Be prepared for some eccentricity in the more advanced tests depending on sample rate mismatch. I don't know of any convenient way of doing fractional resampling, unfortunately.
In the past I would occasionally edit the sample rate in WAV headers to an oddball value with a hex editor, which gives a decent amount of resolution. (At 48 kHz, 1 Hz is about 21 ppm.)
To find out what sample rate you need, you would e.g. have the DUT play a 12 kHz tone (the J-Test signal will do, too) and use RMAA's Spectrum Analyzer function with a reasonably high FFT length (say, 131072) and Beta = 20 or so. Then you can zoom in and pinpoint the recorded frequency.
Say you find it's 12035 Hz and recording was done at 48 kHz. Then the new value for fs should be 48000 Hz * 12000/12035 ~= 47860 Hz.
The modified file can then be dragged into an Audacity session at 48000 Hz. Verify that SRC quality for export is set to Best Quality (slowest), and export as WAV 24 or 32-bit int.
(Note: When using Audacity for recording, a "Windows WASAPI" input device must be used for >16 bit samples. You can also custom compile it to include ASIO support. Failing all that, RMAA's General settings have a checkbox "Save resulting WAV file".)
