Questions and Answers 3D (read the first post before asking a question)

[Alvarez]

What do you think about projection painting compared to painting on a unwrapped texture?

http://www.youtube.com/watch?v=maJj6sLOiKM&feature=related

This method seems so easy and less abstract compared to UV unwrapping, but are there any catches?

I was thinking about taking pictures the different axis from the untextured mesh, sketch a texture on them and projectionpaint it back.

[Luther Blissett]

I suppose the main one is lack of painting tools, layers and the ability to move those layers about - though if you got a high-end model painting program, that might not be a problem - but in the free / open source level applications, you're currently looking at little more than "mspaint" style painting tools only. What does work nicely though, is to unwrap the model normally, then 3D paint onto the model to mark out seams and other such areas, then photoshop the remainder of the flat texture to finish everything off - particularly useful for drawing straight lines around a round object, where the front and back aren't next to each other on the texture.

Your suggestion of working from different axis, painting them, then reattaching them would work on the whole, but you'd have some problems with interior sections (i.e. a person viewed along x axis would have a problem with painting the inside of legs and arms). You could however get 3/4 of it done, then finish those bits off "traditionally".

[pistacja]

Some random questions.
-is there room in the rendering pipeline for shaders(per pixel ligth, shine, glow, normal maps)?
-can a full screen, post-processing filter be used (bloom,  ink, blur)?
-is texture compression used? What kind?
-what is the prefered image format(tga, png, dds...)?
-what about alpha? 1 bit (black-white), 4 bit, 8 bit (256 grayscale) or none at all? In what mode is it used? Would two transparent layers be correcty visible?
-how would animated particle-like effects work? One image-one file, all images side by side in one file (atlas), or a gif-like format from fallout?

I also have some nice explosions that I can give away, are there any plans to remake the original blasts?

   

[Lexx]

-can a full screen, post-processing filter be used (bloom,  ink, blur)?

Yes.

I also have some nice explosions that I can give away, are there any plans to remake the original blasts?

Not that I know of any. At least not officially-- we want to keep the game to look like the original.

[Karpov]

Some random questions.
-is there room in the rendering pipeline for shaders(per pixel ligth, shine, glow, normal maps)?
-can a full screen, post-processing filter be used (bloom,  ink, blur)?
-is texture compression used? What kind?
-what is the prefered image format(tga, png, dds...)?
-what about alpha? 1 bit (black-white), 4 bit, 8 bit (256 grayscale) or none at all? In what mode is it used? Would two transparent layers be correcty visible?
-how would animated particle-like effects work? One image-one file, all images side by side in one file (atlas), or a gif-like format from fallout?

I also have some nice explosions that I can give away, are there any plans to remake the original blasts?

   

As far as I know, the game can use .fx shader files, there are some Default .fx files in the data/effects folder. But I don't know of anyone with knowledge of HLSL, and I am afraid not every .fx works.
About textures: since the image size is going to be very small, we wont need MIPmaps. I used jpg, tga, and png. All three worked fine, but I think JPG is the most compressed format, so it would be a good choice when we have many of them.

[pistacja]

JPEG only saves space on the disc, in memory it take just as much ram as a uncompressed BMP (8 bit per channel x texture resolution, something like 0.7 MB for a 512x512 texture).  But there's a compression algorithm that uses the gpu hardware to decompress images on the fly from ram, so a texture can be stored (and send) to the gpu compressed. It can save (in theory) 3/4 to 7/8 of total memory used by textures... but the most popular format (dds) also stores mip-maps so the save on memory is not that big.

[Karpov]

I didn't know that. so...you think dds would be the best choice? what file size does it have compared to the JPG? let's say, a 256x256 texture.  Good info. 

[pistacja]

This is the best info I found:
http://forums.relicnews.com/showthread.php?62051-Decal-Texture-study-TGA-vs-DXT3-vs-DXT5

256 * 256:
DXT: 85.4K
TGA: 256K

[Luther Blissett]

It might depend on the engine, and this info might be a few years old now (and might not be 100% accurate) :

DDS - small size on disc, small size in memory
PNG - small size on disc, normal size in memory
TGA - large size on disc, normal size in memory
JPG - small size on disc, large size in memory, no alpha channel
BMP - large size on disc, normal size in memory, no alpha channel

I can't remember any specific numbers though - but something about graphics cards could read DDS without needing to uncompress it, which meant it used a lot less memory in game.

Within DDS :
DXT1 - alpha on /off (100% or 0% solid - can't have 50% or any other number)
DXT3 - alpha variable (100 - 0% solid - i.e. 50% here, 30% there is okay)
DXT5 - alpha variable (100 - 0% solid - i.e. 50% here, 30% there is okay)

I can't remember the difference between dxt3 and dxt5, or if there are other (or newer) dxt numbers available.
[edit] In pistacja's link above, it shows dxt5 has a smoother alpha

I don't know if you'll need alpha channels, or if the engine supports it even. Could it be useful for death animations (i.e. a hole shot in the chest), or having a damaged / blood layer on the models to show injury?

[Jimmy BoyX]

Sorry for annoying question but how many animations are done? How many are making?

[Karpov]

Sorry for annoying question but how many animations are done? How many are making?

The animations I need are all the Dodge animations. I have it for unarmed and knife already. I disabled the Running, using, and picking up with weapon in hand, just like in the original game, for now of course. I won't make additional or special death animations yet. The main goal now is make it playable, nothing missing but nothing added either. When I'm done with all anims, I will create a full animation demo. That way I can see, hear opinions, and correct all the stuff I need to before converting them to the female structure.
Oh, also reload animations need to be done. Right now I'm using the same animations for Shoot and Reload.

[Lexx]

I disabled the Running, using, and picking up with weapon in hand, just like in the original game, for now of course.

To be honest, I don't even think it's bad. Running with guns and stuff will need a lot more attention to animation and I don't really think it's that much worth to spend time on it right now. Even weapons visible on back - while interesting first - might be better not to be visible. At least that's my opinion on this.

[Eternauta]

I won't make additional or special death animations yet.

Well when death animations start being worked on, please don't haunt us with a new "hole in torso" animation for a eyeshot death just like in the original games

Anyway, congratulations, thank you and keep it up!

[Crazy]

Well when death animations start being worked on, please don't haunt us with a new "hole in torso" animation for a eyeshot death just like in the original games

We need a hole in torso, as well as brain explosion.

[Gray]

To be honest, I don't even think it's bad. Running with guns and stuff will need a lot more attention to animation and I don't really think it's that much worth to spend time on it right now. Even weapons visible on back - while interesting first - might be better not to be visible. At least that's my opinion on this.

Although running with gun is an individual animation, the weapon on the back is just an attached object. It's not even the similar things. Running with gun requires as many animation sets as there are different weapon holding types. Weapon on the back requires almost nothing.