Epifire

Ahhh, see I was combining both old and new methods accidentally. I don't think it causes any issues directly as I they all seem to work without any hitches. For some reason I was under the impression the description line was used for no more than an actual description (not the actual material type declaration). So what I'm seeing is you need to list the material type as the first word in the description? Cause it looks like that and whatever else you want to add after. That's where my confusion was at.

Well I'm getting to the part where I get to tear out the walls and bust things up so here's the random brick debris of the day!

This should be a bit better... https://www.mediafire.com/folder/7q1r8him1y63d/TDM%20Repository Also I put my sig to better use and just have that there instead. New stuff when in full working order will be updated into those same files. With the exception of anything that was built for a specific user (special secret project-ish things).

Oh does it not let you download the whole folder? Hmm a zip should work right?

I could probably do this in a different thread but random people had been asking for my machines I've made in TDM so far. This should be all the available stuff, PM me if there's anything wrong with it. Cheers! http://www.mediafire.com/folder/h27iw21t822we EDIT: I may have a separate thread later that showcases some of the stuff but this is my main file dump for now. New stuff will be updated into the same link!

Well as long as the lightmaps weren't overlapping to active fires that could be put out, each light could have it's on/off state tied to a toggling skin too. This tells me a lot, but just out of curiosity what was your lightmap resolution set at? My whole purpose I want to do is be able to throttle the amount of darkness, back beyond the ambient_light mainly for dark (barely accessible) areas of the map. Primarily corners and low spots that are too far removed from light sources to be receiving much light. What it should end up doing is give me two dark thresholds of light. The darkest being the lightmaps that blend into the ambient world light. It would also help prioritize more important areas of the map to lead the player by light.

The optimization is my biggest concern. Even if this was made of actual brushwork the big problem is that it's all open. You hardly have any visual cutoff from any angle (stairs or the central elevator shaft). With even two lights running shadows I notice a bit of a hit on my frames. Shadow meshes and disabling different portions that don't really need to be taken into account will help I think (certainly did with my other models I made). There is space between the floors as they were built to snap together in XSI (our other name for Softimage) so there is room for BSP to be added in between. My current plans are to demo this space for a very specific scenario so I don't plan on having any NPCs running around as of yet, but monsterclip wouldn't be too hard to add. If you make stuff around one floor it can be duplicated up to another with ease, since each floor was actually modified from a single floor to start with. Here's my workflow. Build the basic blockout (using precise unit measurements) and see how I like the scale of the mesh in-game. Then I start making my tiling textures. Both the stone floor and brick were setup in Softimage to tile correctly, sculpted in Mudbox and then baked back down to a 4k texture in Soft. All the areas I require to tile are ones that will be taken into account for the massive texture painting pass. Once the sections are complete I will split different levels into groups that are good enough to fill 8k worth of UV space without compromise. Here's the interesting part. Once all the areas are set to be painted, all the floors will be loaded into Substance Painter. SP allows for multiple meshes to be painted at once with their own individual maps. Norm, Diff, and Spec all included with it's export. There I can blend stuff, add dirt, colors and more detailed destruction! Now you may be wondering, how am I going to tile the maps? They will actually be loaded into SP and then tiled (not before) since we can't have the classic overlapping UVs as we ordinarily would. Now hopefully between four 8k texture sheets there shouldn't be any noticeable detail loss. Only time will tell. Do indulge me with this information (maybe in PM if it's fairly lengthy). Currently you're looking at two meshes in that whole map since I'm just testing the layout, but I was betting that would change down the road. Alsssso there's a whole lab section I'm building that you all haven't seen yet. But I planned on having that around a corner so I could run it in it's own VisLeaf for sure. That should be at the bottom of the shaft but more on that later!

Random question as always guys. Is there a recommended blend mode to be used for custom lightmaps? Yes this does mean I'm making considerations for manually setup and baked lightmaps. Not sure TDM would even look like it ran on the same engine were it done right. Wouldn't even have to be at a high resolution, but it would look mighty fancy. Depending on the target hardware the constraints are considerable. That and just how easy it is to mimic my light values and positions from DR back into Softimage. My modeling app supports a rendermap feature that allows me to save lighting data onto a texture (either with or without the diffuse). Also something on that same line, would there be a blend function that could be disabled or culled out by direct light sources? Manual lightmaps would be super if we didn't have moving lights like the lantern or NPCs with torches. Otherwise pointing a dynamic light at a dimly lit area would still be dim, even with the lights. Just some more crazy Epi-thoughts of the night.

