How to make a good content pack

[Tiftid]

As you may have noticed, tf2maps.net proudly advertises five content packs at the bottom of the homepage.
These content packs each promise a distinct map theme, and the textures and models you need to create it.
In fact, one of these - the Swamp pack - was even bought by Valve and officially added to TF2 in 2010!

But it seems that people are really reluctant to make maps for these themes. For instance, you see a lot of people using the Frontline metal wall and trench wall textures, but not a lot of people making specifically Frontline-themed maps. Why is that? Is it because these content packs are past their prime and all the buzz has died down? I'm not so sure about that.

So let's dive into:

Section 1: What makes a good custom asset?​

I won't spend much time on this section since it's mostly known info to asset creators, but there are a few common trends between assets you see commonly used:

• Assets that are still useful outside the theme they were built for (Frontline metal wall, Frontline trench wall, Swamp rocks and shrubs
• Props that substitute commonly-used brush lengths (walls with 128/256 width and height) or have well-placed origins (look at the base-game mining rock props and coalmines window props)
• Assets that do something that's not possible to do with brushes or base-game models (Japan pack roofs)

This is why the Frontline metal wall and trench wall are so popular, since they're still very useful outside of the Frontline theme, and they're different enough from their base-game equivalents that using them is still justifiable.
You'll find that the other popular Frontline assets (generator, trees, hedges, awnings, storefront, concrete barriers) all also meet these three criteria.

But there's actually more to making a good content pack than just making good assets.
I find that one of the major reasons that custom props go unused is because they aren't providing a gameplay purpose that wasn't possible with the base-game assets.
As such, you get this problem where even if your pack is capable of creating a unique and amazing rendition of your theme, you really can't account for the skill level of the people who are using it, because we are in fact a community of hobbyists.
So, your theme ends up being known for all the people who build a map with boring gameplay and then slap a few of your assets into it, because your assets don't actually encourage unique and interesting gameplay - they're only a useful tool that, if used right, can create interesting gameplay.

To avoid this, what you could try doing is:

Section 2: Plan the pack differently​

Conventional content packs will do one of two things:

1. Make the pack first, then try to get people to make maps for the theme it encourages
2. Make the pack as a set of assets tailor-made to fit a single map that's going for a particular theme

The first approach is a huge risk, because like I mentioned earlier, you can't account for the skill level of your users, which will ultimately decide how the theme is remembered. You also can't guarantee that people will actually be motivated to make content for the pack, unless you've taken on one or more high-profile mappers to do it, in which case all maps that come after theirs will look disappointing by comparison, so the theme will end up being remembered sourly anyway.

The second approach might work well for the mapping style of the particular mapper who the assets have been built for, but it might cause frustration for other mappers who don't use the same distances. It's also a slippery slope to making props that require very specific brushwork built around them, which is a fast way to create a prop that no one will use. It also constrains the amount of creative spins on the theme people can create, since you're building all your assets for one person's vision of the theme, and probably to fit one skybox and set of lighting conditions.

What I'm proposing is a new, third approach:

Firstly, scan a number of promising alpha maps, keeping your theme in mind.
What you're looking for is a map where the only way to explain the gameplay the mapper built is by artpassing the map with your theme.
That way, you know that if a future mapper builds gameplay similar to this, they'll want to use your theme pack instead of base-game assets or a different theme pack.
Putting the cart before the horse like this is also kind of a loophole - your assets will inherently be useful outside of maps with your specific theme, since anyone who builds this specific kind of gameplay will want to use your assets, regardless of whether or not they set out to make a map with your theme.

Now, pick one of the maps you found and ask yourself "What would I do if I had to artpass this map with my theme?"
Since your assets don't exist yet, you should get as far as possible with the base-game assets. (Don't decompile the map, this is just a thought experiment!)
This will prevent you from creating assets which are made redundant by base-game assets.
Once you've done that, make a list of everything that wouldn't be possible to artpass to the standard you want with base-game assets, and put it in a "potential assets" document. Then, go to the next map and repeat.

