where subcurves are created, they get passed the v range 0 to 1. They should get an intermediate range determined by interpolation of the full curve's range along the length. I'm not sure right now the best way to fix it, but that's where it is.
George
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
New versions (newer than 1.6, which is the latest public release) of liquid should work, preview and all. I've checked in pixie (and OSX) compatibility changes that will get things running. I have yet to check in the makefiles for OSX and so on (those that are there will work, for other platforms, and there's an XCode project for OSX) but if you download from CVS and compile, you'll get a liquid working with pixie. Over the next couple of weeks, I'm hoping to get a full release done with Pixie compatibility in.
Okan,
I'm still investigating the possibilities in terms of libri and libsdr in the same program, but I am having issues with multiple symdefs for libcommon on OSX. Is there any chance we could revisit this?
Currently I solve this by linking against libri and explicitly compiling in src/sdr/sdr.cpp, though it's not really the right way to do it. If someone could try the checked in version with Linux to see if these link issues are OSX specific, I'd be grateful.
Cheers
George
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Sorry about the late response people, the forum monitor was turned off for some reason.
Excellent work Fredrik, can I add this picture into the Pixie gallery ? Is the high rendering time due to raytracing soft shadows ?
I fixed the v problem on the curves in 1.3.17 which will be up next week. It is true that Pixie does not change u along the width of a curve. I find shading a curve twice along the width unnecessary because the curves are supposed to be thin anyway. Otherwise there would not be a difference between bi-linear/cubic patches and curves. However, if there is demand I can change the implementation so that curves are shaded twice along the width.
I will also add some new command-line switches.
George, I will look into using libri and libsdr in the same executable as soon as possible (which is probably towards the end of next week).
Thanks
Okan
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Feel free to put the picture in the gallery.
The rendertimes is deffenitly due to both raytraced softshadows and occlusion. And i havent tweaked the sampling AT ALL ; ). I think it would be nice to be able to use the u coord on RiCurves, but i don't know what the RiSpec says about it....
Regards
Fredrik Brnnbacka
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
RiSpec says that u should vary across the curve - but it's seems a valid decision to not do this, depending on what the performance hit is. I'm not sure that I have a strong oppinion either way. It would be interesting of Okan could comment on the performance hit were it to be added.
George
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I have made some updates on the WikiPage regarding problems with Curve primitives an some commandline switches wishes.
Here is also an image that i rendered with Pixie:
http://www.madcrew.se/Pixie/FrankMustang.jpg
Regards
The Emerald
Thanks for contributing.
It definitely looks like a bug against what RiSpec says for u,v on Curves.
Cheers
George
Hi,
The bug is in curves.cpp
lines 854 and 894
where subcurves are created, they get passed the v range 0 to 1. They should get an intermediate range determined by interpolation of the full curve's range along the length. I'm not sure right now the best way to fix it, but that's where it is.
George
I think there's also another issue, in that u isn't interpolated across the curve's width - possibly a little more involved to fix that one.
George
Wow! What is your modeler?
Max
Maya
Are you using liquidMaya to export? On a Mac or PC?
I'm currently contributing back the changes for Pixie and OSX compatibility to the main LiquidMaya source tree...
Cheers
George
I'm using Liquid on PC/Win2000
Really - you got it (Liquid) working, with preview and all?
New versions (newer than 1.6, which is the latest public release) of liquid should work, preview and all. I've checked in pixie (and OSX) compatibility changes that will get things running. I have yet to check in the makefiles for OSX and so on (those that are there will work, for other platforms, and there's an XCode project for OSX) but if you download from CVS and compile, you'll get a liquid working with pixie. Over the next couple of weeks, I'm hoping to get a full release done with Pixie compatibility in.
Okan,
I'm still investigating the possibilities in terms of libri and libsdr in the same program, but I am having issues with multiple symdefs for libcommon on OSX. Is there any chance we could revisit this?
Currently I solve this by linking against libri and explicitly compiling in src/sdr/sdr.cpp, though it's not really the right way to do it. If someone could try the checked in version with Linux to see if these link issues are OSX specific, I'd be grateful.
Cheers
George
Sorry about the late response people, the forum monitor was turned off for some reason.
Excellent work Fredrik, can I add this picture into the Pixie gallery ? Is the high rendering time due to raytracing soft shadows ?
I fixed the v problem on the curves in 1.3.17 which will be up next week. It is true that Pixie does not change u along the width of a curve. I find shading a curve twice along the width unnecessary because the curves are supposed to be thin anyway. Otherwise there would not be a difference between bi-linear/cubic patches and curves. However, if there is demand I can change the implementation so that curves are shaded twice along the width.
I will also add some new command-line switches.
George, I will look into using libri and libsdr in the same executable as soon as possible (which is probably towards the end of next week).
Thanks
Okan
Feel free to put the picture in the gallery.
The rendertimes is deffenitly due to both raytraced softshadows and occlusion. And i havent tweaked the sampling AT ALL ; ). I think it would be nice to be able to use the u coord on RiCurves, but i don't know what the RiSpec says about it....
Regards
Fredrik Brnnbacka
RiSpec says that u should vary across the curve - but it's seems a valid decision to not do this, depending on what the performance hit is. I'm not sure that I have a strong oppinion either way. It would be interesting of Okan could comment on the performance hit were it to be added.
George