i realize that render time depends on what is being rendered, but for one of my projects, the render time rarely exceeds 3min a frame. So now you can understand how shocking a 45min frame is for me.
everything starts normally on frame 1, but as i get closer to frame 150 or so, each frame will take a few more seconds then the last. Pretty soon, the render times just get unbearable and i have to stop the render and start it on a more powerful computer. I first saw this issue on 2.58 and still see it on 2.59.
As i watch the frame render i can see that the black areas of the screen fill up to make up the frame. The frame will finish and compositing layers will come in, all in about a minute or so, just like normal. At this point, cpu activity will drop from 100% and hover around 30% and the render will just hang there for the remainder of the time. The longest i have seen a frame hang at this point was 45min. Then, raytree and occlusion processing will finish in a few seconds and the next frame will start and end up hanging at the same point for even longer.
I have seen this happen on a number of different projects on 2 different windows 7 computers, as well as my Mac (snow leopard).
I can upload the blend file that gave me the problem last. It seems be inconsistent between different computers though, so im not sure how much it will help. Its about 50MBs.
Do you have a physics simulation (cloth or smoke)? some of them are calculated on the fly, so each frame needs to recalculate the simulation to the current frame?
daren -
i have about 15 particle systems hitting some collisions which kill the particles. But i have had MUCH more in a scene with no issues like this.
Sim88 -
If this issue starts happening and i stop the render and render the frame by itself, the frame finishes in a normal amount of time. However, if i start the animation render again from that point (in video or image sequence), each frame will start to get longer and longer.
It is as if something gets build up as each frame completes
well, the thing is, like with particle emmiters fram 1, 50 particles are emmitted, frame 2, 50 more particles (100 total to deal with) frame 3, 50 more (150 total) so on and so forth… it compounds. Starting at frame 150 begins the particle system at frame 150.
Have you baked all the particle systems appropriately (and I mean ALL)?
I have not tried baking all the particles. But I have never had to. Like i said, i have had WAAAAAY more particle systems and overall particles emissions on projects that i have never experienced this issue with. Because of that, i really dont think it has anything to do with the amount of particles. Plus, i dont let the particles stay alive much longer then 20 frames, so it is not like the particles are just piling up as new ones are being born on top of them. Maybe the the particles are kept in memory even after they die? That makes sense of the symptoms.
Well, I have had this situation once, and it was with a smoke simulation. Never found a solution, it may be a bug. I just got frustrated and found a workaround without smoke.
yeah, i cant really do a workaround to avoid the particles. All our work is based on it. I think it is a bug too. I guess i was just looking for some confirmation from you guys. I will present my case to the devs and see what they come up with.