We current a novel differentiable rendering framework for joint geometry, materials, and lighting estimation from multi-view photographs. In distinction to earlier strategies which assume a simplified atmosphere map or co-located flashlights, on this work, we formulate the lighting of a static scene as one neural incident gentle area (NeILF) and one outgoing neural radiance area (NeRF). The important thing perception of the proposed technique is the union of the incident and outgoing gentle fields via physically-based rendering and inter-reflections between surfaces, making it doable to disentangle the scene geometry, materials, and lighting from picture observations in a physically-based method. The proposed incident gentle and inter-reflection framework might be simply utilized to different NeRF programs. We present that our technique cannot solely decompose the outgoing radiance into incident lights and floor supplies, but additionally function a floor refinement module that additional improves the reconstruction element of the neural floor. We display on a number of datasets that the proposed technique is ready to obtain state-of-the-art outcomes by way of the geometry reconstruction high quality, materials estimation accuracy, and the constancy of novel view rendering.