Demystifying the Use of Compression in Virtual Production

Vibhoothi Vibhoothi, Julien Zouein, François Pitié, Christopher Nash, James Bentley, Philip Coulam-Jones, Anil Kokaram

Virtual Production (VP) technologies enhance on-set filming and live concerts through high-resolution, high-brightness LED panels. Due to technical challenges like synchronization and color balancing, the industry prefers resource-intensive lossless compression for displays. Modern lossy standards (AV1, H.265) could maintain quality while reducing resource demands, yet no agreed methodology exists for assessing their in-camera impact. We present a method comparing lossless and lossy compressed footage displayed through VP screens and recorded in-camera. Our assessment compares HAP/NotchLC/Daniel2 intermediate codecs with hardware Nvidia Encoders (NVENC) using AV1/HEVC/H.264 compression at bitrates from 2–2000 Mbit/s with various GOP sizes using different perceptual quality metrics. Results demonstrate hybrid codecs achieve equivalent quality to intermediate encoders at significantly lower bitrates and storage requirements.

Print ISSN
Electronic ISSN
2160-2492
Published
2025-05
Content type
Original Research
Keywords
virtual production, compression, hevc, h.264, av1, hap, notchlc, daniel2
DOI
10.5594/JMI.2025/MJPY7062