This is all a bit vague, so let's try it with a couple of recent alphas and the Space/Moonbase theme.

Spoiler: cp_havoc_a3

Immediately we can see a very symmetrical and angular design which would be very easy to translate into a space-base theme.

All of these circular structures are perfect, since rounded or beveled edges are a common space-base trope.

These open courtyards with very low-to-the-ground structures are also perfect, since pipes running along the moon/planet surface are common in space-base concept art and could easily be imitated here.

This area combines a number of these ideas - there are pipes running along the ground, a small circular structure contained within a much larger one is great, and the white trim highlights how amazingly useful a generic "light strip" texture could be.

Current "potential assets" list:

• Circular wall pieces (256 radius, 32 width, lots of grates and internal pipes/wiring)
• 128x128, 256-tall cylinder prop; perhaps more than one (RED and BLU skins?)
• Light strip texture for edges and trims

List of things that could be achieved with default assets:

• Wall details (numbers, vents, windows)
• Pipes
• Various angular shapes (brushwork)
• Spherical storage tanks

Spoiler: cp_edge_a6

This BLU spawn looks very like some kind of starship. Maybe there could be a "starship spawn room prop", except rebuilt to act as more of an effective spawn room (e.g. with two side doors that go to catwalks on the sides of the hull so that the rotate time is shorter for players exiting spawn than prospective spawncampers)

Additionally, a space skybox with a planet in the bottom texture (or a generic space skybox + a few planet props to stick in the 3d skybox) would be really useful for maps like this, where getting sucked out into space would be a hazard.

A generic 512x512 or 1024x1024 "floating platform" prop with some kind of stabiliser machine under it could be very helpful to mappers who want to build underneath a platform without having to build supports - the stabiliser on its own could also be useful for mappers to place under floating brushwork platforms

A particularly tall stabiliser machine could be useful as cover in situations like this.

This shipping container with the health pack in it could be replaced with some kind of escape shuttle, which could act as both a cover prop and a nice little health cubby, similar to a gravelpit shack.

Current "potential assets" list:

• Circular wall pieces (256 radius, 32 width, lots of grates and internal pipes/wiring)
• 128x128, 256-tall cylinder prop; perhaps more than one (RED and BLU skins?)
• Light strip texture for edges and trims
  
  
• Starship spawnroom prop
• Escape shuttle prop
• Floating platform / stabliser machine prop
• Space sky texture
• Planet props for 3d skybox

List of things that could be achieved with default assets:

• Wall details (numbers, vents, windows)
• Pipes
• Various angular shapes (brushwork)
• Spherical storage tanks

Spoiler: cp_blast_x8

This sort of machine with three cylindrical storage tanks and a catwalk along the top would be great to turn into a prop, especially if the catwalk was also mirrored onto the other side like so

A gigantic mound of space rock or crashed meteorite would be an amazing prop for this area.

A set of circular catwalks and stairs to go around our circular wall pieces would be amazing - also, this bowl shape with the hole cut out of it would be useful in so many situations.

Current "potential assets" list:

• Circular wall pieces (256 radius, 32 width, lots of grates and internal pipes/wiring)
• 128x128, 256-tall cylinder prop; perhaps more than one (RED and BLU skins?)
• Light strip texture for edges and trims
  
  
• Starship spawnroom prop
• Escape shuttle prop
• Floating platform / stabliser machine prop
• Space sky texture
• Planet props for 3d skybox
  
  
• Triple-tank + catwalk cover prop
• Circular catwalks and stairs to fit with the circular wall pieces
• Version of the circular wall piece with a tall window that can be shot through cut out of it
• Bowl with a hole cut in the bottom (could be made of glass and braced with metal to cast interesting shadows)

List of things that could be achieved with default assets:

• Wall details (numbers, vents, windows)
• Pipes
• Various angular shapes (brushwork)
• Spherical storage tanks

And there you go! By doing this, we end up with a much smaller, but more effective list of props - it's almost difficult to use any of these props without forming a unique and interesting gameplay space around them.
The texture suggestions and 3d sky planet props also feed into gameplay somewhat, since the space skybox and 3dsky planets can be used to create a space map that still has normal gravity and death pits as a falling hazard, and the light strip... well, that's kind of just a light strip.
It is totally fine to make assets that are just there to look cool - just as long as they're not too similar to base-game assets.

