View Full Version : 3d generalist reel - advice on what to include
05 May 2010, 09:20 PM
I am updating my demo reel as I will be looking for a new job very soon... I am more of a generalist and have worked on various stuff including motion capture and physics simulation.
I would very much like to have some advice on the overall reel presentation, it's length, the time given to each piece, if I should remove some footage, move things around, etc.
I am also looking for critics on the actual content as I could rework some of the scenes.
Here it is: http://www.youtube.com/watch?v=RVXx2aQva_M
05 May 2010, 08:57 AM
I like your work!
I might suggest shortening your reel. Not re-editing it, just cutting if off. The children are showing off clothing sim that you have already shown with two other models. And, the things past the children on the run way don't really have any thing to do with the first part of the real.
I would cut it off after the girl with the shirt sim, and not at the but shot, but a bit before where its obvious you're showing off the shirt cloth sim.
Also, the first short dress sim you show where she is comped into the real life environment isn't that good of a composite.
I would get some more advice on that or cut it out.
I'm sorry I'm not bigger on compositing and can't offer suggestion on how to make that part better.
Overall i think its a nice physics and simulation reel!
all the best,
and I love the little shack being destroyed! hehe
05 May 2010, 06:27 AM
Dude, you call yourself a generalist but, cut it at 2:08 and you're a dynamics and simulations expert. You should really push this aspect further as I hear they are in high demand. To be honest none of the comping, rendering, texturing etc. look amazing, the real standout is the dynamics. I especially liked the way the house on the cliff fell apart.
Cut it at 2:08 or maybe a bit sooner, I dunno, I think that cloth sim stuff is very marketable.
If you want to polish this a bit more, I'd look into some more complex particle animations as the clouds of dust you have in a few of the shots could be improved.
05 May 2010, 11:11 PM
Thanks so much for your feedback.
I am working on a second version basically cutting off lots of the end part. I was pretty keen on keeping my demo around 2 minutes but couldn't figure out what to leave out. Seems pretty clear now!
I think I will also remove the "girl in the garden" comp as I agree with you Bergquist it's not the best, but I don't have the original footage as this was a collaborative project...
Morlankey, thanks for the suggestion about more complex particles... I was thinking of learning more about those and maybe some Houdini as well... so that comforts me in my choice!
Thanks again for the comments and I'll be posting the second version soon enough...
05 May 2010, 01:39 PM
Yea, this looks like it should be a dynamics/FX reel
Are you final on the current content or are you able to go back and tweak still?
I think mainly that while you have some interesting dynamics scenarios none of them look entirely convincing. I'm not referring to the models or lighting which is really secondary but you HAVE to nail the physics if this is what you want to apply for.
The tower in the first shot looks like a miniature. Given it's mass it would not move and lean that far to the left before breaking up. Whatever hit it would cause a huge radial "crater" and the tower would only move at the point of impact, crumble around this area and cause the tower to fall to the right. Your broken pieces do look good and pretty much to scale but to make it more interesting you should consider emitting particles of instanced geometry debris from the cracks to "pad it out".
I would also leave out the rotation around the tower. We can already see the important stuff from the original angle.
The house in the second shot looks too spongy and elastic. Sure there would be some give in the wood but not that much. Again having particles emitting (or rather exploding) from the cracks with splintered geometry would lend it more power. It basically needs more detail. Don't even worry about the lighting or texturing, it's ALL about the motion.
The third shot looks quite fine but could again benefit from more debris particles. That last chunk that falls on the wall and hangs there for a second should break up give it's weight and mass.
I'm not sure what a DMM realtime sim is so I'll skip those.
Now the Rube Goldberg idea is great! This is a fantastic way to show off a lot of dynamics and FX in a short span and you should really push this idea to the max.
Now, we've seen you smashing rocks and wood in the first shots so I'd leave that out. The glass is a good idea but you need to execute it better. Look at some real videos. It would be cool if the glass shattered radially when the ball flies through. Leave catapults, ropes and bouncing stuff to an absolute minimum. They are easy and won't impress. And the ones you do keep in, again, you need to nail the physics like gravity, drag, inertia, etc).
