Terrain creation suggestions

[Paul]

Can anyone suggest terrain editing software and textures?

Currently I'm using L3DT standard but having difficulty making the terrain how I want it.

I also need some suggestions on textures, I'm using the ones that came with the leadwerks engine and the texture pack but on my terrain they look an eyesore; I can't get my terrain to look natural like the ones in the Gallery.

[macklebee]

MG and Pixel both have used DEMs to create terrains with good results... but maybe if either would be oh so gracious enuff to give a small tut on how they acheived such good results... hint hint

[Marleys Ghost]

MG and Pixel both have used DEMs to create terrains with good results... but maybe if either would be oh so gracious enuff to give a small tut on how they acheived such good results... hint hint

 

 

Well TBH DEM data is harder to work with than L3DT, and again TBH the results can be very similar.

 

Test Of DEM ( Pixels Test of DEM )

 

Test of L3DT

 

There is a tutorial on the wiki for using L3DT by Rstralberg, I can't sing its praises as I have never used it. His one on creating skybox's in terragen is very concise and easy to understand, so I would imagine the one on L3DT is too.

[Paul]

Test of L3DT

Is this being shown in the leadwerks engine? Did you get L3DT or Leadwerks editor to place the textures?

 

I usually make something in L3DT, export the heightmap to Leadwerks editor and use the terrain paint settings to automatically place the textures. I end up with some stripey or splotchy mess, and the textures don't blend in well together.

[Marleys Ghost]

Is this being shown in the leadwerks engine? Did you get L3DT or Leadwerks editor to place the textures?

 

This was a screenshot taken from the LWE 2.3 editor, with the application of the editors post processing FX. Using two textures, a base layer with no normal and the standard smooth sand texture that comes with the SDK. Leadwerks editor was used to place and scale the textures, L3DT was used to create the Height map which was loaded into a terrain, set to the defaults 1024x1024 2MpT.

 

use the terrain paint settings to automatically place the textures. I end up with some stripey or splotchy mess, and the textures don't blend in well together.

 

I rarely use automatic placement of the textures, its ok for the base or to get a "rough" height of say, sand on a beach but you can't beat a bit of good old fashioned "hand" painting.

[Paul]

One problem is the scaling, too low and the texture looks unnaturally smooth in the distance, too high and its pixelated at close range.

[Pixel Perfect]

I tend to agree with MG, unless you specifically want to replicate a known bit of terrain then you are probably better using tools like L3DT than using DEM. With regard to texturing ... using the right textures, along with scaling of the textures, is really important. Never more so when applying rock type textures. I invariably end up scaling them right up to get the effects i'm looking for. Try varying all the settings, experiment and learn from the visual feedback you get. You'll soon get the hang of it and your terrains will start to look a lot better.

[Marleys Ghost]

One problem is the scaling, too low and the texture looks unnaturally smooth in the distance, too high and its pixelated at close range.

 

 

Requires a balance "of sorts" to be struck and texture resoltuon and mipmap fading can help.

[Pixel Perfect]

One problem is the scaling, too low and the texture looks unnaturally smooth in the distance, too high and its pixelated at close range.

The unnatuarally smooth in the distance problem might be caused by the fact that the Leadwerks Texture Pack uses a technique that fades the texture to the average color over the mip maps giving the effect of no texture to them at distance, this helps get rid of obvious repeating texture patterns but to my taste does make it look unnatural. Try high resolution textures avoiding this technique to help with that!

[Marleys Ghost]

The unnatuarally smooth in the distance problem might be caused by the fact that the Leadwerks Texture Pack uses a technique that fades the texture to the average color over the mip maps giving the effect of no texture to them at distance, this helps get rid of obvious repeating texture patterns but to my taste does make it look unnatural. Try high resolution textures avoiding this technique to help with that!

 

 

Which is also true, and again a balance needs to be struck, low scaling less pixelation but more visible tileing, I use textures that do and do not have mipmap fading but as you said Pixel, its best to try varying all the settings, experiment and learn from the visual feedback.

[Paul]

Requires a balance "of sorts" to be struck and texture resoltuon and mipmap fading can help.

I'm trying to go for a FPS with an open area kind of style, my biggest problem is with forest_grass which I think is from the textures pack. The terrain is 1024x1024 1m.

[Marleys Ghost]

I'm trying to go for a FPS with an open area kind of style, my biggest problem is with forest_grass which I think is from the textures pack. The terrain is 1024x1024 1m.