Section 3: Remember your Roots​

The final step to creating a good content pack is ensuring the quality of the theme itself.
Firstly, you have to have a strong visual idea. This comes even before you select maps to draw gameplay ideas from. Ideally, we would have concept art, but Valve's concept artists are busy with their real jobs, so we can't ask them.

So, what people choose to do instead is build a moodboard of reference images from other media or real life. This is great and all, but the problem is that other media, real life and TF2 tend to not inherently be compatible. For instance, I see a lot of people these days making office spaces, and lovingly furnishing them with authentic shiny wood walls, and objects that would be plausible to see in a 1960s office.

But the thing is that not everything that's 1960s is inherently TF2. TF2 as of 2007, and most importantly as of Valve's concept art, focuses around a bunch of dusty, dirty industrial buildings. This is true of the deserts of 2007, the alpine lumberyard of 2008 and the snowy viaduct of 2009. Conversely, even the fully industrial foundry of 2012 still has its dust and dirt, its paint, stains and grime on its textures.
Now, you don't have to specifically imitate industrial areas - that would be boring. But you have to remember that TF2 focuses around almost cartoonish ridiculousness. You don't have to put your buildings on stilts, but rejecting an idea because it makes it hard to build the space in a realistic way is just not using TF2's design toolkit to its fullest potential. For instance, electrical boxes and crates would look pretty out of place in a real office, but in TF2's universe, would anyone bat an eyelid to see them there?

It's also important to keep a strong colour palette and the right amount of colour contrast in mind. I see too many themes fail to get off the ground visually because their scenes are all effectively a single colour. Conversely, a theme can have too much colour contrast, and just give every building a different colour of plaster wall to produce a result that really isn't visually appealing.

Typical TF2 maps will strike a good middle ground by having really intense colours, but a low raw number of colours.
Here's Dustbowl as an example:

The sunlight is quite intensely orange, and the surfaces do nothing to reduce that, being all yellow or brown. However, all this is contrasted with the strong blue of the sky, creating one of the most simple yet effective two-tone aesthetics possible: blue and orange.

Many custom themes fail to preserve this, either by failing to create enough contrast with the blue sky, or by using a sky which isn't blue and ends up having a colour too similar to the sunlight, again ending up with not enough contrast.

It's important to note that blue and orange isn't the only "strong colours but few colours" aesthetic that TF2's capable of. For instance, my favourite map, Gravel Pit, has a grey sky. So how can it have contrast?
Simple. The sunlight is REALLY brightly orange-yellow, and the shadows are REALLY blue:

The sunlight and shadows contrast strongly both against each other and the sky. The slight purplish tint of the sky helps it contrast against the sunlight, which is tinted more towards yellow than orange.

So, no matter what skybox and lighting conditions you're going for, you have to remember to have a strong colour palette.

I'll leave off with one final shot of Gravel Pit that I think demonstrates the most fundamentally important aspect of crafting a TF2 theme - the need for every scene to feel like a painting.

 

[Freyja]

I think you are proposing a solution to a problem that doesn't really exist (or when it does, it's not that big of a deal?), and your message gets lost half way down and turns into a lecture about detailing, rather than the development of a custom theme itself. The assets in those images of gravelpit and dustbowl can be used in good compositions or bad compositions or lighting nearly as easily as many of the assets from custom themes.

