Apologies if this is not possible to do but what I think this could be a good idea.
If i want to record at the cine crop, at the moment the cropmarks give me an indication of how to frame the video, but the whole image is still being recorded. I am not suggesting that the recorded video size is changed (i realise this would be much harder if not impossible to do) but that for the pixels covered up by the crop marks, are turned black. This way, the bitrate for the non cropped areas could be relatively higher because a portion of the field will have no detail visible.
A bit like how the qscale doesnt go very high if recording a shallow DOF because there isnt much detail in the scene to be encoded as its soft. If the cropped areas are fully black and so have almost no detail, the total amount of data per frame is less.