First off I want to say the doppler effect implemented, when showcased properly from a static camera on the ground or as another train passes by, is fantastic. Absolutely great. However when moving any of the external cameras, even slowly, you'll get the following result which is absolutely terrible and completely immersion-breaking: I think a good adjustment to make this less cringey would be to just disable the doppler effect for sounds being emitted from my own train for the external cameras, or reduce its effect the quicker the camera is being moved. And to showcase how good it is when showcased properly (it's so good):
In addition to this, I want to point out that for some reason Rapid Transit train signals are not subject to Doppler effect for some reason.