Neural Radiance Fields (NeRF) enable novel view synthesis of 3Dscenes when trained with a set of 2D images. One of the key compo-nents of NeRF is the input encoding, i.e. mapping the coordinates tohigher dimensions to learn high-frequency details, which has beenproven to increase the quality. Among various input mappings, hashencoding is gaining increasing attention for its efficiency. However,its performance on sparse inputs is limited. To address this limitation,we propose a new input encoding scheme that improves hash-basedNeRF for sparse inputs, i.e. few and distant cameras, specifically for360◦ view synthesis. In this paper, we combine frequency encodingand hash encoding and show that this combination can increasedramatically the quality of hash-based NeRF for sparse inputs. Addi-tionally, we explore scene geometry by estimating vanishing pointsin omnidirectional images (ODI) of indoor and city scenes in orderto align frequency encoding with scene structures. We demonstratethat our vanishing point-aided scene alignment further improvesdeterministic and non-deterministic encodings on image regressionand NeRF tasks where sharper textures and more accurate geometryof scene structures can be reconstructed.