and not actual sound waves.
> The research paper you link to almost surely uses the heat differences to see the jets of air coming out of the mouth, and not actual sound waves.
Regarding the paper, while I can't read it myself, I imagine they chose to look at fricatives precisely because they're the result of turbulent airflow which is probably ideal for Schlieren imaging rather than a different speech sound which would be more wave-like. If the image is the result of heat differences, I wonder if it could be improved further by changing ambient temperature or temperature inside the mouth, or alternatively if the subject could be asked to inhale sulphur hexafluoride first to increase the density differential? (Edit: Or, perhaps, introduce a 'uniform' (laminar) thermal source along the direction you're interested in, so that when a sound wave propagates, the resulting density differences would be much more pronounced? I'm not sure if that makes any sense, as a layperson this is pure speculation on my part...)
I believe this is an image from the paper: http://www.sciencephoto.com/media/89151/view
Thanks again for the info though, really fascinating. Hopefully we haven't reached the limits of this technique yet, and it can still be taken further.