With Armageddon coming up on the weekend this week was all about finalising and polishing the build to demo at the showcase. Firstly, the glider level was revamped with some new fire models which have now been used underneath the glider fans in the level. This gives the rising smoke context and provides another visual cue for the player to recognise and aim for. The glider levels lighting was also baked which increased the overall visual appearance of the level.
Other than this most of the time was spent on building and playtesting the final build for the Armageddon showcase. One of the major bugs encountered was the construction of the level select bridges after exiting different levels. Because of the new asynchronous loading system, the saving and loading manager was not functioning correctly between scenes and this was causing the issues when loading into the level select. It took a large chunk of time to find and resolve this issue however, it has now been fixed and is working as intended. We also used Unity’s profiler to try and determine the cause of the major FPS drop in the glider level. Upon doing so we discovered that the trail renderers on the wind particles in the level were causing the game to drop between 120 and 140 frames. To increase the frames during the glider level, the wind particles were changed and as a result we now have frames of up to 200 during the gliding sequence. Apart from preparation for Armageddon, I also started building the new Aztec glider terrain which will act as the canyon for the player to glide through. This glider level will differ from the first one encountered in the wild west as it follows a much more “curvy/bendy” path as opposed to the straight linear path found in the previous level. A new system has also been implemented to allow the player to disconnect from the glider and continue platforming. This allows for more diversity when building levels as it gives us the opportunity to design levels which have multiple platforming and gliding sequences. Finally, the week ended with us displaying the game at Armageddon and receiving lots of feedback on the game. Overall the feedback was very positive with most people finding the game fun and a couple of people playing through the entirety of the game for periods of up to 50 minutes. With the feedback received form Armageddon we are now ready to fix some of the issues and continue developing the game over the next week as we prepare for PAX in Australia.
0 Comments
Much of the work I completed this week was spent developing the level I created and populating one of Zac’s jungle levels. I spent time lighting my scene to improve the feeling of being inside a dark temple. Some blueish light cascades down from openings on the roofs of the temple. I made sure to only place ferns and trees in areas touched by light. Darker rooms could use vases and statues more. One problem I had was that the large stone walls didn’t look good due to the repeating grid texture. I decided to add bricks to the walls to break up the grid and make the walls look more interesting. I also wanted to create the illusion that when the player looks down, it looks like the temple walls extend into an endless pit. To achieve this, I extended all temple walls downwards and used a black fog to fade it out. I began work on Zac’s scene by closing in some walls where necessary and adjusting the structure of the rooms to more closely resemble the rooms in the previous level. This meant adding the opening in the ceiling for the light to shine through. I also began adding some props to this scene but there is much more that needs to be done.
On Wednesday night, I headed out to PAX with other team members to show the game. We received some very positive and constructive feedback. This week is cut short since Zac and i are at Pax from the end of the week until early next week. I managed to sculpt, retopologise and texture a Jungle boulder asset for the Jungle glider section. I textured it using the proceedural material i made for the other rocks/stone assets. I also worked on a poster design using a render of Klepto that Juane made. In doing this I made a mock up of our Klepto logo which I ended up modeling in 3d and then again Juane rendered out for the final poster.
This week has been jungle work and preparation for PAX. Considering we leave tonight, I've written this prematurely to make sure that I actually do a post this week.
Jungle scenes 1 and 2 have been my primary focus this week, debugging and replacing puzzles that I didn't feel were appropriate or worst case boring. Having an entirely symmetrical level with puzzles on each side is a goal this time around, but it was starting to feel too formulaic and boring without some variance. I changed up the puzzles, as well as adding in a 'laser gate' that blocks progress for the crate. This was already planned, I just set it up and added a side passage that players must move down in order to activate it. The puzzles at the end of the level have been finished as well, with a rotated 'electric fence' on one side and a moving platform puzzle on the other. These puzzles are different enough to be interesting while also requiring a differing level of skill. I'm not hugely happy with them but they should suffice as an ending to the first jungle level. Next week we'll incorporate some of the feedback from PAX into the work that we're gonna be doing, and make sure that the jungle levels are being 'prop-ulated' in preparation for submission around end of week 7/ start of week 8. This week, most of the team flew off to Melbourne. For this reason, Klepto was effectively put on hold. We had Put as much time and effort into Klepto as we could up until Australia, and this way we would have a strong build to showcase, and after Australia, we could all focus on our final assignments without having to worry too much about Klepto. The work that went into the game in Melbourne was more networking than actual physical work.
