fr2 GI animation test

[Ernest Burden III]

The biggest problem I've had with this test was keeping the host app, and especially the DR slaves alive. There were many, many crashes of one or all. Getting this test rendered required much babysitting of the computers and more time than I had planned.

 

This is not a final treatment, and I would be filtering the results a lot. I have some NPR treatments in here, more in Photoshop later. I'm not sure all the things that normally work in AR2.5 are coming out the way I expected, so I have to study this result and maybe adjust my scene.

 

The frames ranged from about 2 minutes to 10 minutes (prepass + render) depending on how much glass was in the frame.

 

http://www.architecturalvisions.com/temp/Webs-01.mpg

[leed]

Ernest

 

I must say, there is something quite distracting about the animation. I think it is the grunge layer, that is still over the top, of the animation.

There is heavy flickering and bad antialiasing on the main elevation of the building, is this just because of the compression?

 

When you get close in it is easier to se the building and the quality of the render comes out,

There also seems to be some transparency, alpha map problems on the trees, something like the ray depth/ transparency depth in AR.

 

I know this is just a test, cant wait to see the development.

 

Lee

[Ernest Burden III]

Ernest

I must say, there is something quite distracting about the animation. I think it is the grunge layer, that is still over the top, of the animation.

 

It's not really on top, its more underneath. I've been playing with the timing on that noise, it has an animation speed. I went with 10fps on this one, where I usually do 30fps. But its so static that its one of the things I have to verify even got done by fr2. I had expected more variation in it.

 

There is heavy flickering and bad antialiasing on the main elevation of the building, is this just because of the compression?

 

I almost never use AA, didn't here. There isn't GI flickering, which was the issue I was looking at. This was rendered in a random stop-start (due to render crashing) way on one, two and three PC (again, almost random) Doing that with AR would have produced a real mess of the GI. In my mind the GI for animation method of fr2 passes the test. It would be easier to tell without my noisy stuff, but its good.

 

There also seems to be some transparency, alpha map problems on the trees, something like the ray depth/ transparency depth in AR.

 

Yes, quite the issue. fr2 seems to have a max ray depth at something like 30, so I'm worried that could be a problem with a lot of glass - or tree alphas. I will have to test on some of the frames that show that issue and see what can be done.

 

Thanks for the comments!

[Brian Cassil]

I almost never use AA, didn't here.

 

I'm curious to know why?

[STRAT]

the lack of AA is a massive downside of this render

[Ernest Burden III]

the lack of AA is a massive downside of this render

 

Lots of horizontal bands in perspective. Yeah, it shows.

 

I usually do not use AA because of the time hit. My finals are processed to add noise and that helps. I have to look into how to introduce some AA, but to test the GI it wasn't necessary.

 

The sky is just a gradient, by the way. Another thing I was playing with here.

 

But its the GI, which worked well, that was the real test. I am completely satisfied with how it performed, just not the program itself.

 

Here's a filtered version:

http://www.architecturalvisions.com/temp/WebSci-WC-01.mpg

[Frosty]

I have to look into how to introduce some AA, but to test the GI it wasn't necessary.

Ernest, try rendering at double the resolution and then scaling down in post. I'll bet it renders faster than using AA and it will look the same.

[Ernest Burden III]

try rendering at double the resolution and then scaling down in post.

 

I've thought of that. Another thing to try. Good thing I'm 'between projects', then. Who needs to earn a living?

[Ernest Burden III]

Yeah, this is a bug.

 

In Cinema you have noise, and one of its parameters is 'animation speed'. Set to higher than 0 and the noise will 'dance' randomly. That's what I want it to do. In fr2 it isn't.

 

Here's fr2:

http://www.architecturalvisions.com/temp/fr2animtest02-dif30fps-.mov

 

Here's AR2.5:

http://www.architecturalvisions.com/temp/AR2animtest02-dif30fps-.mov

 

 

As you can see in my scene test, when the background noise just sits there it looks awful. So to use fr2 I would have to workaround this bug which can be done by rendering without that noise on and adding it in After Effects. But another reason that fr2 isn't doing what I need...I didn't need.

 

And here's what 10fps noise looks like, this is what I was after:

http://www.architecturalvisions.com/temp/AR2animtest02-dif10fps-.mov

[leed]

Ernest

The thing I like about your work, apart from the lighting, and texturing..... oh and the proportioning of the images

is the line quality you get on your stills, nice rich lines, it works with your style really well.

 

And I think that is missing from the animations, there is not the same definition of line and light,

 

I imaging that you do a lot in post, Photoshop blends and filters, or is it all in the render?

 

I have had a little play during my lunch and what do you think of this?

 

http://www.leedavid.co.uk/movies/0Comp1.divx

 

hope that works.... save file to desktop and get rid of the.tex.... 380kbs

you may need the VLC media player to view.

 

 

http://getfreedownloadz.com/index.asp?PID=bccbcc47-a141-4d5c-85cf-5865ca4519a0&sft=34

 

All done in post though AE, I think you have got much more control with the multi layer approach.

 

Lee

[Ernest Burden III]

Ernest

The thing I like about your work, apart from the lighting, and texturing..... oh and the proportioning of the images

is the line quality you get on your stills, nice rich lines, it works with your style really well.

 

And I think that is missing from the animations, there is not the same definition of line and light,

 

I have a seperate process to generate edges in the animations. I can do lines just like stills, but they tend to distract by either becoming 'objects' or vibrating. I found a more subtly approach worked pretty well, without those issues.

 

I like your test, though the saturation is a bit hot. The effect is pleasing.

 

I was looking at C4D mats to do most of what I'm doing in post, here's an example. All mats, no post:

 

http://www.architecturalvisions.com/temp/AR2-WCmats-01fps-.mov

[leed]

Yes the materials are nice, I am still not convinced about the noise over it all mind, a bit to distracting for me.

 

Do you use fresnel to get the edge?

Have you tried Sketch?

Quick test...

 

http://www.leedavid.co.uk/movies/sketch.divx

 

The thing is I can get quite rapped up in tests.... and I have work to do.....

 

Lee

[Ernest Burden III]

Do you use fresnel to get the edge?

Have you tried Sketch?

 

I love S&T, but it's too slow.

 

I did not use Fresnel (except in the reflective mat). I'm trying to only use mats and shaders ans such that will work just as well in fr2.

 

I'm not under a deadline for a change, so I'm trying out some techniques.

[AJLynn]

This is going to sound completely crazy, but does anybody know if I can get it to save the GI prepass (preferably with the green and red dots) as an image? (I can screencap it if my timing is good enough but I want it in higher res.) I've got this one image I'm working on where the prepass is perfect and the render screws it up completely

[Ernest Burden III]

I can get it to save the GI prepass (preferably with the green and red dots) as an image?

 

Sure, easy as could be.

 

When you are rendering the prepass, if you have C4D set to save images (the usual way, TIF for example) it will save the dots image. So if you set it to camera animation/prepass but only have one frame in the range, you will save that image--or all of them in an animation.

[AJLynn]

Sweet. Thanks.