PDA

View Full Version : LW10 up and runnning...



Maxx
01-10-2011, 06:42 PM
Big thanks to Lauren Tate and Richard Jacinto (Lauren rocks the customer service and Richard got me access to the HC forums in under 24 hours - these two rock, NT). :thumbsup:

Now that it's running however, I've noticed something. Weighted CC sub-d's are still broken in 10? Is anyone getting proper results while knifing or mirroring weighted CC sub-d's? Please tell me I'm doing something wrong here... please.

kfiram
01-10-2011, 07:05 PM
Still not working.
Should work in Core though...

Sensei
01-10-2011, 07:11 PM
CC edge weighting is not broken, and it never have been broken. But every single tool must be rewritten to support them.. Do you imagine rewriting hundred tools? Have no sense when they have to work on Core..

EasySplit is the only tool which handles cutting geometry when edge weighting was used.
Here is video
http://www2.trueart.pl/Products/Plug-Ins/EasySplit/Graphics/Movies/EasySplit_Edge_Weight_Preserved_1.mov

Maxx
01-10-2011, 07:35 PM
CC edge weighting is not broken, and it never have been broken. But every single tool must be rewritten to support them.. Do you imagine rewriting hundred tools? Have no sense when they have to work on Core..
The implementation is broken. I realize that every tool must be rewritten to account for the algorithms of Catmull-Clark type subdivision surfaces, but there are white papers all over the Internet on the topic. And while I certainly don't doubt that it's both time-consuming and difficult, if you're going to tout the inclusion of the technology, make sure the technology is included. Right now it's not.

EasySplit is the only tool which handles cutting geometry when edge weighting was used.
Here is video
http://www2.trueart.pl/Products/Plug-Ins/EasySplit/Graphics/Movies/EasySplit_Edge_Weight_Preserved_1.mov
Unfortunately, most of my work has migrated away from 3D over the last two or three years, and I was scraping just to justify the upgrade. I got it, and I'm not complaining (though it may sound like I am), but this was a part of the 9.0 program that I purchased that still is not in usable condition now that we've reached the next ordinal release. Honestly, I'm a bit baffled and amazed by that.

Had I the budget, I would purchase EasySplit. At the same point, NT touted LightWave 9.0 as an off-the-shelf suite that included Catmull-Clark subdivision surfaces, edges, and edge weighting. This implies that a user can actually use CC sub-d's in a production environment by doing basic mesh manipulation such as mirroring or knifing said mesh, even if it happens to be after application of weighting. We can't as of version 10.

Sensei
01-10-2011, 07:55 PM
White papers won't help at all. What tool does when it's cutting geometry? It's making point in half, and then making two polygons, and deleting old one. Every single tool modifying geometry is doing it this way. This is even more complicated while dividing polygon to more than 2 new polygons. EasySplit needs 5000 lines of source code to handle everything.

Edge weighting and discontinuous UVs are stored in per-polygon vertex map (which means vertex map is either for point and polygon).. Polygon is gone - no more weight map attached to it..
How they could know while writing it in some '80 years they will need what's needed now..

Maxx
01-10-2011, 08:39 PM
White papers won't help at all. What tool does when it's cutting geometry? It's making point in half, and then making two polygons, and deleting old one. Every single tool modifying geometry is doing it this way. This is even more complicated while dividing polygon to more than 2 new polygons. EasySplit needs 5000 lines of source code to handle everything.
As I said, I'm not doubting that the work is difficult and time-consuming. I'm raising issue with the fact that we (I) were (was) told the technology was completely incorporated into the program. And it still isn't. Belive me, I have nothing but respect for what you've accomplished with regard to this issue.

Edge weighting and discontinuous UVs are stored in per-polygon vertex map (which means vertex map is either for point and polygon).. Polygon is gone - no more weight map attached to it..
How they could know while writing it in some '80 years they will need what's needed now..
And the problem there is that it wasn't written 80 years ago, it was written 4 - maybe 5, my LW9 CD is dated 2006 - years ago. And CC technology had been around for many years before that. Since 1978, according to Wikipedia.

lwaddict
01-11-2011, 07:48 AM
?

Let me see if I understand this correctly then...
according to Sensei...

If I buy a drink at a bar touting it's amazing mix of hard liquor...
pay and then get a drink that's missing the alcohol...
I'm supposed to buy another drink because I should understand how
tough it is to make the liquor?
Really?
:bangwall:

I could have come up with a better metaphor, but it's almost happy hour. :)

MaDDoX
01-11-2011, 08:53 AM
Is anyone getting proper results while knifing or mirroring weighted CC sub-d's? Please tell me I'm doing something wrong here... please.
Personally I've never weighted edges before getting the finished model 100% set in stone, not even in Modo, for multiple reasons (mostly export- and compatibility-related). In any case, just out of curiosity, is there a way / lscript / plugin that allows the conversion of edge weighting to an vertex map, say, a weight map in Lightwave? If that was possible one could theoretically use the weight map operations to preserve, mirror/edit the maps then convert it back to edge weighting after any mesh-changing operation is performed.

Just a thought.

Sensei
01-11-2011, 09:03 AM
Edge weighting is much more data than weight map in just points. Imagine you have quad poly- that has 4 points, then edge weighting is inside of all 4 points in this polygon. If you have array of 4 such quads, point in center will have 4 edge weighting data, because it's used by all 4 quads. Points connecting to it, will have 2 edge weighting data (point used in 2 polys), and 4 points on corners will have just 1 edge weighting data (point used in 1 poly).
The only way would be unwelding everything, and then move per-polygon vmap to vertex map, then do operation and reverse this process..

hunter
01-11-2011, 04:37 PM
Big thanks to Lauren Tate and Richard Jacinto (Lauren rocks the customer service and Richard got me access to the HC forums in under 24 hours - these two rock, NT). :thumbsup:

Now that it's running however, I've noticed something. Weighted CC sub-d's are still broken in 10? Is anyone getting proper results while knifing or mirroring weighted CC sub-d's? Please tell me I'm doing something wrong here... please.

I hear ya. It was promised in 8 wasn't it? Still not really useful unless you can model without mirroring or slicing.

Maxx
01-11-2011, 07:50 PM
I hear ya. It was promised in 8 wasn't it? Still not really useful unless you can model without mirroring or slicing.
I thought it was 9.0, but I could be wrong. Either way, this is way overdue...