Annouce: Mental Ray Render Pass ToolKit 1.0 Released

[Bitter]

The hangup for Autodesk is preserving the workflow for Maya. Renderman and Vray are less concerned about that aspect. (Hence why Renderman fur doesn’t look like maya fur, they are not concerned with reproducing the look or workflow to a certain degree, they expect you to adapt for the benefits.)

I suspect if MR were a plug-in with a similar idea it wouldn’t be where it is now. It would do more of what we want but there of course would be hangups with some things not being translated correctly. (Like no fluids for Renderman.)

It’s a craps shoot on what gets fixed next. And I feel your pain, I am having to resort to render layers and horrible setup when I should be able to do framebuffers easier with a faster rendertime.

However, since I do not control the software we use at the studio, nor do I have the time to learn something new. . .yet, there’s no point in giving myself an ulcer.

Work does that for me.

But for now, people like you have at least graciously attempted to solve the issue instead of just complaining.

[MinaRagaie]

The only thing is it would be nice to have everything native to MR for those situations where you want to send it to a MR standalone renderfarm. But when my programmer gets back from vaccation next week he is gonna start on this stuff and I’m sure we can come up with something.

I’m looking forward for that

-Mina

[razorback]

yeah I forced to agree with you. most post defocusing methods are designed for the ramp to be one way and not ramp in and then back out. I think ZDOF is just more of a pain in the @$$ than it is worth.

The only thing is it would be nice to have everything native to MR for those situations where you want to send it to a MR standalone renderfarm. But when my programmer gets back from vaccation next week he is gonna start on this stuff and I’m sure we can come up with something.

I think the way ZDOF works that way is to make it easier for less experienced FFI operators who dont know how to remap the depth maps to generate post DOF. Ive personally never used ZDOF, but the sample images look pretty much like it.

Not too sure about creating a depth based shader in MR, it seems like your programmer would have to figure out how to rewrite the sampler info in MR.

Good luck with it. Let us know the results

[MinaRagaie]

Not too sure about creating a depth based shader in MR, it seems like your programmer would have to figure out how to rewrite the sampler info in MR.

not really…
I know in Renderman sl you got E (Eye position) and P (Position of the surface) as pre-defined variables. getting the distance between those is a matter of simple vector math.

I’m no shading expert, but since MR shaders are written in C, I bet it would take a couple of extra lines to do. E and P are quite basic and I’m sure there are similar variables in MR.

-Mina

[razorback]

I really doubt its as simple as that, as what you have mentioned is only eye position and position on surface. Depending on what you need to achieve, it could be as simple as distance between. The real challenge is transforming your location of your point in 3D space to pixel location in 2D image space.

I could be wrong tho’

[calmasacow]

I think the way ZDOF works that way is to make it easier for less experienced FFI operators who dont know how to remap the depth maps to generate post DOF. Ive personally never used ZDOF, but the sample images look pretty much like it.

yes it is very obvious in its nature but unfortunatly it is not what many of the post production filters that add this effect are expecting and thus they do not work correctly. In the images that you it is the compound blur being used and you get color bleeding becasue the the depth is not taken into account, and it can’t be becasue it is not correct. For the effect like Lens blur to work correctly the gradient has to be a one way gradient. And as a uuse point out there is another advantage to this in that if you change your mind later you can ajust it and change you focus zDOF you have to re-render becaue you rendered your focus range into the image.

I think the best choice in all honest is a MR version of the maya depth shader. possbilly though we should prolly inclue a script that will connect a locator to the camera that can be moded to set the far clipping plane and ther for adjusting the depth of the gradient. this will provide the affor mentioned “user friendly approach”

[MinaRagaie]

I really doubt its as simple as that, as what you have mentioned is only eye position and position on surface. Depending on what you need to achieve, it could be as simple as distance between. The real challenge is transforming your location of your point in 3D space to pixel location in 2D image space.

I could be wrong tho’

I’m not sure why you would need to transform the point into 2d space… or maybe I didn’t get what you said.
P is the point being sampled and E is the camera… if would try to write depth shader I’d get the distance between them and remap that to 0…1 range based on two user parameters “near” and “far”…

but as I said I’m no expert… I could be wrong too

I think such shader already existis among puppet’s shaders and the source code is also included. you can check that if you want

-Mina

[razorback]

yes it is very obvious in its nature but unfortunatly it is not what many of the post production filters that add this effect are expecting and thus they do not work correctly. In the images that you it is the compound blur being used and you get color bleeding becasue the the depth is not taken into account, and it can’t be becasue it is not correct. For the effect like Lens blur to work correctly the gradient has to be a one way gradient. And as a uuse point out there is another advantage to this in that if you change your mind later you can ajust it and change you focus zDOF you have to re-render becaue you rendered your focus range into the image.

If I’m not mistaken, it wasn’t the intention for it to be used as a depth matte for post DOF, for example : the depth blur in fusion, but for it to be used as an effects matte over a stadard blurred image.

    Matte --> Blur Operator / Node  --> Original Image

I think the best choice in all honest is a MR version of the maya depth shader. possbilly though we should prolly inclue a script that will connect a locator to the camera that can be moded to set the far clipping plane and ther for adjusting the depth of the gradient. this will provide the affor mentioned “user friendly approach”

Most backroom / desktop compositors would agree, but I find that even the normal depth mattes from the render layer default, set to 16bit output, is enough. No need for the fancy in Maya controls. Any comper worth his salt would be able to set white points and adjust gamma.