Sure thing! I'd love to give a good and in depth article on the process. Depending on how well Substance Painter likes my massive (paint the whole friggin level in one shot) method, than I should have no problems for the last step. What's great is that if you have a good hand for modeling it's much more freeing to not have to worry about normal BSP mapping constraints. I can't tell you how much I've tweaked things and pulled stuff around to fit for this project. It's a pretty organic kind of structure with practically nothing being grid snapped (which is infuriating to work on at first). Such as the the major brickwork. It's based off a large brick sheet I sculpted out into a 4k texture for generous resolution. Nice thing is it's big enough that I could stretch the thing out a lot to minimize excessive tiling, as it's used nearly everywhere. I'll be able to paint on top of it for dirt, moisture, and buildup details to break it up even more. What's even cooler is any UV edges I can mask even more with Substance Painter on the final detailing pass. Really excited!

Okay so I told myself I wouldn't post any WIP shots but... oh well. No brushwork is present either. All my geometry and spaces have been planned out in XSI. It's still pretty rough, but the overall form is there. Enough that it was at least worthy enough for some lousy screenshots. Been pretty excited though. Started it about a week back, and it's come a long way in a short time. All highres details, no actual texturing has happened yet. Each floor will have a bit different details and layout. Can't wait to show you guys what the scenario will be, as I actually have a pretty specific plan for what gameplay will be taking place in it. As some of you already know, I'll be taking into account the megatexture theory for this run. All major surfaces (with tiling textures) such as floor, ceiling and walls will be detailed individually. Gonna see how far I can push the texture limits for this one. Still so much more detailing to do!

Greetings Pete! I ask a lot of questions as it's my many questions thread. Currently I've been trying to arm myself with various techniques to tackle different scenarios. Parallax mapping isn't the end of the world, it would just give me one more thing to bring out depth a bit more. So it's not a necessity, it would just be the icing on the cake. However for Megatextures that's a different story. Laying down your baseline tiling textures is a start but lack much realism with their edging. Vertex Colors are a step in the right direction as it gives much more customization to how the materials are blended together (very good for exteriors using large spaces). Megatextures are obviously the most detailed and involved of the two choices. It's more time consuming but provides the best results as it gives the ultimate level of control. I have plans currently to map into 8k textures and section parts of the map accordingly. It's a slightly less mega version but I'm willing to try it out and see what we get. Now you may be wondering why split it down? Well first I can't export a map higher than 8k with Substance Painter. I may also be able to combine four 8k maps into a single large texture as that should be enough pixel space for my current map. Substance Painter gives me a great deal of tools to really detail my meshes. So with the major walls and floors paint-able, it's crazy but it should look pretty awesome when finished. The best part is once I figured out my workflow it's actually not too hard. There's a lot of meshes that wont be in the, "megatextures". Smaller structural details that aren't tiled will have their own texture as usual. But when it comes to the unified painting method for larger tiled areas, the larger the better. It's just easier since it's less divided. Hope that answers some of your questions. Hopefully you'll be seeing some of my work on this next project shortly!

Well that's okay, I just figured since I had some height maps handy I'd put the question out there. Switching topics (again xD) I have another one. After running some tests I found out that Vertex Paints work marvelous when using Lightwave 2015 that is... I would love to be able to spare the cash right now to use it, but I can't justify dishing out that much just to paint in my materials. So that leaves me trying to find a demo (modtools) version to do my bidding. It's a shame all I want it for is a simple Vertex Paint, but it does exactly what I need. That being said, does anybody know of any backwater (old) copies out there available? Even if it's something much cheaper to purchase outright I'd consider, since all I really want to do is this final paint step. I know we have quite a few people here who've used it for a long time, so I figured it was worth a shot!

Doing up some rather large scale level textures to go into this new level and I gotta ask... does id4 support parallax effects of any kind? I have some good heightmaps to use (if it does) but I don't recall it being talked about. My brick wall I'm working on would make great usage of that if it exists in the system.

