I started with the company in 2009 near the beginning of production for NS2.  I was hired as an intern to help with scheduling and tracking the Art tasks for production.  After a short while I started to see some other useful things I could do for the team.  One of the tasks that came up was to create a more interesting cinematic for the next alien to be revealed… the Lerk.  I volunteered to create the cinematic using our Spark Cinematic Editor and got started.

————————————————————————-
The Lerk Reveal

http://www.youtube.com/watch?v=0qTePDz4oTg

At this time the Spark Cinematic Editor was still very new.  It was first created with everything needed to make the original NS2 Teaser video with the Onos in the hallway.  I did not make that cinematic so this was my first real experience with the tool.  I jumped right into it and surprisingly was able to get a very rough version of the cinematic ready in a day or two.

http://www.youtube.com/watch?v=eOMv51NdA3o

Quickly I realized that it was not ideal to create this all in one scene and soon broke it up into 4 parts.  That is my first pro-tip for those interested using the tool… Pretty much every camera cut should be a new scene.  Why is this such a problem in the SCE? Well mostly it’s because you can not adjust tags using the time slider.  Instead you need to manually adjust each key in the Graph Editor of the tool but moving multiple keys often causes the tool to crash.  So by separating the files you won’t have to worry about moving everything in scene 4 if you just adjusted the timing in the scene 2.

Second pro-tip is to SAVE OFTEN as even today the tool is crash happy. The auto-save feature was not available for at least a year after starting this project so I learned this the hard way. heh.

Other than that it was pretty painless.  Once Cory got involved and people started to say “wouldn’t it be cool if…” things got shifted around and the polish phase was extended a few weeks but it was worth it. I’d like to add that you should make sure you give your sound guy some time, as sounds and music can make or break a cinematic. (pro-tip #3?) I believe Simon did all of the audio and re-purposed some of the music from our soundtrack.

Note that this whole video was created fully in SCE as advertised, and was using all stock animations.  The only movie magic that was added was in the final scene where the first person view was a composite of two renders to show different levels of motion blur between the view model and the background.

Everyone enjoyed the results so we began to plan more short cinematics.

——————————————————-
____, Teaser:  

http://www.youtube.com/watch?v=7bpcA0S0pBc

The gorge belly slide came from a suggestion on the forums. We thought it was an awesome idea and added the feature to NS2. To announce this feature we thought it would be cool to make a cinematic and so it began.

Initially it was going to just be a short 30 second video showing a little build up and then the gorge would slide down and bump into the barrels.  We were having a lot of issues with the rendering process which pushed for further development of the cinematic tool and the engine.  You can still see z-fighting issues in the background and some animation popping, but that is what we had to work with at the time. Also with limited options in the curve editor, it was difficult to make any complex camera or character movements.

The most complex move was the transition of the gorge from the angled stairs to the flat ground.  To fix this I ended up making two gorges, one going down the stairs and one going across the floor, and then swapped visibility on the two.  I coupled this with a strategically timed cut to cover up the problem.

Of course I ended up making this whole video in one scene despite my previous lesson on the topic…  By the time sound was added we were getting much farther with the Fade so we decided to add the mysterious shadow tease clip at the end.  Again this was done all in the SCE and all with stock animations.

————————————————————————-
Fade Reveal:

http://www.youtube.com/watch?v=U98Tbr5P2bs

As our final alien lifeform to reveal and after seeing what can be done with the SCE, we were already talking about what kind of cool things we could do for the Fade.  Cory really wanted to kick it up a notch and I was feeling confident that I could do pretty much anything he wanted, but there were going to need to be some changes to the process first.

So, with a longer video packed with action, we were going to need to do some more in depth planning.  Cory drafted up a story board which I put together into this animatic:

http://www.youtube.com/watch?v=DEMC_na0LvE

After that was done I created this test cinematic based on the animatic. I have to share this because it cracks me up:

http://www.youtube.com/watch?v=DpVzm3kREkY

As you can see, changes where made between what we started with and what we ended with, but the majority of the shots remained intact.  For shot #2 in particular I thought that showing the fade at the beginning was too early of a giveaway and takes away from his more dramatic appearance later in the video.  On the other hand, Cory’s original idea was fleshed out a bit and looked like this:

http://www.youtube.com/watch?v=5d1Hj1gIfHo

Eventually we settled on cutting out the first appearance and adding a lot more action to the fight scene.

You might have noticed that at least half of the animations in this video are completely custom.  This is true.  We worked closely with an animator throughout the production of this video. Using an exported .obj from the level editor as reference, he created entire scenes in 3DS Max and I exported everything for the cinematic.  It was a great effort to make sure every character was set up correctly with the right animations loaded but it really improved the animation quality of the reveal. Cory even made custom marine skins with the rookie that gets killed first and the vet who gets knocked into the glass.

