Lauri Kettunen
Well-known member
Be warned that the looks are only an approximation of what you can get in post.
While your argument holds some water, still, this is not in contradiction with what Jim tries to say. This is a common issue in engineering and the standard approach is to solve it by iteration.
In tangible words, one creates first a look in RCX, which is, as you say, an approximation of what the image will eventually be. But, the keypoint is, it's a rather good approximation. So, after processing the file, if one finds the approximation wasn't precisely what he or she wanted, then one can go back and finetune the look. Thereafter one iterates until the end result is satisfactory.
And now, after this process one can exploit the look file with confidence. The key point is, although the preimage in RCX is not precisely the end result, it is close enough, implying if one wants to make further changes to an existing (iterated) look file, RCX is able to reveal in which direction the end result will go. Consequently, the user can be confident on what he or she will gain.
For another example of such a conceptual approach, just think of driving a car. If you turn the steering wheel, the car seldom goes precisely into the direction the steering wheel "points to". Instead, there is a feedback loop and the driver subconsiously makes some minor corrections every time when turning the wheel. This is why driving in the winter is more challenging than driving in the summer time. In the winter the convergence of the iteration takes more cycles. But still, one would not argue the steering wheel is only an approximation of setting the direction.