Burned Models

[Darkness_Falls]

Hm, now I'm confused. On my machine, the flame looks... well, the edges of the texture show. Does anyone else see that? Maybe it's my resolution? (640) I'd assumed it was a placeholder till we had some particle fire better than D3's.

I was also going to bring this up if it didn't get fixed over time; as I also thought it was a placeholder flame. There is a noticeable square'ish shape going around the candle flame. I'm on 800 x 600 resolution w/ Doom 3. You can even see this square'ish shape on the candle in the recent readables image that got posted on thedarkmod.com. I was surprised nobody commented on it.

[Springheel]

Hmm, a burnt skin doesn't work very well after all. The model is too crisp and sharp for a burnt texture to look believable--just comes across looking like a black-coloured book.

 

So maybe some kind of charred book model is the way to go...although again, it won't really match every book model. I wonder if it's even needed? We could just use the monster gib effect to have the book 'burn' up and disappear if it's thrown in a fire.

[Ishtvan]

I think I posted this in the other thread, but I can't find it now, maybe just going crazy from lack of sleep, but:

 

If the skin doesn't work that well, what about a charring decal like what currently happens when you shoot stuff with the fire arrow?

 

Are you doing a normal map with these skins? Maybe a normal map would help convey the illusion of a burned book better, because it could have random height variations, indentations where it charred all the way through the cover, etc. I'm just wondering because you said it looked unconvincing and you wanted to do a model that's more eaten up, maybe we could do what we do for our other models, create a high-poly version and use that to create a normal map?

[Springheel]

It doesn't currently have a normalmap, but I don't think that will help. It's more to do with how sharp and crisp the corners and edges are. It just looks like a perfectly intact, black book. I'm not sure a decal would help either.

 

The more I think about it, the less I'm sure this is even worth doing. What's the point of having ONE object able to burn if none of the rest of them are?

 

Maybe you're right, Ish. A decal or series of decals applied to an object might be the most cost-effective way of doing this...could a system be created that applies charred decals if the object comes into contact with fire? It won't look believable, but it will be something, and it won't require the creation of three dozen new models.

[sparhawk]

The more I think about it, the less I'm sure this is even worth doing. What's the point of having ONE object able to burn if none of the rest of them are?

 

Whats the point of making this mod if we can't get everything perfect?

[Springheel]

There should be some consistancy, however. If a player throws one book into the fire and it burns, but he throws another book into the fire and it doesn't, then that's just confusing.

 

When things don't work like a player would expect them to in real life, they should at least be consistant.

 

It might be ok to have inconsistancy if it were a very important effect, but I don't see how this is very necessary. Unless the player is just messing around, how often do they try to set books on fire?

[Ishtvan]

Hmm, so the edges need to look less intact. What about a skin that applies some alpha transparency at the edges and corners to make it look like chunks are missing? Or if you don't want it to be completely missing, but want it to look like a hole that goes down to a deeper layer, we could put two texture layersi n the model, one below the other, and alpha out the top one to create holes. Don't know if that would look right though.

 

As for whether it's worth doing.. well, we might be gettin a bit carried away. We don't have models for the water arrow, fire arrow, gas arrow, noisemaker arrow, etc, and we're thinking about making new models to support the player sitting there and throwing stuff in fire to see what happens. This would be a pretty low priority compared to existing stuff that we don't have modeled yet.

[sparhawk]

There should be some consistancy, however. If a player throws one book into the fire and it burns, but he throws another book into the fire and it doesn't, then that's just confusing.

 

Well, that's a problem of the mapper I would say. And I already explained that we don't need individual bruned versions for every book. If we have some extremly charred models, which are not to be identified as individual books, and a few charred models which ARE identfiable, you can buncht them together and make it believable without so much effort.

 

It might be ok to have inconsistancy if it were a very important effect, but I don't see how this is very necessary. Unless the player is just messing around, how often do they try to set books on fire?

 

Nothing is neccessary. In fact we can stop right here, because the whole mod is not neccessary. We are doing it for fun, not because we need to achieve or proof some important point. Making the gameworld more immersive is always worthwhile, especially when this can be achieved rather cheaply. Also it is worth, because this way we can show what our system can do if properly used. We may not only consider how something is used in an actual map, we should also consider examples that shows mappers what they really can do, which is much more than with TDS or even with T1/T2. Everybody knows that the books in the game don't burn, but we can easily show that we can create such an effect rather easy. The bruning of books is not neccessarily the primary intent, but that we can do much more is.

[Springheel]

Nothing is neccessary. In fact we can stop right here, because the whole mod is not neccessary.

