Why OpenGL?

[Marleys Ghost]

Tunnels are working, but terrain_arctic, train, desert are crashing. Thats what i wanted: to have a possibility to load it.

 

Well if the X1950 is pretty much the same as the X1550 I used on a POS test system for another thread, then the crashes when using a terrain are because it can't perform texture lookups in the vertex shader.

[VeTaL]

Interesting thread, thanks.

So, the question of all this thread: is there a "big red button" to disable

the crashes when using a terrain is because it can't perform texture lookups in the vertex shader.

 

I'll be happy if there would be added possibility of rendering terrains without SM 3.

[Marleys Ghost]

Interesting thread, thanks.

So, the question of all this thread: is there a "big red button" to disable

 

 

I'll be happy if there would be added possibility of rendering terrains without SM 3.

 

 

I think (and could be wrong) its also used for skinned meshes

[Josh]

We're going to offer a low-end renderer in the future, but it won't look like what you think of as Leadwerks. It will be just as bad as every other low-end renderer.

[Rick]

We're going to offer a low-end renderer in the future, but it won't look like what you think of as Leadwerks. It will be just as bad as every other low-end renderer

 

I think the people asking for that expect that it won't look as good. I'm guessing it's the programmers who ask for this more than the artists. Programmers use Leadwerks because it's easy to code with. It's very straight forward and requires little to no advanced math. The fact that it looks so good is just a bonus. Personally I wouldn't care if it looked like any other indie engine out there. I've tried almost all of them and Leadwerks is the easiest to program in. That's why I like it and I'm guessing I'm not alone.

[VeTaL]

We're going to offer a low-end renderer in the future, but it won't look like what you think of as Leadwerks. It will be just as bad as every other low-end renderer.

Thats good if you are going to do that Of course, it cant be as beautiful as Leadwerks on modern computer, but...

 

 

Rick, you're absolutely right: after 6 years of work with 3dGamestudio, Leadwerks is really like a candy

[Rick]

I feel like Josh sometimes forgets this. It seems he gets so caught up in the visuals that he forgets that one of the biggest (in my view) pluses that Leadwerks has is that it's easy to program with. That should be part of it's promotion. Get amazing visuals with little to no math and a clean and easy API. The math statement alone would probably get him a ton of people coming over here. Entity programming is 100% easier than straight math. I would recommend any programmer new to 3D to start with Leadwerks because it makes making games fun and not just 1 giant math problem.

[Marleys Ghost]

I think the people asking for that expect that it won't look as good.

 

 

and they will probably be the first to complain about that

[Rick]

I think as long as it's easy to programmatically switch between features people won't complain at all. I think ideally we would be able to query the graphics card and see what kind of features it can support, and then setup the game according to that everyone will be happy.

 

When I was getting into DX a long time ago I know this was how they did it to determine if something had to be software supported or hardware supported. The ability to query the graphics card would be helpful.

[Marleys Ghost]

we will see

[VeTaL]

Yep, very good idea

[Josh]

This was something I conceptually struggled with, but I think the best solution is define statements in the material file:

texture0="scarymonster"

#ifdef LW_RENDERER_OPENGL_1_2
cpuskinning=1
shader="mesh_diffuse.vert","mesh_diffuse.frag"
#endif

#ifdef LW_RENDERER_OPENGL_3_2
texture1="scarymonsterdot3"
shader="mesh_diffuse_bumpmap_skin.vert","mesh_diffuse_bumpmap.frag"
#endif

 

The biggest question is how far back do we want to go? No matter what, someone is going to complain when their GEForce 2 isn't supported.

[Marleys Ghost]

would that mean having to go back and alter all current .mat files?

[Josh]

I think this should be easily adjustable in the model editor, so you don't have to edit text files. You can just choose from a bunch of presets, since we have almost every effect under the sun already. If you want something new, you can either specify a custom shader, or add your own material/shader preset to the list of available presets.

[Rick]

Nobody said it was going to be easy or exciting

 

My personal preference on this would to not support older stuff anymore. Instead but these in now for the current versions. Then when the engine upgrades to new versions just add if defines statements as you go. In a couple years time everyone will be happy and todays tech will be "old", but at least LE will support "old" tech going forward.

 

With each year the complaing of supporting old tech dies a little more.

