That50'sGuy reacted to Radix for an article, Static Prop Combine in CS:GO
Static Prop Combine in Counter-Strike: Global Offensive
A step by step guide
thanks to @untor
What is Static Prop Combine?
Static prop combine, or informally speaking "autocombine", is a new feature in CS:GO's VBSP.
It allows VBSP to merge together multiple static props into a single static prop, either automatically or with user-defined rules.
What is static prop combine good for?
Static prop combine is another feature to optimize your maps. Most people might think that "the less geometry rendered the better". So if you use small props, it's easier to hide what is not visible.
That's not wrong. But there is a problem:
In Source, there is one draw call per model per material. And these draw calls are very performance-hungry.
That's where static prop combine comes into play:
By combining models sharing the same materials, less draw calls are performed, which greatly helps optimization.
Valve has stated that Nuke runs 40% faster after they implemented static prop combine.
How do I use static prop combine?
The static prop combine feature was added in 2016 with the release of the reworked de_nuke. But since then it was not (?) used by community mappers, there are no (?) guides on the Internet except this documentation.
@untor helped me to make static prop combine feature do its job. So we decided that it's time to publish a step by step guide how to use static prop combine.
We presuppose that you are already familiar with the creation of props
0. Backup your CS:GO folder (optional)
We do not take responsibility for any damage done to your files. So it's time to backup your game files now if you have not already. In general we recommend to duplicate your "Counter-Strike Global Offensive" folder, so you can use a separate installation of CS:GO for mapping while keeping the other one clean for playing.
1. Source files
You must have the source files of the models you want to be combined. Usually 3 files for each prop:
*.qc reference mesh (supported formats are *.smd, *.dmx and *.fbx) physics mesh So if you want to combine props made by you, you should already have these files.
If you want to combine props made by Valve, you will need to decompile them first. And then change the names - otherwise, the version of the prop that is packed in the VPK would overwrite your version.
In this guide we will use two different pipe props:
You can download the example files here (contains the *.qc and *.smd files) :
Browse to "...\Steam\steamapps\common\content\csgo\"
Create a folder "models". In our example we have another subfolder "example". Save the model source files there:
These are our QCs:
Restrictions for the *.qc:
Only the first $body is recognized. $model is not recognized. $appendsource and $addconvexsrc are not recognized. You can only use $upaxis Z or Y.
2. Compile your props
Your models have to be compiled from this directory now:
Open your model compile tool (I use Crowbar)
Then browse to "...\Steam\steamapps\common\content\csgo\models\example\" and compile the QCs.
The compiled model files should be in "...\Steam\steamapps\common\Counter-Strike Global Offensive\csgo\models\props\example\" now.
Browse to "...\Steam\steamapps\common\Counter-Strike Global Offensive\csgo\scripts\hammer\spcombinerules\"
There you will find "spcombinerules.txt". In this file the combine rules for Valves props are defined. It is a standard KeyValues-formatted text file. Each entry follows the format below.
Rename it to "spcombinerules_valve.txt" (or whatever you want) and create a new text file "spcombinerules.txt".
Then copy and paste the following into "spcombinerules.txt" and save it.
4. Stub QCs
Stub QCs are QCs which contain a base template for the QCs which static prop combine generates. Generally, they should only include:
$staticprop $surfaceprop $cdmaterials Any $texturegroups used by the models. Browse to "...\Steam\steamapps\common\Counter-Strike Global Offensive\csgo\scripts\hammer\spcombinerules\qc_templates\".
In our example we create a new subfolder "example", open it and then create a text file and rename it to "pipe_combine.qc":
Copy and paste the following into "pipe_combine.qc" and save it:
5. Compile your map
Add some of our example props to your map in Hammer and compile the map.
In our example we use the following compile parameters for VBSP
Compile parameters (full list here) :
-StaticPropCombine: Merges static props together according to the rules defined in scripts/hammer/spcombinerules/spcombinerules.txt. This lowers the number of draw calls, increasing performance. It can also be used to lower the number of static props present in a map.
-StaticPropCombine_AutoCombine: Automatically generate static prop combine rules for props that VBSP deems should be combined. Note: This does not write to spcombinerules.txt.
-StaticPropCombine_ConsiderVis: Instead of using the distance limit, combine all props in the group that share visclusters.
-StaticPropCombine_SuggestRules: Lists models sharing the same material that should be added to spcombinerules.txt.
-StaticPropCombine_MinInstances <int>: Set the minimum number of props in a combine group required to create a combined prop. Tip:Valve had this set to 3 for the new Dust 2.
-StaticPropCombine_PrintCombineRules: Confirm: Prints the combine rules?
-StaticPropCombine_ColorInstances: Instances of combined props get colored.
-KeepSources: Don't delete the autogenerated QCs and unpacked model files after finishing.
-CombineIgnore_FastReflection: Combine props, even if they have differing Render in Fast Reflections settings.
-CombineIgnore_Normals: Combine props, even if they have differing Ignore Normals settings.
-CombineIgnore_NoShadow: Combine props, even if they have differing Disable Shadows settings.
-CombineIgnore_NoVertexLighting: Combine props, even if they have differing Disable Vertex lighting settings.
-CombineIgnore_NoFlashlight: Combine props, even if they have differing Disable flashlight settings.
-CombineIgnore_NoSelfShadowing: Combine props, even if they have differing Disable Self-Shadowing settings.
-CombineIgnore_DisableShadowDepth: Combine props, even if they have differing Disable ShadowDepth settings.
The combined props look exactly like the single props. So how can you be sure that the static prop combine process was successful?
- Once the map is compiled, the combined props will be packed into your *.bsp automatically.
