3 replies to this topic

[zephyr707]

hi there,

I'm trying to figure out the best way to accumulate color via an attribute transfer node using a SOP solver.  I've attached a simple scene with a null animating across a wall grid and using a SOP solver to accumulate transferred color onto the grid.  It works fine, but if the sub-steps are too low there are gaps in the path and I want a straight curve/line that follows the null's path even if it jumps great distances between frames.  Is increasing the sub-steps the best way to accomplish this faithful color transfer of the null's path, or is there some other more efficient method?  I started off using particles emitted from the null point and sticking to the grid, but that also required some oversampling.  I've read about the timeblend node, but I think that's not going to help me here.    

Any tips would be most appreciated, thanks
z

Attached Files

•  nullPath_sop_solver.hipnc   126.73K   21 downloads

[bloomendale]

zephyr707, on 02 June 2012 - 04:55 PM, said:

hi there,

I'm trying to figure out the best way to accumulate color via an attribute transfer node using a SOP solver.  I've attached a simple scene with a null animating across a wall grid and using a SOP solver to accumulate transferred color onto the grid.  It works fine, but if the sub-steps are too low there are gaps in the path and I want a straight curve/line that follows the null's path even if it jumps great distances between frames.  Is increasing the sub-steps the best way to accomplish this faithful color transfer of the null's path, or is there some other more efficient method?  I started off using particles emitted from the null point and sticking to the grid, but that also required some oversampling.  I've read about the timeblend node, but I think that's not going to help me here.    

Any tips would be most appreciated, thanks
z

Maybe just create trail, resample and transfer color from it?
http://dl.dropbox.co...rResample.hipnc

[zephyr707]

bloomendale, on 02 June 2012 - 07:50 PM, said:

Maybe just create trail, resample and transfer color from it?
http://dl.dropbox.co...rResample.hipnc

that's a cool idea, thanks for sharing that!  I guess I should have added more complexity to my example file, the null will be moving away from the wall and towards it as well.  When I add this dimension to your file the proximity of the line to the grid for the attribute transfer sometimes creates flickering of the red.  It is not as noticeable when the ball is just moving on the flat grid plane, but there is a little bit of flicker as the line is trailed and re-sampled.  Let me try this in my larger scene and see if it workable, thanks for the new method!

Attached Files

•  3ak_attribTransferResample_y.hipnc   268.44K   23 downloads

[fxrod]

zephyr707, on 03 June 2012 - 08:05 AM, said:

that's a cool idea, thanks for sharing that!  I guess I should have added more complexity to my example file, the null will be moving away from the wall and towards it as well.  When I add this dimension to your file the proximity of the line to the grid for the attribute transfer sometimes creates flickering of the red.  It is not as noticeable when the ball is just moving on the flat grid plane, but there is a little bit of flicker as the line is trailed and re-sampled.  Let me try this in my larger scene and see if it workable, thanks for the new method!

Anoter good way to get sub-frame positions is to use the Cache SOP on your animated object. Set the 'Index' parm to $FF for floating frame. Then, turn on 'Blend Position'. In your 'Global Animation Options' turn off 'Integer Frame Values' and set your 'Step' value to something like 0.5 so you can scrub the subframes. Now, if you're using the 'Solver' SOP, you can set the 'Sub Steps' to something like 2 and you should get higher fidelity attrib transfer.