Exploring Dynamic Novel View Synthesis Technologies for Cinematography

TL;DR

This paper investigates dynamic novel view synthesis techniques, particularly Neural Radiance Fields and Gaussian Splatting, for cinematography, demonstrating their ability to generate new shots and effects in film production.

Contribution

It explores the application of dynamic NVS methods in cinematography, providing insights into model selection and showcasing practical results with a short montage.

Findings

Demonstrated the potential of NVS models for cinematic effects

Compared different NVS techniques for dynamic scene synthesis

Provided a practical framework for model selection in film production

Abstract

Novel view synthesis (NVS) has shown significant promise for applications in cinematographic production, particularly through the exploitation of Neural Radiance Fields (NeRF) and Gaussian Splatting (GS). These methods model real 3D scenes, enabling the creation of new shots that are challenging to capture in the real world due to set topology or expensive equipment requirement. This innovation also offers cinematographic advantages such as smooth camera movements, virtual re-shoots, slow-motion effects, etc. This paper explores dynamic NVS with the aim of facilitating the model selection process. We showcase its potential through a short montage filmed using various NVS models.