Well at the moment UE4 does not have a clear design pathway when it comes to animation systems in general but does have the building blocks with the inclusion of blueprints where more complex animation migration pathways can be added in parallel with the main project build. Most engines requires a more direct bottom up development based on the need to solve problems today with out the overhead requirements of top down design that requires the developer to think in terms of what is “needed” say four or five years from now.
With in context though our project does not need facial animations as a primary requirement but as part of our overall development design agreed that adding a level of personality would be a nice design detail hit that as an option should not effect forward progression that does not effect the bottom up development that can be retro fitted as to structure developed with in a top down approach. In other words the design pathway had to fulfill our needs that exceeded the current demonstrated ways and means but created a sound foundation that the rest of the development team could work with little or no interruption.
That bit of an intro out of the way and not knowing the full bottom up requirement of your project our starting point was not about the requirements for a video game but what tools and production values are available as to animation requirements for feature films and getting it done is not about tools that are unique to Unreal 4 but rather tool usage as part of the development pipeline as to the need to author the required data set. For example my tool of choice is MotionBuilder so any questions I have as to data creation is best asked with in Autodesks MotionBuilder forums as to figuring out how to use the data can be figured out the migration as part of the top down requirement and via the discovery process.
To get a quick start we started using Daz Studio and the G2 frame work but with the introduction of Genesis 3 we were so impressed what we could do via discovery that we optioned to purchase a developers license as the base rigging fit the needs of our coders and added an extension to ways and means as part of our character development pipeline. At the very least it’s a good place for you to start to figure things out based on your own questions that can only be answered via proof of concept and discovery.
As far as proof of concept goes here is an example of what we have so far as long as it’s keep in mind that the “proof of concept” is about the work in progress in proving the theory will work disregarding the ideals as to the needs and demands of the simplicity of a video game.
The conclusion we came to is with blueprint availability and an FBX pipeline there is no difference between the needs of animation assets needed for animated film and a real time video game with the exception that for film the art asset is disposable once the render is acquired and once you have a workable pipeline and foundation in place you can add what you want later on with out having to restructure .