- If you add -keepsources to the compile parameters, you can also find the combined props in "...\Steam\steamapps\common\Counter-Strike Global Offensive\csgo\models\props\autocombine\*name of your map*\"
and their QCs in "...\Steam\steamapps\common\content\csgo\models\props\autocombine\*name of your map*\".
- If you add -StaticPropCombine_ColorInstances to the compile parameters, instances of combined props are colored in CS:GO.
7. Additional notes
- You can manually disable static prop combine for individual props with the "Disable Prop Combine" keyvalue.
- Prop scaling (Uniform Scale Override) is not supported yet (?)
- If the original props don't have a collision model, you will have to set collisions to "Not Solid" in the properties. Otherwise the combined prop will be solid (automatically generated collision mesh; causes problems).
- If the props differ in specific keyvalues, in most cases the default (e.g. Alpha) or the higher value will be used (e.g. fade distances)
- Props that differ in the below keyvalues will NOT be combined, unless manually overridden with the appropriate VBSP option:
Render in Fast Reflections (-combineignore_fastreflection) Ignore Normals (-combineignore_normals) Disable Shadows (-combineignore_noshadows) Disable Vertex lighting (-combineignore_novertexlighting) Disable Flashlight (-combineignore_noflashlight) Disable Self-Shadowing (-combineignore_noselfshadowing) Disable ShadowDepth (-combineignore_disableshadowdepth) - Props that differ in the below keyvalues will NOT be combined:
Skin Color Disable Flashlight
some fps tests with an actual map! which gives better results: " -StaticPropCombine_ConsiderVis" or prop combining based on distances? Is there a console command to display the number of performed draw calls/props? ...
That50'sGuy reacted to will2k for an article, Displacement Vs. Func_detail - A comparative fps study
What is the question?
Ever since the dawn of humanity, this question was the center of a colossal debate. Greek and Roman philosophers tried to solve it to no avail. Alchemists in the Middle Ages gave it a go and failed miserably. Even Industrial Age scientists touched on the subject with no big breakthrough.
Luckily for everyone, I am here today to answer this question and put an end to a centuries-long argument: What is better in terms of fps, func_detail or displacement, in the context of the Source engine? If you were expecting an existential question, I am deeply sorry to disappoint you but hey, life is full of disappointment.
This is going to be a short but sweet article; fewer words, more numbers and screenshots. The study is pretty straightforward and systematic. To make things fair and square, I will create 2 exactly identical test maps: In one, everything will be turned to func_detail while the other will have everything switched to displacements. I will then proceed to record the localized fps in these maps from a preset location and compare. Pretty simple, isn’t it? Well, it should be as the whole purpose of this study is to compare func_detail vs. displacement in absolute terms while keeping all other parameters constant.
The first map to test is the one made of displacements. Here is the screenshot showcasing the fps.
The map itself is very simple consisting of 7 identical houses placed at predetermined locations and surrounded by 4 walls. The houses are detailed enough to put some slight pressure on the rendering engine. For the skeptics among you, here is a wireframe in-game shot to show that everything is made of displacements.
To refresh your memories, in Source engine wireframe mode, green is displacements, pink is brushes (world, func_detail, brush entity, etc…), blue is props, and yellow is decals/overlays. The recorded fps in this map is 289. We now move to the second map, the func_detail version to check how the frame rate is faring. Here is the awaited screenshot.
Surprise, surprise. The fps is 330, much higher than the displacement version. Here’s the wireframe shot to put your mind at ease.
Honestly, I was thinking the figures would be more on par as the engine handles both details and displacements pretty well, but in the end, Source is about BSP so I guess brushes would get a slightly preferential treatment over polygon meshes (conspiracy theory ensues).
The question that forces itself now is: Should we rely solely on func_detail in our maps? Of course not. Both func_detail and displacement have their advantages and inconveniences and leaning exclusively on one will inevitably lead you to a dead end. The best thing to do is get the best of both worlds by using them together.
In our little test map, how about we mix things up in a third version: let us make the house walls out of displacements while having the doors, windows, frames, and roofs made of func_detail. Incoming screenshot, brace yourselves.
Much better, isn’t it? We have now 311 fps, a very nice middle ground between the 330 fps of func_detail and the not-so-bad 289 fps of displacements. The mandatory wireframe shot follows.
So, what can we learn from all this? Well, apart from the obvious places where displacements are mandatory for the organic mesh sculpting (rock formations, cliffs, bumpy/twisted roads…), it is a good idea to spread some more displacements around your map to alleviate the total brush-count that you will inevitably hit the maximum in a highly detailed map. Your fps will remain high and you will enjoy the margin to keep adding structures to your map without fear of reaching the maximum allowed total brushes (substituting brushes with models/props is another viable solution that is not in the scope of this article).
I’m a man of science and I know that one example is not enough to draw conclusions. That’s fine, I have a second test map to investigate what we established before. The concept of having 2 identical maps is still the same, however, this time, we will spice things up by adding some static/physics props and some decals here and there. We will start with the displacement version.
230 fps, not too shabby. Let’s check another angle.
220 fps, more or less, on the same level as the previous number. Now for the wireframe shot.
The tree cards in the background are func_brush in both maps (the detail and displacement versions), so it’s a level playing field in this case.
Now for the moment of truth you all have been waiting for: will the detail version have better fps to support my earlier findings or will I be publicly embarrassing myself? A screenshot to the rescue.
I knew I was right, never breaking a sweat (apart from the nervous cold sweat I just wiped off my forehead). 255 fps for the first location A. Let’s check the other angle or location B.
