I am curious about the "the camera settings ate my nice sunset colors" idea. I have seen this myself, you can "get rid" of those nice colors of a sunset by neutralizing the white balance.
What I don't understand, is do cameras have a setting where white balance is not touched, i.e. a WB setting which is faithful to the light that was there, as recorded by the CMOS sensor, without interpretation / free parameter called WB ?
Good question. This goes to one of the core differences between RAW and jpeg. White Balance ["WB"] is nothing more than the camera's best *guess* at what colour temperature (a) existed when the shot was take and (b) the photographer wanted. It is a fancy way of saying, "in this scene, what is white/grey/neutral?"
In RAW, the camera stores this WB as a number attached to the file. Because the RAW file is raw, it is just a series of values read off every sensor. Nothing has been done to that data.
In your RAW converter, the software reads that number - think of it like a recipe sent home by the grocer with the food on a separate piece of paper - and then applies that information to decide how to convert and interpret the colour data coming off the sensor. (This is a massive over simplification). But basically, you can follow 'the note', or do whatever you like, because the note was simply a suggestion, and the raw ingredients (the data) has not been touched.
Because 'correct' WB is by default thought of as daylight (about 5600 degrees kelvin), the converter will interpret the colour data to try to make things look like the scene was lit by daylight.
In turn, because daylight is much 'cooler' (ie: less light on the yellow/orange end of the light spectrum) than sunset or sunrise, a sunset photo converted in this way will look bland and washed-out, because the software has interpreted the scene as having much less yellow/orange light than was actually there.
This, in a nutshell, is why most basic workflows *start* with a neutralized or boring conversion of a sunset.
In Raw processing, we can simply change the way in which the data is interpreted, by changing the WB or colour temp settings. Either warmer or cooler, until we get the colour rendition we like best.
With in-camera jpegs, all of this is done by the camera, according to whatever WB has been set on the camera at the time of taking. If you tell it it's a sunset, it will make the picture warmer. However, because the processing is done, what you get is the camera's interpretation of the colour in the scene, with much less ability to change that after the fact.
The recipe has already been applied to the food, if you will follow my earlier analogy.
Hope that gets at your question.