HL2:RA and Project 17 Status Update
Welcome! It's been some time since our last dev blog, but we've been very busy and have a lot of exciting news to share! First and foremost is our new collaboration with the Anti-Citizen team!
In the time since our last dev blog, we reached out to their team about working together in order to make sure what we create is useful to the entire community. To that end, we've begun working with them on our NPC framework, and they've already helped us make tremendous progress! This is our longest devblog yet, with a lot of news regarding HL2:RA Part 5, and a Project 17 status update for you as well, so let's hop right into it!
The Faces of Progress
We've been working on our characters since nearly day one of development, and we're finally comfortable revealing our work to the world. These are assets that our team has been working on in some form or another since nearly day one, and they've proven to be by far the most difficult hurdle to overcome thus far. While we expected these characters to be labor-intensive to get right, we never could have estimated the difficulty that comes with updating an aging character asset while preserving their likeness.
Dr. Wallace Breen was the first character sculpt we began, with the first piece of progress being shared among the team in August of 2020. Since then, we've been iterating on his likeness regularly, until we've reached the point we're at today. Dr. Breen is one of the most iconic characters of the Half-Life 2, being the primary antagonist, so we feel he deserves special care and attention when it comes to preserving his likeness. Even with that said, Dr. Breen is still far from finished, still needing major work on his hair cards and other areas.
Playing an equally important role as Dr. Breen is Dr. Kleiner, another character that we began very early on in development. While earlier in the pipeline than Dr. Breen, even missing his hair cards entirely at the moment, we still want to reveal our take on this character due to his significance in the Source community. Kleiner's assets have been in development nearly as long as Breen's, and we still have a long way to go.
With Dr. Breen and Dr. Kleiner's development well underway, we've still only scratched the surface of the work to be done. A long road lies ahead for these characters and others, and if the community is going to get their hands on these in any reasonable timeframe, we need your help. If you have any experience with creating character sculpts, facial flexes, or character grooming, we urge you to apply via our form here or contact us directly via Email. With all that said, we're still very excited to reveal our characters for the first time, and we of course have plenty more we've been working on in parallel!
Metrocops, Beatdowns, and More
Of course, with characters, the way they behave within the environment plays a major role in how 'grounded' they feel in the game's world. This holds doubly true in VR, and so one of our main goals for our npc's animations is to make their choreo sequences as indistinguishable as possible from their AI-driven behaviors. This entails a focus on not only smoothly transitioning in and out of choreo, but also making every action blend seamlessly without interrupting existing movement. In practice this is achieved through a number of techniques. Firstly, rather than rely on different animgraph paths for each behaviour, every gameplay system is treated additively, layered on top of a base idle. With this approach, metrocops and other characters are able to perform almost all of their player and environment interactions without interrupting whichever idle animation they may be playing. This system allows seamless transitions in and out of scripted scenes or animations, while still allowing the AI-driven behavior to play. For example, a metrocop shoving the player while coming out of a citizen interaction, or a citizen speaking to the player while grabbing a ration.
To compliment the enhanced choreo-to-AI system, we've put a lot of time and effort into designing a workflow that allows for fast iteration on our animations. Characters walking around driven by AI with layered gestures didn't meet the standard of quality that we're aiming for. This, coupled with the lack of any proper faceposer or choreo tools in the Half-Life Alyx Workshop Tools, meant we needed to create our own way to create high quality choreo sequences in a timely manner. In our case, motion capture has proven to be the clear choice.
While raw motion-capture data isn't viable for a shipped asset, it at the very least provides excellent reference for our artists, and when properly cleaned up can provide a very solid base to work off of. This has allowed us to quickly iterate on our animations, and opened the door to exploring how we can expand our characters' interactions with their environment beyond their original sequences. For example, exploring different way metrocops and citizens might interact with each other.
Working with motion capture is not a simple process. Creating these animations still requires significant work from our artists and is a time-consuming process. By the end of the pipeline, these animations are mostly hand-crafted with the motion capture data simply providing the backbone. While some motion captured performances may be useable out of the box, they are certainly the exception. While we're using the flexibility provided by motion capture to explore our possibilities, we're not neglecting the rest of what makes a character feel 'real'.
For any character, audio is just as important as any animations or sculpts that help to sell the illusion. In the case of the metropolice, there's simply no match to Valve's original recordings. For now, we're focusing on recreating the functionality of Half-Life 2's idle systems, while adding more variation. Rather than use Valve's existing (but simplistic) idle dialogue system, which involves nothing more than simply playing random audio files, the logic for choosing voice lines for the metropolice runs through animgraph. This allows us to incorporate the full range of randomness that the Half-Life 2 dialogue includes, while also giving us the freedom to blend in gestures for specific voice lines, or even other contextual behaviors.
In addition to all of the work being done for individual characters, our AI framework itself has received an overhaul, in order to work more seamlessly with existing Half-Life: Alyx NPC functions.
