3delight tests

Plugins linking to this thread: (hide)

Blob StrokesAuthor: Anto Matkovic
Here's a small ICE compound used to generate these 'blob-strokes'. I had something like this before, but I really wasn't so excited by the speed of generation. Now with 3delight, it's definitely feasible, updates are performed immediately. The compound creates particles on curve list, evenly distributed, count is relative to particular sub-curve length (still, you'll need to take care of curve re-parametrization). It's also able to perform 'write on' on one-after-another subcurve (by sub-curve, or normalized by sub-curve length). In short, you can 'draw' 3d strokes with it, by animating the slider value. There are a few additional parameters. [..] contd. under si-community link

local backup: Emit Blobs From CurveList.xsicompound
3DelightAuthor: dna research
v4.0.41 updated Feb 21st 2015. 3Delight is a fast, high quality, RenderMan®-compliant renderer designed to produce photo-realistic images in demanding production environments. The renderer was introduced to the public in the year 2000 after being used for more than a year as the sole renderer in a sister production company. It is now widely used and earning a reputation as a benchmark in rendering technology. [..] (cont'd on product page)

DNA research recently released the 4.0.41 upgrade along with making the free license now support up to 8-core CPUs. Highlights from the changelog: Volumetric Smoke and Shards Any shader as light source projection Exposure, Gamma Controls Mesh Light support ESC to stop render. Last year's big 4.0 update introduced features such as: support for deep EXR files Native MARI textures support Improved sampling of environment maps Added the physically plausible 3Delight Material "Continuous rendering" is now enabled by default Up to x2 acceleration in hair/fur rendering Memory usage is down 30% when using the path tracer Skin shader for 3Delight for Softimage Motion Map property is now supported

A list of movie project rendered with the 3Delight renderer. As mentioned the product is free of charge for the first 8-core license. Additional quad-core licenses are available for 400$ each and unlimited multi-core licenses for 900$ each. Yearly support and updates 190$/290$ resp.


Comments on work in progress
Kzin
Posts: 432
Joined: 09 Jun 2009, 11:36

Re: 3delight tests

Post by Kzin » 29 Jan 2012, 12:43

1 raydepth reflection is mostly enough for wide glossy stuff, so this is really helpfu feature.
gi rendertime looks also really good, especially for dispmap. have you tryed to use this method for ibl rendering?
a coarse env geometry representation should also render faster with this method?

User avatar
Mathaeus
Posts: 1778
Joined: 08 Jun 2009, 21:11
Location: Zagreb, Croatia
Contact:

Re: 3delight tests

Post by Mathaeus » 29 Jan 2012, 15:57

Hi Kzin,

for getting the environment map into mix, still without activating the "Image Based Lighting" in global render options, yet another hack is required... Like in pic, at the end of xsi_material.h in RSL folder. Not sure is it the same for all SI versions, best way would be to take look at neighboring code for point-based IBL. In this config, "Image Based Lighting" should be disabled, otherwise IBL is sampled twice.

BTW this is a public forum, so I really need to say:

Make backup first, one wrong letter in *.h files could screw up entire renderer ! You doing this at your own risk !

Image

And that's with plain, blurred IBL, plus one standard light for testing purpose. For getting it faster it's always possible to use a very high shading rate for global
illumination. However, this forces a low mip-map resolution for textures which deals with shading normal (displacement, bump, possibly normal map), but still works fine with "color modulators" (diffuse, specular, reflection color...) - all that only in areas lighted exclusively by GI. Bellow standard light, global shading rate is used for everything. I think this is especially noticeable on skirt.
Standard light is just added to shading, no bounces here - because all "SurfaceColor3Delight" CAV properties are black.

Render time is about 3 minutes on Quad Core, global shading rate is 0.4, GI shading rate 32, 256 GI samples. GI fallof is 6. For me, still it behaves like standard AO.Samples/falloff ratio, level of detail of "target" geometry, this seems to be key, here. Displacement doesn't add that much into account, I think.

Image

User avatar
JPWestmas
Posts: 69
Joined: 04 Feb 2012, 22:35
Skype: Jason Westmas
Contact:

Re: 3delight tests

Post by JPWestmas » 16 Feb 2012, 04:11

Excellent! I really gotta try this out. I can't wait to get back into Softimage and thanks for posting these tests. How does 3Delight work with 2012, Any glaring issues?

Kzin
Posts: 432
Joined: 09 Jun 2009, 11:36

Re: 3delight tests

Post by Kzin » 16 Feb 2012, 10:08

Mathaeus wrote:Hi Kzin,

for getting the environment map into mix, still without activating the "Image Based Lighting" in global render options, yet another hack is required... Like in pic, at the end of xsi_material.h in RSL folder. Not sure is it the same for all SI versions, best way would be to take look at neighboring code for point-based IBL. In this config, "Image Based Lighting" should be disabled, otherwise IBL is sampled twice.