250 fps. Bam, sweet victory…sorry I got carried away a bit. Ahem…Let’s get back to being scientific, shall we. Here’s the wireframe proof.
Let’s recap all the action and numbers in a nicely formatted table.
You can notice the fps gap between the func_detail and displacement versions in both test maps whereas the “mixed” version considerably narrowed this gap. The numbers have spoken.
The bottom line
The bottom line is, if you rely only on func_detail, you will hit the maximum brush-count allowed in Source and severely limit your map and creativity. You might also run into T-junction issues as well as parts of your geometry flickering and disappearing from certain angles in densely func_detail’ed areas.
On the other side, if you stick to displacements alone, then you will have lower fps than a func_detail map version. You might also run into visible seams and un-sewn displacement issues.
Having a clever distribution of both func_detail and displacement in your map is the way to go. You will have high fps, better lighting around the edges, and organic sculpting while not getting anywhere near the total brush limit; the best of both worlds.
That50'sGuy reacted to Radu for an article, 2018: Mapcore's Year in Review
Keeping with tradition, I'd say it's about time we took a look at what our community has achieved throughout the year. If last time I was saying how 2017 was a year of immense growth, then 2018 was surely one of significant change. And it hasn't been without its troubles and anxious moments. No change ever is, but I believe it to be for the best. We've seen some of our friends become parents, change work fields or get their first job in the industry. We've even seen a few pursue their dream projects. And for that, we have to applaud them. It takes courage to keep moving forward and to realise when it's time for something new. In the meantime, I hope this article inspires you and I wish everyone
2018: Mapcore's Year in Review
SteamVR - Gulping Goat Space Farm
by @Steve, @marnamai, @The Horse Strangler, @Sersch and others at Scraggy Rascal Studios
produced in collaboration with Valve
"Scraggy Rascal has been working with Valve to create all new SteamVR content, we've been given a lot of liberty to create these locations. Our goal was to create interesting and fun locations for the player to explore. These projects, over the last couple months, have been a crash course in Source 2,VR, project management, delivering within deadlines, working together as a team and personal growth. It has been an invaluable experience and great opportunity ... and we're just getting started!" - marnamai
Darksiders III - Art
by @The Horse Strangler and others at Gunfire Games
"Probably one of the biggest challenges the artists and designers faced on Darksiders 3 was working with both a platforming and fully connected streamed world. This meant that everything exists all the time. While we streamed levels in and out, areas couldn't intersect and we couldn't do the classic "Small exterior, big interior" swap. This was especially challenging because of how much verticality our design must support. We had a few "vistas", but for the most part every aspect of the level was accessible. If you can see it, you will likely be able to get there, jump on it, fight around it, etc. Fury, the main playable character can double jump, swing, float, glide and even rocket jump over 10 meters high. Personally for me it completely changed how I looked at art filling up a space. Every single mesh we placed impacted design. Art was design, and design was art." - The Horse Strangler
"Europa is a relaxing narrative experience. The goal with this game is to offer just enough challenge that its rewarding to get from one area to the other for more than just the visuals by using environmental hazards, platforming sequences and light puzzles that you can beat by exploring.The game is split into linear sections and wider areas, that's at the core of the game and as you play, you keep improving your characters moving ability, which will further exploration and give you the ability to solve newer light puzzles. There's none of the typical character upgrading systems, rather, the levels will offer the incremental challenges and the sense of progression. Europa's main focus lies in environmental storytelling and immersing the player in it's universe with passive storytelling, evoking awe and bliss with colorful watercolor-like art and music." - Helder Pinto
Counter-Strike: Global Offensive - Turnpike
"For a while the "Highway Restaurant" theme has been sitting in my little Concepts.txt file. When the Wingman Contest was announced, it felt like the perfect opportunity to turn this idea into a map, as its relatively small size would be fitting for the Wingman gamemode. The casual nature of Wingman made me add some elements that I would not normally add to, let's say, a Defusal map, like the TF2-esque team color coding (albeit subtle), the moving vehicles and the silly bomb target. Additionally, since the playable space is (almost) completely indoors, making it nighttime felt right, as it both emphasizes the interiors and makes for an atmospheric blorange background." - Squad
Dying Light - A New Hope
"A full-fledged custom single player campaign that ties in to the original story of the main game. It will see the main protagonist, Kyle Crane,leaving the City for the countryside to search for a specific elusive medicinal herb and bring it back to Dr. Camden who believes it could be the cure to the Harran Virus. This campaign is a one man show as I’m doing everything myself: level design, environment art/detailing, story creation, scripting/quest creation, custom dialog, custom audio, custom materials/textures, custom foliage systems, custom brushes for terrain painting/sculpting, lighting, manual nav mesh tuning, scripted NPCs…" - will2k
by @General Vivi and Michael Voeller
"Prodeus is the first person shooter of old, re-imagined using modern rendering techniques. Oh, and tons of blood, gore, and secrets. Creating Prodeus has meant a lot to us over the last year. It feels great to finally be doing something for ourselves. It can be pretty ambitious at times since there are just two of us, but I’m confident we can pull it off. Keep an eye out for the end of February for a big announcement." - General Vivi
Counter-Strike: Global Offensive - Ruby
"When I was on vacation in Portugal years ago I was so impressed by the city Lisbon that I really wanted to build a map that has the same vibe. At the time I was already working on different projects so I decided whenever I got enough time to work on a map this size I would go back. So early 2017 the moment was finally there, I went back to Lisbon to shoot (~2000) reference photos then made a list of things that are iconic for Lisbon and started working on Ruby. Adding a lot of height differation, warm colors, tile patterns and ofcourse trams was essentiental to get the Lisbon vibe." - catfood
