[Cosmasad]

Good morning everyone,

Following up on another thread regarding creating light fixtures with the Rhino plugin, I would like to probe a little deeper into what our options are:

1) As per messges with Jeremy, if you want to use an IES file then you have to attach it to a Rhino spotlight. Would it be fair to say that the same thing would apply to a Rhino Pointlight (if you want to create a table lamp that radiates light in all directions)?

2) If you don't want to use an IES file will the Rhino plugin allow you to create a light fixture using one of the presets? And if so, if I am creating a recessed canlight with a floodlight bulb in it, would I--
a) Want to create the can, (since this is the part of the light fixture that is visible to the eye), install it in the ceiling and install a 1cm diameter sphere in it, or--
b) Create a can, install it in the ceiling, and create a realistic looking light bulb inside it that you then attach the preset emitter to it. Or --
c) Simply create a round circle, turn it into a surface and attach the preset emitter material to it.

The important thing is that the can lights look realistic in the ceiling. In this respect you should be able to see their rim and then a bright hot spot inside it and, of course, light coming down.

Thank you.


Cosmas Demetriou

[JDHill]

1. Yes and no; you could assign an IES emitter material to a point light, but it would always point straight down, since a point light has no directionality. As to the example of a table lamp radiating in all directions, this would really depend on the IES file used, which is why I mention that it would always point downward. You would more likely want to use a non-IES emitter material with a point light, and this would generate a mesh sphere which would radiate your light.

I would not recommend this though; it would be best to just use explicitly-created mesh geometry instead of a point light, which gives you direct control over the light's size and shape. The auto-translation is provided only as a last-ditch fallback, and you will always be able to beat its performance using geometry created for your specific scenario.

2. In this case, yes you would want to model everything accurately, with the exception of the actual bulb. Generally speaking, it is best to keep emitter objects to a minimum of triangles, and also not to throw away calculations which cannot contribute to the scene meaninfully -- namely, the top half of, say, an emitter sphere which is being used in a can light. It is therefore generally best to use a mesh plane directly (i.e. not a NURBS object) so that you have direct control over the mesh topology. Naturally, you will need to balance out this optimization with how directly visible the light in question is; if you are doing a closeup of this can light, and you can readily tell that the bulb is not modeled as a real bulb, then you would need to sacrifice some optimization for visual accuracy.

[Cosmasad]

Jeremy,

I appreciate your patience. I feel like we're almost there, but some more questions:

1) To create a table lamp, when you say "it would be best to just use explicitly-created mesh geometry instead of a pointlight, which gives you direct control over the light's size and shape" are you saying that I should create a bulb (say a full-size pear-shaped natural looking incadescent bulb) inside the lamp shade, or should I just create the lamp shade (since it would be the only part of the light fixture that is visible to the eye) and then attach an emitter material to it and not bother putting a "bulb" inside?

2) Can you explain what you mean by "It is generally best to use a mesh plane directly (i.e. not a NURBS object) so that you have direct control over the mesh topology"? In the above example, are you saying that the lamp shade should not be a NURBs object? What happened if it was? How come I can attach materials to all the other NURBs objects in my Rhino model (such as the walls and floor) but not to the lampshade?

I think that these questions are important to understanding how Maxwell works with Rhino.

Thank you.

Cosmas Demetriou

[JDHill]

1) There is no 'correct' answer; it comes down to choosing how you prefer to do it. On the one hand, yes, you can create a physical object shaped like a light bulb, and you can assign an emitter material to it. In this case, I am saying that it is best if you create this geometry as a mesh object, rather than a NURBS one, so that you directly control how the mesh is created. As you may or may not know, Maxwell does not (same as many render engines) directly render NURBS; it renders what is know as the 'render mesh', which is a mesh-representation of the NURBS object, and which is, incidentally, also what you are looking at in your OpenGL viewport when you are in Rhino (graphics cards cannot display NURBS surfaces directly). You can control how NURBS objects are meshed using Rhino's Object Properties page, but it is sometimes just better to skip the middle-man and model an object using a mesh in the first place.

Alternatively, as you guess, you could definitely skip the bulb altogether and put an emitter material directly on your lampshade. Here, you could use a normal emitter, or you could use an image-based emitter (i.e. use the emitter's 'Image Emission Texture' mode in the plugin material editor) assigned to the lampshade. As I say, there is no 'correct' answer here. Different strategies will be better suited to different scenarios, and the determination of which method to use comes down to experience.

2) What I mean is that it takes longer to render when you have more triangles in your emitter meshes. As I detailed above, using a mesh object directly, as opposed to a NURBS object, allows you to more directly control how that mesh is built. On the question of why you cannot put a material on your lampshade, I really cannot guess; you would need to describe a bit more about the situation. There is nothing specifically prohibiting you from doing so, but maybe you have found a bug of some kind. Let me know what you meant.