Thursday 27 November 2014

Week 8: Sentry Gun Post-Mortem
15:19

Week 8: Sentry Gun Post-Mortem

Project:
For this project, I had to design and build a sentry gun. I’ve been given total freedom on the style and design choices which is the total opposite of the previous project. I had a limit of 5000 tris which seemed like a hell of a lot when I built the first model to use for paint overs, but I found myself quickly running out of tris when building the final thing!

Summary:
After the project had been set, I started my research by creating a board on Pinterest full of various things that interested me and had quite cool shapes. Once I felt as though I’d collected a sufficient amount of reference, I started creating moodboards in Photoshop to aid me in creating a bash kit. When creating my bash kit, I used the lasso tool to select all of the interesting shapes and filled them in using block colour so I could later place them together in order to create new assets in a fast and enjoyable way.
From here, I then went on to ask my peers to choose their favourites so I could create iterations and hopefully become that little bit closer to getting a final design!
I decided to take the third idea forwards as I really liked the shape and I felt that it’s something that I’d be able to iterate on even more and still have a fairly simple yet interesting design. After choosing the design, I made a quick 1000 tri block out in 3DS Max to aid me with the paint-overs in order to get the proportions correct. I also did a few iterations on the neon lights and added a few extra assets - as shown on the final design on the left.





















From here, I then went on to ask my peers to choose their favourites so I could create iterations and hopefully become that little bit closer to getting a final design!


I decided to take the third idea forwards as I really liked the shape and I felt that it’s something that I’d be able to iterate on even more and still have a fairly simple yet interesting design. After choosing the design, I made a quick 1000 tri block out in 3DS Max to aid me with the paint-overs in order to get the proportions correct. I also did a few iterations on the neon lights and added a few extra assets - as shown on the final design below.


Beauty Screenshots:




Here’s the final turret in UE4! I keep forgetting that 3DS Max doesn’t support PBR and always end up feeling disheartened at just how bad everything looks. UE4 seems to make everything look so awesome, I love it! I’m actually really pleased with how it looks in engine – definitely makes the messing about with UV packing worthwhile!

Technical Screenshots:


I used my favourite method of modelling again for this project – strip modelling. I find this much simpler than box modelling as there’s much more freedom with topology and edge flow. As my design is based on the curvature of Art Deco appliances, I had to work out which areas needed to contain the most amount of tris. Looking back, I realise I could’ve removed some edges from the base and had more in the barrel as some parts of it look slightly blocky in comparison to the rest of the pitch.


 Unwraps:


I always seem to enjoy unwrapping and packing UVs, but for some reason, I didn’t like it at all for this project. I think I re-packed it all about 6 times before getting fed up about too much unused space and left it for the day to play around with the model in engine. I had another go the following day and enjoyed it a lot more and managed to get a fairly efficient layout. I’m still not 100% happy with it but it’s much better than what it was!





 Additional Information & Screenshots:
In addition to the set brief, I decided to set myself a stretch goal to really push myself and escape the dreaded comfort zone. I decided that if I had enough time, I’d try and animate my emissives. As I’ve never attempted them before, this was quite a daunting task - so setting myself the goal of making them flash seemed horrific at the time, but was something I wanted to do to try and achieve full potential. I was surprised at just how easy it was. I had all of these visions of having a breakdown because everything was going wrong, but for once, everything went right. Which worried me even more.
I also got set the task of creating a poster to advertise my sentry gun. This involved me creating a whole new Pinterest board full of awesome advertisement related stuff. I’d based my gun on the curvature of Art Deco appliances mixed with the bright neon lights found in Las Vegas so as you can tell, the moodboard is extremely colourful.






From this, I took a variety of screenshots in UE4 and applied different filters to them in Photoshop and adjusted the hue.
I was planning on using these for the background of the poster, however I decided that these seemed a little too all over the place and wasn’t keen on the overall design. So I scrapped them, and started from scratch! I started off by using the lasso tool to create basic shapes and ended up with these.
I then took these further and added gradients to them to give them more ‘depth’.
I’d taken inspiration from retro film posters and the atomic/nuclear designs of the 1950’s which fits in well with the style of my sentry gun and all the initial research I produced.




In addition to the poster and glowing emissives, I decided to change and create a sound effect that matched the style of my gun. Again, I had help from other classmates but the process was surprisingly simple. I must admit, I was worried about changing the blueprint and breaking everything at the last minute, but luckily I plugged everything into the correct places and the sound worked! Blueprints are awesome.

Conclusion:
Overall, I’m really pleased with how this project went but there are still certain things I’d like to change.
I loved the freedom I had over this project and I enjoyed the whole thing from start to finish! Getting to grips with the blueprints in UE4 was proving a challenge at the best of times, but they were really fun to try and figure out what part of the turret they were affecting, and how I could change certain components to make my gun act the way I wanted it to.

If I were to do this project again, I’d probably choose a more complex design or do many more iterations of the one I chose. This was because I found the texturing really simple – although it looks good, I have a gut feeling that I haven’t done enough.
However, I did complete my set stretch goal and even did some extra work on the side too which I’m really happy with. I feel like I’ve learnt so much on this project in comparison to the others – mainly engine based knowledge.
In hindsight, I’d also liked to have spent more time on the concepting process. I think I panicked about putting it into engine so much that I didn’t allow myself as much time as usual on the concepts so my design probably isn’t as good as it could be, but I’m still really happy with it!

0 comments:

Post a Comment