Although it's far more flexible than PNG but IMO it has a tremendous amount of bloat that really doesn't need to be there. I'm sure some program can encode layer information into a TIFF file but again, like the Fireworks for PNG, it'll be a per program thing and whatever reads it from there will need to know the magical incantations to read it all back in.

So basically, Adobe PSD, JASC (Now Corel) PSP, Corel Photo Paint, Corel Painter and whatever else out there that supports layers, their native formats have the concept of layers already implemented in their image formats. Most of the other formats don't.

From what I know of PNG's history, it doesn't have a native application. When you say Fireworks, are you meaning MacroMedia's (Now Adobe) Fireworks? PNG format does allow for multiple images in one file I think and if Fireworks manages to store layer information in a PNG file format the I would think that it is more of how Fireworks has used the PNG format and that it may not be common. LibPNG.org's API doesn't specifically line out layer handling or higher level image stuff like that. From libpng web site I found the following bit of info. # Fireworks [Macromedia] (Win32, Mac) - all versions; read/write (Fireworks' native file format, in fact); full alpha support, including good quantization to RGBA-palette mode in version 2.0; no gamma support; partial interlacing support; version 1.0 reportedly writes 1-bit images as 8-bit and includes spurious, incorrect sBIT chunk indicating 8-bit precision; commercial. (As of version 2.0, the quantization from 32-bit RGBA to 8-bit RGBA-palette appears to be perfect; Greg will do some personal testing soon. Even without that, this seems to be the best PNG-supporting image editor available as of mid-1999, though The GIMP is quite close. Be sure to use "Export" rather than "Save" to create final PNGs for the Web! The "Save" versions include Fireworks object and state information, and not only are they huge, they also crash Windows Explorer and MSIE 4.0.) So basically Fireworks tosses in specific information so that it can read back in the layer information. This will be Fireworks doing and not the more general or open libPNG's doing. So if some other program wants to use that information they will have to know what Fireworks is doing to the file. It just goes to show the flexibility and potential abuse that the format allows. Any other paint program can read in a Fireworks PNG and import all the layers or is it just limited to Fireworks?

Not really a PNG feature. As far as I know it isn't possible. PNG's origins was to replace GIF (and possibly TIFF.) So it's a simple format just to hold an image. Not a complex one that has ties to a given program like PSD.

Message edited on: 06/14/2005 18:41

I've hit that a few times as well. I then start naming selections and hiding stuff. Although this isn't ideal when you are trying to position with respect to the stuff your are hiding. :/ In that case, temporarily move the points into clear space and then back. Still not all that simple at times.

That is "shouldn't make a lick of difference..." I should really proof read more.

1/ I can't say too much about the lead thing. What does the shader look like. There isn't duplicate faces in the same ares I presume. 2/ The pattern is a texture map I take it. I haven't hit this sort of case before really. Although I've noticed that using the UV editor in Carrara you may need to rotated the geometry before uv mapping the piece so that you know where the left edge or the texture map should end up. Booleans should make a lick of difference on uv mapping IMO.

Yeah. Seeing some of the other contests they've held, it's a seriously tough challenge. Well, good luck thought.

Well done! Good idea about the light cone thing. I'll have to remember that. I'd be interested to know how you do in the competition.

Woohoo! :) Thanks everyone! It was fun image and topic to work on.

The problem with successively smaller objects, like in the case that Shonner has done, is that Carrara is a limited ray depth of 25'ish. I say ish as I've tried this with one of my images and I had less because Carrara would stop after a certain point. Probably due to some tolerance on transparency. This sort of limits how 'translucent' you can get. Make the steps to the smaller objects too much and it'll be noticeable. Make it too small and there won't be much depth to the translucency. Also it can be really hard to make sub surfaces if they aren't simple like a cube or sphere. Another draw back is that if you look close (or if there is something behind the object) you'll see through to what's behind the 'flesh'. Not really want you want but not much to get around it. However if Carrara did allow for a significant increase in ray depth, super fluffy clouds could be made but that's a different issue.

Everything done with the vertex modeller. Some simple boolean subtraction for some parts but most are actually quick and dirty modelled. Meaning, I didn't bother to make them perfect; the black rook for instance, there are actually a couple of pipes intersecting itself instead of one clean mesh. You'd probably see that if I made the solid glass or something.

A bit of a shame you don't seem much of the tressle work under the board. Ah well. Small part of the whole anyway.

But man, those tiny screews on the corner really slowed the render down. :/

Message edited on: 04/25/2005 18:05

Kixim, small problem with your set. King and queen are in the wrong place and they are lined up on the wrong side. Ok, being fussy. Both look good. Sdnobles looks very elegant. Sdnoble, how did the close up switch the king and queen? Yet it's fine in the first render. Still working on mine. Never made a chess set before in 3D. I guess it's about time. Is there a time limit of 30 min? If so I've blown it easily. :P

Thanks all. I think I've settled on veloute with some tweeking with brightness, contrast and bump height and I'm getting close enough to what I am looking for.

Actually on another forum (3dxtract) someone was smart enough to set 1 to the vertical brick count. The black I/we are seeing is the groove of the first brick layer seam (grout). I feel dumb. :P Mystery solved in my opinion. However with 1 for the vertical you get a seam that runs the length of the object. Not what I want so it would appear brick isn't an option for me to use in this case. However the formula case for bump mapping still doesn't look proper. Maybe that is something Eovia could look at. Thanks all!