I made a new release, containing 4 and 5-phase generation:
https://github.com/diglet48/restim/releases
To use this features you need a stimbox with 4 audio channels. I will explain below how the setup works.
Attach the wires as:
Electrode 1: Channel 1 black.
Electrode 2: Channel 1 red + channel 2 black.
Electrode 2: Channel 2 red + channel 3 black.
Electrode 2: Channel 3 red + channel 4 black.
Electrode 5: Channel 4 red
it also works if all red/black are swapped.
At this point there is no recommended electrode setup, but I recommend ordering from top to bottom just to make sharing scripts a bit easier.
Configure MFP like this:
Go into settings and configure the audio device. I recommend using WDM-KS, I saw windows audio do strange things in other modes. WDM-KS opens the device in exclusive mode, so it will only work if there are no other programs using the audio device. The channel map depends on your specific audio device, if you click on the button it opens a dialog which lets you easily search for the correct channel map (have a multimeter handy).
Make sure 5-phase is selected on the main screen:
You can calibrate the signals on the 4/5-phase calibration tab. If you want more power to a specific electrode, increase the skin resistance. Want less power, decrease the skin resistance. (wording may change in future release). The transformer resistance affects how much signal from one electrode bleeds through to the other electrodes. It is very difficult to judge the correct value, I think it should be somewhere between 1-2x the average skin resistance.
You can use the sliders to change the power (funscripts will overwrite the slider values). The phase diagram and the pattern generator are not functional for 4/5-phase. There currently is no visual indicator for the signals.
I made a few patterns (download the zip with test patterns in the release). The test patterns are very simple, a simple rotation. Each file repeats the same pattern over and over again, the files differ in intensity and speed. I recommend renaming a random video to match the calibration pattern name, and then opening that video in OpenFunScripter, so you can see what's going on. This also enables you to make your own patterns.
The 5 funscript axis are simply the 5 electrode power levels. If you try do so something that's impossible, restim will project to the most similar valid signal.
At this point there is no funscript conversion, need to do a lot more testing to figure out how that should work.