Instead have the ball go through flaming hoops (fire, smoke FX, trails, etc), smash through or bounce on different materials (paper, dusty surfaces), get cut up, melted or whatever. Just as many interesting FX scenarios as you can think of.
Your cloth looks quite nice but it's not really anything complicated. You should try to show your understanding of cloth by layering stuff, have it slide over eachother, show different materials interacting and maybe even tearing.
The reel is on it's way but needs some polish before it's really competitive.
05 May 2010, 07:34 PM
So here's an edited version.
What do you think?
I had not seen your post Wiro, so the content has not changed... yet. Thanks a lot for your comments. I cannot rework all of the shots as I don't have some of the source files and/or access to in-house software anymore but I will definitely try to improve what I can!
I can rework the first tower simulation so I will try what you suggested.
The house simulation I cannot redo, but I will try adding some particles to it. Do you think I should remove the first shot as it feels too rubbery?
For the Rube Goldberg thing, the only purpose of the shot was to show off DMM for an IGN article (which is why their logo is everywhere...) so I didn't use any other FX than DMM. But I really like your suggestions! Thanks! I might not have the time to finish it for the first version of my reel (it has to be ready for the 6th of June) but I'll definitely start playing with those ideas...
As for all the cloth sims they were done with an in-house software developped by the lab I was working for at the time so I can't change them. But I have started to test out ncloth, so I will keep your suggestions in mind while working on that...
So... back to work! And once again thanks for the feedback!
05 May 2010, 04:44 PM
So I've been working on the tower simulation. Here's my latest playblast:
Also about the edited version of the demo: I'm not too sure if I should keep the figure drawings. As simulation and motion capture are more technical stuff I wanted to show something a bit more artistic but I don't want it to seem totally weird. What do you guys think?
Thanks again for all crits!
05 May 2010, 04:58 PM
That looks so much better!
Pretty good weight to it all. Works for me.
I think you're right about the drawings. For an FX reel, yea probably not necessary.
06 June 2010, 02:40 PM
Just realised I hadn't posted the "finished" reel. I'll still be working on it but I had to have a version to show at the Annecy Animation Film Festival... So here it is (http://www.youtube.com/watch?v=GgqP2WzWrpQ)!
Thanks for the help!
PS: If you're interested in a higher res version please visit my website: www.clemilo.ch (http://www.clemilo.ch)
06 June 2010, 08:39 PM
just a few things to take into consideration, try to have at least 3 peices of ur reel that are mind blowing, and wat i mean by that is simply good renders. something where an employer could ask "what movie was this shot in". u have some dynamics' skills however as an fx artist my self i'll just point some mistakes out for u and will leave the the other areas to other artists specialized in it.
i noticed in most of ur dynamic sims there r wrong physics, the shot where the rocks fall on a house, if u think abt it, when a rock that sizes falls on a house in that size it would completly distroy it. in other words the first rock falls and u c the wood kind collapse and push back, well that's not what would happen in real world. so u can c clearly that the pieces of the house are acting like rubber and not wood. and it continues on the whole shot. it seems like most of dynamic sims have that rubbery feel to it, so u need to address that, another example is when that rock that's tied to a string goes throw a peice of wood, u can clearly c that the lower half of the wood after the impact they move forward to break apart, the push back to their original spot.
anyways just to keep in mind, getting a job as an fx artist is much harder than getting a job as a generalist, so i would say stay away from dynamics, or if u want to show dynamics then do so but u have to be good at it, or else u would open up a whole new era that u wern't aware of
06 June 2010, 09:25 PM
Thanks for the comments.
I would definitely like to have some movie quality shots on my reel but unfortunately I have only worked for technical companies, so I don't really have that kind of stuff...
I agree on the "rubbery" wood thing. If I get the chance I will try to rework it. I don't quite understand your second comment though... You say that the lower part of the wood acts weird when the rock goes through it... In what way?
Most of my work has been dynamics, so well... there's lots of that in my reel obviously! I'm not really an FX artist as such because I also did all the rest of the stuff on most of the shots. (I was the only 3D artist in the company for more than a year). But from what I heard in the industry there's actually a demand for FX artists. It's much harder for modellers and animators... I've only just started sending out my resume and reel so well... here's hoping!
06 June 2010, 09:25 PM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.