Chiefly, our AI framework has been upgraded to use a 'faction' system rather than relying on programmers to set a character's relationships. The previous implementation meant that it wasn't very easy or even possible for a Level Designer to change a character's relationship. Now, our AI is capable of being assigned a 'faction', which will update their relationships accordingly. These 'factions' are based on the original Half-Life 2 relationship files appended with Half-Life: Alyx relationships, meaning that anything with a 'rebel' faction you can expect to have the appropriate relationship toward a character with, say, a 'combine' faction. While this system is still a work-in-progress, it already makes working with the custom characters much easier and more flexible than having to track and update each of your character's relationships yourself. This also means that we've been able to make very quick progress on our different custom NPCs now that we have a foundation set for their behaviors.
With this foundation in place, we've found ourselves making rapid progress on all of our different custom NPCs, including our turrets, cameras, and even some other entities. Using this new 'faction system', our NPCs can be quickly iterated upon and respond to the other characters in their environment from the earliest stages of their development. As such, significant development time has been redirected from laying the logical foundations to putting on the several layers of polish that our NPCs need. New animations, particles, sounds and behaviors have been added to all of our NPCs we've previously shown, with some of them already beginning to function beyond their Half-Life 2 counterparts.
Of course, we've still got much more in store for you beyond just NPCs!
HL2:RA Part 5 Development Update
We're excited to finally have a solid foundation to build new behaviors on, and we continue to explore the possibilities for new NPCs going forward. But of course, we still have plenty to discuss for the immediate future with Half-Life 2: Remade Assets Part 5!
Our biggest motivation for going this direction was seeing value in giving more variety to the Half-Life: Alyx modding community, and enabling the creation of environments that lie far outside City 17. For Part 5, we deliberately chose to address an asset 'blind spot' in Half-Life: Alyx. Given Half-Life Alyx's setting, there aren't many assets for content creators to choose from if they wished to create non-urban environments. As such, we hope you'll find that Part 5 helps to fill that gap!
Of course, you can't mention the coast without antlions, and with antlions comes the thumpers. We're happy to say the combine thumpers are planned to be included in Part 5 as some of the 'hero' assets! You can expect to see not only this thumper in Part 5, but it's variant asset as well.
For our thumpers, we're happy to say that not only is full Half-Life 2-like functionality planned, but we've already developed a working proof of concept! Our current system closely mimics the functionality found in the 'Sandtraps' chapter of Half-Life 2, and is still very early in development, being only a proof of concept. However, despite that, the functionality we've engineered is already fairly complex, with a visible area-of-effect visible in our test map that correctly affects only antlions, scaring them off appropriately.
We're happy to say that not only is full Half-Life 2-like functionality planned, but we've already developed a working proof of concept! Our current system closely mimics the functionality found in the 'Sandtraps' chapter of Half-Life 2, and is still very early in development, being only a proof of concept. However, despite that, the functionality we've engineered is already fairly complex, with a visible area-of-effect visible in our test map that correctly affects only antlions, scaring them off appropriately.
And all of this isn't without purpose! This functionality is being created for a reason, meaning you can expect the thumper to play a major role in Part 5's showcase map! While we're not quite ready to reveal much about the showcase map yet, you can absolutely expect to hear more about it in a future devblog. We're very excited to develop this system further and share more with the community once it's ready!
Part 5 will, of course, also include several different environment materials for developers to use. To be more precise, we're creating a significant number of blends between different grass, sand, rocks, mud, and gravel materials!
A significant amount of work is going into making these as easy as possible for developers to use, while remaining flexible. As you can see, the materials we're tackling here can be used in several of Half-Life 2's environments. Quite a few of those are mostly used in the coast section, but things like grass variation and blends have been a long-time request for those of us working with the vanilla Half-Life: Alyx materials. This is something we identified ourselves when working on Project 17, and decided pick it up for community use.
Part 5 will be shipping with a number of 3D assets as well, again themed around the coast chapters. But of course we've also taken the extra step to add completely new assets to the list for content creators to work with! That said, while the internal list is pretty substantial, it's still subject to change. So this part of the showcase will only be touching upon a few of them. There’s still a lot for us to do on that front, and you can expect to see even more in future blog posts!
While this is just a taste of what we're planning to include in Part 5, we've gone out of our way to create each of these to be as flexible as possible. As such, there's a number of nuances to each one. Some will include separate skins like the hula doll, or even a prefab, like the scales and wagon. Keeping in line with our previous work, all of these assets will be made interactive as well!
Certain assets are even receiving a total overhaul in order to support interactions, or even just to fill in visual gaps that might have been overlooked for the sake of performance or technical limitations in Half-Life 2.
And finally, as an aside, we've taken the time to fix a simple, but annoying coincidence in the Half-Life: Alyx files. That being Valve's watermelon not being gibbable. We've entirely redone the original watermelon, complete with subsurface scattering on it's gibs. This may seem like an almost unnecessary addition, but on top of it being a fun asset to make, we felt like the addition of a gibbable watermelon was valuable, due to the confusion caused when the original watermelon couldn't be splattered.
If you're a 3D artist or a material artist and are interested in our work, we're currently accepting applications, and you can check out our closing section for more info!
A Look Into the Future
Looking beyond Part 5, we’ve decided to make this section a bit special and show you a little bit of behind-the-scenes on what we've been experimenting with internally.
