A gun of substance

So I kicked off the week by baking my normals again. I cannot stress how stressful this was, because something would always go wrong, but hey, I managed to fix it all in the end, it just ate up more time than I was expecting it to. I did however learn a lot about baking in the process of things wrong, so there's always an upside.

I intially only baked using X-normal, but didn't realise it needed to have specific settings for hitting more complicated objects (I'd only ever used it to bake simple information onto objects-like creases). I also tried baking in max, but this only gave me a 2048 resolution, and whilst that is the one our textures need to be, you get a better quality bake if you bake at 4096, then down-scale it to 2048, which you can do in X-normal.

In the end I made a cage in max, then exported it and used it as a cage in X-normal. Creating the cage as a mesh was useful anyway, as it allowed me more control over what faces were being covered by the cage.

After doing my main bake and fixing the map however, I realised that I was still missing parts of my normal detail. I didn't want to do the whole bake again, so I ended up just re-baking the parts I'd missed and replacing them on the main map.

I also recieved feedback on baking the hose part of my model. I was told I'd get a neater bake if I made a cross-section of the detail, baked it onto a plane, and then put the bake over it's corresponding UV Island. If I had the time, I could also have done that, and the skewed the Uvs after the normal map had been applied, to get them to be horizontal lines, as skewing the lines after they'd been baked would keep them clean, whereas baking them skewed would create a bit of pixelation. I also learnt about how to use smoothing groups to make sure that detail only baked onto the places I wanted it to.

Here's how I baked the hose detail:

In the end I think I created more work for myself because I was rushing, and so kept making silly errors, that then took up time to fix. To be fair though it was also my first proper bake (with this much detail), so errors were bound to be made as part of the learning process.

I also ended up having to make lots of corrections on my models as they were another reason certain things didn't bake properly. But again, it's taught me a lot about what I need to do to have a successful high-poly bake. One thing I did learn especially, which I hadn't realised until after I'd done my baking, is that it doesn't matter if highpoly detail is hidden by the low poly, as it'll still bake on. For example, if I've modelled a dent into the high poly, which is then hidden when the two are overlayed. This detail will still bake. I made the mistake of assuming it wouldn't as it couldn't be seen, but hey, it's a useful thing to know from here on out.

So here are all my normal maps:

And here is the final one:

The next thing I did then (because it needed to be done, and I wanted a small break from 3D), was do my value iterations, and final colour pallettes to end up with a finalised design, I could take forward into a polished final concept.

A third year helped me do my value iterations really quickly by using a clown and lighting render from max. Essentially I split all the parts of my model up into what would have different values, then applied a different value to each part (using a standard gradient-white-mid-dark, overall I think I used about 7 different values in max, though on reflection I think I should have used a standard 9Value scale for consistency). I then rendered my model out and put the renders into photoshop. This meant that I could easily just select and fill the different values with new ones to create really quick value iterations.

I then picked my favourite/most popular value combination, and did one with the detail added to see how the final design would look before I put colour on it. The tricky part of doing the value iterations was that the different metal pipes would have to be the same value in each iteration, it was only the things around the pipes I could change the value of. However I think I ended up with a design that reads quite clearly, and is nicely broken up, although colour will certainly help in breaking it up just that bit more (as the seperate pipes will be more obviously different then).

I then used my finalised colour scheme (I picked all the colours from the most popular colour concepts and created a swatch from them, as well as choosing set colours for the metal and wood), to do a few colour concepts. I only needed a few as I already had a very clear idea of the colours I wanted to use, and where they'd go on the design. The patterns I put onto the design were inspired by cells, bacteria and popular retro patterns, as well as from portal's very own distinctive patterns, as shown on the moodboard:

The most popular colour scheme was selected, and now, at long last, I have a final concept.

Here are my value and colour iterations:

And here is the finalised concept at it stands:

I found out this week that apparently our deadline is the 11th of January, not the 18th of December. Quite a big difference, and a huge relief. Because this now means I can spend up to the the 18th of December finishing as much of the 3D/engine stuff off as I can, and then spend over christmas doing my promotional artwork (and any stretch goals I don't have time to start/ finish before we break-up). Hurray, so it looks like I'm not as behind schedule as I thought I was.

Anyway this means that thats where I'll leave the concett until christmas, as next week will be mostly engine stuff now.

Right, so onto the dreaded texturing, made far less painful by the beautiful substance painter.

I've never used substance painter before, (apart from my material iterations) so getting used to the programme took me about a day, but once I was used to it, my life became a lot easier. I have to credit this tutorial:

https://www.youtube.com/watch?v=IhBVsn2tfGc

(and a few third years) for teaching me the programme's basic functions. I then created most of textures by having a play around and cannibalising smart materials. For example I looked at how they made their rusted copper, and then re-created it using my own effects, values, etc. It took a while to get confident with the work flow, but eventually I got the hang of it, and ended up with some pretty nice textures by the end of the week. (I managed to finish all the metal parts of the gun, but I'll need to spend monday doing wood and other bits and bobs-but the bulk of the texturing is done now-yay).

The plan is to use substance to lay down most of the textures, and then anything I need more control over, I can edit in Photoshop. I suspect I'll need to do some editing anyway, as substance textures won't look the same in engine as they do in substance (unfortunetely).

But so far so good. I'm pleased with how my textures are coming along. I ended up baking my own AO for the turret, as substances default didn't work for me. I also had to edit the curvature map abit. But it was fun to learn about a new map, and start really thinking about my texturing and how to make it as good as it possibly can be (as thats one of my weakest areas at the moment-I think).

If I did this project again, I'd definitely spend more time getting the bake right from the get go- as it would leave me with a lot less constant editing to do.

I didn't get a chance to make glass in engine, but as next week is now engine stuff anyway, it won't be too much of a set-back. I did however start looking at tutorials on fireballs and forceshields.

Here are the tutorials I've found, which I'll watch and apply hopefully next week:

https://www.youtube.com/watch?v=enx6ShyKUL8

https://docs.unrealengine.com/latest/INT/Engine/Rendering/Materials/HowTo/Fresnel/index.html

http://www.tomlooman.com/the-many-uses-of-custom-depth-in-unreal-4/

As a useful set up for next week, this week in Game Production we went over blueprints and how they work. This will definitely come in handy when getting my turret to work in engine.

In visual design we received feedback on our desings (this was where I recieved feedback on my unwrap and pack).

In Ciritical Studies we went over what we'll be doing for our marks in the module. It was really useful to get a heads up on what we'll be doing, and hopefully everything will work out okay when we start doing them.

So okay, next weeks action plan:

• Finish textures/check in engine-fix any issues that may arise

• Make Glass shader

• Put final gun into engine and get it working with the blueprints

• Do Projectile and Cartridge particle effects (as mine will be effects not models)

• Do Muzzle flash particle effect

• Make the liquid in the beakers move

• Create the force-field shiled

• Tidy up-finalise what I can before X-mas

• Do emissive for sun-lamp

• If I have the time model and unwrap the anode, cathode and wires connecting them to the tops of their respective beakers (as I only just realised I missed them out).

I think next time I work on a model like this I'm going to keep a note of what details and parts it should have, almost like I did to plan for the previous project. That way I can keep track of what needs to be there, and I won't keep on realising 'oh I forgot to model this, or bake this, or texture this, as it's all too easy to do when there are lots of components and little parts to your model.

Well I'm feeling a lot more positive this week, and I'm looking forward to next week, as engine stuff is something I particularly enjoy. Hopefully I'll get the bulk of the work done next week, but if not, there's still christmas! Happy December everyone.