I believe you are heavily blurring the line between poor use of custom assets, and poor custom assets.

Yes, I believe there are many asset packs that have much in the way of difficult to use assets, either technically or in way of finding a place for them, or assets that are likely never to be used entirely. Frontline and Swamp come to mind as examples here, but I feel those are examples of content packs that have been made in an attempt to be generic, rather than being developed in tandem map/multiple maps

Making a content pack without making a map alongside, as you propose, is simply much harder and IMO gives less reliable results. It requires a lot of mapping/level design experience already, and is kind of unrealistic. If not just for the fact that usually when you get mappers or a group of them to make a custom theme, they want to see it used in a cool map, because that's kind of the thing we're here for.
Not to mention that asset development and map development go hand-in-hand, or at least work best when they do. Never seeing your asset used in context means getting no feedback on it until it's too late. Things like the scale, visual noise, mismatch with stock assets (or other assets in the pack) and ease of use are very hard to suss out from the get-go, unless you're already a very experienced mapper.

I would be interested in discussing this further at length with you, but I don't think the mapping resource forum is really a suitable location for that. I don't think what you've said is wrong, exactly, but I don't think presenting this as the true method is kind of misleading.

 

[Aapelikaeki]

I think it's a little silly to equate "assets that would get easily used by everyone to get you lots of asset credits" with "good".

A content pack is good when it accomplishes what it sets out to do with high quality: for some, it's having useful and easy to use props with guides and an asset zoo vmf to match, for others, it's great when they make just one map shine bright - especially when seeing your assets used to really elevate a map is often the motivating spark to create things in the first place

 

[Tiftid]

I think you are proposing a solution to a problem that doesn't really exist (or when it does, it's not that big of a deal?), and your message gets lost half way down and turns into a lecture about detailing, rather than the development of a custom theme itself. The assets in those images of gravelpit and dustbowl can be used in good compositions or bad compositions or lighting nearly as easily as many of the assets from custom themes.

I believe you are heavily blurring the line between poor use of custom assets, and poor custom assets.

Yes, I believe there are many asset packs that have much in the way of difficult to use assets, either technically or in way of finding a place for them, or assets that are likely never to be used entirely. Frontline and Swamp come to mind as examples here, but I feel those are examples of content packs that have been made in an attempt to be generic, rather than being developed in tandem map/multiple maps

Making a content pack without making a map alongside, as you propose, is simply much harder and IMO gives less reliable results. It requires a lot of mapping/level design experience already, and is kind of unrealistic. If not just for the fact that usually when you get mappers or a group of them to make a custom theme, they want to see it used in a cool map, because that's kind of the thing we're here for.
Not to mention that asset development and map development go hand-in-hand, or at least work best when they do. Never seeing your asset used in context means getting no feedback on it until it's too late. Things like the scale, visual noise, mismatch with stock assets (or other assets in the pack) and ease of use are very hard to suss out from the get-go, unless you're already a very experienced mapper.

I would be interested in discussing this further at length with you, but I don't think the mapping resource forum is really a suitable location for that. I don't think what you've said is wrong, exactly, but I don't think presenting this as the true method is kind of misleading.

Hi Freyja,
I'll start by saying that it was never actually my point to say that a poorly used custom asset makes a bad custom asset.
I think more positively than that. My purpose, from the beginning, was to analyse what traits are present in successful custom assets, a.k.a the factors which are most likely to be what creates a good asset. Building off that, I created a purpose that could be considered opposite to the way you stated it - I don't want to say that a poorly used custom asset is bad, but I do want to say that a really good custom asset will actually inspire people to use it in ways that aren't boring, and custom assets like that can exist and not actually be that difficult to produce.