In the end we had 29 separate scene files to work with.  This did make it much easier to develop in some cases, but added a lot more setup time on my end.  Again, camera limitations made this quite difficult but I was beginning to develop a new technique of animating the camera in Max and importing it as a view model. For the Fade Reveal I only used this technique as a zooming pose parameter for the slow-mo spin around shot, but in future videos it is used much more.

Another feature that was added during the development of this video was the ability to load a cinematic within a scene. With this I could add effects like bullet impacts, tracers, and the Fade blink just by loading one file instead of copying the entire effect in manually.  This was exceptionally useful since we were still tweaking the blink effect for the Fade and I was able to see the tweaked effects in my scene without needing to make any changes.

————————————————————————-
Onos Vs Exo Launch Trailer:

http://www.youtube.com/watch?v=re9sqgeWFDM

With launch looming and the Exosuit being developed, we wanted to make one more big cinematic. Cory busted out the story board rather quickly and I put a 2d animatic and 3d animatic together:

http://www.youtube.com/watch?v=wWC-piCFyAY

http://www.youtube.com/watch?v=QZe2rS0qkqE

http://www.youtube.com/watch?v=NVWV3HAJpeQ

For this cinematic we decided to do everything in 3DS Max and then import it in.  In most cases I exported a group of characters together as one big animation file. For example, all of the skulks in the last shot were exported together as one model and animation set. Most of the cameras where exported directly from 3DS Max so that they could be animated there instead of using the limited controls of the SCE.  Having the cameras in Max also helped the animator see exactly what was going to be in the shot so he didn’t waste time on unseen animations.

http://www.youtube.com/watch?v=zG6y69w9kdo

One new thing that I pushed to have added was the ability to attach a cinematic to a model.  Attach points are used in the game already to indicate where a muzzle flash is, so why not use them in the cinematic editor?  Trying to animate the position of an effect manually is a real challenge for shots like the Exosuit shooting everywhere.  Now all you have to do is add the cinematic event to the model, assign the attach point, adjust the base time, and then the effect will follow the motion of the animated object. I should mention that our VFX guy, Sylvain, was a large part of why this video looked awesome.  He spent a good bit of effort creating custom effects for each scene which was something we had not done before.

————————————————————————-
Since then, Cory has picked up using the SCE for his own shorts (like the cinematics in the Gorgeous update) but we’ve mostly switched our focus to using more Gameplay footage in promotional videos. In summary I feel like the tool has a lot of potential for creating fun videos and I plan to put together better tutorials for those that are interested in doing so.  I’m not sure what anyone will get out of this post if anything, but at least I think it was a good excuse to dig through old files and show off some cool things I found.

-Brian

The post Spark Cinematics of Natural Selection 2 appeared first on UnknownWorlds.

Brian espousing wisdom during the ModJam debrief

First up, what’s happening with the mods! You will be able to play them very soon. At the start of the modjam, we decided to use an internal build of NS2 that is not yet public. That build is currently being playtested and prepared for release. It is scheduled for release on Thursday, though such is the process of releasing builds that we cannot be sure. Once it is released we will put all three mods up on the Steam workshop, and you will be able to play them to your hearts content!

http://www.youtube.com/watch?v=1gT_CQ41qOA

Of course, they are not quite finished. So far from finished in fact that the real fun that will come from their release will be what you choose to do with them. Along with the Steam Workshop release, we are going to release the mod source code. Hopefully, some clever community modders will take on the challenge of turning a mod like Last Stand into a game that really shines.

During ‘meetings,’ this dooverwhacky connects those in the UWE office to those around the world

During the jam, many of us in the San Francisco office spent lots of time working closely with those offsite. This was a wonderful experience, as distance can sometimes be a barrier to collaboration. By using tools like Teamspeak and even Twitch.tv, offsite programmers, artists, mappers, and designers could fully participate in the jam.

http://www.youtube.com/watch?v=B-8I8ZqxTTE

Small game developers like UWE typically encourage people to be ‘T-shaped,’ which is a fun way of saying: ‘Be very good at one thing, but understand and be effective at lots of things.’ This philosophy is amplified during a jam, when people have the opportunity (and are encouraged by their peers) to dabble in areas of development they do not normally participate in. For example, programmers creating levels, designers creating game code, or artists designing games.

Charlie makes a point during the ModJam debrief

A verbose and corporate way of describing this situation would be ‘multidisciplinary cross-pollination.’ When people better understand what their team-mates are doing, they can work together better. Programmers that understand the workflow of artists can build better tools, and artists that understand the limits of code can create better art. The process also breaks down the traditional programmer / artist / designer / other divides, that can appear outmoded and inefficient at a small, nimble studio like UWE.

http://www.youtube.com/watch?v=_STWk7__bhk

Spending a week using our own mod-tools has alerted us to some of their weaknesses. They work well, and work fast. But there are areas where they can cause frustration to modders. Now that we are aware of the problems in the mod-pipeline, we can make changes to the Spark Source Development Kit (SDK) to make creativity easier. Everyone had a blast during the ModJam and we will no doubt have more of them, each time learning new lessons and become better at making games.