[Josh]

I think computers on average are getting slower. Laptops are now more popular than desktops, and I think netbooks will replace laptops. Therefore, the average computer's graphics hardware is steadily getting worse, even as the high-end hardware improves geometrically.

 

In three years, I think the average computer will be a netbook sold at Walmart for $100 with an Intel integrated chip, while the average Steam user's machine will have a GEForce 285 or better.

 

So I'd like to offer both the extreme high-end graphics we always have, as well as the really low-end simple stuff.

[Rick]

I think this should be easily adjustable in the model editor, so you don't have to edit text files. You can just choose from a bunch of presets, since we have almost every effect under the sun already. If you want something new, you can either specify a custom shader, or add your own material/shader preset to the list of available presets.

 

This almost sounds like we would have to define what we want to use in the material file? Is that programmable? If these defines can't be done in code then I think that wouldn't be a very robust system.

[Rick]

No, computers on average are getting slower. Laptops are now more popular than desktops, and I think netbooks will replace laptops. Therefore, the average computer's graphics hardware is steadily getting worse, even as the high-end hardware improves geometrically.

 

In three years, I think the average computer will be a netbook sold at Walmart for $100 with an Intel integrated chip.

 

The "average" person buying that $100 "average" netbook isn't worried about playing 3D games. So I think you're fine there. You won't be able to win those people over. It's the "average" 3D gamer that you need to win over. This group group has realistic expectations of what a $100 netbook can do and what a $100 graphics card should be able to run.

 

You sound like that Google exec. "In 3 years desktops will be dead." It's not happening for gamers, and those are the people you are trying to reach.

[Marleys Ghost]

Just seems I spend more time re-doing stuff.

[Josh]

You will just use the editor to select what kind of material you want, and it will copy from a set of presets that already have the defines written for each platform. Or you can do it by hand.

 

The defines aren't code, they are just statements that tell the engine to skip parts of the file when loading.

 

Just seems I spend more time re-doing stuff.

This is all very theoretical right now.

[VeTaL]

Josh, i think, you may just add alternative SM 2.0 version for those features, that use SM 3.0, so they would work on 99% of users (i hope everyone have SM 2.0 card)

[Josh]

I'd design it around whatever Intel's low-end offering supports.

[macklebee]

sheesh... really? kinda defeats the purpose of even using LE if you want to make a game to run on a SM2 card...

 

but whatever, as long as everyone else gets to request ridiculous things, I want LE to run on my old atari 1200... I just cant believe we aren't considering this market... its crazy to leave them out as possible users for my game

[Rick]

You will just use the editor to select what kind of material you want, and it will copy from a set of presets that already have the defines written for each platform. Or you can do it by hand.

 

The defines aren't code, they are just statements that tell the engine to skip parts of the file when loading.

 

Yeah, I guess I was thinking if someone wanted to use 1 feature of version A and not another of the same version. This comes down to more than just the shader model doesn't it? There are other features that OpenGL has in different version right? I'm just thinking in DX where there are different hardware features in the library that you can use depending on the graphics card, which is why they allow you to query the card to see what it's capable of doing.

[Rick]

sheesh... really? kinda defeats the purpose of even using LE if you want to make a game to run on a SM2 card

 

 

Again, I think people get lost in the visuals of LE and that's all they think about. LE is more than just visually appealing. The API is one of the easiest to work with that I (and many others) have ever seen. So by doing this he will attract more programmers. Those programmers who want to make somewhat casual games in 3D but not have the high end requirements. For example as a side project I'm making a collapse game. Does it need to be in 3D? No, but because it is it allows me to do some cool effects much easier than a 2D engine which adds a twist to the game and possible lets it stand out from the other collapse games. For example in my version I have 4 sides of blocks. Each side is like it's own game of collapse but they are all going at the same time. So you must rotate the board constantly to keep destroying blocks. I could use a different engine but LE is so must easier to work with than any other engine I've used before. The market for a game like this isn't the hardcore gamer that has a high end graphics card, it's the casual gamer that needs basic 3D stuff. The more I'm able to do different kinds of games in LE and have a wider audience the more likely I'm staying with LE and not looking anywhere else. The more I do that the better I get at working with LE and teh better games I can make.

 

Just saying, these are valid requests. These casual games are making billions a year also. There is a market for them.