To be entirely clear, we do not currently have release plans for all of the assets in this section. These assets are extremely early in development and are the result of us experimenting with the limits of Source 2.
With that said, our team is looking to be more transparent with the community and show more of what we're working on. Of course, we'd also like to take the opportunity to show you more of the amazing work put together by our developers.
Our version of the SMG1, while necessary for Project 17 as an NPC weapon, has been created with the concept that it should be supported as a player weapon as well. As such, this SMG1 design sports a few extra features intended to give some options for developers that might want to make the weapon functional.
Arguably the most important change is the addition of a loading port for the SMG grenades you can use in Half-Life 2. Given that weapon handling in VR is so fundamentally different than in flatscreen, it's necessary to have a physical way for the player to load their ammunition, including their grenades. In our case, that functionality takes the form of a small chamber where a player would load your grenades much like you load the SMG in Half-Life: Alyx. This feature was added in this way such that it would fill our functionality needs while retaining the important design elements of the original.
On the topic of remaining faithful to the original while filling our own goals, reaching this design was certainly a team effort. As some of you may already know, Half-Life 2's SMG1 design was based off of a very specific prototype of a weapon popularly known as the 'MP7'. But even knowing that, the original takes several artistic liberties and ends up with it's own unique design. Given the niche nature of the original reference, our research actually turned into a multi-team effort, with even the Shooter: Revived team pitching in with their own reference. At the end of it, our reference board ended up packed with images taken directly from retired firearms dictionaries and press releases. All of that is to say, we deeply wanted to respect the original's design, while also freshening it up with new VR-specific features, all without becoming 'just another MP7 model'. The team is very happy with how this one has turned out, and we're looking forward to sharing more with you once we have our metrocops all geared up!
Something we've always looked to for inspiration was how Valve revamped several Half-Life 2 designs for Half-Life: Alyx, while implementing new mechanics to complement the updated designs. So of course one of our developers just had to give the process a shot of their own! With the Half-Life: Alyx health station being one of the most obvious examples, it was decided to give the health kit a similar treatment.
Once we got started on this, even though it was just for fun, we just couldn’t help ourselves but to keep 'chasing the dragon' and see how far we could take it. Though this asset is only tangentially related to Project 17 and HL2:RA, we ended up taking this quite a way through the pipeline!
Now of course, it could have ended at just being pretty. But in typical fashion, we couldn't help ourselves, and made several exploratory animations for how this might work in VR with our own mechanics.
Now, these are just exploratory animations that we created only as a proof of concept, and there’s still significant coding challenges to overcome if we want this to actually function with a character's health pool. Since we have a lot of work ahead of us directly relating to Project 17 and HL2:RA, this has fallen to the wayside as we complete our main goal, and thus has no current release plans. However, should the community show an interest in even the non-functional model, we'll certainly entertain the idea. So let us know what you think!
Even though some of the work in this section might not have current release plans, we still wanted to share this work with you in an effort to be transparent. We hope you enjoyed this behind-the-scenes look, and if this is something you'd like to see more of in our devblogs, let us know!
Project 17 Development Updates
Project 17 has made tremendous progress in the time since our last update as well. Much of that progress has been made on the technical and gameplay side however, meaning there's not a lot to show as far as visuals go. We're more than happy to speak about all that's been done, however!
Project 17's environments have had a total overhaul since our last devblog, after our first rounds of playtesting exposed several issues with our previous designs for those of you that have been following us since the beginning, you would know that we have once before reached the 'art-passing' phase of recreating City 17. We knew going into this project that VR gameplay would demand a tighter, more 'focused' playspace than a traditional flatscreen game would, but initial playtesting made very clear that we had underestimated just how much rescaling and re-designing would need to be done.
As such, we've spent our time going through all of our environments, refactoring them with the lessons learned from our playtesting. At this point, we've gone through all of our environments and reached the art-passing phase once more, and secured countless improvements along the way. From visuals and comfort to performance, we're confident that we've made major strides since the first iteration, and we're looking forward to sharing more with you once we make progress on the art pass. That doesn't at all mean we have nothing to show, though! But before we get to that...
We pride ourselves as a community-built and grown project, and we are once again looking to the community for help. As we expand our workload into character animations and choreography, we are looking to expand our network to help us reach a proper release.
Specifically, we're looking to add the following, among others:
We are also open for other positions depending on the need and movement of the project.
Interested? We'd love to hear from you! If you're interested in working with us, please send an application our way via our Google form or reach out to us directly at firstname.lastname@example.org. Or if you'd just like to follow our projects' progress, feel free to join us on our Discord.
We look forward to welcoming you into our amazing team, and until next time, we'll leave you with some of our most recent work.
We hope to hear from you, and until next time, we'll leave you with some of our most recent work.
Copyright © Project 17 Team 2021
Project 17 and/or Project 17 Team is not affiliated with Valve and/or Valve Software® in any way, shape, or form.
Steam, Half-Life, Half-Life Logo and the Steam logo are trademarks and/or registered trademarks of Valve Corporation © in the U.S. and/or other countries.