Short answer: You don't. Ray tracing makes hard shadows, shadow maps make soft shadows. You can play a little with the shading rate of the shadow receiving object, but that's about it. Long answer: Ray traced shadows work on a simple yes/no principle. For every point shaded, the ray tracer answers the simple question "Can I see the light from here?", determining wether or not the point will be bright or dark. Since the lights in Poser are point-sized, there is no "I can see parts of the light" answer, it's straight yes or no - shadow or no shadow. Then there are algorithms to do soft ray-traced shadows, sometimes called area lights or area shadows (Poser 5 does not do that, so this paragraph is about how other programs create soft ray traced shadows). These simulate a non-point sized light source, by pretending there was a finite number of point sources distributed over a certain area. That way, the answer to the question "Can I see the light?" would be "you can see 4 out of 10 points of the light source." Unfortunately, replacing one point-sized light source with, say, 16 point sized light sources has unwanted consequences: Obviously, it'll take 16 times as long to calculate that shadow. Furthermore, the answer is "you can see n points", with n being a number between 0 and 16 - leaving you with 17 levels of "shadow", which will result in visible edges - you'd need around 200 different levels to make the the edges disappear, which would take 200 times as long to render. In order to avoid that, people had the idea of "jittering" the sample points, which would replace edges or aliasing, as it is often refered to, with noise. Noise is something the human eye is much better at ignoring than aliasing, with one exception: Animation. If you use jittered samples to create soft shadows in an animation, the noise will create ugly flickering. So, the bottom line still remains: For soft shadows in Poser, use shadow maps.