You assume work is linear, and vastly underestimate, which is the downfall of many a project manager.

"30x times that size" is not going to be a whole level, it's going to be 30 more empty hallways each with their own slightly different textures. Your basic sci-fi hallways are the easy part of making a map.

Make it 300x more, and you have a slightly more reasonable estimate... but it will likely be even more than that as complexity of the level grows. 

Aww man, you got my hopes up that Spike actually implemented it in FTE somehow. Calling lots of standard textures megatexture is like promising the greatest boobs ever and then they turn out to be fake. :(

Wouldn't texturing like this be very wasteful on memory?

You still seem very limited to the faces so the look is not very unique. It looks very similar to recent Q3/QL. I was expecting some really wild texturing not "standard tileness". For random decals as detail you could just use decals. 

"You assume work is linear, and vastly underestimate, which is the downfall of many a project manager. "

Very much so, yes. Assuming a linear level of productivity over the life of a project is just not realistic. People will get bored, tired, distracted, etc. You won't have consistent levels of output.

And even then, doing an entire game with levels don't like this would require an astounding level of discipline where levels won't be changing once they are textured. That has NEVER happened on any project I've ever been a part of.

That's why things like decals are preferable to painting in splatters and such that cross multiple walls. The walls can change, the decal can move, etc. It's flexible and inherently far more productive than repainting the decal every time you decide that a pillar should be 16 units further to the left. 

Having said that, I understand that actual megatexture games don't have all those limitations ... they have a source file that bakes down into the final so you CAN move things around. It just requires baking everything down again.

Doing this for Quake with existing level design tools feels very limiting. 

Just keep doing your thing KP, I dig.

This. 

I went to bed last night with the essence of Lunrans comments fresh in my mind. I realize that my time/work assessment is about the most optimistic it can be while still remaining quasi-realistic. I choked down a few bits of harsh reality, some of which has been posted mentioned by Scampie and Willem.

Work is not linear. Nor can I expect anyone to BLEED this project until it's done. Just because I could abandon life for a year and a half to live at a computer desk doesn't mean I could expect that of anyone else.

My outlook is often optimistic, it hasn't failed me yet, in terms of personal projects and work. But I understand going by a plan that is very much "best case scenario" is unwise.

After taking a more cynical, or realistic, look I would say using standard, but diverse tiling texture for a majority of the map with key areas uniquely textured is a goal that would require little bloodshed to achieve.

As for now, I'm going to have to let this topic rest. I have more stuff to work on. However, I'll continue to weigh all points raised here. 

For a hobbyist project, I think you can do whatever you want. Who cares? It's your time, spend it as you see fit. :) I was speaking from a professional perspective.

I think uniquely texturing areas is an awesome idea but the key is to make it worthwhile. Not just, as Lunaran pointed out, a unique number on the walls or custom scratches. But really stretch it ... look at Rage for inspiration. It's very obvious in a lot of their environments that the textures are unique. 

And their geometry often makes it so that custom textures are required to make that geometry sing. 

It looks good, but the architectural style of 'future base' is designed to be built from prefab panels and girders, the generalised look of which doesn't lend itself so well to megatexture.

Take what you've got and stretch it as Willem says - explore the benefits of megatexture within the Quake setting.

I dunno. How about turning the centre of the base into a gigantic crater? Or twisting parts of it into organic shapes that have been warped by demonic / slipgate energies.

Anyway, keep going - its great to see stuff like this. 

I'm a pretty lazy mapper, although I do try to incorporate the feedback I get... but yeah, whenever I cut corners I used to think "well, it *is* free" but these days I kind of have to in order to get it released at all.

I loved how maps in the Myth series of games were basically megatextured whereas other RTS games at the time were using tiled images (and were mostly just 2d as well). But for FPS, even for Rage, I don't think unique textures everywhere is an efficient use of artists' time. My $0.002. 

I mean I saw the Rage editor videos - these guys were painting onto a 3D world map, just as you would paint onto a 2D canvas in Photoshop. It gives the ability to use decals for example, but without unnecessary layers underneath [at some point in the rendering pipeline], and then do little touch-ups here and there.

I like the fact that you can use tiling textures and then touch parts up to make each part seem more unique.

Imagine if SleepwalkR put this ability into Trenchbroom! Imagine an editor that would allow you to pixel-paint onto any surface, and then created the necessary .mips and generated a wad when you export the .map file....

Good idea, huh, huh?

I bet there's some ready-to-use code out there that would save the effort of creating a painting app from scratch, it could be hacked into it.

Then doing what Killpixel has done wouldn't be as time-consuming... 

I really geeked out over Myth maps... the placement of goals and terrain features, the available unit selections, how that all worked together. Some of those big texture maps were cool and iconic; I even used them as desktop images for a while. But they were also the barrier that told me if I wanted to make my own Myth maps I could either become a texture artist or fuck right off. So that was a bummer. :-) 

One could import a map into blender. Also, Photoshop CS6 has a greatly improved 3D editor. I'm not sure how well it would handle a map, or even sections, though.

