This is the final week we have in order to complete our sentry gun and I can't wait to see what sort of character we have to make - it doesn't feel that long ago that I was saying the same thing about the film scene project! Making the sentry gun has been a lot of fun, but as with everything, I wish I had done more development and refinement. This is a common problem which I hope will be at least improved by following the method that Chris has told us about this week.
With this method, I hope to improve my workflow and how I go about my creative decisions in order to end up with a refined design that I am happy with.
Monday's class with Craig was so helpful to me that I'd still have been 1 step behind if I didn't ask a couple of questions. One of the questions I asked was about Z-brush. I know this is rather late, but I haven't used Z-brush yet to make any of the models I've produced for the duration of my course. I simply asked about how to go about importing a low poly mesh, subdividing it, how to sculpt on it, then export it. This is such a vital process for getting low poly meshes with good looking normal maps, creating a nice, organic feel to the model, and therefore, more realistic. I am so happy to have a very basic understanding of this, alongside a baking method also demonstrated in this lessson.
Another question was about how to get lights to attach to a mesh and move with it as it moves through space. Even though we didn't get to figure this one out, it was very good to just relay comments which actually made me decide to just change the emissive map which wasn't actually right (solving a problem).
During Monday, I also managed to get into cascade in UE4 and mess around with the particle effects which were already in the file given to us. I know I didn't personally make the textures that make up the particles, I still learnt a great deal and made progress in the world of particle effects. I changed the lifetime, size, light emitted, layered particles, and changed the colour of them. The particle effects window is something I would have previously been scared of, but I now feel excited to see it open and I really can't wait until I start making my own textures to be used for different effects.
I more or less finished off the main textures for everything on Monday, but tweaked them slightly on Tuesday along with the particle effects.
Tuesday was a little bit different this week because we had a lecture from Jeff - experienced in life drawing. We had our usual lessons, just squashed into the morning sessions. Life drawing this week focused on heads/shoulders which I found very useful to do as I really struggle with drawing the face & capturing the essence with not a lot of detail.
Not the best drawings, but it's good that we had this time to only look at faces, I think that having done that, I will be more likely to actually put a half-decent looking face in my life drawings instead of leaving it to the last moment and potentially ruining the drawing.
Later on during this week on Thursday, we had a little session to introduce us to the next character project. There was a lecture to brief us all on what we'd be doing, and a seminar to go into some depth. Within this, we did a few quick 2min sketches of the people around us and tried to emphasise some of their features (or in my case, just simplify the face).
The brief for this project is to design/concept 2 characters who belong together, but have completely different characteristics/personalities (Also known as dichotomy). For example, Shrek and Donkey, are 2 characters who are almost nothing like each other, and yet, they seem to work very well as a duo:
The brief is to design two characters, but leaning towards more humanoid ones. By the end of the project, we should have modelled and textured at least 1 of them, and perhaps with an idle animation.As of this moment, I have no idea where I want to go with this one, but the only way to find out is to research... The first part in Chris' design process.
Throughout the rest of the week, I just kept adding into the Unreal Engine file, trying to do more and more things with my sentry gun. With the help of fellow pupils, I managed to get some sound working whenever the sentry gun is fired. I solved the problem where I wanted to attach lights to the moving mesh, and, I also managed to get a laser beam to work and follow the player as it rotates:
Some work which I previously forgot to post:
In addition to this, I produced a little poster for my sentry gun, plus, a physical model of it:
Here's some beauty shots of my sentry:
The screenshots really don't do it justice, it looks so much nicer when you can see it in motion with the mussel flash+laser.
In summary, a good week again, I just wish I'd been able to get a better start on the character project because I need as much time as I can get on the initial 2D/concepting stage. I'm excited to be doing this dichotomous character project, but the idea generation is the part that unsettles me a little (at this moment in time).
Sentry Gun Project - Post Mortem
Problems faced
with 3D modelling:
-Tri count: I had to eliminate around
900-1000 tris in order to make the whole model under the budget. It even
required me to bake on an edge loop for the ammo just to get rid of a few more
tris (190 to 118 for the bullet):
Problems faced
with texturing:
-I’d say that there wasn't exactly any problems
with the texturing as such because the way I planned it out made it so that I
could end up with some carbon fibre, metallic gun, but also have room to play
with colours, grunge, and overall feel of it. This drove me to make some shiny
alien-like skin which came out fairly good but it could be better.
-The emissive maps aren't hard to make, but I
have to admit I found it a little challenging trying to get the kind of effect
I imagined in my head. It still isn't right but I think all of the texturing
just need refining a bit further to take it to that next level.
Problems faced
with Unreal Engine 4:
Small problems with easy fixes:
-Sentry not firing = The collision mesh on the barrel/body
was interfering with the ammo spawning.
-Sentry all backwards x2 = Imported facing the wrong way as
well as the pivot point’s transform+scale not reset.
-Barrel was floating as it rotated = I put the pivot point
in the wrong place… It slipped my mind.
-The mussel flash not spawning where I wanted it to = Move
the placement in the game engine.
Summary of the
project:
In my opinion, the project was very successful because I
fulfilled the brief, and created a fully functioning asset which aesthetically
is around the ballpark I wanted to get in after I realised what design I wanted
to follow. There were a few things I didn’t get to do e.g. Z-brush sculpting,
but it wasn't completely necessary. I did however learn the starter interface for
Z-brush, so I am excited to use it in future projects. The normal map wasn't
very good, but the overall effect of the sentry gun works well; it looks mean,
it hits hard, and the silhouette is very distinctive.
What I’ve learnt:
-Baking normal from high poly to low poly meshes - Basic
interface for Z-brush/how to export & import meshes in order to bake them
(I only used baking for the bullet, but I still played around with Z-brush
which is not seen on the final model - Will be used for my future projects).
-How to use the cascade particle system to a certain extent
in UE4.
-To explore the blueprint functions and find what you’re
after to work out what exactly it is you want something to do. E.g. Make a
sound file play when an event happens.
-How to take ‘beauty’ screenshots properly in the game
engine.
-That the modelling/unwrapping time really doesn’t need to
be as long as we think – I.e. don’t ‘long it out’, set a deadline to finish it
and stick to that deadline. This principle applies to all aspects of the
project.
-The sooner you can get something into the game engine and
test it/iterate it, the more refined it’ll be at the end of the project; not a
rushed thing at the end.
What I would do
differently:
-Spend more time on the creation of the normal/roughness
maps to really refine my work - make it look a lot more polished.
-Give the concept stage a bit more development - try messing
around with more pages of colour schemes to really try and define what look I
may like to do instead of going with the first cool idea (I did do
experimentation during the texturing stage, but it isn't very evident other
than that the final model has a different effect than the page which described
what materials I wanted to use.
-I think I could have done with some more 3d experimentation
too. I modelled almost right from the orthographics and altered the design a
little bit, but It would have been wise to try different shapes and firing
systems - It was only towards the end that I realised how nice it would have
been to have three barrels instead of just the one.
-I would like to have really defined what look I want to go
for using some describing words. This would have helped me in creating a
certain feel which would make it more cohesive as a design and therefore, more
effective. E.g. If I said "mean, alien, gnarly", I would have been
more inclined to perhaps include things which make it look much less like a
man-made machine, and more like something which has a life of its own.
