Thank you all for the feedback! That went smoother than I had feared
Iโll focus on finishing the VST plugin scanner this weekend.
As a few mentioned, Multi-Threading is in fact needed for lowering peak CPU usage, and thus will likely solve crackles on slower setups. However, it will also add roughly one buffer of latency. This is not as bad as it sounds: Often you can lower the buffer size a notch (for lower latency) because of multithreading efficiency, and this way multithreading makes up for itself.
Since we are trying to use very low buffersizes, itโs important to make sure the hardware and audio interface runs smoothly to avoid any clicks and pops in the audio. You can check your PC for DPC latency which can occur even on the fastest PCs and make it unsuitable for real-time audio, because of any unoptimised hardware component/driver, using LatencyMon (freeware):