[icecoldduke]

So I'm almost finished with my vox2obj converter. This takes KVX/VOX and KV6 and converts it to OBJ. The obj conversion works(with proper UV coords), just need to get the texture wrote out properly. I'll release the tool when I get the texture output working right. With PolymerNG the mesh tool with just read in the obj file and the model will get used like any other mesh.

[Lunick]

So... is this something players have to use if they want "voxels" or the engine will do it on its own...?

[icecoldduke]

 Lunick, on 15 July 2016 - 07:49 PM, said:

So... is this something players have to use if they want "voxels" or the engine will do it on its own...?

If I can get the proper permissions I'd like to just include the game_meshes.payloads file with the converted voxels in the release. Outside of that I have a batch file that will go in and convert all the voxels to obj's then build the game_meshes.payload file. This takes about 2 minutes to execute and your up and running. That would definitely be a option to have the engine do it, but that won't be in the next release.

[icecoldduke]

If anyone needs a voxel converter, here is the first release of vox2obj. I'm not sure if the texture output is right, but this will get the geo converted. First command line arg is the voxel file, a obj with the same name will get saved when the conversion is complete. Or you can run the batch file and all files in the folder will get converted. The batch file is pretty stupid, it assumes the only files in the folder are voxel files.

**SEE UPDATE BELOW**

[Hendricks266]

Why not have the user manually pre-process ART files too?

[icecoldduke]

 Hendricks266, on 15 July 2016 - 09:20 PM, said:

Why not have the user manually pre-process ART files too?

Art tiles are not currently getting processed by the payload system. They have such a low memory footprint I don't feel its necessary to DXT compress that data. The image manager will load the art tiles if there isn't a high res albedo texture in the payload file. Can you think of a good reason to DXT compress the art tiles? If not the load times for the art tiles is nothing.

[Hendricks266]

</sarcasm>

[icecoldduke]

 Hendricks266, on 15 July 2016 - 09:32 PM, said:

</sarcasm>

Ah . The reason for having the voxels go down the mesh payload pipeline is 1) I don't have to special case the voxel code and 2) the voxel load times are higher then I would like them to be. Remember I want seamless loading; no load screens.

[Hendricks266]

You could convert them all at once upon the game loading SWVOXFIL.TXT.

[icecoldduke]

 Hendricks266, on 15 July 2016 - 09:34 PM, said:

You could convert them all at once upon the game loading SWVOXFIL.TXT.

I could but I really don't want to special case the voxel assets. I would like all mesh data to come through the mesh payload system. Special case shit always causes problem somewhere down the line . What I might try is having the mesh payload system run on first run. It only takes a minute to execute then its done.

[Hendricks266]

 icecoldduke, on 15 July 2016 - 09:39 PM, said:

Special case shit

Welcome to BUILD.

[icecoldduke]

 Hendricks266, on 15 July 2016 - 09:40 PM, said:

Welcome to BUILD.

Yeah and that's why its taking me longer then I'd like the find and figure out certain bugs . You think you know the codebase until you find another special case codeblock for the special case code block. With PolymerNG I'm trying not to add to that mess. If this becomes a issue I'll look into another solution, but for now I feel this is the best way forward.

[icecoldduke]

I fucked up the texture coord output. I somehow forgot to write out "vt" and instead wrote out "v" :/. I also didn't realize the shadow warrior hrp came with MD3's. Anyway here is the updated fixed tool. This will only affect a couple of you.

Attached File(s)

•  voxels.rar (31.43K)
  Number of downloads: 175

[icecoldduke]

I have updated the PolymerNG public github repository to P4 rev 476. Previous update was rev 395. This includes all the graphical work done this far. There is spot light code in there but it doesn't work yet, and this also includes the removal of the Microsoft math library I was using. PolymerNG is now exclusively using the IntelCPUT math library. This also includes some light shader refactoring.

[icecoldduke]

When I was finishing up the lighting in $bullet the alpha shadows on sprites look fantastic IMO(even though the sprite is facing away from the light, but thats here nor there ). The last shot the ninja is black because its the invisible ninja.

[Paul B]

 icecoldduke, on 16 July 2016 - 09:21 PM, said:

When I was finishing up the lighting in $bullet the alpha shadows on sprites look fantastic IMO(even though the sprite is facing away from the light, but thats here nor there ). The last shot the ninja is black because its the invisible ninja.

I'm by no means an artist or have a graphical eye like most of the people here however, i find the shadows that are cast on the sprite enemies kind of darker then they should be. They seem to be pretty sharp and hard as oppose to softer and more transparent. Not sure what's involved with tweaking that but I'm enjoying following your progress!

