Sunday, 5 February 2012

Mr Potatohead comic


I AM NOT SORRY.

(I really am)

Awful puns aside, I tried delving a bit deeper into Softimage than just using the basic transformation tools. I toyed around with some of the deformers and modifying tools. Feeling brave (and perhaps a bit stupid), I dipped my toe into the vast pool of materials and rendering. Mostly just for fun.

The first thing I noticed was that when moving and positioning the completed figures, selecting all the components and attempting to rotate them led to things flying all over the place. I think this is probably because each individual component (eyes, nose, body, limbs etc.) each has their own pivot point — so rather than rotating them on one global axis, they all get shifted independently. To get around this I created a transform group for each figure.


A transform group basically takes all selected objects and gives them one shared pivot or centre point, allowing them to be moved and rotated as a group. This made moving my models into position much, much easier!


To pose the models I altered the centre/pivot point of each limb that was to be moved — in most cases this was just the arms, mouth and eyebrows. Softimage has a couple of ways of doing this but the quickest method was to simply hold the Alt key (with the translate tool selected) and drag it into the new position. This isn't a permanent solution however, and once you deselect the object the anchor point will go right back to where it was before.


The first panel was pretty straightforward. Once I'd posed the figure using the new anchor points and the rotation tool, I simply adjusted the camera to give me the desired framing/perspective of the shot and set a 'camera memo' so that I was able to quickly hop back to that exact point after making adjustments.

I wanted the shot to have some dramatic lighting so I dropped a very basic 'spotlight' from the "get primitives" submenu. Using a combination of the viewports I was able to position the light roughly where I wanted it at the base of the figure and angle it upwards using the rotation tool. By using the camera memo I saved earlier I was able to pan around the scene quite freely to check the position/rotation of the light without fear of having to re-align my camera every time.

Once I was happy with the light, I maximised the viewport and took a high-resolution screenshot to take into Photoshop when the time came to assemble my comic.


For the last panel I wanted Mr P wearing sunglasses — I made a hamfisted attempt at modelling my own using the primitive shapes available but it looked absolutely horrific, and so I rapidly scrapped the idea. Fortunately, there are some great archives of 3D resources available for free, so I nabbed a nice pre-made model from turbosquid.com and imported the .obj file into Softimage.

Using the Explorer I was able to quickly select and locate the imported model and bring it into view with the rest of the figures.


I was then able to scale the sunglasses up and rotate them appropriately so that they (mostly) fitted his head — being shaped like a potato, it wasn't going to be perfect, and I did have a crack at using a 'bend' deformer to wrap the arms of the glasses around his head, but I couldn't quite get the hang of it. 




To get the reflection of the glasses, I simply selected the lenses (which were separate objects/shapes) and went to Get > Material > Architectural, a relatively expansive material type that allows you to mimic all sorts of object types, from reflective chrome to frosted glass. Being relatively new to the whole materials thing I stuck to the safest options — Illumination and Transparency/reflectivity.

The small box you see over Mr Potatohead's eyes is a 'Render Region' — it allows you to select a small portion of the scene and will generate a rough, low-res preview of what the rendered version will look like. This is extremely useful when working with materials — such as the lenses of the glasses here — as it updates in real-time, allowing you to see the changes as you make them without needing to do a full render each and every time.

I simply dropped the transparency slightly and upped the reflectivity, changing the colour to a darker grey/black to make them more sunglasses-y.



The last-ish thing that I did was use the polygon selection tool (found at the top of the MCP — shortcut Y) to grab the lower half of Mr. Potatohead's arm. I wanted to bend it at the elbow very slightly to give it more the impression that his hand was on his non-existent hip — trickier than it sounds. His arm is relatively complex and consists of more polygons than I realised, so I had to muck about with the camera position to see 'inside' the core of his arm to select the remaining polygons that I'd missed!

Once you have polygons, edges or points selected, you can transform them using the same tools as entire objects. I simply rotated it very slightly and pushed it back a little with the translate tool.

For dramatic effect I then dropped a light into the scene and positioned it somewhere between the two potatoes so that it would give a suitable reflection. To use the prettied-up render shot in my comic, I maximised the viewport and hit the 'preview' button under the render menu, then copied a screenshot into Photoshop.

Some unused panels:


He was supposed to be holding an eyebrow, but it looks more like a liquorice allsort!

No comments:

Post a Comment