Pixel Perfect

Hi Sytto So long as your programming skills are reasonable you should have little problem working with LE. Support from the community is good and people are happy to respond to questions. Buying the full LE licence will get you 2.3, the upgrade is just for those already on a previous version. Beauty is in the eye of the beholder they say and I'm sure it’s true. From my personal experience I think Leadwerks is a fantastic engine and in terms of graphics is better than most. You should have no issues developing a large scale game but perhaps you might want to elaborate on exactly what you have in mind, as the devil is often in the detail

I have had no issues with Windows 7 what so ever. If you don't like UAC then turn it off, just don't blame Microsoft if some hacker's process then runs riot on your PC with full privileges Microsoft are damned if they do and damned if they don't. It makes me laugh. The only reason it is continuously under attack by viruses and trojans etc is because it is the most popular OS in the world and hence the biggest target for the sad assed geeks that write this **** when they could be doing something wonderful with that talent!

I find myself nodding and agreeing with everything that Josh stated in his first post. To put everything in place that is needed, even given the huge helping hand we have been given by Josh and all the technology behind LE, is a big undertaking and I believe that is very hard for individuals on their own to achieve. Having said that there will always be exceptions and Naughty Alien appears to be probably the furthest down the line at this point and the best of luck to him! My strategy from the beginning has been to aim towards producing a playable demo of my FPS game idea consisting hopefully of about 1 hours game play with less than AAA quality but hopefully to demonstrate the potential for it to be ... should artists, modellers etc feel inspired enough to join up and form a team. I don't honestly believe with my limited time, resources and skills that I could do much more than this on my own. In answer to Andy's original post I do believe generally that teams are necessary and the members need to be highly motivated individuals working to a well defined development schedule and time scales. I do not believe your average forum type open projects ever really achieve anything much as fun as they might be initially. People lose interest and drift off and inevitably the projects fall apart.

This may seem arcane but it is for the reasons given. Most generalised questions are posted in the General Forum which is open to the general public and as such you are free to search this and ask any questions. They will be answered by either by Leadwerks or the community at large. I think you will find the majority of us very honest in our answers. There is nothing to be gained by the community pushing potential buyers into a product that is unsuitable for them and we would not do so. So feel free to ask.

Yes. Milkshape has a good rep for animating too. I have a copy but I can't say I've ever used it for animation. I'd have used blender too but I found it too dificult to get into although I know its capable of great things!

All of 12 minutes for me. I answered one question and I seem to recollect it rebooted itself several times. Easiest and fastest installation I ever did. Plus it comes with its own disk imaging now (see backup) so once you have it all setup and your critical apps installed burn an image of the boot drive to a spare drive for a quick recovery. Most people recommend you do a clean installation though rather than upgrade from Vista etc.

Take a look at Fragmotion, I find it's a great little package for rigging and animating and is a very reasonable price.

Thats correct, as opposed to frustum culling which is the culling of any object with lies completely outside of the view port. Note: In Leadwerks occlusion culling does not operate on terrain, so placing a hill between you and a load of models will not prevent them from appearing in the rendering pipeline.

[ No, it works for undulating ground like terrain (that is ... a single surface). In the example you give I would use three grids joined by linking waypoints on the stairways. I have the concept of navAreas which are just square areas defined by placing two marker points in my editor representing the diagonal. These can be comprised of A* grids or 'navigation waypoints' (useful where you want to follow paths or roads in large outdoor situations) or a navArea can be comprised of both allowing you to switch from one to the other. The navAreas are linked via one or many 'linking waypoint' pairs. You can have as many navAreas as you want and mix and match the types. In addition to this 'info waypoints' can be placed anywhere in a navArea to store information for AI. So requesting a path from a location in NavArea A to a location in navArea D would result in the AI Manager initially finding a path between the 'linking waypoints' (nodes) to determine the sequence of navAreas traversed and then a path is computed through each navArea in turn. In practice in a fairly dynamic world you will only compute part of the path as by the time you have got that far things may well have changed and a new path request issued

Cool ... so this is done with a frag shader then. Certainly the static pics show nice looking waves although it's a bit hard to get a feel for how good the effect is without seeing some video. Looking forward to seeing more!