BTW this is a public forum, so I really need to say:

Make backup first, one wrong letter in *.h files could screw up entire renderer ! You doing this at your own risk !

Image

And that's with plain, blurred IBL, plus one standard light for testing purpose. For getting it faster it's always possible to use a very high shading rate for global
illumination. However, this forces a low mip-map resolution for textures which deals with shading normal (displacement, bump, possibly normal map), but still works fine with "color modulators" (diffuse, specular, reflection color...) - all that only in areas lighted exclusively by GI. Bellow standard light, global shading rate is used for everything. I think this is especially noticeable on skirt.
Standard light is just added to shading, no bounces here - because all "SurfaceColor3Delight" CAV properties are black.

Render time is about 3 minutes on Quad Core, global shading rate is 0.4, GI shading rate 32, 256 GI samples. GI fallof is 6. For me, still it behaves like standard AO.Samples/falloff ratio, level of detail of "target" geometry, this seems to be key, here. Displacement doesn't add that much into account, I think.

Image
looks great and render time is also really good. i hope i will have some time on the weekend to run some tests with 3delight.

RichardCulver
Posts: 34
Joined: 07 Jan 2012, 18:21

Re: 3delight tests

Post by RichardCulver » 18 Feb 2012, 10:12

Fantastic Thread. Actually a big part of what brought me to 3Delight.

I have a question maybe someone could answer about a shadow problem.

This is a simple scene to test interior with lots of dark areas. The area of question is there the skylight creates a pattern in the floor. There are no lights in the scene. Only an environment shader. The darkness and the lighting is intentional. I just want to know why this is creating a strange stepped edge around the light pattern on the floor.

Image

The particulars are here:

http://www.3delight.com/en/modules/PunB ... hp?id=3312

And here is the scene file:

http://dl.dropbox.com/u/15928817/3Delig ... DLight.scn

Newcomer (<20 posts) alert: please use the URL tags - HB

User avatar
Mathaeus
Posts: 1778
Joined: 08 Jun 2009, 21:11
Location: Zagreb, Croatia
Contact:

Re: 3delight tests

Post by Mathaeus » 18 Feb 2012, 17:19

Thank You for kind words!

Honestly, almost all stuff in this thread is created with 'make it faster' in mind, without too much worrying about physical accuracy. Forcing a whole sphere to cast a light through a small window, that's probably last thing I would try to do. In few pages above in this thread, I've used a few directional lights with shadows maps for getting a look of "soft spread". Not accurate at all, but fast.

If you're using Image Based Lighting, without activated GI, the look form this picture is somewhat strange, usually you get grainy noise with low sampling, not this.
What exactly happens, I really don't know.

For now, I think there are a three way for getting the IBL lighting:

1: Environment image/MIA sky, Image Based Lighting, GI disabled: in this case I think 3delight uses shader called Envlight2 plus ray-traced shadows from brightest part of environment image.
2: Environment image/MIA sky, Image Based Lighting, point-based color-bleeding enabled, photon mapping disabled: that's a point-based indirect diffuse, quite different method than 1.
3: Environment image/MIA sky, point-based color-bleeding enabled, photon mapping enabled: Photons are emitted form environment, finally they are gathered with point-based indirect diffuse. That's the only way which provides a multiple bounces.

The best example of photon-mapping in 3delight for SI I've seen so far, on vimeo

cheers

RichardCulver
Posts: 34
Joined: 07 Jan 2012, 18:21

Re: 3delight tests

Post by RichardCulver » 18 Feb 2012, 17:43

Thanks for the reply. I know this may sound strange to you. But I like the way my image looks as a basic test. I have my reasons. Not worth going into here. But I just want to figure out why the funny shadow. This should work - theoretically. If I can get this working then I have something to build on.

I saw that lighting test before. And it is definitely not the look I am after. So I am avoiding all of the usual things. As I said I have my reasons. Perhaps I'll go into it at a later time. As for IBL I really wish there was another name for that because it is really not what I am after. Just so happens that is what I have to enable to get the scene to light from the environment rather than a light. In LightWave for instance I can set it to Backdrop and just use flat polys to emit light if I want. Or in mental ray I think you can Use Final Gather with the same basic result. IBL usually makes people think about HRDI and ultra realistic lighting. Not what I am after.

Just want to figure out that pesky shadow.

