Recent advancements in novel view synthesis, driven by Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS), have significantly improved rendering quality and speed. However, the reliance on accurate camera parameters, such as intrinsics and extrinsics, remains a challenge, often leading to performance degradation. This survey explores emerging methods that eliminate the need for precise camera parameters, aiming to improve both practicality and performance in NeRF and 3DGS applications. We highlight these approaches as motivation for further research.