Light Field Docs

D3.1 Asset and capturing plan [M9]
This deliverable is part of the reporting for work package 3 “New Technologies for Asset Creation” and hence deals with the capturing plan for lightfield assets. Lightfields within SAUCE have two major purposes: a) To demonstrate the enhanced post-processing possibilities by capturing a partially volumetric object information and b) by introducing a new multi-dimensional perspective onto the spatio / temporal content of captured scenes. Consequently, the deliverable describes the methodologies for capturing 5D lightfields, the availability of the respective tools within SAUCE and the plan to when and what to capture. Lightfield assets will, unless explicitly stated and required due to usage restrictions on the asset itself, be made available to the public. Some, but not all, assets are required to be available M18 (compliant to D3.3 “Lightfield Assets for SAUCE”), hence the current deliverable will focus on the period between M10–M18.

D3.2 Analysis of Requirements and Prototypes of the Tools
This document contains the implementations and qualitative reports of the baseline tools for smart asset transformations for LF processing, based on the requirements of the partners.

D3.3 Lightfield Assets for SAUCE [M18]
This deliverable is part of the reporting for work package 3 “New Technologies for Asset Creation”. It introduces the lightfield assets created under work package 3.1 “Smart Assets via Lightfield Capture” and additionally includes those parts of the pre-processing pipeline required to make these lightfield assets available to and usable by the project partners (covered in work package 3.2 “Transcoding of Lightfield Assets”). In addition, the deliverable describes the asset storage implemented at USAAR and the sustainable storage maintained by DNEG to document how assets are preserved for further enrichment and usage. Since in the first half of SAUCE the emphasis has been on consolidating the work of the partners into a running lightfield pre-processing and processing pipeline, the majority of the described assets will be of the type 4.5D lightfield as defined in D3.1 “Asset and Capturing Plan”. The report, however, also documents first results in capturing 5D lightfields.

D3.4 Accelerated Tools for Creating Smart Assets [M24]
This document contains the implementations and reports of the accelerated tools for smart asset transformations for LF processing, based on the baseline tools described in the deliverable D3.2.

D3.5 Transcoders for SAUCE assets [M30]
This deliverable concludes the reporting for work package 3 “New Technologies for Asset Creation”. Aspects required to make lightfields applicable to current production workflows under the task 3.2 “Transcoding of Lightfield Assets” are introduced. The deliverable describes optimization and evaluation of several state-of-the-art image and video compression standards for both lightfield still images and videos [4D, 4.5D, 5D lightfields]. Specifically, the implementation of pseudo-temporal sequencing of lightfield assets compliant to H.265/HEVC encoders and extracting high resolution lightfield assets compatible with industrial standards such as OpenEXR and PSD are dealt with. In addition, the deliverable discusses experimentation using several recent formats such as JPEG Pleno, AV1 and XVC. The majority of the testing is performed on the assets defined and created in D3.1 “Asset and Capturing Plan” and D3.2 “Smart Assets via Lightfield Capture” respectively.