Wall Street Journal is reporting that DreamWorks and NetApp began collaborating on predictive analytics tools for artists working on animated features. These tools can help make production of digital assets less time-consuming and pricey.
DreamWorks CTO Jeff Wike told the Journal that the tools can help businesses drive down costs while growing the business. Currently, artists have to render digital assets from scratch, which can demand up to half a billion files per film or 600 terabytes of storage and 120 million hours of rendering per movie, says the article. The companies are looking to predict what assets artists will need by building a dataset of the rendering process, including data on what assets are created, what infrastructure they are located on and how much storage was needed.
According to the piece, DreamWorks currently employs 800 artists worldwide and 200 IT and software development staff. Its on-premise data center consists of 48,000 processing cores. It also has a backup and recovery data center in Las Vegas and accesses data center space at a few Comcast Corp. facilities and uses various public cloud services such as Amazon’s Web Services.
The dataset helps reduce files that are stored but never used in the cloud, improve latency in data transfers and estimate compute power and storage needed in cloud and on-premise data centers. “From a pure data and computational aspect … you get nothing for free. Every blade of grass, every eyelash, every drop of rain has to be built from scratch, time and time again,” said Wike.
DreamWorks isn’t the only company that has been adding new capabilities into its portfolio. Disney Research, Pixar Animation Studios and UCSB also worked together to speed up rendering time with AI. Technology experts trained a deep learning convolutional neural network to eliminate “noise” in images and make them appear as if they were computed with more light rays. As AI and analytics capabilities evolve, open source editing and rendering tools have also increased the pool of professionals working with and improving tools outside of the big studios.
You can read the full report here.