 

 

Well how much "detail" with regard to "grass" would you see anyway at a distance?

 

I use that very texture in my test scene. forest_grass is from the textures pack and is (I believe) mipmap faded.

 

[Canardia]

It seems your 3rd person cam is missing freelook mode (usually activated by hitting the middle mouse button once), you can copy it from gamelib

[Marleys Ghost]

It seems your 3rd person cam is missing freelook mode (usually activated by hitting the middle mouse button once), you can copy it from gamelib

 

 

Its not actually missing anything it has exactly what it was programmed to have, and I am busy writing page numbers and dont have time to debug gamelib code as well

[Canardia]

But still you have time to rewrite gamelib from scratch

I know, it lacks of a updated BlitzMax version, and I can only blame the people who use BlitzMax for that, and don't convert all C++ libraries to BlitzMax.

My dream is to have a simplistic and efficient programming interface for LE, which works in all languages.

[Marleys Ghost]

But still you have time to rewrite gamelib from scratch

 

not sure I follow that ?

[Canardia]

Well, what I see often in the forums is that people ask questions and write code which already has been standardized. I would rather use the standardized code and help to improve it, than rewrite the same code over and over again each time.

[macklebee]

I think its more than some people just dont think we need to boost your ego to ridiculous proportions by using that mess you call gamelib. And if I remember correctly, you were the one that dropped support for gamelib. Why would any of us update your gamelib for bmax? Your constant advertisement and answer to everything is to use gamelib. You didn't invent code lumooja so stop acting like anything anyone else contributes is not up to par with gamelib.

[Marleys Ghost]

Well, what I see often in the forums is that people ask questions and write code which already has been standardized. I would rather use the standardized code and help to improve it, than rewrite the same code over and over again each time.

 

 

Maybe in writing the code comes the gift of being able to understand it and then go on to write ones own code to ones own standard or needs, maybe not everyone wants to use standardized code?

 

but I am not sure Pauls thread on "terrain creation suggestions" is the place to have such a conversation.

[Canardia]

Yes, I dropped support for BlitzMax, since it's a community project anyway, and I don't have time to support all languages. Nobody needs to use gamelib, but I think it's just a waste of time and resources to reinvent the wheel each time again. If you make a better library, go ahead, then I can stop developing gamelib also and use the better library. But so far I haven't seen anything better.

I just want to use libraries which can be reutilized, to avoid rewriting the same code over and over again, for each person. And I don't understand why some people don't want to do that also.

[Marleys Ghost]

I should have added Paul the Forest_grass texture was scaled to 3.0 and used as the base texture.

[macklebee]

EDIT- MG's right, Paul's post is not the place to discuss this.

[Paul]

Well how much "detail" with regard to "grass" would you see anyway at a distance?

Just the smoothness and the darkish color looked out of place, then again I didn't have any post processing FX.

[Rick]

My main reason now for not using gamelib is because Thingoids is all the craze and those are what I'm making. Before 2.3 I did use gamelib, but I wasn't a fan of the structure. Honestly, just because you have the code that doesn't things doesn't mean people will like using it. I think you put more time into functionality than making the code enjoyable, structured, and scalable.

 

Lumooja, the reason gamelib won't make it anywhere is because of how stubborn you are. People have time and time again said how messy and NON STANDARD (there is no standard for a game library so please stop calling yours standard) it is, yet you don't listen to them. If you want people to use your library then you must listen to the people, but you have this idea in your head of how the world of coding should be and almost no one shares that view.

 

So now I'll give my advise. Make it OO. Use classes instead of structs. Don't make every single variable public. Get a class diagram going so people can visually see the components to it. Make each component as independent as possible. Make it so if I need to use a Tag class to put name tags over my characters I can easily use that class and it works. Every class should be like that. Every class should allow to be hooked by other classes you make OR from plan LE code, but this needs to be more eloquent than just making every class variable public.

 

The readability and usability of a library is more important than the functionality. If no one uses it because they have a hard time figuring out how to use it, then it's worthless to them.

 

[EDIT]

EDIT- MG's right, Paul's post is not the place to discuss this.

 

My bad. I was writing my post when you did this. We should discuss this though because I think gamelib is a good idea just done incorrectly.

[Marleys Ghost]

My bad. I was writing my post when you did this.

 

20 hours to write a post?