A typical test would consist of using an inexpensive digital VOM to test audio bandwidth of a transceiver.
First, you must calibrate for the deviations of the sound card and VOM from a 0dB reference. To do this, play back the tones INTO THE DESIRED LOAD, and note the voltages at each tone. Now, you can tell how much loss at discrete frequencies is being added by the relevant audio circuitry.
Impedance has a dominant effect on absolute voltage level. However, to give one an estimate of existing voltage, the following table lists values in dBm vs. voltage at 600 ohms.
|dBm at 600 ohms||Volts RMS|
Of course, you might calculate things yourself by realizing that:
dB = 20*log10(V/V0)
W = 10^((dBm-30)/10)
V = sqrt(P*R)
The important thing is typically the amount of level change over frequency, rather than absolute level (unless you can’t get proper output on any frequency). Some data from years ago on a narrowband (12.5kHz) FM radio is shown below:
real audio bandpass of 12.5kHz RF bandwidth FM transceiver
I haven’t made any effort recently to test similar audio coupling circuits. If your audio response is significantly worse than this graph, you might expect to have trouble with digital modes, especially digital SSTV, which uses 2.3..2.5 kHz bandwidth. You also would experience power dropoff on PSK31, which should be compensatable with audio level changes.