wraitii

This is not true if you specify it separately in the JSon, you can just sort by group name directly.

I'd do the opposite, like Brythonic (Celts) Gauls (Celts) Ptolemaic (Egyptians) Spartans (Greek) Athenians (Greek) In the end we could have a selection screen kind of like AoM.

This isn't too complicated, and I would even suggest a last "billboard" level of detail, but it does require some technical changes. Since instancing isn't showing too much promise, I might try LOD first and see if instancing helps after. This is kind of technical, but OpenGL supports far more than "tiling", such as "mirror", "mirror on X axis", "repeat", "clamp", and is probably a better option. I don't get what you mean by "blend" and "order". See above. OK now I get what you mean... I think this should be handled in the material and the shader itself. You'd supply a "snowTex" and pass some sort of define "USE_SNOW" and it would load a custom shader. Perhaps for this we ought to multiply the number of shader files rather than just add techniques. It's up for discussion. I'm not sure what you mean here... There are actually three issues on reflections: what to reflect (skybox, environment map, real?), where to reflect and how (fresnel based or not). This could be added on top of the current shader code or in a separate file but that would need some reorganization (which I'd like to do anyway). Done tomorrow if you want. Per-model or per actor basis? Also not sure why you'd want that. Not urgent in my opinion and probably annoying to code. I don't think we'd need that for a translucent shader. Maybe. I had never thought of that but it should actually be codable without too many tricks.

Thanks Enrique. Updated it with bigger eye zones so you can make them out at normal zoom (doesn't matter beyond that). Redid the animation, I think it actually looks pretty good now but I've got some looping issues now and then. Posting the current .Blend for comments. Hyena.zip

Kicking Bird, for someone who claims to have known the game for awhile, you seem awfully misinformed. 0 A.D. simulates the time period circa 1 A.D. to 500 A.D. The concept of "phases" simulates the increasing resources a booming city will have. This is, to take a completely unrelated example, similar to how Cities: Skyline handles cities: you start with limited stuff and a small map and as you grow you have more "resources" (in a very general sense) and you have access to more features and a bigger map. We want our art to be as close to history as possible, not the gameplay itself, or we'd have to start coding in some latrines. I think re-evaluating what you are doing here would indeed be a good idea.

I utterly disagree with you that 3D stairs are a waste of memory. It's been proven time and time again that 2D stairs using normal maps look terrible. About the persian set not being instantly recognizable: you're taking huge shortcuts here, particularly as the silhouettes tend to be on desert ground and thus not that noticeable. The forum is instantly recognizable because of its center alley. The market is because it's got the most tents. The temple is tall and has all the columns. The only ones where there actually are issues are the fortress/special building/that other blocky one which I don't know of. Houses are voluntarily indistinguishable between themselves. I think your house model looks pretty good but looks too modern. This is not Tangiers. And unless we have some historical proofs for wooden roofs, we won't do it.

Well everything is feasible, the question is "how" and "can it be fast enough?". I'm not an expert in reflections and I don't think we can use what we do for water here, so I'd have to look into it. Could be driven by anything really. I was thinking alpha of the specular map or the normal map. Looking pretty good! It could be slightly improved though, I'll be working on that a bit. Do you use A18 or the SVN version? For your latest post: it is certainly possible to code a new component that does that, I'm not too sure about existing ones.

The obvious thing with bushes is that they're generally the exact same color as everything around them. Except in some rare cases: One solution would be to make them much darker. Another would be to have much redder, bigger berries (or berry packs) as currently this red color isn't very visible which really doesn't help (I mean bushes will generally have to remain green). Lighter bushes would make the green stand out more, or quite darker ones. We need to care about R/G daltonism though so playing on contrast would be important too here. My personal opinion is that it's actually a mesh problem as currently they're just round blobs and the edges really blend a lot.

That seems like a good idea.

We are using fixed-point numbers for anything that requires synchronisation, which fixes the issue.

Based off this: Hyenas are weird. I don't claim it's perfect, but it didn't look too bad to me. I'll check again tomorrow to see if I still like it.

What I got fiddling around: It's kind of a hack right now but if you like the effect I can make it work better and put it in-game after A18 is released. Current system can tint the skybox so it looks perhaps more realistic (could probably be made to only color rays that look towards the ground). I also use an optional fresnel so you can make it fully chrome or only reflective at sharper angles. Edit: yup. Can probably look right on maps where terrain won't change color much.

