ZBrushCentral

XSI 5.0 Wheeeeeeeeeee

Community
#6

last time i used xsi ( 4.2 ), the displacements in xsi were rather horrid to use.

the setting up of a shader to give you accurate results to what you had in zbrush usually needed 2 math nodes/ lots of tweaking. then sometimes it wouldnt like to read tiffs if you chose certain settings wherever you saved the tiff out of. sometimes it would just crash because mray felt like complaining.

then there was the bug where if you had subdivided your model once in the viewport, xsi would take that as 1 subdivision+ whatever you had in your subdivide @ render time settings, which would crash xsi entirely ( not just mray ).

so we had to resort to using the map format, which was a pain to convert to every single time ( especially when your doing tests ).

in renderman you just load in a map and press render. theres really not much else to it.

so anyway, if xsi has a mental ray version that has fixed its displacements ( because it has been an area that mental ray hasnt been really fixed at in a while, then i dont have much hope for it in terms of displacing geometry.

it can do it, but its a pain.

#7

In renderman you just load in a map and press render? c’mon there’s more to it than that.:rolleyes:

I’ve never had any problems with tiffs as long as converted to rgb they are ok
nor subdividing …there’s subdivision at rendertime(which is what you should look at) and subdivision in open gl noticed no bugs with them

Anyway…4.2 is about to be history…
check out the new videos in xsi 5 feature list … esp the import option for z brush and ultimapper video.:wink:

#8

There’s a new trender tree node for 4.2, ZTIFF.

Plug it into the materials displacement slot, set min and max sub-d level in geometry aproximation, adjust an amplitude slider and voila, painless 16Bit Grey displacements without PS conversions and math nodes.

Cheers
LemonNado

PS:Link is here in the community forum.

#9

Displacement in XSI 4.2 was basic… what are you guys doing wrong?

#10

Displacement in XSI 4.2 was basic… what are you guys doing wrong?

its not hard once you have a shader setup to something that works every time with very accurate results ( well youll still have to tweak displacement amount regardless), but its still very far away from where renderman is.

In renderman you just load in a map and press render? c’mon there’s more to it than that.:rolleyes:

well its not exactly how it goes, but it sure takes 90% less setup time.

set min and max sub-d level in geometry aproximation, adjust an amplitude slider and voila, painless 16Bit Grey displacements without PS conversions and math nodes.

the reason we needed math nodes was because we needed to first of all lower the displacement amount by using multiplication. this way we could get more accurate results without it blowing up on us. then we needed to use a division node for, i forget why exactly. i havent used it since december, but i know theres a good reason why we did the 2 math nodes, and they did give us a lot more control over the final product. not as simple as using amplitude, because mray needs to be setup to work with 50% gray and then some math to adjust it to work properly.

#11

Gee and I thought the process of rendering displacements in Maya/MR would be horrid.

Seems to be kind of the same thing. I’m happy to hear (again) that Renderman is obviously doing a very good job here - Another reason to look into the Renderman for Maya plug-In.

#12

Hmmmm I cannot picture that any renderer can simply take any displacement map and display the geometry exactly the way it looked in ZBrush. I can manipulate to much on the ZB side to believe that there is a ‘auto aestetics finder’ module :D.

And for real, creating the nodes in the render tree takes what, about 3 minutes. And once the tweaking (another 2-5 minutes) is done all is good for that model/material. Now with the ZTiff plugin it takes as much time as any texture takes to be assigned. Pretty much drag and plop.

Cheers
Lemmo

#13

just looked up that that ztiff loader. looks nice.

when i spoke of how we did displacements it was way before this node ever existed. the -0.5 shift it does for you, and allows you to use tiffs out of z without converting to rbg will surely make the process much cleaner.

Hmmmm I cannot picture that any renderer can simply take any displacement map and display the geometry exactly the way it looked in ZBrush. I can manipulate to much on the ZB side to believe that there is a ‘auto aestetics finder’ module

it doesnt. but think of how much this ztiff node helps you in your displacing, and thats all there is to renderman. plus take out ever needing to fiddle around with tesselating, because it always renders your geometry displacement at pixel level, and as fast as a bump map.

#14

I think I read somewhere that RenderMan is lightning fast there compared to other apps… Not a criterium for me as my renderfarm consisting out of an old 1.9GHz not-hyper P4 is making render time relative(sound of old record winding down)…
LemonNado

PS:I am very close to my Dual Dual-Core Opteron PC… Just a few more toy’s sold on ebay and I can order the bit’s. Yehawwwww.

#15

Jeez this got all drawn out.

As far as i know its waaay! more of a task to set up displacements in renderman with maps coming directly out of zbrush than setting up with mental ray in xsi.

It seems its just a matter of user prefs and what we all find works best for us.

Even though speed counts in the long run it is the quality of the work that counts …not the fact that it took 10 hours to set up a render.:stuck_out_tongue:

#16

I could not agree more regarding quality. I was only refering to my render times. Not optimizing… My render times are terrible 8-).
Lemmo

#17

I just fond out that Versio n5 has a ‘native’ way to import ZBrush Displacement maps. Hahaha. It’s getting better and better.

http://webrel2.softimage.com/Videos/NFT50/09_Import_Options_for_Zbrush.wmv?data1=a2e5efe115c1fda70233a5f2e4ca40bb&data2=437w3des1842025d445ss03998671wer45t1735562kfj7002445613dfd324h0838066kj8ff345x&data3=14917&ext=.wmv

LemonNado

#18

now thats the type of ease of setup i been looking for.

i wonder how the original model looked before they imported the map into xsi. to compare results.

#19

It’s here on the site. It’s in THE GUIDE. Looks exactly like that.
Lemmo

#20

i dont remember seeing it though.

the reason i ask is because in the video the final render had a lot of artifacts.
im wondering if the real model ( in zbrush, before it had a displacement map extracted ) had those too. thats why i ask.

#21

Is this true? That would make it handle more poly’s then z :eek:

XSI now handles the tenfold increase in detail demanded by next-generation productions – in tests handling a billion polygons on 64-bit systems.

#22

Check out the XSI videos. The guy manipulates and renders a scene with over a billion triangles. He was running on 4 CPUs with 8 gigs of RAM though. Having said that, rendering such scenes without crashing on slower systems is possible because of the scalable memory management.

#23

Regarding the artifacts, I saw those also and the ZB original doesn’t have them. But I could see the artifacts only in the realtime shader, the mental ray render view was free of them and was looking like a copy of the ZBrush one.

After seeing those V5 features I must say that my hotness towards the next ZBrush release has cooled of significantly as the V5 stuff will more than adequately compensate for the expected ZB features in sight. Of course this is a subjective thing based on my expectations.

Cheers
LemonNado

#24

Going to have to take your word for it about that video lemonnado, it tells me I have no access there. I’m registered there also so I don’t see why I can’t view it. Says access denied. I don’t own XSI…yet…so that may have something to do with it, though I can’t see why I couldn’t see a video showing a new feature.

#25

XSI makes an impressive coming back…

Since I read they put pannel layout looking like Maya, I wonder if I wont spend some time trying it…