ZachBG
-
Posts
1,673 -
Joined
-
Last visited
Content Type
Profiles
Forums
Events
Posts posted by ZachBG
-
-
Hi-ho. Still haven't masked out my son's head yet, but I did re-work some of the animation... and showed more of the story.
-
Nice!
The anti-aliasing terminology has changed from the 8.5 days. The equivalent to turning anti-aliasing off is turning ON Multipass and using a pass value of one. (Looks like that's what you did.) To get anti-aliasing, you can either up the number of passes (16 is a good baseline), or turn OFF multipass, which will be quicker.
-
Animatics I've finished:
NIMMEE AMEE (133-170)
RETRANSFORMATION (318-325)
Animatics I'm working on:
HIPPOGYRAF (263-317)
-
Yoinks, that's pretty gruesome.
His hands look strange to me. Are they supposed to be bare bone, or skin stretched over bone?
-
Another slight update. Animation-wise, I tweaked the fall a bit, but more obviously, I added PARTICLES!! They cut out toward the end because my computer was estimating 48 hours to render 2 seconds, and I didn't want to waste that time on a test render.
Enjoy...
http://www.hash.com/users/zachbg/GiantShaggyDown_v3.mov
(Could this perhaps be the first part of an epic? Stay tuned!)
-
Another version. Mostly worked on the helicopter, though I reworked Shaggy's fall a little bit, too.
Something went weird with the shadows at the end. Don't know why.
-
Congrats!
-
There was mention of how you do it manually. I use After Effects quite a bit, and with that program you use motion tracker to track the motion of the layer you want tracked, in this case the kid running from the library, select a point to track, track it, and then apply the motion track to the layer you want, in this case the giant shaggy. It's pretty straight forward, and the nice thing is that its all keyframed, meaning that if one part of the track is good, but another starts to drift, you can keep what's good and just re-track the part that needs fixing.
But that would only work with 2-D tracking (like when the camera is locked to a tripod), wouldn't it? Or am I missing something?
-
I agree. That's what I'm working on now.
Er, I mean, that's what I will be working on, once my TWO shot is done, of course!
-
And I assume "in front of your eyes" pose transformations aren't needed. A puff of smoke and they're changed?
We haven't discussed it (it's me and Ed in charge of the transformation scenes), but I for one would like something a bit more elaborate. But that doesn't mean we couldn't build special models just for the transformations--there's no need to worry about it in the models you're making now.
-
My only critique would be that the bomb needs some pyro with it. What is it, a push bomb?!?
Oh yes, that'll be coming. I love me some particle F/X.
-
Zach;
I need to go back and review to see the entry requirements. Are you thinking like a hallway or a vestibule construct?
Vestibule would be fine. Actually, now that I think about it, maybe it isn't needed--the important thing is to allow TinMan to have a reaction, and even if he reacts right upon entering the room, that's probably OK. It would be nice to have a corner for him to go around, though.
-
And here's the latest animation. Probably the last update for a while, because if I don't get back to work on Tin Woodman Martin will kill me. Well, hurt me anyway.
Comments on the animation are now heartily encouraged, though there's a bunch of stuff I know I want to fix anyway.
http://www.hash.com/users/zachbg/GiantShaggyDown.mov (1.7 MB, ~15 seconds)
-
that is so cool, how'd you do it?
I have no idea how one would do it manually. I used SynthEyes.
In a perfect situation, you just load a movie into SynthEyes, hit "Go," and export a Hash project. In reality it takes a bit more than that, especially for a hand-held shot like this one. You need to identify a few easy visual points in the film that you can follow throughout the scene (or at least through most of it). In this case, I chose the corner of the trash can, a point where the grass met the sidewalk, and a few light spots in the grass. After I told SynthEyes to track those points, and tweaked the tracking so the software didn't lose them, I ran the auto-tracker to add whatever points the computer could find on its own. Then I told the computer which point I wanted to be the origin, which points were on the floor, etc. and solved the camera matching. And that's basically all there is to it.
With smoother shots, you don't even have to do the first step of manually tracking points.
Then, you open the project in A:M, and animate away.
Unfortunately, I can't at the moment render front-projection in A:M with this project (I've got a Report in the system about it), which is why I haven't masked out my son's face or the foreground tree yet. So when I want to render it, I do an alpha pass and a shadow pass and composite them in Quicktime Pro. (I wonder if one could do it with the new Composite feature in AM? Must try that.)
I've attached the project file output, straight from SynthEyes, for people to play with. Note that it's 60 frames per second, since I was using interlaced video. I haven't included the movie file because it's huge. (If enough people want it, I guess I could post a compressed version...) Also note that there's one glitch--it's around frame 705 at 60fps--which I had to get rid of manually in AM. Which just goes to show that nothing is perfect.
-
There does need to be an entryway for staging purposes. See the animatic.
-
If you put a root bone at 0,0,0 you should be able to scale everything down to 6'
Not if there are muscle poses.
It would certainly be better to have him be 6' to begin with, but scaling in the chor is not as bad a thing as many people seem to think, so I'm OK with it.
He's looking amazing, BTW.
-
Cash, man. Cold, hard cash. Baby needs a new pair of shoes.
hope that's OK...
-
Zach is going to audition for TIN WOODMAN.
......
Zach is delivering start of MEET HIPPOGYRAF next week.
Golly, I'm busy, ain't I?
(I'm actually planning on hitting RETRANSFORMATION first.)
-
I would like to know what the animation directors would prefer or need.
This one would prefer teeth (tongue not quite as essential) for lip sync--but that said, it's certainly doable to not have them.
I really liked the version where the tongue is the same color as the rest of the bag, though. Just FYI.
-
I don't know when or if they'll be made available to the public.
If? Hmmmm. So, what would be released? I thought it was a community project. Am I delusional again? (that's another issue).
Well, of course it's a community project, but these animatics are a) really rough and b ) not necessarily final. In this case I elected to send Bill links to the animatics privately so he can see where the camera will be and what's needed in terms of modeling.
Personally, I have no problems with releasing my animatics to the public, but I don't want to set a precedent for my fellow animation directors.
-
How long did it take you to camera-match this scene, just out of curiousity?
Not too long. Less than a day; probably two or three hours at most.
Also, would you be interested in doing some camera-match-for-pay (or trade... i.e. models, rigging, etc.) work at some point?Yes, yes, o yes.
-
The black line is actually from the original footage; nothing I can do about it.
A second or so more, and different movement for Shaggy. Still no masking yet.
-
Are there animatics available to see? - or is this hidden from the riff raff? I can't find it.
As of yet, they're in the private area. I don't know when or if they'll be made available to the public.
Chris, I'll send you the model, but please don't let it limit you. (It's a pretty yecchy model anyway.)
-
That's cute. Are you going to rotoscope his head back in over Shaggy?
If I can, yeah. There's another 8 seconds of footage that's matchmoved, so there's more to come, but I have to figure out a "story" first.
A little match-moving
in Work In Progress / Sweatbox
Posted
Not sure what you mean... the only thing for which I didn't use A:M was the actual match-moving of the camera, which I haven't a clue how to do manually.
But yes, it is possible to do CG+live action directly in A:M, especially if the camera is locked down.
I should say that, for now, I'm not using A:M's front projection abilities, because I'm having trouble rendering with an image sequence camera rotoscope. (AMReports knows about it, but if anyone else wants to try...) Instead, I'm rendering out Shaggy and the other CG stuff alone, as well as a shadow pass, then compositing everything in Final Cut Express.
Thanks! I still need to "grain it up" so it's not so pristine...