by @dux, @PogoP and others at Unknown Worlds Entertainment
"A mix of Survival, story, mystery, resource gathering, base building with some accidental horror and plenty of deep, deep water. We had not long finished up with Natural Selection 2 and were hungry to develop a different kind of game. During development we were (and still are) a small team but the game kept getting bigger and grew into something far larger in scope than originally planned. So we soon realised that what we had could be turned into something really unique if we put our heads down and just cranked on it." - dux
Unreal Tournament 4 - Chamber
"I used Halo and Warframe artstyle as a reference. The goal of this project was to make fun and cool looking map with 100% custom art that is 100 mb in file size. To achieve that I used several advanced techniques such as custom vertex normals, deferred mesh decals, no bake, tiling base materials and masks. There are basically 5 or so texture maps used in the entire map, most of the filesize space was taken by lightmaps. I learned a lot doing this project in terms of composition, art direction and optimization. Hope you enjoy this map as much as I do!" - Ubuska
Counter-Strike: Global Offensive - Pitstop
by @Quotingmc and Quadratic
"It is not often that CS: GO receives a new game-mode, especially one as competitively focused as Wingman. I was understandably pleased at the announcement of the 2018 CSMapMakers contest for the mode. Pitstop was my entry where I set out to create a thematically bold centre piece for my portfolio. With the help of my teammate Quadratic and support from multiple Mapcore members, I learnt a lot about taking a level from a simple blockout to completion; I can say for certain I’m thrilled with the end result!" - Quoting
Black Mesa - Xen
by @JeanPaul, Adam Engels and others at Crowbar Collective
"While building Xen we had to design, iterate, and iterate (then iterate some more). We took what we thought we knew, and put it to the test. We learned how design and scope work together, and how to build momentum as a team. We are extremely proud of what we have accomplished over the year(s)! Despite the long and occasionally frustrating timeline, it has been a real testament to the commitment that this team and this community have for Half-Life." - Adam Engels
Unreal Engine 4 scene
"So I decided I would step out of my comfort zone and create a small environment in an engine I've never used before, UE4. Although I think I did a fairly decent job at the time there were ultimately many nuances I could have done better, but that is the artist dilemma. This project taught me the value of properly blocking out your environment, gathering as many references as you can and to have patience and not rush through assets, when breaking any of these rules I was punished for it. Stay tuned for my next project which will be a giant mech, coming soon Valve time TM." - Vorontsov
Counter-Strike: Global Offensive - Opal
"My goal with this project was to make a fun and compact defuse map, with a simple level flow, ample verticality, and an overlapped layout! I wanted to have interior and exterior, and break the grid a lot, to avoid having that "90 degrees grid" feel in the layout. I needed to have a vista on one side of the map to help with orientation, so I decided to make it a coastal town, inspired by those found on the island of Skopelos, Greece. Expect more updates in the near future, as I'm not yet satisfied with it. Since this is my only CSGO map, I want to put all my time and effort into it, and focus on quality instead of quantity. Thank you everybody for your support and feedback! <3" - MikeGon
Insurgency: Sandstorm - Precinct
by @Xanthi, @Squad, @Jonny Phive, @LATTEH, @Steppenwolf and others at New World Interactive
"Precinct, was a fun and challenging map to work on. We decided early on to melt District and Contact two of our very nostalgic maps together into a single large-scale urban environment. The goal was to preserve the nostalgic feeling and at the same time create something unique and fresh not just a 1:1 copy. In the block-out stage we started playing with different terrain heights, which eventually was the key to accomplish our goal. Terrain height was a bit of a trial and error process; I remember driving up a hill and not having enough torque, oops!!" -Xanthi
Counter-Strike: Global Offensive - Killhouse
"Killhouse showcases brutal duels, player reaction times, and close-quarters combat. A highly vertical layout ensures the sort of unpredictability and replayability ideal for CS:GO’s 2vs.2 "Wingman" game-mode." - FMPONE
Counter-Strike: Global Offensive - Station
by @Roald and @untor
"All experiences contribute to where I am at this point. I am just a hobbiest but I think I learned alot about level design just by doing it and enjoying it. Overal my goal is to improve myself on level design, but also enviorment art. I think I archieved a goal on level design and it's now time to continue on enviorment art. This is where untor morozov comes in. I have met untor a while ago. He made this map 'Waterfall' which was pretty populair. I liked his designs and added him as a friend. When I had this wingman map going on with positive feedback I just contacted him again to work on it with me and since this moment we have had a incredible teamwork. I am gameplay orientated and he is art orientated so we were a great couple. We just enjoyed work on this project and respected eachother and had alot of fun." - Roald
by @Yanzl and Sara Lukanc
"The Gap is a sci-fi thriller first person narrative exploration video game. You play as Joshua Hayes, a neuroscientist trying to figure out what happened, barely remembering anything about his past. It started as a project for our BA thesis and has now grown into a standalone game. It's also my first "real" indie game project, helping me learn a lot about Unreal Engine 4 and game development in general." - Yanzl
Counter-Strike: Global Offensive - Alexandra remake
"My first successful map was born 10 years ago for CS1.6. It was done in just 4 days. Since then it has been ported/improved several times on CS:S then finally on CS:GO. It always had a "dust" theme. Initially i wanted to remake it with an "inferno" style but when the new dust2 came i switched the plan to use the new assets. The map was and is frequently played on public servers especially in Eastern Europe so i had plenty of feedback to improve it. For some it's just another "dust" map, but for me it's my dust2." - Serialmapper
Far Cry 5 - Wetland Turmoil
