Texturing the Destiny Ascension

After some searching for a decent metal texture, I decided it would probably be simpler to make one myself. I found this tutorial, and quickly mocked up a good texture for the Destiny Ascension:

http://www.photoshoptutorials.ws/photoshop-tutorials/drawing/metal-texture/

(I'm just glad my computer is capable of running an internet browser for reference images, Photoshop, and 3DS Max simultaneously. This process would be a lot slower if it weren't.)

This is a simple test, without any lights or anything fancy - it's not quite right, but I think the blue is close enough. Hopefully it'll look right with lighting - it may not be as shiny as in the original, but it certainly looks something like the same shade. It's probably the best I could get it without hours of fiddling. The next step, obviously, is to texture the model with something other than one uniform shade, and add in the details.

I can tell in advance that the lights on the source material are far too complex - there's no way I'm going to be texturing in each light individually - instead I'll probably create a simple strip that can be repeated over and over again. It won't be as good as the original, but I could easily work for weeks on trying to replicate the lights exactly.  

Comparing my attempts to BioWare's

The difference in the angle of the spindle here is merely the model being rendered in a different rotation - since it has no clear "up" beyond this, it's easy enough to fix with a simple rotation tool. For the most part, I am pleased with this model - it clearly looks like the Citadel, albeit with a few obvious (albeit relatively minor) errors and parts I'm not happy with. The most obvious and glaring issue is that BioWare has significantly greater resources than myself, working alone on my laptop. Given that I don't have a team of animators, or any sort of budget, I'm overall somewhat proud of this model. I'm not sure the central ring is quite wide enough, it doesn't quite seem to match the source image. The small triangular sections on the end of each of the Citadel's arms aren't quite at the right angle - they seem a little too steep when compared to the original model. I'm not quite sure the arms are large or curved enough either - although in the first image, it looks a lot better than the second. Aside from the potential error in the curve, though, I feel I got the shape and length of each arm correct - I'm very proud of this. I also think the spindle is the right angle, size and shape.

I am much, much happier with this model, although I do need to upload some better pictures of it to this blog to properly show it off. This one took me significantly less time than the Citadel, and yet I think it's significantly better. I think the proportions are almost entirely good, although the squashed tube that forms the middle hole may not be quite wide enough, though. I could probably have used a few more polys, it appears to be a little blocky, but the angles are good, and I'm struggling to find any major problems with this model. I'm sure there are some, but I am quite happy with it.

Mass Relays are harder to model than you'd think

 When modelling the mass relays, I learned an unpleasant lesson: The number of polygons in a cylinder matters. I began the top and bottom halves with a simple cylinder, which I converted to an editable poly and extruded into the correct shape. Unfortunately, I was unable to extrude parts of the model - this cost me a lot of time, as it wound up being not entirely circular, and I probably lost at least an hour of my time by trying to fix this - I wound up having to add in a whole other cube, instead of extruding a section for the bottom. The result doesn't quite look perfect, especially towards the loop end of the relay, where it doesn't match up to the rest of the part very well, but it would take far too long to re-do the entire shape of the relay.

Learning from my mistakes, I ensured the number of polys in the second cylinder matched up nicely - this side of the relay progressed far more quickly. Lesson definitely learned: Plan ahead the number of polygons in a primitive before converting it to editable poly.

Some amount of modelling later, I was left with this:

Clearly something is not quite right here. The very top and the very bottom of the curve is smooth, and the rest is not.

http://docs.autodesk.com/3DSMAX/16/ENU/3ds-Max-Help/index.html?url=files/GUID-FF7D7633-03AD-4427-821A-65F8AC484CDD.htm,topicNumber=d30e177574,hash=WS73099CC142F48755-1257E12111BF108800E348C

Some research gave me this link, leading me to the conclusion that the smoothing groups had the model smoothing over the shallow curves there, but not the rest of the model. The top half of the model was easy to do (Auto Smooth worked perfectly), but of course the tricky bottom half of the model wasn't as easy.

The shape looked a bit squashy here, so I did some editing, and when Auto Smooth didn't work, I had planned to turn on smoothing groups at random, and try to work out a pattern until I got it right. Luckily for me, the first two I turned on sorted the problem out entirely, leaving me with no need to investigate further:

The model is still not quite finished - there are some small parts left to model, but they won't take long.

After this, I did some research on the spinning rings in the centre, since they'll inevitably need animating - and since it's a simple rotation, it shouldn't be hard to figure out. I used the following link to analyse the mass relay:

https://www.youtube.com/watch?v=bGr9T1k8QOY

This provides a good close-up of the spinning rings - the inner one spins on the Y-axis in the screenshot. Meanwhile the whole lot spins on the Z-axiz, as shown in the following screenshot:

This will require a bit of planning, and two separate animations - I know how to do it in Unity, unfortunately that's not much help here. I'm sure I can work out how to do it in 3DS Max, though - a simple matter of going over last semester's notes on animation. Thankfully, the lovely slow-motion close-up of the Alpha Relay in the Arrival DLC for Mass Effect 2 gives us a lovely image of the two rings spinning. 0:50 in the video is the best example of this.

The model is currently missing only a few small spindly parts, then I'll call it complete. These shouldn't take long at all to model - a few cylinders ought to do the job. After that, I'll provide a full analysis of what I've done right and what I've done wrong (plenty) in creating this model, along with the previous two. My progress so far: