#i can do some walk cycles for the dinos while working on concept for whichever project is next highest priorities but otherwise i'm out

[cjgladback]

[ID: A grid of four images of the same 3D scene, each toggling between two animation frames. The two sets on the left are consecutive frames, with a brown dinosaur in the middle of tripping on the tail of a blue dinosaur. The two sets on the right are further apart, with the second frame showing the brown dinosaur fallen on its back behind the blue one, who now faces front with its eyes open. The focal length and position of the camera change through the sequence, so in the second frame on the right flowery bushes are out of focus in the foreground. The lower images have more flowered bushes placed throughout the foreground and midground. End ID]

So far so good! The camera-based flower color variation is about as stable as I can get it with educated guessing, a 3D grid of our instanced flowers, and some trial and error when reality is unintuitive.

[ID: Material nodes from the 3D program Blender, showing the camera data's view vector being split and combined to the X and Z coordinates with its view distance for the Y coordinates after some transformation through math nodes and a map range node with animated inputs. These XYZ coordinates are used for the vector input of two noise textures, whose scale is also animated before one is passed through a color ramp and the other is mixed on top of it with a soft light blend mode. End ID]

The background flowers are probably going to be the weak link there, based on the fact I can see the furthest back bush so far swapped some colors over 60 frames. I'll definitely still need some more tests with all bushes in place in case any shifting is more distracting than not. But! Especially after some of my tests yesterday with only two bushes were pushing 22 minutes per frame, it's encouraging that the current scene with 8 bushes took only a little over 16 minutes per frame.

I'm still considering whether it'll be worth the further effort to try getting a no-translucency, yes-alpha, yes-displacement, flat gray render to "paint" with relatively few high quality frames and EbSynth. The fact the geometry will have to be exactly as complex and the transparency passes will need to be just as high makes me doubt it'll be significantly faster...but if I am greying out everything, it'll be easiest to duplicate the whole final file and mess with the materials there, so that decision will be put off 'til the bitter end.