Page 4 of 15

Posted: Tue Jan 16, 2007 9:31 pm
by zuegs
Hi Wedge,

just as i saw your post in other forum... the "<gain>" tag for textures is no more valid for v0.7t1 !
Please see Onos ChangeLog for v0.7t1:
* removed gain from texture elements, using a, b, c polynomial coefficients instead
all textures (albedo and bump) use the <b> tag for gain. Final color/bump is calculated by polynomial from "pix" pixelcolor in image:
col = pix^2*a + pix*b + c (not absolutely sure if this is right)
So <b> is the linear scale, <c> is an offset (defaults to zero), don't know when using <a> :wink:

Posted: Tue Jan 16, 2007 10:18 pm
by Wedge
Thanks for bringing this to my attention. I have overlooked this and its the only thing I have left on my todo list. I'll have a look into the example posted by Nick, your example, and into the testscene and manual to see what else I can find.

Update: Opps, it appears my todo list has the blend material on it, which appears different than the gain thing. But I'll have a look at both.

< embarrassed > I also updated that post I made explaining that you brought this to my attention and to ignore my information. < / embarassed >

Update: Also, I feel it is my duty to tell you that Kosmo said the screwdrivers are made in indigo 06. :) I'm not sure if you are aware. :P

Thanks again!

Posted: Wed Jan 17, 2007 1:33 am
by afecelis
ahhhhhhh! that explains this!:
Image
cool! 8)

Posted: Wed Jan 17, 2007 1:41 am
by OnoSendai
Read the updated manual in the 07 test 1 thread for some info on the a, b, c stuff.

Posted: Wed Jan 17, 2007 1:48 am
by afecelis
great Ono, thanks! You mean your updated pdf, right?
I also got this other error:
Image
which I guess is U3dreal's rotation matrices question. But what causes it? I didnĀ“t have any rotated geometry in blender!

ps. Sorry for hijacking the thread a bit. :oops:

Posted: Wed Jan 17, 2007 3:10 am
by Pinko
it's possible put in this post the correct .py file for export 07...
Tnx in advance
Pinko. :cry:

Posted: Wed Jan 17, 2007 3:40 am
by afecelis
well, there's no such thing as a "correct" python script yet since this is Wedge's work in progress and we're practically beta-testing the code here to report bugs and problems. Still, he has updated the script 9 times since he started it out and has reported each of the improvements he has done to it. You'll always find the latest version available on page 1 of this thread :wink:

Posted: Wed Jan 17, 2007 6:36 am
by DaveC
Ok. What's the deal here? The first time I copied the script into a text file, it worked. It exported. Now, I copy the same part of the first entry (which has been update, I know) and paste it into exactly the same file (over the top of the old text, obviously) and I get an error!

Code: Select all

  File "<string>", line 47
    Takes module, class, list, dictionary, or string. """

IndentationError: expected an indented block
Am I doing something wrong or is it the script that is in error?

Posted: Wed Jan 17, 2007 6:40 am
by Wedge
zuegs wrote:@Mute:
ensure every objects scale is equal for all 3 directions X,Y,Z - all uneqal scaling should be applied to the mesh (not the object) - use Ctrl-A to apply objects asymmetric transformations to the underlying mesh in your case
hope it help, regards, zuegs
Afecelis, the roation matrix fix has been added from zeugs since the early stages. I am posting his quote from the front page when he included more info about that error. Maybe this will help you? Thanks again Zuegs..

Thanks Nick I am reading the 07 manual more fully now.

<gain> tags removed

The SSS crash I found was indeed with SSS and sun. SSS with the rectangle meshlight/blackbody at least gets past the fatal error when starting indigo and starts to render the scene. So Nick has also said this in the thread titled: "absorption and scattering coefficients for liquids" in the Indigo renderer general section, it is a bug he knows about.

I am now in the process of updating the code at the front page and moving all relevant directions to the first post. Will be done very shortly, I'll update this message.

Update: Code updated, directions for new SSS in front page with code, yanking all other directions from earlier posts to make it clearer.

Posted: Wed Jan 17, 2007 6:43 am
by DaveC
Wedge, I would like to volunteer to host your script for others to download in a zip file. If you email your updates to me at dodgydavec at googlemail dot com I will upload it in a zip file and provide a link.

Posted: Wed Jan 17, 2007 7:26 am
by Wedge
DaveC, it is best to copy/paste the entire scrip to avoid whitespace errors.

Thanks for the offer to host but right now I'd like to keep it like it is. (since I have been updating it very often) Hopefully, when U3dreal/other devs get ready with 07 we can merge it together for the download.

The script is GPL all of its terms apply. :) I don't want to take anything away from that so feel free to do whatever as long as its within the license. :)

I, however, will stick to updating in the front on the post.

I rewrote SSS in this update. Update 2: Moved all directions to first page to keep confusion down.

Posted: Thu Jan 18, 2007 11:43 am
by DaveC
I am, as of now, having absolutely no luck whatsoever getting the exporter to do SSS materials. I make the sss (for example milk.med) and the material (ball.milk) as instructed. The resulting xml doesn't even have a medium tag in it anywhere. I need some step by step instructions. Can anybody help?

EDIT:
Scratch that, I didn't have RayTrans for the medium switched on.
How do you set how much the light is absorbed in the medium?

Posted: Thu Jan 18, 2007 2:09 pm
by afecelis
@Radiance: check page 1 of this thread. Code is updated there. Copy and paste it into a text editor. :wink: Save as .py

Posted: Thu Jan 18, 2007 7:05 pm
by Wedge
Radiance, I wouldn't use this exporter for anything other than tests.

SSS is in partly, but probably not right and it is also confusing. It is likely not going to work as expected and is still in the process of a rewrite.

In the material post by Afecelis I have posted some ideas on how to think about implementing it in a clearer way. I will likely try rewriting it but I'd prefer to get some other people's ideas about how they think it can/should be done.

Update: But as far as I know, all of the other added features are working good. I recently did a bigger render of a ball with a reflection using the normal smooth. Looks nice! So the exporter appears to be in good shape if you only use diffuse or phong materials. It will output some mediums/SSS if you follow directions...and maybe good enough for you to manually edit the xml? I don't know for sure...you will have to test for yourself.

Posted: Fri Jan 19, 2007 2:31 am
by jurasek
How to control bump value in texture if there is no <gain></gain> tag anymore?

greetz,
jur