Wet Map innacuracy

Discussion about fluid simulation in Realflow 4
User avatar
deckard
Posts: 51
Joined: Wed Sep 26, 2007 9:46 am

Wet Map innacuracy

Postby deckard » Sat Feb 13, 2010 8:20 am

So I got it in my head to attempt to recreate the Shining blood elevator scene, and this is the first time I have tried out wet maps. But the wet map is putting splotches in places where there are no particle collisions. Why? Am using an atlas UV in Lightwave. In RF the texture res is 2000, and otherwise default settings.

Here's an early, rough sim where you can see the problem on the walls and floor.
http://www.box.net/shared/static/65fpoj29n3.mov

The wetmaps work fine on a simple cube with an atlas UV. I tried subdividing the hallway into smaller sections of polys which reduced the degree of error. I have since tweaked the model and remapped the UV 11 times now, but the problem persists. Anyone else have this problem?

Below are the UV, and a top view of the floor problem. It seems odd because there are wet splotches corresponding correctly with particle collisions on the same polygon where other splotches are occuring with no corresponding particle collisions. How can it be simultaneously interpreting and misinterpreting?

Attached files


shaun_michael
Posts: 4766
Joined: Sun Sep 10, 2006 8:04 am

Wet Map innacuracy

Postby shaun_michael » Sun Feb 14, 2010 6:16 am

Usually this is a sign of a dodgy UV map.

I would steer clear of the Atlas unwrap in Lightwave and use a traditional unwrapping method either using the PLG tools, or something like 3D coat (i.e. something external). Plg tools work really nicely and there are some tutorials knocking about for it. The atlas unwrap is just a mess and likely to lead to the sort of unpredictable results tha you're getting.

Shaun

User avatar
deckard
Posts: 51
Joined: Wed Sep 26, 2007 9:46 am

Wet Map innacuracy

Postby deckard » Wed Feb 17, 2010 4:06 am

Thanks Shaun. It does not seem to be a problem with LW, but with how I am constructing the collision object and RF being a little finicky. I revised the object yet again and a low rez test looks correct. Will see what happens on a hi rez.

User avatar
deckard
Posts: 51
Joined: Wed Sep 26, 2007 9:46 am

Wet Map innacuracy

Postby deckard » Thu Mar 04, 2010 6:33 am

Have never used PLG, it doesnt seem to have any docs and I dont know what I am missing.

PLG UV generation keeps saying 'there are some charts that can not be developed' (?) and it then makes a mess.
http://img715.imageshack.us/img715/2760/plguv.jpg

LW atlas is making more sense
http://img43.imageshack.us/img43/296/lwuv.jpg

RF keeps inisting on putting splotches where there should not be any.
http://img638.imageshack.us/img638/2214/realflowwetmap7uv.jpg
Am I really the only person to ever run into this?

shaun_michael
Posts: 4766
Joined: Sun Sep 10, 2006 8:04 am

Wet Map innacuracy

Postby shaun_michael » Thu Mar 04, 2010 6:55 am

PLG UV generation keeps saying 'there are some charts that can not be developed' (?) and it then makes a mess.

I take it you've marked up the seams properly and everything? There's a tutorial on the Newtek boards (I think) that runs through how to use it. I'll see if I can find it and I'll post the link.

The problem with atlas UV mapping is that it rips the entire model apart poly by poly and simply lays that out into UV space as your vertex map - hundreds of seams and all! If you're using this method, it's no wonder you're having problems. If you can post up the model, perhaps someone else can do the UV map for you to get you going while you investigate the plg tools.

In answer to your last question about having problems with wetmaps, well yes, but only if the UV map isn't very good :)

Shaun

User avatar
deckard
Posts: 51
Joined: Wed Sep 26, 2007 9:46 am

Wet Map innacuracy

Postby deckard » Thu Mar 04, 2010 8:01 am

Thanks Shaun. Yeah, I found the PLG tuts and am finding it pretty tedious. Edges are refusing to select.

But I just dont see anything all that wrong with this UV. For instance, the biggest problem I am having is with the 'floor' I have selected here. Its all one continuous, flat section of the UV. What appears so confused about it?

Attached files

shaun_michael
Posts: 4766
Joined: Sun Sep 10, 2006 8:04 am

Wet Map innacuracy

Postby shaun_michael » Thu Mar 04, 2010 8:59 am

Here's a quick video showing you how Make UV (one of the PLG suite of UV plugins) works:-

http://www.youtube.com/watch?v=fHKjCD3xTTU&feature=related

You'll see a bunch of other related videos when you're on this page too.

LightWiki PLG Tute

Lots of useful info in this Newtek thread

Shaun

shaun_michael
Posts: 4766
Joined: Sun Sep 10, 2006 8:04 am

Wet Map innacuracy

Postby shaun_michael » Thu Mar 04, 2010 9:02 am

Its a little difficult to tell. Can you maximise the UV layout window and post a screenshot of that by itself, as large as the forum will let you post. Looking at it in this size, it does look like you have overlapping polys but it's impossible to tell at the res you get with a quad window display.

Yes, UV editing/creation is always tedious, but I assure oyu, without paying for a third party app, like 3D coat that has excellent UVing tools, not using PLG in Lightwave is far far more tedious.

Shaun

User avatar
deckard
Posts: 51
Joined: Wed Sep 26, 2007 9:46 am

Wet Map innacuracy

Postby deckard » Fri Mar 05, 2010 4:05 am

So here's a larger view of the LW atlas. There are no overlapping polys (all tripled, aligned, points merged) and it renders fine (other than wetmaps). I've carefully remodeled this thing several times. I dont get why RF has a problem with it.

It seems odd because all colliding particles have correctly corresponding wet map splotches. The problem is mainly with the floor where it adds a lot of extraneous splotches at some feet out in front of the actual particle collisions.

http://img638.imageshack.us/img638/2214/realflowwetmap7uv.jpg

Attached files

shaun_michael
Posts: 4766
Joined: Sun Sep 10, 2006 8:04 am

Wet Map innacuracy

Postby shaun_michael » Fri Mar 05, 2010 9:55 am

UV mapping can often be a pain in the bum (when it comes to Lightwave), but is a necessary evil . Your model looks pretty easy to unwrap properly and to illustrate a point regarding the difference between your UVs and what I would do, I have attached a comparison between an Atlas unwrapped UV and a cylindrical unwrapped UV, which is what your model would require - see attached pictures.

You can see that I have one seam in my map, hidden up in the top right corner of the mode, whereas your model (and my Atlas unwrapped model) has multiple, very visible seams.

Do excuse me if I'm teaching Grandma to suck eggs though. I'm just trying to say that if you want the results you'll need to start by sorting out the your UVing method and unfortunately, take the longer road.

Secondly, and I'm just checking through possible solutions, you're running the latest version of RF? You attached an rgb image to your model in LW, in the colour channel (having saved the model with the image map) before exporting the SD?

What's your substeps? Are the calculations accurate enough in RF?

The last possibility might be that the UV map isn't being exported correctly with the plugin - which I'm about to check.

Shaun

Attached files


Return to “RF4: Fluid simulation”

Who is online

Users browsing this forum: Google [Bot] and 1 guest