AoD Lighting sucks. Badly.

[Nick]

Hey EvoG,

Our artist are using this technique (SkyLight+LightTracer) every day, but only for preview purposes.

We tried different options when we switched to 3D, including baked lightmaps stuff, but that increases texture size tenfold. Since AoD's primary distribution is download, we must keep the size down, even if that means having inferior graphics.

Multitexturing is not supported in TGE, so that's not an option either.

Of course, we can improve the visuals using baked lightmaps, but a much bigger download size will probably turn off more people than slightly prettier graphics would attract.

 

[hiciacit]

Of course, we can improve the visuals using baked lightmaps, but a much bigger download size will probably turn off more people than slightly prettier graphics would attract.

I wouldn't be so sure about that. What differences in file size are we talking about?

 

[vazquez595654]

A lot of people have broad band now days, so it wouldn't be such of an issue, unless your talking about gigs instead of megs.

 

[Vault Dweller]

A decent game would be broadband users only anyway. One of the indie projects that I know if is 2GB. We are trying to keep AoD size in MBs.

 

[Bluebottle]

Would it be possible to offer the less efficient lightmap system as an optional download for those who care? If it was released as a patch a couple of month after release it'd be a good excuse to get the game in newsposts again, keep the hype-train runnin' and all that.

 

[Nick]

I wouldn't be so sure about that. What differences in file size are we talking about?

Something like 50Mb vs 300Mb, I think.

 

[Hazelnut]

A decent game would be broadband users only anyway. One of the indie projects that I know if is 2GB. We are trying to keep AoD size in MBs.

Well if you think it's gonna be broadband users only, then 500Mb, 1Gb or 2Gb doesn't really make a lot of difference surely. As long as you're not paying per minute then just leave it overnight... of course you will need to ensure there is a way of resuming broken downloads...

Can you do two versions with no extra work but a re-build?

EDIT: 300Mb is really not a lot these days.

 

[Nick]

EDIT: 300Mb is really not a lot these days.

That's a very approximate estimation, of course. We didn't work on LM version, so I can't say precisely. 300Mb is just to give you idea what numbers we're talking about: hundreds of megabytes.

 

[suibhne]

I d/l 300MB movies all the time over the intartron, so w/e. My big question would be how many modem users you'd be shutting out.

There are a few ways to deal with this. You could take Vogel's approach and design your game specifically for modem-sized downloads, or you could shoot for a larger, broadband-sized file but still make an effort to keep it accessible for determined dial-up users (e.g. hosting a reliable torrent, printing detailed download instructions for dial-up users, etc.).

As a broadband user, my only concern is how much dev time any improvement would take. I'd rather see new content added than better graffix. If different people work on those different tasks, tho, I'd love to see both. :wink:

What you want to shoot for imho is publicity from outlets like IGN, Gamespy, or Gamespot. Yeah, they suck in many ways, but at least the first two regularly highlight shareware projects and can steer massive clicks in your direction. In fact, see if you can catch Tycho's eye at Penny Arcade while you're at it; you'd get a veritable eff-ton of interest. It's a delicate balance to strike, because additional deep game content won't help you get that kind of attention - whereas niftier graphics could definitely help with this sort of mainstream publicity, to pass the initial acid test of first impressions. The goal isn't for people to say "OMG SOIL EROSHUN", but just for them to not dismiss this game out of hand.

 

[Fryjar]

I definitely have to agree with suibhne on this issue.
If you want to cater to both, the modem users and the broadband crowd, you should consider releasing two different versions, one with additional lighting and one without.
Otherwise you should probably stick to a bigger download size, including the lighting.
First and foremost, most of the players should have a broadband connection even though the average cRpg player might in other respects not have a top notch pc.
But on top of that, you really have to bear in mind, that although there are many Crpg lovers, not all of them visit "underground" sites like the codex.
Most might still stick to Gamespot and therefore coverage on these sites will surely boost your sales tremedously.
The issue with this is, as we all know, that they only cover good looking games.
But the difference between your game and 99% of the other indie games is, that already now your game looks much better than theirs.
So probably with this new lighting your game will cross the border of their requirements of coverage. In my opinion it would at least be worth a try.

 

[HanoverF]

I'd have to see the difference in graphix between the current planned release and whatever huge graphics whore version would be (but seriously I d/led a 300meg E3 video for Bioshock in like 20 minutes), but a graphical upgrade patch isn't unprecedented. If its not too much time involved it might placate the steaming masses

 

[sheek]

A decent game would be broadband users only anyway. One of the indie projects that I know if is 2GB. We are trying to keep AoD size in MBs.

How about making an additional "Texture pack" available for those who want it? It would add to the appeal and if you already have those additional graphics why not? Maybe even charge $3 more for it... :wink:

edit:
What Hazelnut already said.

 

[Claw]

I also agree with the Hi-Quality version or update notion. I might be willing to pay extra for eye candy, and file size isn't an issue.
I download 600mb game trailers and similar stuff on a regular basis. Fuck, my download folder is several Gigs in size, I seriously need to clean it up but it's easier to just buy a bigger HD these days.

PS:
Of course it depends on how great the improvement is vs. the effort required / price. The game looks good enough for me, but I can't say graphics mean nothing to me.

 

[Elwro]

I think it will be a challenge even to let modem users know about the game, since they don't spend much time on the net. I'm not sure what is the proper way of handling this difficulty...

 

[xedoc gpr]

I really don't care about file size. I'm only going to be downloading the game once. Leave it on in the morning, go outside for the day, come home at night, it's done. I really don't get why people would complain about download time.

 

[EvoG]

Funny, but what xedoc gpr (??) said is what I told VD my last PM; if people want your game they're gonna DL it, regardless.