Some interesting reading Vanished one. Though sadly I can't get any of the links for the tool to work. But as I can see id4 can import huge texture resolutions correct? If so I may do some smaller (slightly less mega) textures for unique building sections. I think I may try to demo some of these themes into my new level I have in mind. Let's just say to start with I'll need to create a lot of stonework and structural details from scratch. The theory is extensive so I'm not sure how well this will work but I have to try. The goal is to create mesh sections as I always do, highpoly with lowres baked finals. 1: The repetitive pieces shall be created from modular assets, then positioned in XSI to their respective places and all saved with their unique world positions for a final export scene. Then I'll load each connected space (within reason as there's a considerable performance hit) into Substance Painter. There I'll work my magic and do all the detailing I need. 2: Once finished with all the required sections I will then export (with as large and compact sections as possible in separate maps) depending on the size of the space I will do for TDM. Here's the tricky part. If my tools are exact enough I'll go back into the final versions of each level section in XSI and move the whole UV sheet to only take up a portion of the final sheet. For instance if I have four 8k textures, I would therefore split the UV space into fourths. Each mesh section would occupy one of the four sections. 3: The last step would be to combine my four texture sheets into a single 16k image in Photoshop. Not sure what a 16k DDS texture would look like but any kind of file compression would probably help this instance immensely for file sizes. For a whole unified level structure I think it can work, permitting the engine can load larger textures with no problems. Now for my current project I may only section out floors for this usage and try to keep the rest as individual assets. This is largely due in part to my lack in tools to batch process/bake my whole level into a single sheet.

Some interesting reading Vanished one. Though sadly I can't get any of the links for the tool to work. But as I can see id4 can import huge texture resolutions correct? If so I may do some smaller (slightly less mega) textures for unique building sections. I think I may try to demo some of these themes into my new level I have in mind. Let's just say to start with I'll need to create a lot of stonework and structural details from scratch. The theory is extensive so I'm not sure how well this will work but I have to try. The goal is to create mesh sections as I always do, highpoly with lowres baked finals. 1: The repetitive pieces shall be created from modular assets, then positioned in XSI to their respective places and all saved with their unique world positions for a final export scene. Then I'll load each connected space (within reason as there's a considerable performance hit) into Substance Painter. There I'll work my magic and do all the detailing I need. 2: Once finished with all the required sections I will then export (with as large and compact sections as possible in separate maps) depending on the size of the space I will do for TDM. Here's the tricky part. If my tools are exact enough I'll go back into the final versions of each level section in XSI and move the whole UV sheet to only take up a portion of the final sheet. For instance if I have four 8k textures, I would therefore split the UV space into fourths. Each mesh section would occupy one of the four sections. 3: The last step would be to combine my four texture sheets into a single 16k image in Photoshop. Not sure what a 16k DDS texture would look like but any kind of file compression would probably help this instance immensely for file sizes. For a whole unified level structure I think it can work, permitting the engine can load larger textures with no problems. Now for my current project I may only section out floors for this usage and try to keep the rest as individual assets. This is largely due in part to my lack in tools to batch process/bake my whole level into a single sheet.

Just from what I see it seems rather out there even for a mod. But I'm actually really skeptical of mods basing themselves as a sequel. Part of me really understands where Valve is coming from but I can't help but feel the pain from either a fan or developer's perspective. The main vibes I get is that they've been waiting on developing new tech to base it on. But also are on the fence about a it altogether. The thing is most people expect something more grand than even Valve can pull off and as I see it, they dug this hole themselves. Even if a sequel wasn't received as well, I just want to see them roll with the Half-Life universe and start pumping out some more games. I'm not near as hard to please as long as it's good content. In all honesty I just want to see a new HL game from it's parent company, with all the care and devotion of the original titles. Least to say it's touching to see a fan runner up, so I guess I'd wait to see what people say about it.

This brings up a lot of different points. What I think of in reference of what I want to do is what you have in the Unreal for Vertex Paints. Pretty awesome results, but I know we're much more limited in what we can do. I think what I'll try to do is take it back a bit and not over-complicate my methods. Can achieve some pretty good results with some more strict methods but fleshing all this out has done me a lot of good to think on. The thing is I'm considering some pretty over the top ideas along the lines of the mega texture theory. Can you make a terrain mesh and have it use a single mega texture only? I don't know how that is implemented within id4 (but some have been saying it's possible). I'm not fully convinced of using it but I'm interested in using smaller (less mega) texture sizes for smaller but intricate outdoor scenes. The production of such a texture would be the interesting part, as I don't think I have any software exactly suited for the job currently.

