Generally, a guideline to expose daylight-balanced color neg film in tungsten (low color temperature) light is to overexpose with (almost) the filter factor of corresponding filter, even if the filter is not used, to place all of the R, G, B exposures on the linear part of curve. If this is not done, blue ends up on the toe, having lower contrast and lost definition in shadows, creating color crossover. This can be quite problematic, because usually tungsten-lighted interiors also have a low level of light, where we need higher ISOs, so exposing a 800 ISO film at ISO 400 or even 200 to get good results may not be possible. Push-processing C-41 works at some level by boosting contrast of exposures placed on toe, but, now I'm chasing for a method to "push" process the blue-sensitive (yellow) top layer more than green and especially red-sensitive layers. I visualize the problem in attached curves. When processed normally, blue is much on the toe and thus lower in definition and contrast. Normal push processing won't help much in this situation, as it boosts also red contrast, driving the crossover problem even worse. I find that a good thing to do is find a processing "fault" that normally causes crossover to another direction, and combine this with push processing. The error should be such that creates bluish cast in the whole image and especially in highlights in the final positive print. But is it easy to make the top layer develop more than underlying layers, for instance, pH, temperature and time variations?