The most time consuming part of the method I used was grabbing the dimensions of each surface. Doing that for an entire map could take a week or more. But once the dimensions are grabbed, a basic, flat texture is made and placed into the map. After that, it's only a matter of drawing on them.

This pic shows some of the basic textures placed onto the map before they were drawn on.

Obviously, if one were to do a whole map (or more) in this way, the artist would have asset libraries to pull from. It's not like every surface would be hand drawn from scratch.

This kinda makes me think of the game Riven. Although pre-rendered, the world was uniquely textured. If they had used tiling textures, the impact of the environments would be diminished, imo. 

When your game's just a prerendered slideshow the whole debate kind of goes out the window because there's a bunch of other tricks that come into play instead - noise based textures, editing and compositing in post, etc. Whole new renderer, whole new point of view.


For a hobbyist project, I think you can do whatever you want. Who cares? It's your time, spend it as you see fit.

Willem reminds me, I should point out that I never meant to tell you you can't do this. It'd be pretty hypocritical of me considering I released a tool-assisted q3 map with fully unique geometry. :P I've just never had any real faith in megatexture as something that adds appeal for anyone other than the people who actually art the environments. I know, I've got the exact same internal artist voice that's always saying "more, more, better" but I know that voice is greedy and utterly unbounded by practicality.

You seemed to be on the same koolaid pretty hard, so I couldn't help but pound the keys. Sorry about that.



Embedding textures in the .bsp does present a neat opportunity for bundling in unique assets, but only if they can be created in a smart way and are only used for the "hero pieces" where the most eyeball value will be derived. Like ijed said, a prefabbed repeaty space hall is the worst place to show this off. I always thought that if there was anything in Rage that did visually benefit from an all-unique approach, it was geology, the domain that suffers the most from tiles. Nature rarely repeats itself.

ugh, now I want to go add compiler-level procedural rock textures and grass blends that just bake huge unique textures into the bsp :| 

Rage did have some of the most beautiful outdoor environments I've ever seen in a video game. 

lol, me too.

Once upon a time I had an idea for a long, story-based "episode" that would take place in several different games, including Myth and Rune. Then I realized that yeah, efficiency is a concern. 

When your game's just a prerendered slideshow the whole debate kind of goes out the window because there's a bunch of other tricks that come into play instead - noise based textures, editing and compositing in post, etc. Whole new renderer, whole new point of view.

I didn't mean as an example of practicality or viability. Simply, that unique textures offer a look that tiled textures could not.

For a hobbyist project, I think you can do whatever you want. Who cares? It's your time, spend it as you see fit.

Heh, a no point did I ever feel I was told I "can't" do this. I don't think anyone here would dictate to anyone else what they can or can't do. Should, well, that's another story... *cue Never Ending Story music*

For contextual purposes, this would be more than a hobbyist project. A potentially commercial effort with a small team of 4-8 people or so. Otherwise, the question is simply "should I spend my personal time doing this", which is something one could answer themselves.

You seemed to be on the same koolaid pretty hard, so I couldn't help but pound the keys. Sorry about that.

It's no problem. Hearing other perspectives was the reason for posting. I already knew what I was in for.

Like ijed said, a prefabbed repeaty space hall is the worst place to show this off.

Agreed, I mentioned this as well.

I'll get a better idea of the logistics of it all as the project evolves. It's still in it's infant stages and at a time when possibilities are on the table and examined 'tis all. 

 

And you got an answer:

discussion goes where ever discussion currently is. over the long run, the thread will go back to it's original purpose.

If you're not interested in intelligent discussion regarding game design, then, well, I have bad news for you. 

actually most of this HAS been directly related to the (numerous, cool, interesting) shots that KP posted.

I'm afraid that there aren't enough people mapping right now for a new 15-20 posts in a thread to be all screenshots (err wait... actually that's over in the Honey map jam thread!). 

Here are some screenshots from work in progrez.
shoot screen shots 

looks great, a few suggestions:
-marble tile floor texture looks out of place, I'd suggest something more like city4_1.
-I'd use a darker, solid rock texture (actually city4_1 might work) with less obvious lines, for the lining of the arch.
-See how it looks with a low minlight (30?) and fog; this might solve the issue of black surfaces next to lit ones.
-make it so rotfish jump out of the water (heh). 

And look at this explanation:

Link

Brushwork looks cool, needs ambience! 

what's going on at the base of the waterfall? is that an mdl?

also, on a scale of "yes" to "absolutely" how much is this your reference image for the first shot?
http://th05.deviantart.net/fs71/PRE/f/2013/082/1/4/a_narrow_and_desolate_ravine_by_crs1009-d5yy6s4.jpg 

Tronyn- brushtex is a bit out of order, as it was more a case of will-it work or collaps.
As it works I was planning to use the rock5_2 tex for the cliff.
The light errors were due to the light1.24 of BJ.

@mfx- thanks for the tutorial, almost gave it up after finding Tyrlight.

@Lunaran- whole waterfall is an mdl, at the base is what my poor imagination could think of turbulence.
On the scale to "maybe" and "sure" I must relay to the inspiration and reference thread like four elephants asking: someone should make this fifth.

Cliff Demons" .OBJ pack has fine eyecandy for skyboxes by Poison.