I really like the main theme .. nice one !

It does look promising! What's the heightmap actually used to manipulate then if there are no vertices?

I wouldn't rush it. If it's going to take a while longer to get it how you want it that's fine, we are patient people

Yes, micro pather is more flexible in that respect. I use a modified version of Patrick's code to do the same when finding a path through multiple navAreas as that no longer fits a uniform grid model. Your approach sounds interesting too and yes I agree with the ray casting, pointless requesting a pathfind if you can ascertain a direct path is available anyway. Your goal of having it work in the editor and the game sounds great, good luck with that!

If you want fast A* you shouldn't be using oop at all, it will just slow it down to much and take much more memory. Confine your use of OOP to higher levels. No one uses it in anything that really requires high performance. No need to store x, y co-ordinates etc because they are easy to calculate on the fly by virtue of the grid poisition. A simple byte array is really all that's needed to either store a 1 or a 0. The memory foot print would then be about 250K. Take a look at Patrick Lesters code A* path finding for beginners and see the lengths he goes to to ensure he gets fast performance. I've used this technique to good effect but there is a trade off in performance with increasing grid size. I have found it's best not to go much over 100 X 100 (10,000 nodes) and have implemented a system of linking grids to extend the coverage to anything you want. My system also speads the work over multiple game loop iterations which also helps maintain good framerates whilst path finding for lots of NPCs. As Niosop suggests, using nav mesh based systems like Recast is undoubtably better but a lot more involved. Thankfully Mikko Mononen has done virtually all of the really hard work for you but I know it still took Chris about 3 months work to get it integrated properly along with steering. I looked at micro panther which is a more generic A* solver but decided in favour of Patrick's approach which really gave me the speed I was looking for. You just need to look at techniques for keeping the memory usage down and the performance up, and then craft a system that works for you and meets your requirements. Oh, and for anyone else who might be interested, don't ever use Lua for the low level stuff, it will simply kill the performance dead! It's an interpreted scripting language and was never designed for jobs like this! Pic of my navArea Editor showing the grid system and auto calculated non walkable nodes (minimal interface I know ... but it works really well and is easy to use): The system works with varying height terrain too, not just flat surfaces.

I've never used the evaluation version so I can't comment on that but in the full version you simply include your models in the models directory and restart the editor to have it pick them up and offer them for selection. No, things like the texture filenames and UV mappings are all preserved in the Leadwerks gmf format A lot depends on what you tools you are intending to use. Most of the major 3D modelling applications have existing pipelines sorted for conversion to Leadwerks. You are probably better asking for more specific information based on the tools you intend to use. But basically you need to convert your mesh files to gmf format as you're aware and convert any textures used into DDS DXT5 format with mip maps. Having done this, make a material file for each texture in which you can specify things like shaders to be used and various other properties. Also normal maps etc if your using those. In Leadwerks transparency is handled by the alpha channel of the diffuse texture and specular by the alpha channel of the normal map. It seems like a lot to take in initially but you soon get used to it. Hope this helps.

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!

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.

A concise guide like this sounds great. I'm sure a lot of people will be very grateful, thanks Aggror.

Welcome Kevin As Rick said there is a great demand for character modellers and animators here so you'll get plenty of interest from the community. I for one might well be interested in commisioning some characters and weapons judging by your work.

Wouldn't it just be easier to implement a simple one level undo function taking the onus off the designer to actually save repeatedly. Seems to work for most other software vendors so it must have something going for it It's just that added bit of protection for when you go ' oh **** ' I really didn't want to do that

You just have to be perfect Gandi

Ok, well the Editor loads ok and the interface appears to be working fine on my XP 32bit machine but unfortunately I'm not seeing any particles displayed, however I suspect that's maybe down to my old ATI graphics driver on this machine as I can't see any particles in the Editor either when I drag a firepit into the scene. So I suspect this is probably working but maybe you need someone else to confirm this as I don't have time to re-install drivers tonight. Looks very professional though Von, nice job!!! Doesn't seem to handle exceptions too well (an incorrect path will crash it) but I appreciate it's early days yet.

I'd test it on my 32 bit machine but I can't download the latest zip file as Rapidshare says its too busy all the time

Sounds great ... downloading now. Cheers Von