Regarding the measurement of the output impedance, the basic idea is to measure the output voltage of the device unloaded, and then (without changing anything else) with a known load connected. From the reduction of voltage with the load, the output impedance can be calculated as Rload * ((Vnoload / Vload) - 1). Therefore, you need:
- a splitter (1 female and 2 male 1/8" jacks) or other suitable cable to be able to drive some load and measure the voltage at the same time
- some kind of test load; ideally, a resistor, but even a headphone can be suitable, if tested at a frequency where it has mostly resistive impedance (like ~2 kHz for most full size dynamic headphones)
- something to measure the voltage with. It can be a cheap digital multimeter (note that some of these only work reliably over a limited frequency range), or a sound card line input, which you should already have. It is not necessary to be able to measure absolute levels accurately, since only a ratio of voltages will be needed
In case you only have the splitter, and no resistors or DMM, connect the two male jacks to the headphone output of the card to be tested, and the line input of the card used for recording. Make sure that the mixer settings are correct for loopback recording, and that you can play a test tone at a reasonable (not very low or high) volume. Generate a sine wave at 2 kHz with some suitable software, and create a loopback recording of it, with and without the headphones connected to the female jack of the splitter. Use any audio editor of your choice to measure the RMS levels of the recorded tones. You can then calculate the output impedance with the above formula (Rload = the impedance of your headphones).