Peace

-K

[razorback]

I’m not sure why you would need to transform the point into 2d space… or maybe I didn’t get what you said.
P is the point being sampled and E is the camera… if would try to write depth shader I’d get the distance between them and remap that to 0…1 range based on two user parameters “near” and “far”…

but as I said I’m no expert… I could be wrong too

I think such shader already existis among puppet’s shaders and the source code is also included. you can check that if you want

-Mina

Hehe, I don’t think I explained clearly either

Anyway, what I originally ment was that having an E and a P as in your original post would not be enough information to simply write a depth shader, as E and P are still values in 3D space. The one more function that you need is how to project those values onto a 2D plane (the image that is rendered out). If this (generating eye rays based on the location of the pixel) has already been handled in the API, then its probably smooth sailing. If not, ouch…

Eitherway, I’m probably quite confused, and have hijacked poor ghostlake’s thread. I’ll stop yabbering now…

[sevver]

Hi
very very nice
but I need just like them for lunux and for maya 2008
ctrl-buffer isn’t for linux
shader_p isn’t for maya 2008 sp1
please help

sorry for my bad english

[MinaRagaie]

I’m probably quite confused, and have hijacked poor ghostlake’s thread. I’ll stop yabbering now…

ouch… me too… I didn’t notice that. I’ll stop here
Sorry ghostlake.
-Mina

[calmasacow]

Well the 64 bit version of ctrl_buffers is out so now it is just a matter of finding a replacemtn for zDepth. We have not had time to work on it here as we have seen a dramitic increase in our workload and are all working 80+ hour weeks. Currently we have been using FinalRender Stage-2 just becasue the pass system is already there and “just works” we are hoping to come pback to MR and get something working with this toolkit once we get our plate cleaned off a little bit and get to catch our breath.

[y-art]

but I need just like them for lunux and for maya 2008
ctrl-buffer isn’t for linux
shader_p isn’t for maya 2008 sp1

Hey,
shader_p is available for maya 2008 linux sp1 / ex2

http://www.puppet.tfdv.com/download/shaders_p_3.2b7_maya_linux64.zip

[MinaRagaie]

Hey,
shader_p is available for maya 2008 linux sp1 / ex2

http://www.puppet.tfdv.com/download/shaders_p_3.2b7_maya_linux64.zip

is that out for windows yet??

*edit: nevermind… just found it
http://www.puppet.tfdv.com/download/shaders_p_e.shtml

[ghostlake114]

A great news, Max from ctrl.studio have compile the 64bit version for ctrl.buffer, so my pack is 32 and 64 Bit Win now, mean much practice :drool:

The 64bit of ctrl.buffer
http://forums.cgsociety.org/showpost.php?p=5100007&postcount=767

[chuckie7413]

hey ghostlake114

excited by the fact that your render pass toolkit is now compatible with 64bit. I have been waiting for a while lol.

Downloaded and was following the install process but noticed that two of the shaders that need to be downloaded (JS_FlatColour and zDepthOF) have not been compiled for 64bit. Will this be a problem, im assuming it will be?

[system]

hi, when i render i got this:

API 0.0 error 301031: call to undeclared function “buffer_write”
API 0.0 error 301032: while defining camera “perspShape”: undefined shader “glbufferwrite”
API 0.0 error 301031: call to undeclared function “buffer_api”
API 0.0 error 301032: while defining instance “glcube”: undefined shader “glbufferapi”

how come? any ideas?
thanks

[ghostlake114]

hey ghostlake114

Downloaded and was following the install process but noticed that two of the shaders that need to be downloaded (JS_FlatColour and zDepthOF) have not been compiled for 64bit. Will this be a problem, im assuming it will be?

Oh sorry, I forget about JS_Flatcolour, the fact, this shader is for misss_pass only, so If you just want to extract pass from mia_material, just use my pack as normal. I will soonly make a n update to the misss_pass so it use another shader instead of JS_FlatColor, that is 32 bit.

About zDepth, it is a extra thing to create a custom zdepth pass for faking DOF,current version is 32, I have try to ask the author but he is not in the mood of making a 64 bit :shrug:

---

So currently, only mia_material is supported in 64 bit. But that is not so bad, I am using this pack day by day and it really save time, also the embed gamma of 0.454 in mia_material is fit for current MR rendering with linear workflow

API 0.0 error 301031: call to undeclared function “buffer_write”
API 0.0 error 301032: while defining camera “perspShape”: undefined shader “glbufferwrite”
API 0.0 error 301031: call to undeclared function “buffer_api”
API 0.0 error 301032: while defining instance “glcube”: undefined shader “glbufferapi”

You are missing ctrl_buffer, download and install it

[system]

i did install the ctrl buffer, i wonder if i did something wrong and if anyone actually managed to get it working on x64

[calmasacow]

I have never been able to get it to work just becasue there has never really been any good documentation or tutorials. there were a couple that come people had put on some wbesite but they are all gone now. Ctrl.studio are the kings of releasing unsupported stuff. who knows when we will hear from one of them again. I did read in the main crtl.buffers thread that several people have gotten it to work though.not sure about with the RPTK it is still missing the Zdof. wich perfonaly I have not use for as it’s format is incorrect for what nost compositing applications expect for a z pass.