Attached Link: http://brenthomer.tripod.com/truck.mpeg

Ok here is something for everyone to look at this weekend. I did a sample animation to find out if depth of field would get rid of the dots. In a word..yes. However depth of field also has its own problems (I'm not complaining...I swear!!!) to deal with. I assume if I would make the DOF very wide then this would go away but as it stands in this animation I made the DOF pretty small so the artifacts are huge. It almost looks as if the background is on fire (or just has lots of heat waves?). The other artifact I noticed is the way Carrara renders. It periodically takes chunks out of the truck when applying DOF. If I used Raytracing instead of the Hybrid engine, would that fix things? On the positive side, if I just render a still picture then DOF is awesome...it just seems like it doesnt like animations. the link is correct and working...tripod just sucks...and the animation is 1.4megs, 10sec, 30fps BTW: thanks for the story below AzChip...I & a co-worker laughed a lot about that...I swear everyone in production, no matter how big (or small) the budget is, is complaining about the same crap :)

Anyone going to take a stab at this? :)

Havn't been able to download the mpg yet. I'll set up some DOF animation tests and get back to you. Mark

I can't get to Tripod right now, but I'll try again later.... I have noticed that in RDS the hybrid renderer tends to be much more artifact-laden than the full raytracer. (Same was true in RayDream 3D.) If you're working on some lower-res stuff for the web, the hybrid might be fine, but for video or if you're planning on a pro print from your image, I think you just have to knuckle down and eat the time to use the raytracer. (These artifacts might also be responsible for the mysterious dots in the distance, too.) Since the underlying code in RDS and Carrara are basically the same, I'd take it as pretty likely that the hybrids work very much the same. As for DOF in an animation, I think that'd probably smooth out a bunch if raytraced, too. The hybrid seems to have some attributes similar to compression, and if each frame is "compressed" or hybridized individually, there may well be some problems with dancing pixels growing in importance. Have you considered doing two renders and compositing them in post-production? You could do the background in one pass with the truck cloaked, then the truck in the next pass with the background cloaked or deleted. If you render as a sequenced photoshop file (or a QT if your in the mac environment), the alpha channel should be preserved and allow a clean, easy composite in AfterEffects or Premiere. By rendering the BG seperately, you could force a blur in post. Just an idea.... (And that shot in Creepshow was particularly difficult because I was in a wetsuit in very cold water in Prescott, AZ for almost the entire time we were shooting it. Of course, there were no wetsuit shoes big enough for me in Prescott, so I wore high-tops. All my body-warmed water just shot right out my ankles.... I know, bitch, bitch, bitch!)

--------------------- The hybrid seems to have some attributes similar to compression, and if each frame is "compressed" or hybridized individually, there may well be some problems with dancing pixels growing in importance. --------------------- I think this is what the problem is. DOF is a post effect. I think b/c it applies to each frame individually rather than as a group, things get messed up. I also like your ideal of two renders..which sucks of course, but I will try and set one up tonight..that way I can blur stuff in AE. It will not be as nice b/c carrara takes the foreground object pixels into account, but carrara is so sloppy at it, it could be worth it. I think I need to start thinking of Carrara as a 3d after effects, and that may help me integrate my ideals better. Now hopefully I can make the alpha's work :) btw: Now that I think of alpha's...b/c I will be turning off the ground and sky, wont all the shadows disappear off the truck and other objects?

Yeah, shadows and reflections might be a pain. Something else to throw at you, although I haven't tried it yet myself, is the distance G-buffer. If I'm reading the manual correctly (RDS, mind you), you should be able to create a distance mask with the g-buffer. Then, import the single render you've already made into AE. Place your current render in the AE composition twice (two layers). Put the distance mask in and create a transfer mode layer based on the distance mask. Use this to guide your gaussian blur, and your should be able to force depth of field that way. AE's gaussian blur is better than what I've seen in RDS or Carrara.... Just a thought.

I saw mark put up a similar question on the elist...antoine had a few suggestions also. Heres the quote: ---------------------- Have you tried to uncheck "Light smart sampling" ? Light smart sampling accelerates the rendering of images but can introduce inaccuracies. As a general advice to the Dreamers: it is always a good idea to uncheck the two smart sampling options when doing a final render. --------------------- For the last few nights I have been trying to get carrara to render its alpha's to my likeing. Once I get that down I'll start on your suggestions.

I used Antoine suggestions of unchecking the two 'smart' options for final animations renders. The 'sparkles' remained when Wind was selected in the atmosphere settings. When Wind wasn't used, the renders came out flawlessly. Mark

hahaha...it had to be my favorite plug-in that was messing things up :)