OctaneRender® 1.025 instancing preview test build

[bepeg4d]

@matej: this stuff is just an excuse for studying the c4d node scripting system (something similar to XSI ICE) called xpresso. it's just a temporary solution waiting for a more professional and easy one.
Since my mathematical knowledge is really basic, i've found very hard to menage matrix and vectors; every time i've tryed to multiply a matrix i've lost myself in few minutes
since i must decompose the global matrix in the 12 digit scheme, it's more easy, for my old brain, finding the right formula for rotation and scaling for the relative nine digits.
it's the more complex node script that i ever builded and i'm very surprised that it works pretty well
so, in conclusion, i've learned a lot about xpresso and i can play easily with octane instance scatter node, not a completely waste of time
unfortunately i can't rotate an object in all the axis simultaneously for now
ciao beppe

[matej]

@Beppe, my point was that it's actually easier to manage transforms with pre-constructed matrices, than diving into the structure of the matrix, term by term. I'm not good at math and frankly I'm totally lost in all that individual rotation matrix data you posted... That's why using a linear algebra library (it such library exists) is waaay easier to manage transformations, because you work with individual transforms, not individual matrix terms. For example in Blender scripting system you would do:

Code: Select all

rot1 = Matrix.Rotation(radians(90.0), 4, 'X')
rot2 = Matrix.Rotation(radians(45.0), 4, 'Z')
scale = Matrix.Scale(2, 4)
t = rot1 * rot2 * scale

This creates two rotational matrices (90 degrees by x, 45 by z) and a scale matrix that scales the object x2. The last line creates a matrix that incorporates all this transforms together. Then you just multiply your objects with this matrix. Easy as pie to understand and manage.

Learning purposes aside, creating the matrices by hand will become exponentially difficult when you will have to correct transforms locally / globally, pre-multiply / post-multiply translation, etc.. a mess to understand. Just saying, cause, you know... you write a simple script but then you want to add this and that functionality, and it soon becomes a larger project..

[bepeg4d]

i agree with you about the more simplicity of a matrix calculation, but unfortunately i can't find any info on what kind of operation i need to do in order to transform the c4d global matrix into an octane one, the Z axes is inverted
the unique informations that i was able to find are relative to the trigonometic decomposition of the values of the matrix, so i've decided to decompose the c4d global matrix, take the translation as it is and recalculate the 9 digits of rotation via formula.
but don't worry it's more easy of one of yours beatiful procedural octane materials
it's made by nodes, so is extremely easy to trasform and elaborate, and they are also groupable
ciao beppe

[face]

If you can read and write an object global matrix, there is no problem.

Seperate each value and multiply the z position with -1.
After that, connect each value to the new matrix.

Here a screen from Softimage...

face

[face]

If you want to mirror the global transformation matrix on axis z, multiply the matrix with an other on with negative z-scaling.

Like this:

face

[bepeg4d]

many thanks face, those screenshots are really useful
i'll try to reproduce them with xpresso, thanks again
ciao beppe

[face]

Why a geometry group can´t render from the command line?
Have build this setup:

The scene has the name test.
Exported mesh is test.obj.
After starting Octane, the rendertarget node is used with the -t command, but only the test.obj is rendered, before the placement.
Commandline is:
-m test.obj -r "G:\Octane\ProjectA\test.obj" --cam-up-x -0.153499 --cam-up-y 0.902506 --cam-up-z -0.402395 --cam-target-x -0.0449646 --cam-target-y -0.0413809 --cam-target-z -0.0114845 --cam-pos-x 0.113271 --cam-pos-y 0.170481 --cam-pos-z 0.403327 --cam-fov 53.638 --cam-aperture 0.01 --cam-focaldepth 5 --imager-exposure 1 "G:\Octane\ProjectA\test.ocs" -t "RenderTarget"

I know, camera and imager aren´t used in my Octane setup, but it should render the same scene as when i select the RenderTarget by hand...

face

[bepeg4d]

many thanks to matej for the constructive criticism and face for the ICE screenshots
it's obviously more simple with matrix math.
now it seems to work correctly, i need to do some more tests and fine tuning but now i can finally rotate on all the axis
i'm really happy, thanks again
ciao beppe

[steveps3]

Not sure if anyone has asked but can it be changed so that if you are using a cvs file to drive the scatter then it still uses the csv when you reload the project. At the moment the cvs gets hard coded in the project meanin that you have to reimport if you update the cvs. This ends up being a bit annoying after you forget a few times.

[matej]

^ agree. Load .cvs on scene load, dont embed it into the .ocs (cause that usually means outdated data) and also give us a reload icon for every resource.

@bepeg4d, no problem, I'm glad you made it work