I'd be interested in respective pointers to actual vendors of such devices. Searching the net I found lots of sound velocity measurements for solids and liquids, but not for gaseous media - apart from general mentionings of the technique.
A high-precision device for gases was described here:
M. B. Ewing, M. L. McGlashan, and J. P. M. Trusler, The speed of sound in gases. I. A cylindrical interferometer and the speed of sound in argon and krypton.
J. Chem. Thermodynamics 17 (1985) 549–559.
Whether this is useful to you depends, among other things, on what you call “high temperature”. Which densities do you expect? If the density is low and your gas is not too far from ideal (and does not undergo chemical decomposition), calculation from an equation of state may be a reasonable alternative.
thanks for the pointer to the scientific publication. But I was rather researching actual vendors of such systems - sorry that I did not elaborate this intention more directly. I change the description accordingly in a moment!
The following information may be useful for your research
In-line ultrasonic flowmeters for gases by the transit-time differential method-
The same principle as for liquids The principle of transit-time differential measurement for gases is the same as that for liquids, except that the sound frequency is lower (typically 100 kHz).
REF: Fundamentals of Ultrasonicflow measurement for industrial
Commercially, (ultra)sound velocity can be measured with equipment for non-destructive testing. The manufacturers of non-destructive testing equipment often have high-temperature probes and delay lines with cooling water.
I would not like to advertise any of them, but I hope that you can find needed solution on the following pages: http://www.olympus-ims.com/en/applications/high-temperature-ultrasonic-testing, or http://www.ge-mcs.com/en/ultrasound/transducers/1221-spclapplhigh-temperature.html.
Naturally, one can think of many additional (smaller) problems to be encountered in your measurements, e.g.if the temperature range of probes is not completely satisfactory then the use of delay lines with cooling water is needed; or if the gas is in container then the some kind of impedance matching is to be expected, etc.