 

That's the kind of exaggeration that really moves discussions forward.

 

We've already agreed that creating books that have pages that turn in the wind is not worthwhile, even though it would be very cool and realistic. How is this different?

[god_is_my_goldfish]

Because this is a hell of alot easier to do. All you're doing is creating a new texture for the old models, add a bit of code that says fire='blend and change skin/object' (probably doable within stim/response), and bam, burning books.

 

Pages turning in the wind would require a hell of a lot more work to do.

[Springheel]

A new texture won't look like a burnt object though. I tried making one for the books and it just looked like a black book. It would have to be a separate model. And if we're basically talking about some kind of 'pile of ash' model that can be reused for any kind of burned object, then alright, that -is- easy. But if we want to make a burned version of all of our models, that's a lot of work.

[SneaksieDave]

Gotta leave something for the community to do as add on functionality. Should it really be considered important for core, base func? Especially at this stage?

[New Horizon]

Could the model be made to look more...'chunky'...with the use of alpha transparency to create the illusion of parts of the book missing? You could likely blend that as well.

[Springheel]

I don't know. The transparencies I tried didn't work that well, as you then were left with a hole that you could look through to the inside of the model. Maybe someone who knows more about alpha blends could come up with something. I agree with the others that this is really low on the priority list, though.

[sparhawk]

That's the kind of exaggeration that really moves discussions forward.

 

Just like defeatism and extreme conservativity ...

[sparhawk]

I don't know. The transparencies I tried didn't work that well, as you then were left with a hole that you could look through to the inside of the model. Maybe someone who knows more about alpha blends could come up with something. I agree with the others that this is really low on the priority list, though.

 

I could accept if you say you can't do it, but saying that it's not needed is really a crap argument.

[New Horizon]

I would be happy if the books looked scroched up with decals or something. Perhaps if they burn for an excessive amount of time, a generic charred up, lumpy burned book model could be swapped in? You could have one for different thicknesses....wouldn't have to match exactly, since the books edges would be burned away pretty badly.

[SneaksieDave]

Yeah, decals seems the most obvious and easy choice - that way, anything can be given a 'burnable' property and they would become charred. It's been done in games already.

[god_is_my_goldfish]

I would be fine with that, but can we add a normal or specular map to the decals? It would be nice if burnt things had that nice rough and semi-shiny surface that you get when carbon-based things get burnt.

[sparhawk]

Decals are going away after some time, right?

[SneaksieDave]

Good point, not sure. But I think corpses retain their damage areas, so it should be the same, I'd guess?

[sparhawk]

I don't know, because decals add up over time. That's IMO why they go away after some time in most games (like bulletholes, bloodsprays, ect.).

[Ishtvan]

I just tested it, and decals on models do not go away over time. However, there's a limit to the total amount of decals you can put on the model. When you go above that limit, as you add another one, it removes the first one added. So it is definitely optimized to limit the number of decals on an entity.

 

You've probably seen this in games like Far Cry where if you keep shooting or knifing an AI, the previous marks from damage start to go away.

 

You might think this would be okay for burning, but there is a problem: Say the player faces one side of the book to a fire, drops it at the very edge of the fire's effect, and the book starts to burn, but only on that side. Now that side is charred. If the player were to pick up the book, turn it around and drop it so that the other side is in the fire, the other side will burn and get decals on it, but because of the decal optimization, the side that was previously burned will get the decals removed and magically "un-burn." Not good.

[Springheel]

could accept if you say you can't do it, but saying that it's not needed is really a crap argument.

 

Asking whether or not something is important is totally valid, and we do it all the time when someone comes up with a new 'cool' idea for making the world more immersive. I don't see any difference here. Everyone who has commented on it agrees that this is, at the very least, a very low priority thing.

 

If you want to make it your pet project, that's great, but then you don't need anyone to make a model for you. If you can model a piano stool, you could certainly have modelled a generic charred object by now.

[sparhawk]

Asking whether or not something is important is totally valid, and we do it all the time when someone comes up with a new 'cool' idea for making the world more immersive. I don't see any difference here. Everyone who has commented on it agrees that this is, at the very least, a very low priority thing.

 

Apoparently you don't get it, that such features are not only there to make the world more 'cool' it should also demostrate to the mappers what they can do. This is quite different from feature crepp, which we should avoid, because this is NOT a feature creep request, it can already be done and I simply need some models to show it.

 

If you want to make it your pet project, that's great, but then you don't need anyone to make a model for you. If you can model a piano stool, you could certainly have modelled a generic charred object by now.

 

I can model moderately well, but I can't texture and this is primarily about texturing, not about modelling.