Tuesday, April 12, 2011

EDW So Far

Ok so in this post I will show some of the progress I have made so far mainly learning how to properly key out a green screen. In all of my attempts I have been made using this



It's pretty awfull footage as you can see from this screen shot of one of the early frames. A large part of the action occurs off of the green screen which means to use this footage I would have to rotoscope. The reason I have continued to experiment with this clip is because it is so problematic. While I can not get an alpha channel of the entire action sequence via keying I can practice keying things that were filmed on a horribly lit green screen which should make me a little better at this in the long run.


So the first week of EDW the group of people that were interested in green screening; me, Tex, Aarron, Nick, and a few other people met up and Arron showed us how to set up the green screen, and light it evenly. Aarron then showed us the basics of keying in After Effects using Key light a plug in made by The Foundry. We had pretty good success with the actual outline of the character useing Keylight the only problem was the coat I was wearing when filmed had light blue in it that kept being keyed out by Keylight.

Second week Tex asked Ciara and myself to act for him in from of a green screen for his project. He rented out the green screen and He and I attempted to light it evenly. It seemed pretty good until an actor actually stood in front of it at which point they would cast a ginormous shadow. We some how missed this and proceeded to record some 30 clips partially off screen and with horrible shadow :D. I guess now we know what not to do.

Week 3 I got some of the footage from Tex to practice on and after watching all of the After Effects tutorials found here http://www.videocopilot.net/basic/ decided to start trying to key the green screen out with Keylight I had a small ammount of success getting an alpha channel of the actors parts that were on screen but the shadows were a huge problem for Keylight. At this point I reaized I prety much would have to refilm something if I wanted and footage to work with that I could have any real success with.

Week 4 while doing that thing where you procrastinate on a project by doing "research" I decided to watch a lecture on Lighting given by Christos Tzeremes http://www.christostzeremes.com/lighting.html this lighter from Blue Sky a few weeks ago at RIT. One of the things he kept bringing up is how After Effects is not the industry standard as far as composting goes, actually most studios use Nuke a node based composting application made by the foundry. He goes on to say if you come out of school knowing Nuke you will be a lot better off. At that moment this project became so much more exciting to me. I immediately downloaded Nuke and started scouring the internet for any tutorials I could find I decided I needed to learn how to use this program and up intill this point I didn't really have any drive as far as this project went. I came across some mind blowingly cool stuff in about a 5 hour time span. Probably the most inspiring resource I found was not any sort of video showcasing Nukes capabilities and what I would be capable of doing If I put the time in learning it. It was this forum right here http://www.vfxtalk.com/. There are people just like me posting their noobish questions but the people who give them feedback are industry professionals. This shit was just mind blowing to me. There is a community on the internet where you can post your work and questions and people who worked on Star Trek, The Matrix, Lord of the Rings; pretty much any movie that I every was inspired by, will give you comments and feed back. What I'm saying is:



It just gave me a totally new perspective on the industry and what might be possible for me in the future. It also made me even more content with the career path I have chosen in CG knowing that while I'm sure there are some asses in the industry there are some really helpful people. Since this "epiphany" I've been working pretty hard trying to get the basics of nuke down pat and really understand what I'm doing rather than just regurgitating things I see in tutorials.

From week 4 in to Week 5 I have been doing tutorials and researching what kinds of things I should be learning and producing if I want a good enough demo reel to become a Junior Compositor out of college. Yuya had a large set of tutorials on Nuke on one of his hard drives and I have gottan through all of the basic ones. I have still been trying to key that one crappy clip as best I can but this time using nuke. Nuke offers two different keying approaches, You can use Keylight and try to key out one specific color or for uneven green/blue screens you can use IBK_Keyer which allows you to attempt to create a clean plate from a image to compare each frame aganist. Basically it plays the What is different about these two images game :D.

Here are a few images showing some of my attempts at keying in Nuke.





You can see it works extremely well with the uneven lighting on the green screen. In the second image the transparency is caused by the lack of a core matte. I am not entirly sure how to create a core matte but apparently some people use a node called primatte and just make a big blurry shape. Sadly Primatte is not included in the PLE version of nuke so I will have to wait until I procure a full version.

All in all I am pretty happy with my research thus far. My next Post will contain renders of the 3d set I am creating and maybe even some parts of the dummy set :D.

EDW STUFF

Hey so this Blog will now also be used as my Blog for my Experimental Workshop class with David Halbstein. My goal for EDW is:

A complex 3d scene and a rough greenscreen duplicate set for actors to interact with. Things I want this scene to include.

-Stairs the actor will walk up.

-A railin gthe actor will rest on.

-Something for the actor to pick up.

-Somthing for the actor to Swing

-A force for the actor to react to.

The whole point of this project for me is learning how to properly composite because I want to know as much as I can going in to senior year to use on my thesis.

Saturday, March 26, 2011

Crude ragdoll test

So I just finished my first test using the ragdoll tutorial linked in my last post and it seems to work really well. The rig is very rough at the moment but its still seams to work pretty well. Looking forward to seeing it on a character with painted weights.

Tuesday, March 15, 2011

Week 1 Research

So after deciding I probably needed a way to add some dynamic motion to my stunt doubles I decided to research a few ways I could possibly make a rag doll in Maya. With some Google searching I came across this thread on 3dbuzz http://www.3dbuzz.com/vbforum/showthread.php?136330-Ragdolls-in-maya . After reading some of the posts I got super excited about using a product called Endorphin http://www.naturalmotion.com/endorphin made by natural motion who is the company behind the Euphoria Engine http://www.naturalmotion.com/euphoria which might be the most amazing behavioral simulator available. If you ever wondered why every time you almost run someone over in GTA 4 the stagger away in a completely different way than you have seen previously it's because of Euphoria.
They have a Learning Edition for free download but it lacks the ability to export and buying a commercial license is $13k. I just emailed the company to ask about pricing on an educational License of the product and I am waiting to hear back. Assuming the educational license is still a few thousand dollars I will probably have to default to this tutorial http://www.goldenxp.com/tutorials/ragdoll/index.htm on making a rag doll in Maya capable of being effected by forces. it wouldn't be anywhere near as fancy as Endorphin but it will get the job done.

Intent

This blog exists to serve as a place to post the progress I make in my Production Pipeline class at Rochester Institute of Technology in Spring of 2011.

Over the course of this quarter I plan to create am ambiguous stunt double to be used later in place of real actors in composite scenes in my thesis next year.

By Ambiguous I mean I want it to be a generic enough human form that it can be used for multiple actors after a few changes maybe made through blend shapes.I also want to include some sort of rag doll fuctionality where the skeleton of the double can be affected by forces in maya. I would like to add cloth and hair to the model but that will be time dependent as I have never worked with either of those things before.