Week 7 Part 1 - Nose and Basic Cleanup

-

 This week I concluded my work on the face retargeting... or so I thought! The nose was missed out during the eye section! I spent some time working on it as I had made a note last week that the nose being static was throwing off some of the mouth movements. The nose doesn't move a huge amount but I made sure to key in where the nostrils flare out; though I'm not sure the limited number of nose controls really allowed for the most accurate movement. I kept to the established eye keys for this as they generally matched the biggest movements of the nose.

At this stage I felt I had done all I could in the retargeting area so I put all of the retargeted sections of the face together and previewed the final result.


FINAL RETARGETING

Generally I thought this came out quite good all things considered! The biggest problem areas I could spot was the eyeball tracking and the tongue; both of which were advised to be manually animated anyway. Next I went through and retimed it to match the 24 fps framerate of the video. I resampled to a 1.0 timestep because there are lots of movements across the face and I didn't want to lose the data associated with it and risk it looking too floaty and smooth. I set the playback speed to 24fps to get a more accurate look at the speed relative to the video clip.


RETIMED




Comments

Post a Comment

Popular posts from this blog

Week 2 - Brow and Eye Tracking

-
Image
 Phew! What a tricky week! We jumped right into the facial tracking component of the performance capture process this week; beginning with tracking the brows and eyes.   In my scene, Miri's brows pull a variety of expressions - lifted high, settled low, and an anxious lifting of the tops. The tracker points on the eyebrow tails were placed on her temples and as a result were not visible for most of the shot, so I elected to track an eyebrow hair instead. It was a tricky and time consuming process. The eyes were less challenging, though I am considering returning back to fine tune them after some feedback if possible. All trackers painted on were visible except for the upper eyelid, but I was able to apply trackers to the eyelash root on both eyelids and it seemed to work well. The main issue I encountered was that the upper eyelid felt like it moved side-to-side a lot and I genuinely couldn't tell if it was a result of the eyeball moving around a lot or due to a drifting track...
Read more

2025 Week 1! New Year, New Everything!

-
 One project down, another one immediately begins! Such is the way of university, but seeing as this is our last year I know I need to go in with 200%. For this project we are reworking our whole workflow to incorporate/entirely replace our old software with Shogun and Unreal Engine 5. I am TOTALLY new to this - I have opened Unreal a total of ONE time(s) and it crashed on me within minutes. I'm a VFX student so our use of game engines has only really just kicked off and I'm slightly behind the times in that regard.  Overall though, I am really excited to have some way to not only use the more advanced technology at our disposal (the new suits, the new screen and camera setup, the Shogun capture tech etc) but to learn how to use Unreal 5 in any capacity; for motion capture or any other use. If I have discovered anything about my learning style it is that I do much better learning how to use new software by having exercises or projects to use it for, so this presents a fantasti...
Read more

Week 9 - Trimming, Knees, and Feet

-
Image
 This week I moved on to the home stretch - cleaning the body. I found the process of trimming the frame range straightforward, though very time consuming to check all of the keys and make sure nothing was missed. I was really lucky to get a shot with relatively little foot and knee movement - she doesn't walk around in my shot and stands on the same spot the whole time. I actually ended up moving straight to the feet as her knees had no glaring issues and the floating feet were throwing me off more in terms of not being planted on the ground like hers. Before cleanup For feet I began by shifting the model so that she was standing at zero, and then I added a grid to determine how much clipping was occurring. Using a split screen with perspective on one side and orthographic on the other was super helpful as it allowed me to cross check all of the technical changes (exact number values etc) with the visual reference video. I added the necessary lift around frame 1920 for the last al...
Read more