3Delight for XSI is out

[Leonard]

DNA Research has released 3Delight for XSI.

3Delight for XSI is a fully integrated plug-in for SOFTIMAGE®|XSI®. The plug-in uses the latest rendering API advancements in XSI 6.5 to achieve a seamless integration and without any compromises in work-flow. Among the most noticeable features (http://www.3delight.com/en/index.php/products//3delight_xsi_features):

XSI shaders are automatically converted to RenderMan® shaders and compiled on the fly
Integration in the render region and support for most output formats
Full geometry support (including smooth subdivision surfaces, NURBS and curves)
RIB export and import
Advanced pipeline options such scene caching for rapid lighting and networking caching for render farm rendering>> Check it out << (http://www.3delight.com/en/index.php/products/3delight_for_maya/3delight_xsi_overview)


http://www.3delight.com/en/uploads/images/skins/gallery/car_3delight2.jpg

http://www.3delight.com/en/uploads/images/skins/gallery/fly.jpg

[Intuition]

I have seen the PRman GI renders when being used with XSIman.

see near bottom of thread....

http://www.xsibase.com/forum/index.php?board=12;action=display;threadid=20306;s tart=255

Any thoughts on getting 3delight to expose the PRman GI controls?

[Sosoyan]

There are a few things, without wich we can't use your product yet. There are`

1. Motion blured objects are reflected BAD
2. No area lights
3. Texture transfered on a hair, is not rendering (BAD BAD BAD)
4. Hair instancing is not rendering
5. AO is noisly and takes mutch time then MR
6. ...

We hope you'll correct that very soon, Thank you! :sailor:

[Adrian Lazar]

From the demo videos (I'm still waiting for a 64 bits demo) it looks incredibly fast in the areas that mr is slow (displacement, mb and dof) so i think that can be used in production but maybe not for every project as your main render... that is if you/the studio ca afford buying a secondary rendering solution.
I just hope that what sosoyan pointed out will be corrected asap.

[Leonard]

There are a few things, without wich we can't use your product yet. There are`

1. Motion blured objects are reflected BAD
2. No area lights
3. Texture transfered on a hair, is not rendering (BAD BAD BAD)
4. Hair instancing is not rendering
5. AO is noisly and takes mutch time then MR
6. ...

We hope you'll correct that very soon, Thank you! :sailor:

I'm not speaking on behalf of 3Delight as it's not a product that I'm involved with... but I can answer that it's really a renderer at the moment that complements MR in XSI. 3Delight renders pixel for pixel accuracy with MR, so it can be rendered out as another pass which can be comped in later.

Point being, yah it has some deficiencies compared to MR, but some of the things it does very very well and will save productions a lot of time and frustration. MR also does some things very well, and is a great all round renderer.

Use whatever works for you. :)

Leo

[wurp]

It's true that 3delight complements mental ray but even if the renders anti-aliasing will match up you should think twice about mixing the renderers as exactly the things that 3delight are good at will not work when comped with mental ray renders, a pass with displacement, motionblur and dof from 3delight wont match up with a pass using the same effects rendered in mental ray so that kind of kills the point. I guess the best way to mix them is to bake as much as possible of the things that require heavy raytracing using mental ray and then render the final stuff in 3delight.

overall i have to say im very impressed with 3delight for xsi, it even beat the maya version in terms of integration so far and im sure it will get even better in the future.


I'm not speaking on behalf of 3Delight as it's not a product that I'm involved with... but I can answer that it's really a renderer at the moment that complements MR in XSI. 3Delight renders pixel for pixel accuracy with MR, so it can be rendered out as another pass which can be comped in later.

Point being, yah it has some deficiencies compared to MR, but some of the things it does very very well and will save productions a lot of time and frustration. MR also does some things very well, and is a great all round renderer.

Use whatever works for you. :)

Leo

[mi_echo_log.txt]

FYI - dna research : the developers of '3Delight for XSI' have a dedicated forum where you can ask specific XSI/3Delight technical related questions.

http://www.3delight.com/en/modules/PunBB/viewforum.php?id=10

[aghiles]

There are a few things, without wich we can't use your product yet. There are`

1. Motion blured objects are reflected BAD


Just enable the ray-traced motion blur (in 3Delight's object properties) OR use deep shadow maps (volume shadowmaps)


2. No area lights
3. Texture transfered on a hair, is not rendering (BAD BAD BAD)
4. Hair instancing is not rendering
Won't take ages to implement.


5. AO is noisly and takes mutch time then MR
I get the same noise level than M.R, could you send a sample scene or image with comparable quality settings ? Regarding speed: unplug the shadow material slot to get much faster renders. We are working on a fix for this.


6. ...
Yes ? :)

-- aghiles

[Sosoyan]

Thank you about AO, but actually now, its primary advantage is not ability to render texture hairs, because of a rare unrelated uses only one-dimensional hair color, or instansing, which is not yet rendering too, but anyway thank you DNA vor developing 3delight for xsi :)

[wurp]

I've heard from several places that point-based occlusion is much faster than raytraced one, is there any way we could use this method in xsi with 3delight? I've also heard that you can use the same technique for glossy reflections, is that true? Apparently point-based glossy reflections can be less accurate but i find most of the time you just want to blur the reflections slightly, not have them super-accurate.

let us know

[aghiles]

Hello,
Windows 64-bit and Linux-32 bit demos are available for download:
http://www.3delight.com/en/index.php/products//3delight_xsi_download (http://<a href=)

Thanks,

-- aghiles