Vray vs. C4D Physical Render Engine


#1

I have a few projects on the horizon which would warrant some upgraded render quality and workflows.
I recently ran some tests with the Physical Render engine in R13, but was stopped in my tracks when I discovered that there is an issue (apparently also in R14) with rendering out .RPF / .RLA files with z-depth and normal (etc.) data.

So…

I’m wondering whether Vray might be a viable alternative?

A large portion of the work I do is product visualization. And small products, at that - tabletop studio stuff.
Is Vray going to give me an immediate, and noticeable speed advantage over Cinema’s Physical engine?
Obviously, it can make some high-quality work - but so too can C4D’s render engine. I’m simply stymied by the lack of post control right now.

Can Vray render to such formats as .RPF and .RLA? Obviously, that’s critical to me.

Thanks. And I was hoping to post this at the Vray4mac forum, but it’s only accessible to Vray owners.


#2

What channels do you need from RLA/RPF that you aren’t getting from the standard multipass with other image file formats? z and normal are always available.

VRay is in general faster than Physical Renderer, especially at GI and blurry effects (blurry reflections etc). It’s got a great crew who offer excellent support on C4D. It’s multipass side though is somewhat different and as I recall nto as integrated, so if your’e using the bridges to e.g. after effects or nuke then it’s not quite as tight as the inbuilt engine in terms of dsending over all the passes, however as your’e using RLA/RPF I assume your’e using an older compositor that’s not listed among the standard bridges.


#3

Thanks for the reply.

Specifically, I’m using Z and normal channels in After Effects. So, having those in the .RPF files are mandatory.
Can you elaborate a little on what you mean by “always available”? Rendering a depth pass is not the same as the z-depth data contained in a .RPF image file / sequence. But perhaps I’m just not understanding what you mean.

And what is a “standard bridge”?


#4

How is the depth channel not the same as the Z channel? What can you do with Z that you can’t do with a depth pass?


#5

There are a number of plugins which specifically take the z-depth data, and allow you to manipulate parameters to an unlimited degree. With the basic “depth” pass export, you’re stuck with the settings you established with the camera in C4D. Additionally, that depth pass is a smooth, anti-aliased image. But that does not work for certain post effects (like with certain DOF plugs). Finally, it’s an extra step, and an extra asset to manage. As opposed to a single .RPF file - for instance. And not just for depth. You can have the normal data included, and then use that with a variety of 3rd party plugs in AE. No need for separate passes.

But the z-depth data is the biggest.


#6

I reported the broken RPF/RLA bug quite some time ago, but it seems to have fallen by the wayside.

Ideally we’d be able to just render the appropriate passes (correct z depth etc) to a multi channel exr (better fidelity, lossless compression) but the multichannel EXR stuff is a niche enough need that there are issues with that workflow too. Specifically the blend channel only stores luma information for some reason, so you can’t use it with them. That one has been reported for a while too with no lovins.


#7

Still not sure I’m getting it. You aren’t “stuck with” anything if you set up the depth pass correctly, and it works with DOF plugins. AFAIK the depth pass provides the same Z-depth information as the RPF Z depth pass. You can also output a separate normals pass if needed. Granted, it may be less convenient, and the bug should certainly be fixed.


#8

VRay’s multipass will be significantly overhauled in the upcoming update so I’d be weary of basing any decisions on the current version.

I don’t know how well used RLA or RPF are anymore (I can’t get Nuke to see their layers)?
An EXR depth map that covers the scene gives you the flexibility to apply DOF to any object and adjust the focus to any point. As part of a multilayered EXR we find it works very well (from VRay at least, we don’t tend to use the standard renderer).


#9

You’re not getting it, but it’s a little hard to explain. The Z-depth data in the .RPF file stores all the raw data necessary to shift focus (in post) from the film plane of the camera, all the way to infinity. You can shift that at will, depending on your post needs - and particularly based on client insanity. Because that channel data is non-anti-aliased, it provides the mandatory foundation for certain DOF (and other) effects. I repeat, that hard-edge data is mandatory. You can NOT just use the depth pass out of Cinema. The nice anti-aliased edges will wreck the way those edges are solved in the filter. So, it’s a moot point whether the depth pass is set up correctly in C4D. It simply isn’t going to afford any flexibility in AE.

Here’s another simple example. Try using your depth pass with the built in “3D Channel” plugins in AE. It’s not going to work. Those plugs entirely rely on z-depth data, normals, etc. which one gets from a .RPF file.


#10

I just tried an OpenEXR frame from Cinema. That seems to contain no extra info, whatsoever. AE doesn’t see any depth, normal, etc info. Am I missing something with that?


#11

I’m not sure that AE reads multilayered EXRs (seriously Adobe!) although I believe there are ways of doing it.
Nuke is obviously fine with them.


#12

You have to use a plugin with After Effects.

And yeah, Cinema’s Depth pass clips wherever the DOF settings on the camera dictate, and by default they are usually nowhere useful. Really needs to be fixed.

Here is the AE plugin.

http://www.fnordware.com/OpenEXR/


#13

Thanks. EXR looks quite cool. So, it essentially combines all the value of the included data of an .RPF file, but also 32bit color, AND high-degrees of compression? What are they talking about with the “fully layered” render file? I don’t see any special options in the C4D render settings for EXR.


#14

What AE plugins are you using for this?

Thanks!


#15

Unlike the RPF where it has built in options for what type of data, EXR then uses the multipasses that C4D generates. Your going to be unhappy because of the antialiased edges again, but that’s also a flaw in those older filters, particularly anything dealing with depth, because depth is the easier blur to handle an acceptable result with transition pixels than say motionblur. No matter what if your doing post one of those blurs is going to have to come after the other, and motion blurring first is always going to be the ideal for minimal discernible artifacts/inaccuracies. If you depth blur system can’t handle an AA’d image then it certainly won’t handle a motinblurred one and then what will you do?

That said I do hate that we still have no control over which passes have AA or not.


#16

Yeah, but also to base you decision on Vray update coming any time soon. It’s been a long wait already, and there is no date that has been set. Could be another year for all we know.


#17

“Could be another year for all we know.”
no, even though no public date is set, this is not very likely. - it was much work indeed, but we have most of the work done now.
stefan


#18

Cant wait! Thank Stefan


#19

Is there going to be a demo version available? I need to try out the render system in my normal workflow, in order to see whether it will provide what I need.


#20

Have you tried using the Zblur2 plugin to generate the depth map? It does a much better job than Cinema and the Frischluft Lenscare plugin for AE lets you “Shift depth focus at will” using it. I’ve used it on a lot of projects creating shifting focus from foreground to background and hard edges haven’t been a problem. It does create an extra asset to manage but it’s much better than nothing.