robcat2075

When you are in a Smartskin window and in Muscle Mode you can use the |<< (Previous Keyframe) and >>| (Next Keyframe) buttons to jump between Smartskin keyframes. They don't always sequence in a logical order, you may need to press the buttons several times to access all of them This will work if the bone you made the smartskin for has no constraint controlling its movement. If it does, you will need to open the Pose with that constraint and temporarily Edit the constraint's Enforcement to 0%. Do this before you open the SmartSkin window and remember to set it back after you edit your SmartSkin. Never save over old work. If you make a mistake you can't undo , you want to be able to reload your previous version to try again.

Those look good! The video looks warmer than the stills, but it's all the same renders?

I was imagining something like a doctor with the mirror on his forehead but yours looks like the existing camera icon http://www.politedissent.com/images/jun10/camels.jpg

Here's a bit more animation using a revised workflow...

This seems like a shortcoming. There is "Use Front Realtime Light" in the Chor but nothing like that for the Modeler or any other window. Perhaps this could be extended to the modeler as a general parameter in the Options window. If it had a button,what would the button look like?

Here's the Rubik's Cube in a Cornell Box...

Rodger, make sure the Options>Cast Shadows>Darkness of your light is set to 100%. Although new lights default to 80%, 100%is the more typically useful setting.

Rodger, could you post your sample PRJ that demonstrates the problem? The effect in your animation is very convincing!

Where may we do this?

Here is the Rubik's Cube that we rigged at LAT today... Now that we've thought this through once, there are several steps we did that I think could be simplified or omitted. I also think @Tom's suggestion about a set of Pose sliders might be doable. Perhaps we will re-visit this topic at a future LAT.

Hair can render faster if it has Cast Shadows OFF it has fewer Control Points it has Accurate Transparency OFF it is lit with z-buffered lights instead of ray-traced lights

Displacement maps are tricky to get good results from. Perfectly vertical sides will be impossible. The edges of any form will need to have some slope. The more slope the better. Straight lines and sharp edges tend to show off the worst of displacement maps. Zero width lights seem to be a problem. Use a ray-traced light wide enough to cast a discernible penumbra. I'm not sure what causes the wavy lines but my work-around is to render at 2x or 3x or 4x the normal res and then scale that down to normal size in post. I used the font wizard to model a "1" with a wide bevel and added a flat plane to serve as a background. I put a black-to-white gradient on it and shot that with an orthogonal camera to get my displacement map. I used OpenEXR but PNG16 will work also. I decaled that to a sphere. This is rendered at 480p with 16 passes This is rendered at 1440p with 2 passes and scaled down to 480

Hi Teo, Can you post a sample case we can look at?

A better description of this true/false second field is "Position taken from another CP") If 1, (true) then the fourth field gives the CP number this CP's position is taken from. If 0, (False) then the CP position information will be explicitly given in the fourth and succeeding fields

You could have him swoosh his head from one side to the other on "alllllll"

A 1 in the 512 bit place means the CP is "locked"

If I replace 262145 with 1 that means the CP has 0000 0000 0000 0000 0001 in the first field instead of 0100 0000 0000 0000 0001 With that first 1 missing the default bias through the spline is parallel to a line drawn through its two neighboring CPs. I believe that is an old style that previous versions of A:M used, now named "perpendicular normals"... With the 1 in place the default bias is derived somehow differently. It is not parallel to the neighboring CPs. This modern result is called "biased normals"... The choice of Biased or Perpendicular normals can be selected in the modeler with the same-named buttons which can be added to a toolbar from the Tools>Customize>Commands menu...

The first field is some set of binary flags The second field signifies "attached" (T/F) to another CP (EDIT: A better description of this true/false second field is "Position taken from another CP") The third field is the CP# After that the fields are data and can be of varying meaning In line 26 the 4th, 5th, and 6th fields are X Y Z locations However, in line 27 there is a 1 in the second field (attached = true) so the 4 in the 4th field tells what other CP is it attached to, no X Y Z is supplied the space-dot-space-dot seems to mean "apply defaults for CP bias"

262145 in binary is 0100 0000 0000 0000 0001 262144 in binary is 0100 0000 0000 0000 0000 that is too much of a coincidence for that to not be a binary number where each digit stores some true/false value. The 2.5 byte size would odd, probably not the full possible extent of the value. The real value may be a four byte integer

Hi Charles, Very cool stuff in that video! Very impressive all around. I recall doing some slight investigation these numbers... The first number seems to contain flags about properties of the CP. It may be a binary number with on-off flags. The last digit indicates peaked/not peaked the dot dot is a placeholder for CP bias info when there is none

I guess you are stuck. Since no program presents everything you want and since you have to make some accommodation for what it doesn't, you will have to choose the one with the work-arounds you can live with having to do. The grid in the modeler can be set and re-set to any increment you want for precise CP placements with Snap to Grid. The "Measure Distance" menu option finds the distance between any two CPs.

You should make the stem wag.

Hi Nate! I recall Martin saying years ago that A:M wasn't built to be a CAD modeler and it wasn't a niche they intended to pursue. You've been warned! None-the-less A:M does have strong OBJ and STL exporters and a number of A:M users are using A:M to create objects for 3D printing*. True modeling Booleans for Splines is a big leap. If you need it, I suggest modeling in A:M as if you were going to use its render Booleans, export the base model and the "cutter" as separate OBJs, then import both into one of those programs that people use to fix polygon models for 3D printing and do the Boolean cut in that. *Our Image Contest Medal is modeled in A:M, 3D printed in resin, then the resin print is used to cast medals in pewter... https://forums.hash.com/topic/48234-the-image-contest-medal-and-certificates/?tab=comments#comment-413200

Sorry you're having trouble, Benjamin! I'll inquire.

Try these. I reconstructed these from screen captures...