"I wanted to try working with location design in an (imaginary) open world game for the first time, so I made this backwater cabin neighborhood. At the time I also wanted to see what the limits were in Farcry Arcade and how far I could push it. The level has fixed spawns (a limitation of the editor), but I toyed with the idea of making it work regardless from which direction the player would have approached it. The pathing and player guidance is more or less shaped like the number eight, with the church acting as an outlook. Your task is to eliminate all the bad guys. In the end I wanted to do so much more, but couldn't due to technical limitations. All in all it was a fun experience to make it." - grapen
Counter-Strike: Global Offensive - Trailerpark
by @OrnateBaboon and @Skybex
"We wanted to make a map for CSGO, using a theme that had not been seen in any previous version of Counter-Strike.The map had to incorporate everyday plausibility, provide for enough variety so that things remained visually interesting, but also be flexible enough to allow for the use of low geometry for easy grenade strategies. Being able to immediately recognize a theme in a map is always important, so with all this criteria in mind, A trailer park fitted the bill perfectly. There is still some way to go before a full release, but 2018 was a great year for progress on this project." - OrnateBaboon
Unreal Engine 4 scene
"I was inspired by games like stalker and the last of us. The goal was to make something photoreal with a lot of foliage. It took a couple of iterations but I think I achieved the goal in the end. While making this project I've had to learn a lot about Speedtree to make all the foliage, it was a really cool experience. Right now I'm in the army so unfortunately I can't make any more scenes right now, but after I'll come back I'll try to make more scenes like that." - Corvus
Overwatch - Busan
by @Minos, @[HP], @PhilipK, @IxenonI, Phil Wang, Lucas Annunziata and others at Blizzard Entertainment
"Busan was a challenging map to make. Due to the game having 12 different heroes on screen we have a somewhat limited memory budget for maps, that includes all models, textures, effects, collision data, lighting information, etc... Fitting three radically different areas (Downtown, Sanctuary and MEKA Base) into one single map budget required us to find new ways to optimize our work. In the end, we were even squeezing kilobytes out of collision data to make it all fit, no kidding! But the result speaks for itself, the map was fun to work on and we are very proud of what we accomplished!" - Minos
Counter-Strike: Global Offensive - Highlands
by @ElectroSheep, @El Moroes and @'RZL
"We wanted to make a map in Scotland because, thanks to dishonored 2, we were browsing a lot of references froms this area and we really loved it. I also went myself here in holliday after that. We asked one of our close friends to make some special props, like the police van, the taxi, the phonebox and some others. Unfortunatly the hard development of Dishonored 2 put us in a difficult state where we weren't able to work on the map. So we lost motivation. Then RZL contacted us because he didn't want the project to die so we gave him the keys. And RZL became busy too ^^. Life sometime say NO I guess, hehe. Now Highlands Is my only advanced project I still didn't finished and I'm ready to give it a try, I hope." - ElectroSheep
"Highlands...is this map is a joke? Certainly no but we can say that the development is quite longer than what we expected. Perhaps we learn well how the famous "Valve time" works? :p No seriously I think we can explain that with the motivation. Of course we were motivated to create something cool with this map but with the time and, I think, with what we live in our life we never took the time to do it correctly...I mean we never had a constant rythm on the map. This (and other personal things) led to the current statut of the map; a still "work in progress" map started in 2014. But ElectroSheep came back and his goal is to finish it, and because he's right, I'll come back too to help him. Just, be patient (again) ;)" - El Moroes
Battlefield V - Fjell
by @Puddy, @Pampers and others at DICE
"Fjell was an explosive experiment which paired a new Battlefield dynamic, planes and infantry only, with an epic gosh darn mountain top. Tackling this design combination was like dealing with a bear after you've kicked it in the balls. It was a fun challenge and even though its extreme gameplay is quite polarizing when compared to more middle-of-the-road maps, I am happy that we went there!" - Puddy
Counter-Strike: Global Offensive - Iris
by @BubkeZ and @Oliver
"Iris was born out of a shared interest in the TV-show "Seinfeld", funnily enough. One day BubkeZ noticed I had changed my Steam profile picture to a photo of "George Costanza" and just like that the wheels were in motion! In the beginning, BubkeZ had the vision of an old city environment with lots of dirty alleyways and brick architecture. We didn't want to fall in the trap of making the map look too bleak, so we came up with the idea of making a mid-century town set in autumn. While the map certainly have visual elements from the 50's, I would say the overall theme of Iris is american auto-industry. Making the old cars was definitely my favorite part of making this map!" - Oliver
Unreal Engine 4 scene
"I have always been a fan of retro and vintage, so this was like a dream to me. After watching the first season of True Detective, I immediately fell in love with the office set and the way the series was shot. I have definitely learned a lot from this project, mostly lighting techniques that can fill your scene with a story. The goal was to recreate their environment in my own style, and I'm pretty satisfied with how it turned out. I definitely wasn't expecting this much of positive feedback and I'm really thankful for this community. I want to do something with the environments, not just as a portfolio piece, but make a short film or make a small adventure game out of them." - Brightness
Counter-Strike: Global Offensive - Insertion 2
"Being the follow up to the first Insertion it will have the same overall concept with the spawning and open-world like layout. However this time it will be a more urban setting and overall higher quality art assets. I always love to make environments that feels real. And that are familiar. Its all made up. But the details and various elements in Insertion 2 is from my childhood basically. Friends that grew up in the same place I have recognizes it aswell." - Oskmos
The Door Challenge
Designing Highly Replayable Stealth Levels for Payday 2
Level Design in Max Payne: Roscoe Street Station
Effect and Cause - Titanfall 2 Level Breakdown
2017: Mapcore's Year in Review
Hurg smiles upon you all!
