Rendering Week, and the Final Stretch of Year 1
This is the last stretch, and it's been a long one. Rendering has naturally been the main part of my work this week, but I've also tweaked lighting, edited, and set up a few more pieces of animation and two extra shots.
End of Week 11/Beginning of Week 12 Playblast:
(I'm happy enough with this as an animatic, but I would have liked to tweak it a bit more if time wasn't an issue. Rigging and other more complex tasks were also an obstacle.)
Firstly though, is the rendering. Our scene is so heavy, (It's about 600MB) that anything we render takes at least a day, and can take up to 3 or 4 to finish. the main issue with this is that when there's a problem, it's a big setback. And there's been a few problems that have set us back considerably. The first issue is that some shots weren't long enough. I'd already started rendering last week because I was worried about the length of the render times (A few tests had taken up to an hour per frame). Unfortunately, the character animation Ronan made lasted longer than the frames I'd already rendered. I couldn't figure out how to match the new renders with the old ones without having a jarring change, so I had to completely re-render the third shot in the sequence (the one with the normal deer near the water.
Another render issue had to do with glitches. A few frames glitched out on the second shot (where the hunter is revealed), and it took me a really long time to fix. I really struggled to figure out how I'd originally rendered the sequence, and I must have rendered about 30 times trying to match a new frame to the old sequence. I guess this should be taken as a warning to note down your lighting samples, your render settings, and everything else before putting a render together.
Finally, there were a few shots that just didn't look right, and had to be redone to match the playblast and the idea better. A good example of this was the shot of the hunter as he falls over to reveal the deer hunting him. this was rendered four times, for various reasons. Firstly, there was a camera movement that shouldn't have been there. This happened twice. Secondly, there was a problem with the render samples, and the image turned out grainy. I had to render it one more time as well, since I'd been rendering the character as completely static until falling over, and it was really apparent, since the noise on the character's render didn't move at all.
I think the main thing I've learned this week is one simple rule; Finish the animation and camerawork before rendering. Mixing up this production pipeline really just wastes time. I'll keep that in mind next time. Hopefully I won't have to undergo this degree of setbacks in future projects.
Rendering is difficult in a team environment, especially for me. I really feel the need to track the renders so we don't miss anything, but at the same time I keep taking charge in the entire process, which is hard on me and not too fair on the others. I think the strategy of other groups, which is to make each member handle a set amount of frames, is the best solution. Otherwise, there's not much incentive for the other group members to keep up, and I end up controlling the project a bit too much.
I'll have more reflection in a new post, as well as in my reflective statement.
Comments
Post a Comment