@okh
the .png uses a palette with partly transparent colours(?) as there seemed to be no transparency in the .mtl file.
Yes. And the reason for that, is to enable the user to determine the amount of transparency he wants. With a predefined transparency in the model itself, the user can not choose to have an opaque lampshade.
My own preferences on how I would want a lamp to work, is to be given the possibility to change parameters for every element. By making the bulb a separate, editable element, any colour or texture can be applied to it, making it easier to achieve the desired result. Darkening the bulb, and tinting it with the colour of the lampshade, will produce a softer glow. Giving the bulb a semi-transparent texture, might even soften it more. I will experiment further when time allows.
if you want to create beautiful renderings. One will need to cheat
Yes, of course we need to cheat!
... or?
Since we are hanging in the bar, why not get philosophical (or quasi-philosophical) for a moment :
I think we can agree on the fact that a 3D-model must be rendered as a photographic image in order to be revealed, and to create a photographic image, we must use a camera.
Using a real camera photographing an interior or an exterior in the real world, we will try to create a photographic image that gives us the best visualisation of our object.
Now, in the real world, using a real camera, we can change the exposure by adjusting aperture size and shutter speed. But even that is not enough to emulate the perfect image we get by just looking at an object with our own eyes. A camera does not have anywhere near the capacity to handle the vast dynamic light-range offered by our brain.
So, in order to reduce the dynamic range, so we can capture our image photographically, we must use extra lighting to compensate and enhance our photographic image – and especially when trying to emulate the way our eyes are perceiving interiors.
A camera can just record light – it cannot predict, and it does not have any dynamic selectivity to offer us, like our eyes have when we are scanning the world around us. We are constantly adjusting the aperture, dependent upon the light conditions of the object we are watching.
Rendering photographic images in SH3D, we can't even adjust aperture or shutter speed, so how can we expect a rendered image to be anywhere close to resemble the image we get from looking at a real model in the real world?
Every photograph in brochures and property prospects are artificially lighted – as well as being enhanced during the post-production procedure.
Every photographer is cheating.
But then, what is really cheating? After all we are using 3D modelling software to visualise an idea. If you look at it from a philosophical viewpoint, the idea of cheating cannot be applied to only one single part or event in the 3D-modelling process. Either everything you can do within SH3D is cheating, or nothing is. The lamp is not real, the light sources are artificial – and we are viewing our created image using a computer that has been mathematically calculating every shadow we see on the screen.
The way I see it, cheating is the only way to create an illusion – and you can quote me on that.
cec