The post ModJam – After Action Report appeared first on UnknownWorlds.

Steve and Charlie cranking on their mods

It is never possible to finish everything. The mod can never be everything you want it to be. The temptation to spend ‘just a bit more time’ on your creation is absolute. Many of us struggled with the solidity of the brick wall that we faced at the end of the day: We had come so far, why did Last Stand, La Fauche, and Fighter have to stop now?

Dushan and Charlie discuss getting custom sounds into their mods

To struggle thus is to miss the point of the ModJam. As the hours ticked by, we could only see the flaws in our creations. The missing models, animations, features. The bugs and suboptimal game mechanics. What we should have seen, and what appeared on the livestream at 1700, was a wonderful expression of teamwork, creativity and technical ability.

Brian led UWE through the ModJam, as well as the La Fauche team through a mod

Offsite, onsite. Employees, non-employees. Throughout the past five days, people have mixed in ways they have never mixed before. Everyone has learnt new skills. The contribution of some community members was above and beyond the call. The line between being on the ‘team’ and being in the ‘community’ was further blurred, and for the better.

Steve and Dushan discuss spawn mechanics for Last Stand

In our haste to begin creating games on Monday, we did not quite think through to Friday. Each of the mods was built off an internal codebase, not yet present in the Steam build of Natural Selection 2. With everyone on the team exhausted, the public release of this code will have to wait until next Monday. At that time the source code and playable versions of all three mods will be released.

Prepping to livestream the mods at 1700

In the mean time, you can catch gameplay by watching the replay of the afternoon livestream. Along with source code, screenshots, anecdotes, and full credits for each mod will be released here on the UWE blog next week. For now: Sleep.

The post ModJam Day 5: H-Hour appeared first on UnknownWorlds.

Cory lands a high kick in Fighter team’s prototype build

The fighter team managed to overcome most of the technical difficulties that plagued them on Wednesday. By mid morning, players could connect to matches and being punching, kicking and dodging. Cory is setting very high standard for the visuals in the game, and had painted lots of original artwork.

Charlie testing out some new gameplay code in Fighter

Charlie is drawing on his vast, hitherto untapped knowledge of fighting games to try and create an authentic experience. Having Andi (otherwise known as the Machine, the Beast, the Code Wizard) on the team has allowed rapid development of the fighting code, and together they are making strides towards having the mod ready for play on Friday.

Steve working on equipment selection code

The survival team, creating ‘Last Stand,’ ran up against some limitations of NS2 gamecode in the morning. In NS2, players rarely have to pick out specific weapons and equipment from large piles. In Last Stand, such precise manipulation is common place, especially at the start of a round. Steve therefore spent quite some time reworking the equipment pickup code to be much more precise than in NS2.

The office ‘Board Room Table’ became a pad for drawing Last Stand equations

Every time an alien dies in Last Stand, the player is offered a choice of three new aliens to instantly spawn as. Over time, the power of those aliens increases. Coming up with game logic to calculate which aliens should be offered occupied the minds of Dushan and Hugh for quite some time. Eventually, a solution was found and alien spawns are now working well. Check out this Last Stand playtest posted on NS2HD:

http://www.youtube.com/watch?v=BM5cpAhZmzk

The volunteer playtest group has graciously donated lots of time and feedback to Last Stand mode. With the addition of some final UI elements and balance tweaks on Friday, it will be ready for public consumption!

Brian working on a player model for the Heist team

The Heist team has successfully implemented artificial intelligence and many gameplay mechanics. For example, laser alarms, button alarms, cameras, and guards are now all working. The challenge for the team on Thursday has been putting all these elements into gameplay scenarios.

Max and Brian discussing scenario creation for Heist

Brian has found that creating a scenario that is not too hard, but not too easy is a tricky task. Each scenario must require some combination of actions by multiple players. Make the scenario too hard, and players may lose interest. Make it too say, and the game will not elicit that feeling of camaraderie and achievement that the team is striving for.

Brian examining camera monitors in Heist

With only one day left, all teams have much to do. The ModJam has been a great success so far, with everyone learning new things every day. Artists coding, coders drawing, designers animating, the cross-pollination of disciplines is helping everyone understand game development from new angles. No matter if any of the games actually work tomorrow, we can be sure that throwing this ModJam will help UWE make its products better.

The post ModJam Day 4: The Business End appeared first on UnknownWorlds.