Towards Precise Scaling Laws for Video Diffusion Transformers