User avatar
JPWestmas
Posts: 69
Joined: 04 Feb 2012, 22:35
Skype: Jason Westmas
Contact:

Re: 3delight tests

Post by JPWestmas » 18 Feb 2012, 19:42

It really looks like a shadow map that is too small in resolution Richard. But I think most renderers use lights to cast shadow maps but maybe not in 3Delight. Hope to be more of a help in the future. ^.^

RichardCulver
Posts: 34
Joined: 07 Jan 2012, 18:21

Re: 3delight tests

Post by RichardCulver » 18 Feb 2012, 20:23

Yeah, that is exactly what it looks like. But there are no lights in the scene. Well there is the physical sky shader, and it has a sun that is "casting" light. You can change the position of it and it will shine at different angles and effect the position of that shadow. So I tried settings there and none of them are for the shadow quality. So I am at a loss.

j3st3r
Posts: 121
Joined: 11 Jun 2009, 09:13

Re: 3delight tests

Post by j3st3r » 06 Mar 2012, 10:57

Hi

I tried the 3delight again, and I couldn't make nothing close to MR (2012 SAP). Unfortunately I couldn't have my fur look fluffy, shaded at all, the fur with the factory hair shader looked very flat, no volume, and shading at all, while with a blinn it was rendered correctly ( as blinn shaded strands)


I use Kristinka 3, used 3delight 3.1.0, standard hair shader. I had a spot light with shadowmap (4k res), and volumic shadow ON. I also tried the shadow trick Matheaus suggested, but in vain, no result. So I returned to MR but rendertime is extremely slow...

User avatar
Mathaeus
Posts: 1778
Joined: 08 Jun 2009, 21:11
Location: Zagreb, Croatia
Contact:

Re: 3delight tests

Post by Mathaeus » 18 Mar 2012, 20:03

Hi,

yet another of SurfaceColor3Delight hit mode. I call it
'SI community in waiting for SI release'

Image

User avatar
JPWestmas
Posts: 69
Joined: 04 Feb 2012, 22:35
Skype: Jason Westmas
Contact:

Re: 3delight tests

Post by JPWestmas » 19 Mar 2012, 00:11

Hi Mathaeus, really nice renders as usual. I particularily like this one because the reflections look really nice. In your opinion, after optimizing your settings, are these renders fast enough for animation projects with just a few computers to render with? Are we looking at one hour per frame or much more than that when using 3Delight for reflections that may or may not include GI.

Thanks for posting.

User avatar
Mathaeus
Posts: 1778
Joined: 08 Jun 2009, 21:11
Location: Zagreb, Croatia
Contact:

Re: 3delight tests

Post by Mathaeus » 19 Mar 2012, 01:20

JPWestmas wrote:Hi Mathaeus, really nice renders as usual. I particularily like this one because the reflections look really nice. In your opinion, after optimizing your settings, are these renders fast enough for animation projects with just a few computers to render with? Are we looking at one hour per frame or much more than that when using 3Delight for reflections that may or may not include GI.

Thanks for posting.
Thank you for kind words. Well this one is based on hacked MIA material, adapted to take this special 'SurfaceColor3Delight' hit mode. Finally I had to change the sl much more than I liked, on price of killing (possible) glossy refraction.
Back to question, imho all these tiny procedural textures (about four-six of scalar fractal or cell per material) won't play nicely in sequence. But I'd believe I'll find a way to keep the similar speed-quality ratio with a bit more baking. Render time for top one is 15 minutes on quad core machine and 4core 3Delight.
First time I've used 3delight, it was a small game trailer in time of XSI 6. That is, full of textures, already I had a baking workflow. Even the old Affogato translator gave a pleasant life. Affogato had no render-tree connection at all, I've played with external shader authoring tool. Anyway all that mess I made, rendered nicely, what I've set in evening, regularly appeared in morning.

User avatar
JPWestmas
Posts: 69
Joined: 04 Feb 2012, 22:35
Skype: Jason Westmas
Contact:

Re: 3delight tests

Post by JPWestmas » 19 Mar 2012, 02:03

15 minutes sounds pretty darn fast for precedurals and reflections. How would it not play nicely in a sequence? Or do you mean 15 minutes after baking the procedurals to texture? Anyway, this sounds pretty cool if it's just a matter of baking in the textures.

User avatar
Mathaeus
Posts: 1778
Joined: 08 Jun 2009, 21:11
Location: Zagreb, Croatia
Contact:

Re: 3delight tests

Post by Mathaeus » 19 Mar 2012, 02:17

JPWestmas wrote:15 minutes sounds pretty darn fast for precedurals and reflections. How would it not play nicely in a sequence? Or do you mean 15 minutes after baking the procedurals to texture? Anyway, this sounds pretty cool if it's just a matter of baking in the textures.
That I know, procedural cell, when is set to some really sharp 'flagstone' mode, could do visible jump to neighboring resolution level, with camera zoom or so. Bitmap plays sooo much better.
15 min is render time. Here, baking is MR AO to vertex colors, this takes a few minutes, even there is really small contribution of 'next bounce', supplied by baked vertex color.

User avatar
JPWestmas
Posts: 69
Joined: 04 Feb 2012, 22:35
Skype: Jason Westmas
Contact:

Re: 3delight tests

Post by JPWestmas » 19 Mar 2012, 02:27

oh wow, baked vertex colors. So it's kinda like Zbrush where you bake the vertex colors to bitmap?

Post Reply

Who is online

Users browsing this forum: No registered users and 32 guests