However, it should be noted that anything that impacts the final scene, despite being off camera, should remain unhidden. Anything off screen – body parts, other models, parts of the background environment should be hidden before any rendering takers place.
Summaryīy removing anything in our scene that won’t be shown through the camera viewport in the final render we can save a lot of time just be hiding it. Sometimes we’ll save less, sometimes more.
Now we won’t get the same 16.64% saving in every scene as the final saving we actually get will be dependent on the assets in each individual scene. In this example with this model and scene we’ve managed to save 3m 29 seconds over the 1000 iterations we ran the render for or 16.64% time saving without any difference in our final image. Now that we’ve removed all of the geometry that isn’t going to get rendered – either because it’s off camera or hidden under clothing – we can rerun the render again and can see the saving of time that hiding geometry grants us compared to leaving all geometry unhidden in scene. So when we start our render – only her head, neck and jumper will be rendered and, importantly, calculated while everything else will be ignored. Even if we remove the jumper she is wearing – we can see that she is just a disembodied head. The closed eye icon next to each bone shows that it is hidden. We’ve achieved this by turning every bone that isn’t going to be rendered within the models skeleton to hidden. So if it’s not rendered yet still calculated – why not just hide it completely? However, Iray still needs to include this off camera or unseen geometry when calculating the bouncing of light rays, thereby slowing down the render. And, often missed, her torso and arms are hidden underneath the jumper that she is wearing. Likewise, her jeans and trainers are not rendered. The models legs, feet and hands are not rendered. However, even though the geometry which we can’t see isn’t rendered, it still needs to be considered when Iray is calculating iteration to iteration. If we take a look at the render above we can see that the lower half of Margot’s body is not displayed in the camera viewport and therefore, not rendered. It’s worth noting, for this test, we never let the render go to completion but stopped the test instead at 1000 iterations. When we get Daz 3d to render our image out we can see that it took a time of 20 minutes and 56 seconds to render. In this case – I’ve set the camera up to show the upper half of Margot’s body. Once we are happy with our model and her pose, we can set up the camera to our liking. The lighting used in this scene is LI Incandescent Portrait Lighting.
The scene that we are working with consists of a single character model – Margot for Victoria 8 – standing within a small studio.
0 kommentar(er)