The other thing (and codexers were definitely insightful on this) is yea, offer the lightmap pack if you think its REALLY going to matter to those few modem users, but dont limit your product to the lowest denominator when there's no evidence they'd even care themselves (vocal minority does not count).

Now, Flashback, we need to be clear on terminology before I can address this:

If you BAKE down lighting into a preexisting texture, it adds ZERO to texture filesize. You're not creating a new map, you're authoring the lighting into the texture you already have. 128x128 is still a 128x128 texture bytewise with or without lighting. Since you can't do mulitexturing (that sucks btw), then this is really your ONLY course of action.

Now, as I told VD in the PM, if you're using ONE wood texture for all your "wood" objects, then yes, you're going to have to assign each unique object(remember a duplicate of that object is NOT unique, its an instance) its own wood texture if you want to then bake in lighting, but this really isn't much of an issue when again, compressed 32x32 - 128x128 textures are tiny. At around 3 - 4kb 10, 100 or a 1000 new textures, its really negligible in the grand scheme of many hundreds of megs.

All in all, I think it could dramatically improve the look with little effort. Will it make the game more fun? Perhaps as people enjoy exploring pretty areas, but for the most part not really. Will it help get the attention of people that love these kinds of rpg's but expect a certain level of production value today? Absolutely. Remember, its about effort:reward, and in this particular case, there's really no great reason not to use lighitng as its simple to implement, not a huge addition to your bytesize(assuming you really are that concerned with DL size) and could attract those people that are of a like mind, but have particular expectations.

I'm only offering help btw, not trying to influence or insist. :D


Cheers

 

[EvoG]

Oh and remember, I'm talking about baking AMBIENT OCCLUSION, not bounce light or shadows. That requires a particular method of geometry construction and is not conducive to using instances, as instances reference the source geometry's vertices and uv's. Materials could still be assigned on a per-instance basis though.

Ambient Occclusion is the occlusion(blocking) of ambient lighting on an object by its topology; fine details, and areas not directly in the path of light occlude light, so they are darker. The tighter the area, the darker it is(less light can reach those surfaces).

Sample from Nvidia's site.


Every character we're making now is utilizing an ambient occlusion layer along with perpixel dynamic shadowing. The end result is stunning to say the least (look at Gears of War for example). So if you can offer half of that equation, its inexpensive but looks great.

Now one last thing I have to mention. You cannot tile a texture that is using baked lighting for obvious reasons (you'd be tiling the lightmap as well). So in this instance, there isn't an easy solution for ROOM geometry, but to use a larger baked map, which isn't a horrible thing and I'd be willing to demonstrate this for you. If you look back at the two images of my gas station, thats exactly what I'm doing. None of those textures are tiled but rather large 1:1 baked diffuse and lighting. I was going to go further by adding grunge detailing, taking advantage of all the pixels at my disposal.


Cheers

 

[Vault Dweller]

Thanks for the help, EvoG.

 

[EvoG]

Yea I apologize if I'm overstepping my bounds here.

 

[hiciacit]

Will it help get the attention of people that love these kinds of rpg's but expect a certain level of production value today? Absolutely. Remember, its about effort:reward, and in this particular case, there's really no great reason not to use lighitng as its simple to implement, not a huge addition to your bytesize(assuming you really are that concerned with DL size) and could attract those people that are of a like mind, but have particular expectations.

I agree. And a few 100megs is hardly in issue these days.

 

[The_Pope]

The graphics are good enough that they won't stop me buying it, but I certainly would like some more eye candy. Opening some of the textures in photoshop and doing auto levels would probably help the washed out look a lot, and that isn't exactly something that takes million dollar budgets. By the way, here's a pretty nice screenshot of a Torque engine game from a few years ago

 

[Hazelnut]

A decent game would be broadband users only anyway. One of the indie projects that I know if is 2GB. We are trying to keep AoD size in MBs.

How about making an additional "Texture pack" available for those who want it? It would add to the appeal and if you already have those additional graphics why not? Maybe even charge $3 more for it... :wink:

edit:
What Hazelnut already said.

Hey, we agree - :shock:



I think if two versions are done, then the large one should be developed and released first, then spend a couple of days/weeks building a smaller modem version if you feel you must.. you should premiere with the best gfx you can for reasons given above by others. And yeah, getting a mention after release on PA would be a very good thing for sales!

 

[DarkSign]

They've added the lighting kit as stock to the Torque Game Engine now and even made even MORE rendering improvements to it lately. It's a great choice for an engine.

We're using the Torque Shader Engine for the advanced shaders, multileveled terrain (we can make caves without meshes! yeah!) and a few other things.

VD, this is coming a long nicely. The OP + EvoG were right and it seems you've listened. Good work.

 

[Veracity]

I doubt a few hundred MB is out of the question even for whoever's still using 56K, but a download broken into ~50MB chunks ought to address that, if you think it'd put people off (simply ensuring they can suspend and resume ought to cover it, but breaking it up might be wiser, since people recoiling from a half-gig, or whatever, might not realize that's possible). Distributing via torrent's a thought, too, I guess, but I would think that'd be more to reduce your bandwidth costs than accommodate potential customers with low download speeds - do people on 56K use torrents? It'd give them built-in resume, allowing them to download over however long it takes, but I'm not sure how it'd help with acquisition speed.

As far as I recall, Fate's basic installation download was tens of MB, distributed alongside an optional 100-odd MB download to bump up teh shiny. Very different game, I know, but that seems the cleanest compromise in distribution model to me, if you want to ensure a small required download size without having to cut out things you'd otherwise include.

 

[FrancoTAU]

Poor people (AKA modem users) are going to steal your game anyways.

If you truly care about this, than use WinRar to break down your game into pieces for the dial up folks so they can d/l it piecemeal and put it back together when they finish all the parts.