[The Commander]

Yeah, I can't seem to them quite as good as the ones that have been done so far. But im giving it a go any how for learning for future mod's etc.

[Parkar]

The Commander, on May 14 2009, 04:45 PM, said:

Yeah, I can't seem to them quite as good as the ones that have been done so far. But im giving it a go any how for learning for future mod's etc.

Note: That you don't need to have a specular map to get secularity. The one you posted here seems to just be a black and white version of the diffuse so the only thing it will do is darken the highlight in the already dark spots. Only time it should be needed is if different parts of the texture need different power and scale values to look right. For instance when the texture consists of two different materials and the diffuse does not supply you with the necessary difference. A perfect example is if you have a dark piece of metal and bright cloth. This one would tend to need a specular map where the cloth is darker and the metal brighter. The cloth should also have a low alpha value to make the specular size bigger then on the shiny metal. All this together with tweaking the specular power and scale values should give you metal with small bright highlights and cloth with bigger and darker highlights.

[Tea Monster]

There are two schools of thought on the web. One says that you can only get decent results from building an actuall 3D model and baking off the normals. I personally think we are going to be waiting a long time for an HRP update if that were the case!

The other view is that you can fudge it or fake it by making a height map (black is farther away from the camera, white is closer) and then feeding it through something like Crazy Bump, PS or GIMP to get your normal map.

The problem with feeding ready made texture images through an automated normal mapping process is that the software just basically often gets it wrong. Usually it goes by the simpler the image, the better the results.

I'm not an expert in this but from what I've seen, it looks like we are going to be using a mixture of different stuff to get the results.

[Hendricks266]

It seems I got it wrong. Whoever made the original renders did them incorrectly. Compare the highres tiles with the originals. comparison GIFs, set at the exact same speed:

Original HRP Order | 8-bit | Correct HRP Order


1) They must be centered
2) They must be exactly 4x the original resolution
3) There must be no empty space
4) The shots must be at these EXACT angles: 0°, 22.5°, 45°, 67.5°, 90°, 112.5°, 135° (that's 7 frames looping)
5) Keep them RGB please, keep palettes out of HRP work.

[NightFright]

Well, what's right now? The edited version you posted? According to your last post, it seems that way. At least changed version it looks fine ingame.

[Mia Max]

I really don't know if I'm doing it right, but here is my first try of tile0000.





I couldn't test it ingame, but if this is correct I could have found a easy way to make normalmaps.

EDIT:
Here's another example:





I know there are still many little fields that still have to be removed, but it's only experimental.

I've made another one:





It's not finished, but it looks already 3d.

I like playing around with normalmaps

[Danukem]

NightFright, on May 15 2009, 12:15 PM, said:

Well, what's right now? The edited version you posted? According to your last post, it seems that way. At least changed version it looks fine ingame.

Do you notice how the HRP versiondoes not animate smoothly even when the order is correct? I'm guessing that's because it fails one or more of those tests he listed.

[Tea Monster]

If you look at the animation, it 'jumps' as if a frame is missing.
Thats apart from the fact that the specular on the model is way under what it should be.

[Hendricks266]

[Tea Monster]

If it were me, I'd just redo the damn thing.
This would probably look terrible, but you could try taking the next frame up from the ones that are missing, copy them and just shink them widthways to replace the missing tiles.

...

Yeah, thats going to look terrible. Who did the original?

[The Commander]

I always thought there was something wrong with that spinning nuke icon, after playing with the HRP and then pausing/quitting to the main menu, the icon would display a random tile near the end of the animation.

[Hectic-Zombie]

I thought I would post another higher resolution version of one of my older textures. This one is 8x bigger than the original texture, the one in the HRP is only 4x. I'll post more when I can.

[Mark]

Here are textures 443 and 959

[Roma Loom]

Mia Max, on May 15 2009, 11:03 PM, said:

I really don't know if I'm doing it right, but here is my first try of tile0000.

Hey there. Such textures as stones, asphalt, bricks etc don't need any handwork at all (in many cases). Either it's a job for CrazyBump or NVidia normal map filter or the plug for the GIPM here.

