Darby Edelen
Forum Replies Created
-
Primarily you need to white balance the shot. The easiest way to do that is to pick something that should be just about white (her shirt) and look at the current RGB color values. Then apply a Levels effect and set the Input White values for the Red, Green and Blue channels individually to those values. Finally adjust the gamma and white output of all of the RGB values together (the previous step may have brightened the image too much).
It’s possible that you will clip some values if you’re not working in 32bpc and you white balance for a white value that is not the brightest in the scene. For that reason I’d recommend either doing this all in 32bpc (less work up front, better results) or taking some extra time when setting your white input values by padding them a little bit (more work up front, lower quality results, much faster render).
Attached is an example where I took this approach with a still from your sequence.
Darby Edelen
-
Generally I’d use special kinds of lights or luminous surfaces for the majority of what you describe.
[laura plobi] “- On some projects, I need to have lights that are placed along a curve, or on the edges of an object. In that case, I use a cloner to put them automatically. Is there a way around that then?”
I’d say it depends on what you’re after. Do these need to be point lights or are you trying to emulate a long neon or flourescent bulb? If the former, yes I’d use a cloner, if the latter I might create geometry and use a luminous surface with GI if I need detail or use an area light with an Object/Spline area shape (samples might need to be increased).
[laura plobi] “- On some projects, I am creating halogens, led strips or lasers. Each of them may contain 10 to 50 light sources along lines/curves depending on their length, and there are several of those.”
For halogens I’d recommend looking into IES lights for the actual lighting. Then for the visible portion of the light I’d use geometry with a luminous material. No need for GI, the light will illuminate the scene and the geometry is there just to look like a light. For an LED strip I’d use an area light that is roughly the size/shape of the strip and then use geometry with a luminous material. For a laser I’d probably use a parallel spot light with its visibility turned on.
In each of those cases I’d use a single light to achieve the result.
[laura plobi] “- And I also imagine that if I was doing all this “by hand” (rather than automatically with a cloner, which is like a function in the end), I might end up having the same problem of memory space for the rendering, wouldn’t I?”
Yup! Just find ways to avoid using so many lights. I’d also add that since geometry can be instanced you have the option of enabling GI, creating a luminous material, applying it to a sphere (or any ‘light’ shape) and cloning that with render instances enabled. Global Illumination will slow down renders as well but possibly less than using thousands of lights 🙂
Also keep in mind that there are several different components of light. You might get away with creating very few diffuse lights to represent a more complex set of lights and then using luminous geometry to create the desired visibility and reflections.
Darby Edelen
-
[laura plobi] “If I uncheck “render instances” it gets back to normal, but my projects require me to have so many lights that my computer won’t let me render them if I am not using instances.”
There’s your answer.
And from the C4D help files:
Render Instances do however have a few drawbacks:
-Render Instances are exact copies of the original geometry and cannot be deformed using Deformation objects.
-Render Instances can only assume the current state of the original object and not different states of animation of the original object.
-Render Instances should be made up of geometry because only geometry benefits from the corresponding advantages. Light sources, for example, will not work at all.
-Objects that are a Child object of a Subdivision Surface object (i.e. serve as a source object for other objects) cannot be render instanced.
-Sketch and Toon does not support Render Instances.
So what is it exactly that you’re trying to do that requires thousands of lights? 🙂
Darby Edelen
-
[Kelton Bingham] “Do you know how to delay keyframes with any other effectors, or how to apply the inheritance effector properly to rigs?”
Almost any effector can be used to offset the cloned source’s animation. The notable exceptions are the Target, Spline, Inheritance and Delay effectors.
Look for the ‘Time Offset’ parameter in the ‘Parameter’ tab of the effector. Based on your description of the setup I’d start by using a Plain effector with a falloff or a Shader effector. The Shader effector would give you a higher degree of control but may take more tweaking to get working properly.
In cases where the root of the cloned object’s axis is animated you’ll need to disable the ‘Fix Clone’ option on the Cloner or place it inside of a stationary Null.
Darby Edelen
-
Technically you’d want to use the distance along the camera’s z axis. Something like this:
C = thisComp.activeCamera;
L = thisComp.layer("3D layer");
CP = C.toWorld([0,0,0]);
P = L.toWorld(L.anchorPoint);
CV = C.toWorldVec([0,0,1]);
dis = dot(CV, P - CP);CZ = thisComp.layer("_Camera04").cameraOption.zoom;
S = thisComp.layer("3D layer").transform.scale;
S*CZ/dis;
This prevents the layer from scaling as it moves laterally relative to the camera. I may have reversed the subtraction order in the dot() function, writing from my phone and can’t double check it.
Darby Edelen
-
With the layer selected tap uu to reveal all changed properties. You should see some bars that look like layers in the Paint effect, these represent the stroke durations. You can select them and slide their in and out points just like you would with layers.
Darby Edelen
-
This may not achieve what you’re after but you can use an effector to offset the keyframed animation of each clone.
Darby Edelen
-
You’ll need a reference point for the scale but the basic idea is:
scale * camera’s zoom property / distance to layer along camera’s z-axis
Darby Edelen
-
If you use the Custom Shader mode of the Shader effector then the UVW coordinates of the Shader are based on the UVW coordinates of the Clones. Try changing the W(UV)-Orientation value on the Transform tab of the Cloner. Depending on how your Clones are generated one of those options may or may not work for you (Y+ or Y- are likely).
Alternatively you can use a Material with the Shader effector and use any of the texture projection modes (defined in the Texture tag).
Darby Edelen
-
The twist deformer is a good place to start.
Darby Edelen