I changed the texture to be much brighter, more contrasted and use bigger spots so they are visible from farther away (also tweaked the model UV). I think it looks much better that way. I think I'll have to make the eyes bigger too. I've got some animations which I think are acceptable now, will work on them some more.

Except for the part where this is not a work environment but a public forum and work ethics have no place here. And that was an offensive comment. That was an offensive comment because it assumed that as a programmer I don't know @#$% about art, design, and things like that, and you didn't even have the guts to say it straight. You were even assuming I was only a programmer, which I'm really not. Now what you said there ^ is honest, literal, and likely true in general. I have no quarrel with that. Let's be straightforward. Also if you think no-one has noticed that our terrains were burned out or too bright, that our textures were sometimes ugly and hard to read, you're delusional. We just never got around to fix it. I personally think someone ought to redo all ground textures and make them coherent between biomes, but it's a huge task. That's what we aim for. And I'm not saying AOEO has a bad art direction or is incoherent or anything. It's great in that aspect. It's just that they chose the worst possible direction, in my opinion, and taking context into account everything is ugly about it. It's an extreme example of something very very bad done very very well. And well, the market agreed. And about those older terrains: dunno what happened, but if they're no longer here, it's probably because we lost the right to use them. I agree that the parallax grass doesn't really look good though. It was a proof of concept, but it's quite meh.

Yeah, I agree that the texture is too blurry right now, I'll work on it.

Undying Nephalim: I'll give it a shot this afternoon, I know the shader code fairly well and you ought to be able to do something cool.

That's not a game, it's a 3D render. Anno 1404 has a vibe that is quite similar to that, though it it obviously way less good looking.

I actually find that terribly hard on the eyes (not to mention ugly) ^

You're right that terrain Blending in 0 A.D. is an issue, and generally our terrain is not the nicest part of the game. But it turns out that changing it sensibly is not really the easiest thing either, sadly. Our units are slightly difficult to distinguish. This comes sometimes from old models, old textures, similar animations, small props, and also the fact that textures can become quite blurry at a regular zoom level.

It's obviously a high-quality 3D render. The wave foam is the worst thing on that screenshot. Replicating the effect is still basically impossible on current GPUs, except maybe for technical demos on very-high-end GPUS and using a ton of work beforehand (for light maps and so on).

Your example is quite extreme: that specific grass texture is really horrible. Many other textures are actually quite nice. Furthermore, the recognition issues are not caused only by terrain (and well, noisy terrain helps greatly hiding the low quality of some of our meshes). I personally dislike too uniform textures too as they look like a British golf course lawn, which is just bad. I do agree that several of our terrain textures are somewhat ugly, too cartoonish, sometimes of varying quality. Reducing their contrast or saturation might help on this. Changing some will be needed. However, your plank example is ridiculous: on the left you get voluntarily noisy wood, for specific uses only, and on the right you have burned out, washed out planks. Both are legit wood textures and have their uses. Finally: This will be your one and only warning: please refrain from further comments of that kind.

A quick pointer: if you have to preface your argument with "I don't mean to offend anyone", it's because you know very well that you're being rude and offensive. In which situation the sensible thing to do is to calm down and write your thoughts down in a concise, efficient, argumentative manner. People will not take you seriously otherwise. Indeed I have no idea what you are trying to say with this post except perhaps "this game is ugly and you guys suck" which I am quite certain is not constructive.

We have skyboxes which the engine can send to a model (check the technique for model_water in public/shaders/effects), you could use that to give an effect sort of chrome-y (it would require a new shader though but I'm sure something can be thought of). We could somewhat easily support some basic environment map to give a better effect though.

Basically you mean what you want to do is LOD, not new models? Again, read on what we want to do with this game and you will understand why it is open-source. The aim is not to finish it quickly, the aim is to make a quality, free, open-source game. So yeah, we're staying open-source. The project also has definitely not grinded to a halt, in fact in the last 5 years it's been rebooted from scratch. We tried, and failed, crowd funding.

In-game. I've re-changed my mesh to fit my animations using IK, it's mostly working now so I started working on some animations and will probably commit it sometimes next week once I'm satisfied with them and have touched my textures up again.