Animation and Penny’s Excellent Udemy Course

I have been learning even more animation with Penny De Byl’s Udemy class ‘Mastering 3D Animation in Unity.’ I am very impressed with her classes.

Penny starts out teaching you how to use raw mocap data for Unity’s Mecanim system, and adding material physics to objects to affect how the bounce or stick or slide. She goes into depth on keyframing, timelines, and tweening. She shows how to use colliders to trigger animations. She shows how the bones and joints rigging system works with humanoid models. She makes great use of Mixamo, which I have learned to love. I hope Adobe keeps this great resource available.

There’s a lot of useful tips on importing animations and fixing common problems. She shows how to add animations to characters, and to skin them to characters from other characters. She shows the problems that root motion, baking Y, baking rotation, and baking XZ are designed to address, and when you need them. She shows how to mix the translation from root animations, with the scripted translations and how to pass control of the translation around. There are so many techniques such as using anchors to align your animations, and putting animation events at the end of an animation.

Then she goes into blendshapes, which I already know a lot about, even though it’s been years since my doctorate. I spent enough hours on that project that I’m not going to forget the basics any time soon. But what I did learn was using the animation window to add blend shape trees, and strafe sets.

Another excellent use of anchors is with inverse kinematics. This time the anchor is an object the character might reach out to, or follow with his head.

All in all, I can’t recommend this course highly enough. Between this and the AI techniques course she has, I feel like I’ve learned most of what I need to know to develop a VR game. Most importantly, I’ve learned what’s hard and what’s easy. I’ve been enjoying adapting her projects to work in VR, and adding to them of course.

2017 in Review

Wow, again it’s been a year since my last blog post.  So I’ll make this post a synopsis of my year.  I did update all of my apps with the latest SDK, and some improvements all around.  I’m 95% done with my estimation app.  My goal is to publish it by Christmas, so I can focus totally on the exciting new stuff I can do with my new iPhoneX!  I’ve played around quite a bit with Maya and Unity this year, trying to get prepared for AR/VR opportunities, but I have to admit, I took off quite a few weekends to dive, read, dance, take cooking classes, and otherwise unwind.    

Now I have to decide if I want to buy an external GPU so I can do a VR app.  I think it’s brilliant that Apple has standardized the NN models.  There is so much potential here.  The problem with VR is it’s still at a stage where it’s just as much about optimizing the performance to get all those pixels to the screen within 10ms, and it’s pretty low level stuff.  Even with the computer graphics class that I taught last year, that level of detail is outside of my expertise.  I know I need to understand enough of it to program it efficiently, but I don’t really have the desire to learn the low-level gpu performance tweaks necessary to get 90 fps.  So I’m leaning towards developing an AR app as I think it will provide some crucial foundation skills.  

On the professional side, in addition to learning some Kotlin for Android, I created VUI apps for Amazon Alexa and Google Home.  I used JNode on their respective servers to provide the dialog and collect the necessary information, and submitted the design document for our ColdFusion experts to implement on our own servers.   I suggested it be done on our servers since VUI is going to be ubiquitous shortly, and even web browsers will eventually be more voice activated than not, so it makes sense.  And even though the free service sounds like a lot of at 100k requests a month before they start billing you, my company gets more hits than that in a day, by far, so it could get unnecessarily expensive quickly if we have to pay for Amazon and Google server usage.   I really enjoyed working with JNode, and with the Voice assistants in general.  You can throw up little private apps quickly that are only meaningful for you and your family.   To talk to a server, though, I need a secure site, and I’m not quite willing to spring for that yet.   

I am officially my company’s ‘AI Expert.’  I know because I’ve heard them refer to me as that on several occasions, not that I’ve been given any kind of title or official recognition.  Plus they gave me the VUI project to design, so that’s a benefit in itself.  New things are always more fun than old things.   There are quite a few companies touting themselves as bringing AI to the travel industry, so I have another goal of trying to understand the value that they offer.   This is primarily because it will benefit Hotelplanner, and I want to provide as much value as I can to my employer, and I am the most qualified to do it.  Unfortunately, it consists of listening to marketing types overhype and sensationalize what they can do, and I don’t relish those types of conversations.  I suspect VR will transform our industry faster than AI will.   But it will be interesting to find out what really is available underneath all the BS.