robcat2075

no. This does sound like something isn't being done in the more advantageous way. I'm still doubtful about the need of a spherical constraint. But without seeing it, I'll stop there.

In addition to what Mark said, I'll just inquire why you want a spherical constraint? You're probably thinking you need that for a knee or an elbow but proper character rigs almost never need such a thing. If you're just making a rig by putting some skeleton bones in your character you will be making animating it 100 X harder than it needs to be and be unhappy with the result. There are some easy to install rigs in the rigging forum. Have you tried those? Is there a reason not to use them?

1st, You model it's shape like anything else. You probably want to model it with thickness rather than single patch. about 95% transparent with a touch of green or gray. You can make reflections on glass and car paint more realistic by using a gradient combiner ( leave start and end at 0,0,0) to make a reflection material that makes reflections less apparent for viewing angles that are perpendicular to the surface. Real glass reflects much more when viewed at an angle than viewed straight on. Same with car paint.

Welcome to A:M! And if you have questions, this is the place to ask. (The whole forum, not just this thread)

Post wire and i bet that can be made smoother.

It used to be that you could just check the shadow buffer ON and you'd get a separate TGA with the shadow image in it. But now you need to do a 2nd pass Here would be the normal settings to render the object and ground all that as usual: But if you want separate shadows for compositing elsewhere, do this render for just the object and its alpha channel: And this render for just the shadow which is captured in the alpha channel of what appear to be a blank image: Notice I haven't turned the Ground object OFF in either case, it is hidden automatically. Rendering to OpenEXR can get you all these elements in one pass. I think Fuchur has put up a tut on using A:M's compositing with OpenEXR renders.

Something else is wrong, shadow catching isn't a big render hit.

I'd think that a material effector is part of the solution

Multiplane! That works well. William, did you do that original shot of the castle? How did you do the grass in that?

Another classic!

what line?

Fun ideas, Paul. You can get you tube to show your own 3D stuff in any format by uploading a crosseye version and adding a yt3d:enable=true tag. Then you don't have to worry about if the viewers have the right glasses. youtube details

no clue here. you're gonna have to show a pic of something similar.

What is stopping you?

Here's the non-action method: After You got your light constrained to a path: then set the ease to 0% at 0 seconds 100% at 1 second 100% at 5 seconds grab it and squish it all down to 1 frame. Use CP settings to make the last one exactly at frame 1. Then select it all again and >Curve>PostExtrap Method>repeat render with multipass and motion blur set to 20% the guide spline has been set to render as line to make it visible. 5 passes shows obvious stepping but gives a quick approximation 36 passes looks smooth 81 passes is only slightly different than 36

you path constrain a light to a spline you set its ease to traverse that spline in less than one frame and repeat that every frame multipass renders the light at all the different points of the spline creating the effect of a non-point light. The same technique is used for spinning light rigs and works fine for animation. This is really a light "Array" but with the great convenience of not having to create numerous separate lights and the option to do quick tests with fewer passes. Here's a page from a book in which they discuss true area lights and their downside even in a vastly more expensive app.

agep's solution is the way A:M supports that need. "Any spline can be a light" True area lights are time consuming to render and probably wouldn't be any faster than the multipass solution.

newly pressed sheet metal! Does that "perpendicular normals" button make any difference or is that a false lead?

looks good. It would be interesting to go thru it surface by surface and light by light and figure out what it would take to make it look like the photos.

I tried your PRJ and I get a odd result also. Wait a few days for v15h to come out. I put in an AMReport on Paste Mirror a while ago and I believe Steffen has addressed it in the next release. Steffen explained to me that in rigs like TSM2 and 2008 apparently, the CPM algorithm can get confused by hidden rig bones that are in nearly same positions as control bones. Notice that bones like Chest Controller Parent are getting keyed by CPM even though those weren't keyed to begin with and shouldn't be getting a mirror treatment. Wait a few days for v15h to come out. Sorry for the frustration. to the left of those. hover over them and they have names. BTW, have you watched my keyframing options vid inthetuts link in my sig? It covers that stuff.

I just uploaded a zip and it worked. embed the model , embed the action, save the PRJ and zip that.

odd results can happen if bones are not exactly mirrored in the model. Try increasing the paste mirror tolerance, the default is quite small at .01 That said, some fixes are coming in the next A:M release.

I got a corrupted zip file when I tried to download it. You could note this at AMReports at www.hash.com/reports.

Is that not "Mirror Smartskin"?

That looks very successful! One controller? One!?