Still trying to perfect my technique for lighting and shading a sphere. Here's what I got tonight. Suggestions appreciated.(Still having issues with drop shadows exporting accurately with blur. Anyone else?)

When you set your canvas to the drawing size, the blur isn't included in the calculation. So you'll need to manually position things so the shadow isn't clipped.

My suggestion when creating a sphere is that the lighting will follow the shape of the outer edge... hmmm... I'm not sure if I can explain properly without drawing something. Basically, the direction of your white hilight is from the outer edge to the centre. In reality the highlight behave the opposite (assuming a the source of light is a "spot" light) and curve radially around the sphere.

I often find that I make mistakes when I try getting my lighting just right. I guess in the end if I can fool the casual looker then I'm good but I always fear being ripped apart by actual artists. Of ten times I have to model a quick-n-dirty in my CAD system (Autodesk Inventor) which has realistic lighting and materials so I can get a sense of what is more correct...not that Inventor is perfect.

Here is a sphere that I made in CAD and gave a semi-matte texture to it:

I like it. It looks more like it came form a raster program than a vector program. Your sphere is very opalescent, like it had a finish 1/2 between a matte finish and a glossy finish. I think you would get a sphere like that if you had 3 point light sources and one large white diffuse light source. And your sphere had opal (or pearl) like light reflecting qualities. Was that what you were going for? Is the problem your having with the export just the clipping of the drop shadow or are there other problems?

microUgly wrote:When you set your canvas to the drawing size, the blur isn't included in the calculation. So you'll need to manually position things so the shadow isn't clipped.

My suggestion when creating a sphere is that the lighting will follow the shape of the outer edge... hmmm... I'm not sure if I can explain properly without drawing something. Basically, the direction of your white hilight is from the outer edge to the centre. In reality the highlight behave the opposite (assuming a the source of light is a "spot" light) and curve radially around the sphere.

thanks for looking and commenting. the drop shadow is nowhere near the edge of my canvas, so that's not causing the cut off. i'm not sure i get your meaning exactly with the lighting following the outer edge, but i knew mine wasn't right somehow. i'll keep at it and try to getwhat you mean.

heathenx wrote:I often find that I make mistakes when I try getting my lighting just right. I guess in the end if I can fool the casual looker then I'm good but I always fear being ripped apart by actual artists. Of ten times I have to model a quick-n-dirty in my CAD system (Autodesk Inventor) which has realistic lighting and materials so I can get a sense of what is more correct...not that Inventor is perfect.

Here is a sphere that I made in CAD and gave a semi-matte texture to it:

is that a single light source or two? thanks for checking this out.

EarlyBlake wrote:I like it. It looks more like it came form a raster program than a vector program. Your sphere is very opalescent, like it had a finish 1/2 between a matte finish and a glossy finish. I think you would get a sphere like that if you had 3 point light sources and one large white diffuse light source. And your sphere had opal (or pearl) like light reflecting qualities. Was that what you were going for? Is the problem your having with the export just the clipping of the drop shadow or are there other problems?

i was going for a sort of pearlescent look, so thanks. i clearly need to work on my understanding of lighting, though, since i was thinking about just a single light source.

llogg wrote:i was going for a sort of pearlescent look, so thanks. i clearly need to work on my understanding of lighting, though, since i was thinking about just a single light source.

You can get more than you ever wanted to know about lighting sources from wiki:

Ray tracing http://en.wikipedia.org/wiki/Ray_tracing_(graphics)





Specular vs diffuse: http://en.wikipedia.org/wiki/Phong_shading


Still more:
http://en.wikipedia.org/wiki/Global_illumination

http://en.wikipedia.org/wiki/Ambient_occlusion

http://en.wikipedia.org/wiki/3D_computer_graphics

Its not going for photorealism, but i did a tut on shiny spheres a while back...

http://john.cliff.googlepages.com/marbles_tut.svg

I also see that there must be 3 light sources on your sphere, llog. But the cast shadow doesn't quite seem to fit the highlights I see. However, the shadows on the sphere itself seem to work for me.

When I 1st started to seriously work with Inkscape (seriously for me, lol), I used an orb or sphere for practice. But because I needed to start very, very simply, I used only one light source. And for the same reason, my orb was suspended in space, so that I didn't have to deal with a shadow cast by the orb/sphere. Also partly because I don't seem to have trouble with cast shadows of single light sources, so long as they are cast on a flat, horizontal surface. Anyway, to be honest, mine didn't look quite right, I believe, because there's only a single light source. Because usually there will be more than one light source, in reality.

So I can't offer any concrete suggestions for correcting the cast shadows, since I haven't tried to use more than one light source in my own work. But I guess I would start with making sure there's a cast shadow opposite each highlight (or vice versa). After that, it would be trial and error for me, getting it all to meld.

Wish I could be more help But I'll be interested to see how you handle the cast shadows of the multiple light sources.

The best tut I found is on this page:
http://magazine.redhat.com/2007/02/27/t ... ur-filter/
But I used parts of these also:
http://chrisdesign.wordpress.com/2008/0 ... gradients/
http://ryanler.wordpress.com/2007/02/16 ... ll-bubble/

I think you can find the answer to your issue in the svg 1.1 docs:

http://www.w3.org/TR/SVG/filters.html#F ... ectsRegion

"The bounds of this rectangle act as a hard clipping region for each filter primitive included with a given 'filter' element; thus, if the effect of a given filter primitive would extend beyond the bounds of the rectangle (this sometimes happens when using a 'feGaussianBlur' filter primitive with a very large stdDeviation), parts of the effect will get clipped."

I had the same issue and it solved it by adjusting the x,y,width, & height attributes of the filter element. Keep in mind that the values are relative and need to be described as percentages.

~Gary

Not sure if it's any help but I had a go at it (it's 4am and I can't sleep... )