With Armageddon this week and Pax the following week, we needed to make a push for polish this week. Again, I made a number of improvements regarding Klepto’s arm and grappling. I fixed an issue which occasionally prevented people from grappling with the Right Trigger on the Xbox controller. The indicator now scales and fades in when the grapple point comes in-view or in-range. Additionally, the arm has become more stable, thanks to getting some assistance from the people that developed the rope physics plugin we are using. When the player drops a crate, if they are near a pressure plate, the crate will automatically snap to the pressure plate.
I also edited a couple trailers. One was for Armageddon and one was for our accelerator application and is now on our Itch.io page. The reception for Armageddon was generally positive and served as a good stress test as a number of crashes and bugs revealed themselves as the crowds came to play. Throughout the week, I spent a lot of time working on the PS4 port of the game. After many issues with getting a build working, I was finally able to get one with a simple scene. There were issues with shaders that I had to fix and unfortunately, the physics arm won’t work on PS4. After the success with the simple scene, I attempted to port Level 3 to the PS4. The performance was terrible. I am working through a guide online to improve the performance but so far, the improvements have been fairly minimal. I think I will need to add more walls to box in each section of the level so that occlusion culling can work better in this level as it is currently quite open. Using the Unity profiler, I noticed that the massive frame spikes we were experiencing seemed to be caused by garbage collection. I researched how I could reduce the amount of garbage our code was generating. We were already caching much of our data so most of the garbage was generated by creating new strings, concatenating strings and using yield return 1 instead of yield return null in coroutines. After spending time in many of the game’s scripts I was able to minimize the frame spikes caused by garbage generation.
Another performance boost came from using Unity’s occlusion culling. I set up my level to support occlusion culling by using mountains and other assets to break up the level. The resulted in a not insignificant performance boost. A variety of fixes were made to the grappling this week. The grapple indicator was improved with new assets to make it easier to read. The indicator handles onscreen in-range, onscreen out-of-range, off-screen in-range, and off-screen out-of-range. The way that pushing and pulling when the player is too close or too far from the grapple point works has been changed. This improves the smoothness of the motion while preventing players from getting pushed through walls. Additionally, Klepto’s mechanical arm has been improved to be more stable. The tutorial signs got more character by giving the sign the ability to face Klepto when he is in range. I also attempted to make the text feel more like it is coming from the sign by making it the same colour and making it move in the same way. Finally, I also improved Dylan’s respawn loading screen by using a UI mask and portal image. This week was all about level design for me.
In preparation for moving into the jungle scenes full-tilt, I blocked out the entire second jungle level this week. The 'gimmick' for this level is rotating platforms, which come in two different flavours - rotating constantly and 'staggered' rotation, which stops every 90 degrees. The two different platforms are intended to introduce a bit of a timing mechanic, rather than the sometimes careless jumping that we encourage. This level is unique because it represents the first time that i have designed an entirely enclosed level from scratch. All of the previous levels that I have designed previously were designed in an open space, laying pathways and creating open areas within the scene. I like this form of limitation; it keeps the level focused and within scope, and is a challenge to create puzzles and activities for. I have created the entire level mechanically, save for the portal at the end of the level. Either next week or the week after I will decorate it, and it should be fully integrated as part of the jungle world by the start of week 7. I'm pretty happy with the similarities between the first and second levels - both of them use pressure plates as the main facilitator of progress, and the second level also includes some new quirks. I have been experimenting with using 'jump-cut' camera scenes to show players things like transitions, opening/closing doors, etc. I think it makes the level a little more explanatory, and allows the player to quickly understand if they've opened a door. In the future i'll probably use this kind of system in other levels as we move forward. This was a big week when it came to rehashing the character as I started to attempt on re animating the jump and redoing some death animations that we needed to add for Dylan’s level. I am proud of the death animations that I had done but I was unable to implement both animations, as the player animator was much unorganised. This was not a good sign for what was to come, as I would and try to attempt to edit the games jump animation. I was able to edit the old animations without too much of hassle and was able to ask for help from the programmers about the animator. While changing this animation, we were also able to fix animation bugs. Whilst updating these animations, we decided to animate the landing animations as well. This took quite a while to set this up but I think everyone is happy with the result. After this I started on modelling the Viewmaster which would be the level select.
This week I worked on props and environmental assets for the new Jungle levels. The vases were designed to look hand painted. I achieved a good painterly look in substance using brushes and normals under the brush strokes to add weight like paint. The throne and statue were also done a little differently from the usual Wild west stuff ive done. These were both sculpted in Zbrush and then re topologised and textured using a proceedural stone/rock jungle material i made in substance designer. The throne base was also sculpted and then used to bake normals to achieve the tiled stone effect.
|
AuthorWrite something about yourself. No need to be fancy, just an overview. Archives
November 2017
Categories
All
|