That50'sGuy reacted to THE OWL for an article, dz_blacksite - info about use "4wayblend" textures
Next tutorial > [dz_bs - info about use "4wayblend" textures #2]
Hello! On Mapcore for the first time, so I don’t know if you write something like that here. In general, I recently encountered a problem I had been thinking about for a long time, and the solution was the simplest...
4wayblend textures stumped me and I didn't understand how to work with them. Since I could not find any articles about this issue, I decided to write about them.
What if the texture details don't work?
In order for 4wayblend textures to use grass, you must write the path to these details in the map parameters. (The path you must follow is shown below)
Find a button "Map" on top panel in Hammer World Editor > Further Map Properties > Detail Material file > detail/detailsprites_survival (Put this path inside "Detail Material File")
(!) To set the grass, use "Paint Alpha"
How blend multiple textures?
Go to the settings [displacement] and select there [Sculpt], then find the button [Blend] and there select the desired texture. (Use the left mouse button to paint, and clamping the right one will reduce the radius of the paint area)
(!) If the camera mode in 3D view is set to "3D Shaded Textured Polygons", then you will not see the drawn part of the texture. Camera mode should be set to "3D Textured" when drawing
That50'sGuy reacted to will2k for an article, Optimizing An Open Map in Source Engine
An open map?
Source engine, which is funnily a Quake engine on steroids (a bit of exaggeration but still), inherited the same limitations of its parents in terms of visibility calculations: BSP and PVS. This fact makes Source, as was Quake engine before, more suitable to rooms and hallways separated by portals where the BSP shines in all its glory.
Inheritably, Source does not like large open maps where the PVS is of considerable size and the over-rendering is a real issue.
If you work with Source engine, then you already know the importance of optimization in a large, detailed map. Optimization becomes even more imperative when the said map is open.
What’s an open map? Good question. The word “open” is an umbrella term to denote any map that does not have traditional hallways and corridors that connect indoors to outdoors. The map is mostly large, outdoors with an unbroken skyline; in other words, the same stuff that source engine nightmares are made of in terms of PVS and BSP.
In a traditional “hallway’d” map with twisted corridors leading to open areas followed by other hallways, and even if you “forgot” to place hints and areaportals, the geometry itself allows the engine to cut visleaves and limit visibility; granted the visleaves’ cuts will be subpar and messy and the PVS will be in excess, but still, the visibility and fps will be relatively under control. A twisted hallway is a remedy to long sight lines after all.
In an open map, and without hallways and enough geometry to help the engine, the PVS risks to be huge and the whole map could be rendered at once from any point (over-rendering). We are talking here about a severe fps killer and a potential slideshow on a medium to low range computer. Source does not like over-rendering; I repeat, Source does not like over-rendering.
I believe a screenshot should be welcome at this stage to illustrate an open map. I’ve chosen a nice medium-size map from CSGO to showcase the issue: de_stmarc.
The shot is taken in Hammer obviously, and you can immediately see that the skybox is one big unbroken body from one edge of the map to the opposite one. This is the classic definition of open map.
Let’s see this map in 2D view from the side.
I have highlighted the skybox in blue so you could see the continuous sky body all over the map. Please note that an open map can have varying skybox shapes but I’ve chosen the simple and classic one to showcase my point where it is easier to see and visualize the concept of open map.
In contrast, a “traditional” map will have several skyboxes, often not connected directly but rather through a system of indoor rooms or hallways, varying in size and shape.
I will have my map de_forlorn as example here.
I have also highlighted the skybox in blue and you can easily notice several skyboxes for CT spawn, T spawn, and Mid/bombsites. These skyboxes are not directly connected to each other but the areas related to them are linked on the lower levels through various indoor locations, some vast (like garage, tunnels…) and some small (like lab hallway…).
If you are not that comfortable with source optimization or feel that certain terms are alien to you, then please read my previous optimization papers and articles before proceeding further in this article (Previous papers can be found here Source Engine Optimization roadmap).
The necessary tools
I’m not revealing a secret when I tell you that the same tools used to optimize any map in Source are exactly the same ones used for optimizing an open map. If you were expecting some magical additional tools, I’m sorry to bust your bubble.
Since the tools are the same (nodraw, func_detail, props, hints, areaportals, occluders…), it is more about how to use them in open maps that makes all the difference.
So, how to properly optimize an open map? Well, you could always pay me to do so for you (joking…not…maybe…I dunno!!)
If the above option is off the table, then read on the rest of this article .
While in a traditional map one might get away without using horizontal hints, it is virtually impossible to skip them (pun intended) in an open map unless you want to witness single digit fps burning your eyes on the screen. They are of utmost importance to negate the "tall visleaves across the map" issue.
In a traditional map, even if you bypass adding horizontal hints, the damage in fps will mostly be local since the skyboxes are not connected and areas are mostly autonomous in terms of PVS. In case of my map “Forlorn” and referring to the 2D diagram above, if I remove horizontal hints from CT spawn, then only this area will suffer from tall visleaves and over-rendering. Obviously, this is not cool in terms of optimization, but at least the effect will be somehow restricted to this area only.
In the case of “Stmarc”, you can certainly see that not including horizontal hints will have tall visleaves seen from across the map as the skybox is one unit. The PVS will grow exponentially and the over-rendering will take its toll on the engine.
Let’s move on to some screenshots and diagrams, shall we.
