Investigating the Trade-off in Allocating Bits to Texture or Depth Information in Rate-constrained Multiview Video Coding