Background Fundamentals 2020
Taught by Mike Seymour
- Course Number:
- BKD250
- Software Version:
- Original Run Date:
- January 2021
- Duration:
- 3 hours 56 minutes
vfx
production
new
Background Fundamentals seeks to provide you with information and insights on a wide range of topics. Unlike other courses, the classes are mostly stand-alone. In this offering, we look at three primary topics over five videos: on-set LED virtual production, AI machine learning and digital humans, both traditional CGI and neural rendering. This course is aimed at artists and working VFX professionals so there is little maths, but a lot of helpful advice and lessons gleaned from experience.
Taught by Dr. Mike Seymour, this course builds on one of the more popular series of courses that dates back to the inception of fxphd. As always, Mike has been fortunate to be assisted in creating these classes by some of the world’s best teams and experts.
Taught by Dr. Mike Seymour, this course builds on one of the more popular series of courses that dates back to the inception of fxphd. As always, Mike has been fortunate to be assisted in creating these classes by some of the world’s best teams and experts.
course syllabus
Class 1: LED Virtual Production
A discussion of the issues involved with LED studio volumes, or LED sound stages, made popular by productions such as The Mandalorian. This is one of the fastest growing areas of virtual production, with hundreds of stages coming on line or being planned.
Class 2: Face capture
One of the most important aspects of avatars and digital human animation is facial capture. In this class, we explore three different levels of capture from iPhone to high end professional specialist rigs from Faceware and Cubic Motion.
Class 3: Machine learning and Recurrent Neural Networks
An exploration of machine learning and recurrent neural networks using the example of speech synthesis. This class is a sister class to the neural rendering class (class #4) which goes further into Convolutional Neural Networks (CNN) and Generative Adversarial Networks (GANs).
Class 4: Neural Rendering
While traditional approaches to digital humans using scanning and modelling continue to be refined, a whole new area of AI and Machine Learning is sweeping the industry. This class provides a background in Convolutional Neural Networks (CNN) and Generative Adversarial Network (GANs), and the various approaches to inferring synthetic human faces.
Class 5: Traditional Digital Humans
A comprehensive look at the stages of producing a high end digital human from LightStage scanning to final render. Building on the last few years of experience with the Digital MIKE project, we go behind the scenes and explain the theory and the steps involved.