This is our glorious open map in side view. The blue lines denote the skybox, the dark grey one is the ground, and the green rectangles represent solid regular world brushes such as building bases for example. The red starfish little-man-with-arms-wide-open is the player. The orange hollow rectangles denote the various visleaves that the engine would probably create in the map (most go from ground level to skybox level and this is what I refer to as “tall visleaf”).
If you know your optimization, then you certainly remember that BSP relies on “visibility from a region” approach (for a refresher, please consult my papers Demystifying Source Engine Visleaves and Source Engine PVS - A Closer Look. This simply translates to the following: the player is in visleaf A and visleaf A has direct line of sight to visleaves B, C, D, E, F, and G. The PVS for A in this case would be stored as BCDEFG. Once the engine recognizes that the player is in A, and regardless of the exact position in A, it will proceed to render the whole PVS content. Everything in visleaves BCDEFG will be rendered even though the player is at the extreme end of A and has no line of sight to most of this content.
You can immediately notice the extent of damage you will inflict on your open map if you neglect adding horizontal hints: excess PVS with additional useless content to be rendered at all times.
Now that we established the importance of these horizontal hints in open maps, the question remains: where shall I put these hints?
In the diagram above, the most logical places would be on top of the 3 green rectangles.
We added 3 horizontal hints (H1, H2, H3) on top of the 3 regular brushes in our map (the hint face neatly resting on the top of the regular brush while other faces are textured with “skip”). This will create more visleaves as can be clearly seen in the above diagram, and vvis will take more time to calculate visibility due to the increased number of leaves and portals but this is done for the greater good of humanity your map’s fps.
Now the player is in visleaf A1 and the PVS is reduced to (sit tight in your chair) A2, A3, A4, B1, B2, C3, C4, D1, E4, F3. On top of the nice result of a greatly reduced PVS (and therefore content to render), keep in mind that leaves A4, B2, C4, D1, E4, and F3 are mostly empty since they are way up touching the skybox.
Some folks will start complaining and whining: what the hell dude, I don’t have 3 green rectangles in my map; where would I put my hints?? My answer would be: deal with it!!
Joking aside, open maps will greatly differ in size, shape, geometry, and layout. What you need to do is choose 1 to 5 common height locations in your map where you would implement these hints. Medium maps with mostly uniform building heights can get away with 1 horizontal hint, while complex, large maps with various building heights can do with 4-5 hints.
If your map has a hill made of displacements that separates 2 parts of the map, then it is also a candidate for horizontal hints. You just need to insert a nodraw regular world brush inside the displacement to be used as support for the horizontal hint (the same technique can be used if you have a big non-enterable hollow building made mostly of func_detail/props/displacements).
These might not come into play as much as their horizontal siblings, however they could see a growing potential use depending on the map’s layout, geometry tightness versus openness.
I cannot go through all combinations of open maps obviously to show you how to lay vertical and corner hints; what I will do is choose one diagram representing a typical open map scenario with some scattered houses, streets, and surrounding fields. Once you see how I proceed with these hints, it will become a lot easier for you to implement them in your own map regardless of the differing geometry and layout.
Here’s our typical map viewed from top with grey lines being map borders, green rectangles being houses (solid world brushes), and our tiny red player at the rightmost part of the map. The map has a main street that goes in the middle between houses but the player is not restricted to this path only.
The diagram below shows how I would proceed with my hints for such setup.
This is basically what you get when you give a 5-year-old some crayons.
Seriously though, I just gave each hint a different color so you could discern them on the spot, otherwise it would be hard to tell where each one starts and ends.
Most of these hints go from one side of the map to the other while going from ground level to skybox top; don’t be afraid of having big hints that cross your entire map.
Notice that we have both straight vertical hints (shown from above in the diagram obviously) and corner hints; what I did is that I compartmentalized the map so wherever the player is, chances are they will have the least amount of leaves to render in the PVS (this is just a basic hint system and more fine tuning and additions could be done but you get the gist of it).
To get more details on hint placement, please refer to my paper Hints about Hints - Practical guide on hint brushes placement
If your map has enterable buildings, then it is imperative to separate indoors from outdoors using areaportals; this is top priority.
Make sure to slap an areaportal on each door, doorway, cellar door, window, roof opening, chimney, etc. that leads inside the house in question.
What about outdoor areaportals? Good call. In an open map without much regular world brushes to maneuver, it could get very tricky to set up an outdoor areaportal system to separate areas. However, you should always strive to have one, even if it is one or two areaportals across the map. The reason is very simple: the view frustum culling effect, which, coupled with hints, will yield the best results in cutting visibility around the map.
Continuing with our previous diagram, a simple outdoor areaportal system setup could be as follows (top view).
This setup will make sure that the map is split into 4 areas and whenever you are in one of them as player, the view frustum culling effect will kick in to cull as much detail as possible from the other areas.
Let me show you the setup from a side view to make it easier to visualize.
This is the same areaportal that was closest to the player in the top down view diagram but this time viewed from the side. Unlike hints where it’s fine to have one big hint going across the map, for areaportals, it is best to have several smaller ones that tightly follow the contour of the geometry eventually forming one big areaportal system.
Another possibility for outdoor areaportal system is to have a combination of vertical and horizontal (yes horizontal) areaportals.
If your map is a village for example with a highly detailed central square where most of the action takes place, a potential system could be made of several vertical areaportals that sit in every entrance to the square from adjacent streets, and a horizontal areaportal that “seals” the area and works as a “roof”.
For a practical guide on areaportals placement, please check out my article Practical guide on areaportals placement
Props fade distance
This is a really, really important tool when optimizing large open maps. In case you got distracted while I was making the announcement, I’ll go again: props fading is definitely vital when tackling open maps optimization.
