Drill (2)

[Warvinc]

I want your thought too but I think that yes it is way better!

I just want to say that making a perfect normal map for a cylinder is impossible(I read a lot on it because I wasn't sure)so I've tried doing the best I can!

Next time,I'll just need to put it more high-res with more sides if you think we see the hard-edges too much,I'll try that with the armoury :)

[Viech]

Thank you, so the blurr was indeed an issue on our end! With the normal map you just provided, the drill looks a lot cleaner ingame:

Here is the new pk3.

It certainly looks good enough for the release, so we can lean back and don't need to hurry anything. If you find the time, could you maybe try to rebake the normal map, and do the following:

1. Try to get rid of the aliasing artifacts. Maybe there is an antialiasing option you can enable. If there isn't, I could try to remove the aliasing artifacts by hand for the following release.

2. Before you rebake, try to increase the number of polygons around the main body on the high poly model so that it looks cylindrical even if it is rendererd with flat shading. (In other words: Increase the tesselation resolution of the main body.) The issue is that the current normal map has no effect on these regions, while it could help creating the illusion of roundness if the high poly model was really round.

3. Check if the high and low poly geometry of the spacers around the glowing core match roughly. On the normal map, it looks like they have entirely different angles.

[Warvinc]

Like I said before,my high poly is reaaaalllyy high,so it's not the problem.

the problem is that it won't be possible to make it more ''round'' only with a normal map,like the exemple right here: http://www.chrisalbeluhn.com/Normal_map ... age_31.jpg

If you want it more ''round'',I need more poly on the low poly.

I've tried to rebake it but it's always the same,and yes the anti-aliasing was on:)

And I don't understand the problem with the spacers?or maybe I just don't see it?

[Viech]

Let me have a closer look in the next days, currently I'm busy learning for my last exam. Explaining what I mean would require atleast a sketch and also I have to check if my understanding of our engine's treatment of the normal map is correct. I can also provide more screenshots of the drill then.

[Viech]

So I assumed normal maps were applied on top of the surface normals but they are in fact applied on top of the interpolated vertex normals. This means that

1. The flat area where I expected gradients to show up (main body) is correct, since the interpolation already aligns the normals properly.

2. The surface area of the spacers where I was confused about the alternating colors and gradients could be correct, too. (Technically, the normal map appears to transform the interpolated vertex normals back to resemble surface normals.)

So the only thing left from my first look at the normal map would be the aliasing, but I can't spot it on the ingame model so it might be outside the visible set.

One last question: Did you use the same smoothing groups when baking the normal map and when exporting the model, especially on the spacers? The normal map applies shading to them that looks a bit odd. A potential explaination would be that the normal map was baked when the geometry was smoothed along the long side of the spacers and then later the model was exported (or converted) with those surfaces not being smoothed. The normal map would then try to "revert" the interpoation that was applied at the time of baking but isn't used on the recent model anymore, leading to a shadow gradient like the one you can see here:

No normal map (and no specular map):

With normal map:

Could you also double check that the smoothing groups on the recent model are correct, especially around 90° edges? I notice that there is a shadow gradient on the lower part of the main body even in the screenshot with normal mapping disabled, which could result from smoothing being applied over the edge between the lower part and the belt above it:

[Warvinc]

Okay what I'll do is redo every normal map again and smoothing groups,so after that if there is any problem,we'll know there is a problem with export or something,so I'll wait before starting the other task(armoury)but I want it so put it on hold for me :P

It's weird because I did the same thing with the telenode and everything went fine !

And I'll also modify my diffuse to delete some dirt and lower the specular like you did!

I'll resend you everything when it's done!(next week probably) :)

[Viech]

The armoury is yours, don't worry, we won't take it away from you.

Here's the most recent texture set. (They are in webp format, if that's an issue for you I can convert them to tga.) Also have a look at how I made the red light stripes align with the seperators on the model. If you're working on the texture anyway, you could add some generic labeling on the white stripes like you did on the telenode. I also feel coloring the small dents on the spacers (or everything around them) could lead to a visual improvment as the spacers as such look a bit monotonous right now.

By the way, I think the reason why the body doesn't look entirely round is the labels and the additional edges they create (especially the high-contrast edge below the lower stripes), the poly count as such would probably be fine. Not sure how to work around it though.

Ishq and gimhael, could you double check that the conversion process and iqm handling of smoothing groups was correct?

[Warvinc]

Well I prefer reworking the textures because you modified only the TGA but not the PSD including the layers,which mean it's a big problem for me to rework on it!

Oh and yes I need TGA please!

[Viech]

Well I prefer reworking the textures because you modified only the TGA but not the PSD including the layers,which mean it's a big problem for me to rework on it!

Sure, just wanted to give them to you as a reference. Tga version is here. Be aware that since the normal map lost quality at some point, this could be the case for the other maps, too.

While you could start working on the diffuse/specular/glow map now, it appears that we aren't quite sure that the issue isn't on our end (we started using a new model format and we don't have much experience with the export process yet). You could wait for us to do some tests if you prefer.

[Ishq]

That would probably be best. We're also dealing with a newbie person exporting the models :) We'll let you know what the results are. This may entirely be my fault for just messing up the export.