I made this interview with Rampage from Paragon. It is in Polish but it has English subtitles:
I record the video and put open it in Unreal as a Media Player. I placed it on a plane as a backround and put Rampage character in front of it. I had two shots: wide with both of us and close up for Ramapge. Bottom part of the wide shot “the table” I put on a plane in front of Rampage.
I record Rampage voice and add “monster” effect to it. I repeat Rampage lines in front of a camera with Faceware software and Glassbox ue4 client to record my facial expression to Unreal sequencer. I map Faceware bones data to Rampage face bones.
I had to record one sentence at a time to be able to sync face movement with voice. I put every recorded sentence in sequencer. Unfortunately at seams Rampage character was glitching. I had 3 different shots so I could edit off the glitching part.
I render both shots in Unreal sequnecer. Redering wasn’t real time but the video was playing real time. That’s why I had me with very fast movement (Benny Hill style) while rampage with normal movement rate. I had too slow down my video to 30% to match rendering speed. It still wasn’t perfect. I split wide shot in the middle so I could have Rampage on the right side from different shot and me on the left side from another one, that fix the problem since we didn’t physically overlap.
In the video there are mainly two shots: me and Rampage. Backround of Rampage shot was just a still texture and I didn’t put “my close up” shot through Unreal at all.
I use DaVinci Resolve to put everything together but DaVini didn’t work with Sequencer render Interlacing. I had to finally use Premiere Pro.
I finnally did some color correction and voila.
I did it in two weeks that was the trial period for Glassbox software, Faceware have one month trial.
I had a great time doing it, and I think I have a nice message to tell.