During this project I went through a lot off different processes, the first of which came when designing the gun;
Silhouetting, kit bashing, mood boards, splicing objects together, sketching and randomizing with a dice. I chose to use all of those apart from randomizing,
mainly because I wanted to experiment with techniques and learn new processes.
There are multiple iterations of this stage to get you to your final design.
Next was the modelling stage where I used the design I have come up with, and
model it in 3ds max. After this I went on to the colour exploration phase. I
used my final silhouette and overlaid it with colour. This allowed me to
quickly make iterations of colour without going into too much detail, giving
me a rough idea of what looks good and what doesn't. After this I disassembled
my model into the 4 required pieces and imported it into unreal engine 4 to
test that it works. When I knew it worked fine, I moved onto the texturing
stage. Throughout texturing I constantly applied the texture to my model in UE4
to see how it looked. I also went into more detail on the colour exploration
when I was texturing, as you can get a better idea of what the colour looks
like when it’s actually on your model. Along with my albedo texture, I also made
a roughness, metalness and normal map to go with it. While using these
processes for the design of the gun itself, I also used some of them for the
projectile too. Finally I imported my final textures into UE4 and altered the
settings accordingly, setting the desirable speed I wanted it to shoot, the
bounciness of the shell etc.
There were a few things I could have done differently to improve my project. To start I wish i'd have designed something different to model. so I could do more with the blueprints in engine. Mainly because I would have learnt a lot more. others did things with lasers, particle effects, rotations of the mesh and more stuff on illumination. Things I'll have to incorporate into my next projects or work on in my spare time. Next I feel like I didn't put as much design into the gun texture and look as I could have done by designing something else. Because my sentry was based on a series of existing products (Black & Decker tools), the overall look was set out for me, it was nice to do this, but next time i'll go for something more original. And finally if I were to redo the project, I would like to have done more with Z-brush. A program we don't need to use really, but if the time is there, then I don't see the harm in experimenting with it a little and incorporating it into future projects. It is a skill that will be helpful in the future for a lot of us.
 |
| White box renders of my sentry in UE4. |
Overall I feel the project was a success. I
finished all that the brief required. I used a lot of different processes in
the designing stage which I’m happy about, as I learnt a lot. Modelling went
well and I managed to keep to the 5000 tri budget, with a whole 6 tri’s to
spare. Texturing was questionable, as I mentioned above I wasn't sure what to
do for the main body of the gun because it felt plain to me. Plastic doesn't have much texture too it, but leaving it a flat colour seemed like a huge waste
of my texture sheet, and left the gun looking boring. But with a slight grainy
brush I managed to make it look good in engine. I’m still unsure about the
space I used on the unwrap, but it does look good in engine so I’m happy. Along
with this, the rest of the texture sheet looks good too, in my opinion.
My favorite part of the project was putting
the sentry in engine. There’s just so much you can do with blueprints alone, to
bring your gun to life. I had a go at working some additional blueprints in
which I didn't end up using, but I did learn about them and that’s the point.
On a side note, as the project progressed I noticed others had done some really
extravagant things with their project. This left me feeling as if I’d hung back
and done a really boring gun. My aim original was to make mine more realistic.
But some of the guns I saw changed my mind in an instant. When making a
semi-realistic gun, you do cut yourself off to a lot of great things. So next
time I’m going to go a bit wackier with my ideas.
To summarize, the processes I used include
silhouetting, splicing and kit bashing for the initial design, along with
sketching and a bit of painting to help me along to my final design. Doing all
these processes gave me a wide range of techniques and helped me generate ideas
quickly and easily. I did a few iterations of this process. Then I modeled my
design in 3ds max, and changed things I didn't like, or that didn't look right
on the gun. The normal I made had more definitive detail on it than the albedo,
roughness and metalness maps. This had to be done as the texture of the gun
should be pretty plain, it’s a relatively new gun and wouldn't have any deep
scratched or large marks on it, whereas the normal shows the design on the
side, a design that would have at least tripled my tri count to recreate in
geometry, something that you wouldn't do even on a next gen game. I then
imported my textures and meshes into UE4 and set up the blueprints for multiple
projectiles, edited the bounciness and speed of them and I was done.
 |
| Final finished sentry in UE4. |
No comments:
Post a Comment