You normal map (without heightmap) should look like this:

[Parkar]

Mia Max, on May 15 2009, 10:03 PM, said:

EDIT:
Here's another example:





I know there are still many little fields that still have to be removed, but it's only experimental.

Not sure what techniqe you are using but it doesn't look quiet right. It should look more like this (using crazybump and some image edting).



Also just realised that this texture actually needs redoing as the original looks more like flat tiles then bumpy rough stones.

[Mia Max]

Parkar, on May 16 2009, 03:48 PM, said:

Not sure what techniqe you are using but it doesn't look quiet right. It should look more like this (using crazybump and some image edting).



Also just realised that this texture actually needs redoing as the original looks more like flat tiles then bumpy rough stones.

I cleaned the surfaces because I thought that it would look too bumpy. I did the same with tile0000, but that was my first try.
I think I did a quite good job on tile0812.
Don't you think that most tiles just look too rough-textured without handwork?

Also most tiles have already light and shadows. When I've made the normalmaps without handwork, Gimp wrongly defined lighter parts higher than the dark parts. @Roma Loom: That's why I think tile0000 doesn't look right.

[Roma Loom]

I've made 2 shots ingame with grey diffuse to see the normal clearly:

[Parkar]

Mia Max, on May 16 2009, 04:48 PM, said:

I cleaned the surfaces because I thought that it would look too bumpy. I did the same with tile0000, but that was my first try.
I think I did a quite good job on tile0812.
Don't you think that most tiles just look too rough-textured without handwork?

Also most tiles have already light and shadows. When I've made the normalmaps without handwork, Gimp wrongly defined lighter parts higher than the dark parts. @Roma Loom: That's why I think tile0000 doesn't look right.

I think it looks just fine.

[Mia Max]

Roma Loom, on May 16 2009, 05:19 PM, said:

I've made 2 shots ingame with grey diffuse to see the normal clearly:

Parkar, on May 16 2009, 06:39 PM, said:

I think it looks just fine.

Okey, you are totally right!

[ReaperTK]

I checked the TODO list and didn't see anything about this issue.
The SLOTDOOR switch (tile 132) HRP model is reversed The off position should be to the left. Here's a comparison.


I also noticed that all of the other switches have a transition animation whereas this one just changes states. Also, the entire lever isn't drawn (the top and back portion when in the off position).

[TerminX]

None of the switches actually have an animation, it just looks that way because of EDuke32's badass frame interpolation feature.

[chicken]

Can this be switched off for a single model (not all)?

[NightFright]

The switch only has two states like TX wrote: ON and OFF. That's two frames. You can hardly speak of "animation" there.

Anyway, if you want to criticize anything regarding that switch, I'd rather say it's the missing "border" you can see on the original sprite. But besides that...

[Master Fibbles]

No, he is right, it is backwards and that is a problem. One would think that that is something the artist would have been able to check.
It is an easy fix though...change the skin a little and reverse the frames.

[Parkar]

Mr.Flibble, on May 20 2009, 02:33 PM, said:

No, he is right, it is backwards and that is a problem. One would think that that is something the artist would have been able to check.
It is an easy fix though...change the skin a little and reverse the frames.

I made that and never noticed it before. it's a simple thing to fix, just mirror the model. The model will probably get remade now anyway though.

[escapist]

Parkar, on May 20 2009, 06:32 AM, said:

.... The model will probably get remade now anyway though.

.... Really?! ALL the models?

[Master Fibbles]

he is talking about the one switch.

[escapist]

well, just wondering...
how much of the current HRP will have to be remade?

[Tea Monster]

escapist, on May 20 2009, 04:22 PM, said:

well, just wondering...
how much of the current HRP will have to be remade?

Quite a few of them, if not most of them. Nobody has done a head count yet. Way too early for that.

[NightFright]

That was actually the main reason why the Polymer thing was kept under wraps for so long. If it had become public large parts of HRP contents would need to be recreated, some guys maybe might have lost their motivation to go on "old-style". But let's just wait till the new engine is finished before speculating too much.