What you need to do is to set an aggressive fade distance for all trivial props that do not contribute to gameplay. Players will look closely at how detailed your map is when they check it out solo on the first run; however, when the action starts and the round is underway, adrenaline, focus, and tunnel vision kick in, and all the details become a blur.
During an intense firefight, players will not notice small props and details up close, let alone at a distance. We need to use this to our advantage to fade props thus releasing engine overhead; a faded prop is not rendered anymore and engine resources will be freed and allocated elsewhere.
Your map geometry will dictate the proper fade distances, but as a rough guideline, small props could have a fade distance anywhere from 800 to 1200 units (flower pot on a window sill, small bucket at the back door, a bottle on the sidewalk…), while medium props could do with 1400-1800 range (a shrub, a power box on the wall, an antenna on the roof, wood plank, gutter pipe, fire hydrant…).
Be very careful though not to prematurely fade critical props used for cover or game tactics (car in the middle of the street, sandbags, stack of crates, dumpster on the sidewalk…).
Many people forget about this technique which is more than needed when it comes to open maps that tend to have larger average PVS than traditional maps.
I showcased in a previous article of mine the fps cost of cheap and expensive assets (Source FPS Cost of Cheap and Expensive Assets).
Get in the habit of using the low-poly model version as well as the cheap texture version in the distant non-playable areas and the high unreachable areas where players won’t have much of close contact with the environment. Potential candidates could include a distant field, the unreachable opposite bank of a river, a garden behind hedges/walls, high rooftops, the 3D sky…).
Fog/Far-z clip plane
This technique, when correctly used, can provide a big boost to your frame rate as parts of the world beyond the opaque fog won’t be rendered at all.
For this technique to work properly, your map should have a foggy/rainy/stormy/dusty/hazy/night setting (use as applicable) where a fully opaque fog won’t appear out of place. Obviously, if your map takes place in a sunny and clear day, this technique won’t work much and it will look inappropriate.
Using this is simple: For example, if your map is set in a rainy and foggy day, you just need to set the fog end distance while having its density set to 1. You will then set the far-z clip plane to something slightly higher than the maximum fog distance (if the fog end distance is 8000 units for example, the far-z could be set to 8200).
This is another good technique to reduce engine overhead and the cost of rendering.
It is true that the 3D sky is used to expand the limits of your level and decorate its surrounding, however, since it is built at 1/16 scale (and expanded in-game), it is also a nice way to decrease rendering costs. Use this to your own advantage and relocate assets in the non-playable areas with limited player interaction to the 3D sky.
One thing to keep in mind though, the 3D sky’s visleaf is rendered at all times on top of the PVS in the playable area. Do not go overboard and make an extra complex, highly expensive 3D sky or you would be defeating the purpose of this optimization technique.
You thought I forgot about occluders? Not a chance as these are the big guns when it comes to large open maps with little world brushes to use for other optimization techniques.
Let’s clear one thing first; if your map is made mostly of brushwork and displacements with little to no props, then there is absolutely no need to resort to occluders as they’d be totally useless in this case. Only when the map is loaded with models and props in an open setup with little regular world brushes that occluders come to play in force.
To place occluders, you would search for areas where these occluders could make the most impact (low fps, high traffic, props abundance) since they run in real time and are expensive, otherwise their cost would outweigh their benefit in terms of frame rate variation.
Remember that occluders rely on the player’s position and field of view relative to the occluder to calculate what gets culled. You need to place them in a way to maximize the number of props to be culled behind them when the player stands in front of these occluders.
Let’s see some examples.
We go back to our famous top down diagram; the occluder is dark blue placed on the left wall of the large house while the little black stars represent various props and models. The 2 diagonal black lines denote the player’s FOV relative to the occluder. Anything behind the occluder and within the view frustum will be culled.
That’s nice; we are able to cull 4 props but is it enough? It is not optimal as we can still do better. What if we move the occluder to the right wall of the house?
Much better if you ask me. 5 additional props were added to the culling process meaning less overhead and fewer resources to render for the engine. That is why I said earlier it is all about maximizing the impact of the occluder by placing it in a way relative to the player’s position that maximizes the number of culled models.
Here’s another example (still top down view).
The player has moved to the middle of the central street, and beyond that L-shaped house is an open field with a lot of props scattered around. One way to implement occluders is as showcased in the above diagram. Notice how I arranged 2 perpendicular occluders along the walls for the maximum occlusion effect as all of these props in the field are not rendered from that player location.
Another way to arrange occluders in this case would be diagonally across the L-shaped house (split into 2 or 3 occluders if needed to accommodate the nearby geometry; they can be floating without the need to seal an area).
If you’re feeling brave enough (you should be after reaching this far in this article), you could also add an extra occluder along the wall of the house to the left of the L-shaped house to further enhance the view frustum occlusion effect and cover more props in the field.
The most common places to add occluders in open maps include a displacement hill that separates parts of the map, a hedge that stands between a street and a field full of props, a floating wall between a house garden and the street, the walls of a large house, the walls of a tall building, a ceiling when it separates multiple levels…
To read more about occluders placement and cost, please consult my article Practical guide on occluders placement
The foundation of optimization in Source engine will be the same whether it is a traditional map or an open one. You will heavily rely on func_detail, nodraw, displacement, props… to achieve your goals but it is the way you use these tools in an open map that makes all the difference.
One might get away with being a bit sloppy with optimization in a traditional map, however, make no mistake that an open map won’t be any forgiving if you decide to skip a beat in your optimization system.
Talking about different open maps and formulating varying optimization systems for them could fill articles; I hope this article has shed enough light on the open maps optimization approach to let you easily design a system for your own map.