[icecoldduke]

 Paul B, on 17 July 2016 - 07:23 AM, said:

I'm by no means an artist or have a graphical eye like most of the people here however, i find the shadows that are cast on the sprite enemies kind of darker then they should be. They seem to be pretty sharp and hard as oppose to softer and more transparent. Not sure what's involved with tweaking that but I'm enjoying following your progress!

Are finding that shadows in those screenshots are generally very dark for you? I find on some monitors the ambient isn't high enough to yield the same results I'm seeing on my monitors(which is why things might be appearing darker for you). I might have to include a brightness scale screen at engine start.

[Mblackwell]

It looks too opaque.

[Steveeeie]

It is far too dark, its almost black.

[oasiz]

Brightness/gamma is something that can vary radically. Comparing laptop/desktop monitors can already yield wildly different results on the same area.
A calibration screen would be nice, I attempted this with a map earlier.
What I discovered is that doing compensation doesn't really work in a linear fashion and you need like 2-3 test points to form a curve that it gets balanced against.
Laptops in general have really boosted gamma with a low contrast and low saturation, everything looks very washed out but simply making everything darker doesn't always work.
Powerslave is one of the worst offenders here, I had to boost my saturation to around 150% just to match my IPS panel somewhat

Those SW shots, at work on a TN panel, I really need to look at a proper angle to not make the dark areas completely black.
I can however see that it looks better if I really try to look at an optimal angle.

[Tea Monster]

I don't know what options you have for compositing shadows, but multiply works well, with about an 80 to 90 percent strength.Maybe less depending on local lighting.

[icecoldduke]

I'll increase the brightness by default, its wierd on the monitors I have at home, none of those shots have dark shadows. It looks like I need to reconfig my monitors.

[Paul B]

I've tested the images on a CRT view sonic monitor, 2 LCD monitors & a single LED Samsung Flat Screen TV and I get the same across all displays. The shadows are pure black and not very transparent. It's interesting how we all seem to get different results on a static image.

[Kyanos]

 Paul B, on 18 July 2016 - 09:39 AM, said:

I've tested the images on a CRT view sonic monitor, 2 LCD monitors & a single LED Samsung Flat Screen TV and I get the same across all displays. The shadows are pure black and not very transparent. It's interesting how we all seem to get different results on a static image.

The shadow is definitely pure black with no transparency. RGB 0,0,0 checked in photoshop.

I like what Tea Monster recommended, pure black shadows applied as a multiply layer at ~80% opaqueness.

[icecoldduke]

 Drek, on 18 July 2016 - 10:16 AM, said:

...

Shadows are always (shadow * light contribution). The ambient on my main monitor brings it up to 20% ambient. Again its a error with my monitor config. It's wierd because I have three monitors on my main box and all 3 the light contribution is fine. Same on my secondary box with a different monitor.

[Kyanos]

I fail to understand how your monitor would make the screenshot look any different. Your monitor isn't about to draw back in the sprite under the black bar.

████ what does your monitor see there?

[Mark]

I decided to take a closer look at the 3 pics. It seems the closer to the player, the darker the shadow. There is shadow transparency on the monster farther away.

I've also noticed that screenshots seem a little darker than what I see in the game.

[icecoldduke]

So over the past few days I have shifted focus from the renderer over to networking. As a kid I really loved playing the Shadow Warrior shareware over TEN(my dad never bought the full version), and I want to share the same experience with my son. That is the reason I shifted gears a bit. So for multiplayer, I did not use any of the networking code in the eduke32 codebase. I also ditched most of the original shadow warrior network code(the code that synced the pp->input struct). Instead right now we have a client/server model; were the server is full authoritative. The server tells the client the stats of each sprite/sector and wall. This means the only types of hacks that users can do is wallhacks, and this can fixed by running the Polymost occlusion system on each client before it sends down the world update packets. I haven't done this yet, and probably won't do it for awhile. The state of the networking code is both the client and server can get into the same map, sectors/walls and sprites replicate properly, and the client can see the host move around. The client can't move yet. I will get this working soon. I'm also aware there is some guy on the eduke32 team working on the networking code; but I haven't seen any updates on it and I want to play PolymerNG with my family .

[Spiker]

......

[Danukem]

 Spiker, on 20 July 2016 - 11:52 AM, said:

......

I think what Spiker is trying to say is that what you are doing is great but he hopes that you will resume working on the renderer soon and specifically for EDuke32 so that those of us waiting for our shiny new toy will have something to play with.