As you're a much more experienced content pack creator than me, I'm willing to accept you saying that the right way of doing it is by building a map in tandem - and, even while I was writing the original post, I half-thought that.
However, the reasoning for justifying that is strange in my eyes, and almost makes me think you've not read or missed the point of the planning section.
The whole purpose of inspiring the asset creation process from existing maps is that even if you don't have any amazingly experienced level designer on your team, you're observing a scenario where one or more experienced level designers have wanted to use your asset, even if your asset didn't exist yet. You can therefore assume that there's a high chance that the asset will find use in the wild, and be convenient for a designer to use.
It's even better if you have contact with the various designers who made the maps you're basing the assets off, since mappers have all sorts of different grid sizes and workflow, so making an asset for just one person or map is almost never a good idea if you want that asset to be popular.
Also, by basing your asset on existing spaces, and imagining how that space would look if artpassed, you have a very high chance of getting the scale, style, visual noise, etc... right.

I also never said that you should avoid making a map to go along with your content pack, although I now see that I should have said that you shouldn't avoid making one. Obviously it's good to make a map to show off your theme and test out your assets - and preferably not just one! The thing I take issue with and believe sabotages a lot of theme packs is that in this context, your assets are being used by people who are essentially forced to use them. Even if these people do a great job of using your assets, the simple fact is that they had an incentive which people who are considering using your pack will lack.
They've built their gameplay to fit your assets, but most designers actually don't enjoy or profit from being forced to build their gameplay in a specific way.
That's why I recommended building your assets to explain geometry which would be inexplicable/out of place in other themes. You create an automatic market, where people who've built a specific gameplay type will have a strong incentive to use your assets just because no other assets are gonna work for their purposes.

By the way, the original purpose of Section 3 (which seemingly got lost somewhere in the writing process) was to comment on another facet of user error that can ruin a theme:
People making creative riffs on the theme lack any concept art or moodboards that your team had in the creation of the theme, so although they might think their spin on it will look awesome, the customer is not always right and it will frequently fail and make your theme look worse in the eyes of the people. So, Section 3 is more of a guideline for users of a theme pack than for its creators, although the quality of the theme you're making is without a doubt important.
If I wanted to make a pure detailing guide, I would have already done that, and it would have been ten times as expansive as Section 3. There was a reason why I decided to put those tidbits here in these posts instead.

Small tangent:

I think it's a little silly to equate "assets that would get easily used by everyone to get you lots of asset credits" with "good".

A content pack is good when it accomplishes what it sets out to do with high quality: for some, it's having useful and easy to use props with guides and an asset zoo vmf to match, for others, it's great when they make just one map shine bright - especially when seeing your assets used to really elevate a map is often the motivating spark to create things in the first place

This thread doesn't have the kind of scope to be two different guides - one for larger asset packs and one for smaller ones.
It's totally okay to make an asset pack for a single map with the expectation that those assets will never be used again.
I'm a fan of open asset licensing, personally, where people don't get disappointed because whoever created that new "Dark Marsh" map won't release their overabundant custom assets for us to use.
But if you're not, then that's fine.
If you want to make a pack like that, this is where Section 3's second purpose comes in - it's still very important to ensure the quality of the theme. The other factor I'd say is really important is strong communication between the asset pack creator and map designer, but that is again outside the scope of this thread. The most I'll say about it is that even if you think the theme pack is tiny and unimportant, it's still important to have some kind of reference, preferably concept art or something very paintinglike, because that helps the map designer understand exactly how you want your assets to be used.

Finally:

I would be interested in discussing this further at length with you, but I don't think the mapping resource forum is really a suitable location for that. I don't think what you've said is wrong, exactly, but I don't think presenting this as the true method is kind of misleading.

If you're interested in discussing this further at length, then so am I, but I won't do it privately, and I don't think the mapping resource forum being an unsuitable place comes into it at all. Even if we argue back and forth for weeks, it's all contained within one thread - not exactly spamming the forum!
And even if every new thread on the homepage was some back-and-forth argument, it would for me personally make the forums much more interesting to use, since I'm learning something new with each thread I click on, and even each time I'm clicking on the same thread.