  Yeah the Vertex Blending relies on color paints directly on the mesh rather than an actual image sadly. At least within Id4 anyway. This is pretty much what your mesh looks like before export (seen in vertex paint mode)...     Heightmap info is the most helpful for what I'm trying to accomplish. This is easier to process (in theory anyway) as we're not relying on select RGB channels as most painting methods utilize in modern engines. Instead with the black and white coverage we control the blending. Both the Vertex Paints and Height map use black and white. The Vertex Paints shouldn't see any change in my process. Alright, so note in my example the dirt path surrounded by grass...     The dirt is within my InvertedVertexColor block. So that part of the mesh is painted black. The downside is that there is only linear blending at the geometry edge loops. The second step I want to do is assign the B&W height map as a last step on top of the vertex blend. So then the white would contribute to the coverage in the black painted areas giving a refined edge based off your heightmap info. What I'm trying to customize is the B&W coverage down to the pixel rather than a Vertex Color. It would be very precise and since I make sure the tiling all lines up, so would the heightmaps.   I may make a visual example to define this further as it's difficult to explain the power it gives the designer. Here's an instance I could use with this. If I had a stone floor that was supposed to have loose tiles. I could blend a different dirt texture in between the tiles. Now this could also be done with a duplicate texture with a different color of dirt in between, but it's still more restrictive. I could pair several tiling dirt textures for most my needs rather than having similar alternate tile floors with the desired dirt in between the tiles.       This may be just the break I need, as that's exactly what I need to accomplish. I'll have to save that and try to use it when I get back to my desktop.

Yeah the other method of mapping out high/low areas means a lot more polygon supports layed in to make the terrain more dense. Now this still isn't really a problem as the terrain wont be needed to cast shadows. Not for the major low lying portions anyway. All I know is if our coders (or other well todo knowledge goers) somehow found me a way of combining heightmap influence on top of vertex painted areas, I could transform how terrain currently looks in TDM. Megatextures would be beastly, but I really would only combine actual terrain elements rather than the rest of the environment. The other thing too is last time I tried to load a Megatexture (in id5's editor) it took me over twenty minuets to load a single map. Not to mention how difficult it was to run after that even (and my hardware isn't too shabby). Sadly I don't think our technology will catch up to Megatextures for at least another five years. Right now we're really only seeing an entry level usage of it and not even it's full potential. To build on that platform of tech and do well, you have to have the elite of the elite hardware to pull it off. So for now I'm still stuck with my different blending methods for multiple textures, to break up noticeable tiling.

Been up for a while going over this. I don't have my actual material here but this should still suffice. What I'm wanting to do (and kinda talked at this) is to use a height map combined as color contribution to the vertex blend. Basically a two step operation with the height map added last. { blend diffusemap map file/path/image_d.tga VertexColor } I know you can use multiple paths in a single blend diffusemap block but it doesn't accomplish what I'm going for. Other blend modes just layer it over the existing effects present. As I'm trying to target the VertexColor directly with a heightmap. Might be possible, might not?

Man this looks great. I'd love to hear a detailed bit on how you're getting these results. I could use something like this for some effects for my rooftop charger mast. Much better! You are your own judge of what you've got in your head to go for the design. Add as much density to the highpoly as you need to smooth out the normals. On the occasion you may actually want to do the opposite for a sharp edge. So what I do in that situation is split edges (or by XSI terminology disconnect polys). If you break up the mesh you can control the hard edge factor that way. I used to rely solely upon smoothing deformers (or turbo-smoothing) to get quick previews, which is good. But it can also be limiting as to what you can do for manipulating the final highres mesh. But yeah now the big thing is to make sure your cage mesh (lowpoly game version) will look like. It needs to conform to the highres shape as closely as possible. This should be very easy as you have a simple shape. Often for most of my technological works I have to blow the mesh up many many times to get rid of baking artifacts. That only happens when you have reference faces (low resolution areas) that have their bake radius intersecting with other highres portions. The bake radius (or rays so to speak) fire directly along the lowres normals in towards the mesh. The radius basically just tells how far out from the mesh surface will be taken into account for the bake. So in tight corners you can get bake rays clipping through the corners. If the corner is an acute angle that is. If it's obtuse, even by a hair then the rays wont project into highres geometry. It's an annoying thing to get around but when you model complex meshes a lot it eventually becomes standard to think that part through.