max transparency RLA

[BrandonD]

No, the Wavefront RLA format does not support a transparency buffer, but RPF does.

[CHRiTTeR]

Isn't RLA a format devolopped by Kinetex (formar makers of 3ds max)?

I was just wondering ... :shrug:

[baby]

that means not only no DOF but you can't also put a 3D fog in after effect if you have transparent objects...

Sh.t !!!!!!!!!

Thanx !!!

[BrandonD]

Actually the RLA format comes from Wavefront. Kinetix and Discreet Logic initially used it to add communication between MAX and Paint/Effect. Then Discreet (the two companies merged) created a more capable format called Rich Pixel Format RPF.

[baby]

by the way,
very impressive demo Brandon !!!

:beer:

[ilasolomon]

baby, i don't know if the RPF does what you need or not
but if you want to do it i recommend you hide the helmet's glass
& render out the z-buffer to a regular rgb format like
tiff then use this as z-depth source in AFX.

[BrandonD]

He's correct. There's definitely the "elegant way" of doing it vs. the "make it work" way. Usually the former is best for demos and the latter for production...hehe

[baby]

that's what I was doing...


render without the helmet

use this image with the Z-depht in after


then render the helmet with all the lights but the rest of the scene in black...
then composite in screen mode with its own zdepht for the helmet DOF...


a bit complicated for a simple stuff, not sure it will work...very stupid !!! :(((



something I don't understand...if you render as a separate element the zdepht in max...he will calculate a Zdepth with an antialiasing !!! :surprised wich is not normal, and which creates problems in After !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

[ilasolomon]

if you render as a separate element the zdepht in max...he will calculate a Zdepth with an antialiasing !!! wich is not normal, and which creates problems in After

:) indeed, that's very stupidly...it is useless & just helps for
preview of z-depth
but do this:
make everything black or a simple material & turn off shadows
& reflections & atmoshpere & ... (for speeding up the rendering)
now render out the scene to RLA format , then in AFX, import
these footages & apply effects>3d channel>3d chanel extract
& adjust the black & white point to achieve the proper good
looking z-depth image, now render out these seperated z-buffer
& use them!

[baby]

hopefully I don't have reflections on the helmet...

I would have been...f...d...




another stupid thing...

when U render a RLA file, you have a a gradient from white (first thing in your scene) to black (far thing in your scene).

when U render as separate element you have to do it (give arange) manualy !!! :( :thumbsdow (well it can be usefull, but if you forget !!!!!!!!!!!!!!!!!!!!!!!!)

another thing I don't understand !!! too bad !!!! when U render a RLA and watch it in max, you have as I say a perfect white to perfect black gradient !!! basic...already done...alone...

Why the f...do I have after that to adjust it in after FX !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! it's not hard...but it's not very precise !!!!!! or too precise !!! why can't it take what max show me !!!! pffff

arggg !!! grrrrrr

[baby]

I have an Idea !!!!

I could make my own Z-depht !!!! :)))

I put everything white !!!!! I keep my transparency objects !!!!

and then I create a black fog...that start from 0% and finish at 100% where I want my zdepht to finish...
and I desactivate for rendering the light AND !!! the antialiasing !!!

what Do U think of that ?!? could it work ?!?

[BrandonD]

Yes, use a Falloff Map with DistanceBlend turned on. You could put it in the Self-Illumination channel or Opacity channel.

BTW, thanks about my reel. I missed your earlier post.

[ilasolomon]

baby, you don't notice!...there in no logical reason that
transparency could be in z-depth channel!
the z-buffer tell us that what pixel is in front of another & if you
render the glass transparency in a 'fake z-depth' image then
you'll get the wrong data about depth of the glass' pixels
try & see yourself! :)
btw, if you need such a thing, follow my giude, hide only the
glass & render the z-depth only footages...or use RAYTRACE
DOF that is very slooooow!

[baby]

my test with fog seems to work very well...I'll see the result tomorrow...

my test is for lenscare demo...think U know it ;) which doesn't work like combustion or after DOF.

if you think of it...when a pixel is transparent, you get the Z-buffer colour of the object behind with my technique...
so what is going to be blured in the final frame is this pixel behind...and not your transparent pixel...


My helmet is opaque/almost white only on the fallof side...So lenscare, if my focus is on the helmet will let the helmet side sharp, but will "gradiently" blur the pict when you leave the fallof side...

if my test is ok I'll put a pict here...but with the demo noise it's a bit hard to see clearly...


if it's not I'll see for the other solution...

[baby]

an important thing to do is to put an exponantial black fog...(which means change the camera parameters)

if not the glass will attenuate/darken the object inside and behind !!!
which is not good at all, because it would be like if the head was much more further than the helmet around !!!! :((

now after this hours of intense search ZZZzzzzzzzzz

[baby]

to find how to render fog and volume lights with this technique...


a volume light in front of a far background should keep its sharp edges even if the background is very OOF :surprised (hum out of focus)

[baby]

the result is not too bad, it could be enough for me, much better than if I didn't do it...but it doesn't seem to work well as I though...
The blut behind the helmet is good, but the helmet is not blured correctly everywhere...

I'll try the second solution...(I wanted first to find another solution as my sequence was already calculate a long time ago, so now I would have to calculate it again separating various elemts)

[electricfx]

baby, you might want to try assigning the helmet a unique object id, then rendering the scene twice: pass#1 would be RGBA with the transparency turned on, and pass#2 would be RGBA + Z Depth with the helmet made opaque.

Then in AE you can extract the Z Depth from pass #2 for pixels matching the unique object id only, and use the resulting image map to drive Lens Care or other lens effects applied to pass #1. That way the lens effect will hit the helmet but nothing else.

In AE you can do the extraction with a combination of the Pro Bundle ID Matte and 3D Channel Extract plug-ins - or you could try our new eFX Extract3D plug-in, which I think is easier to use and often does a better job (totally biased opinion, of course ;))

Regards,

john

:: John Crawford :: info@electricfx.com
:: Use 3ds max 3D channels in After Effects with eFX Extract3D
:: Limited time intro price $65 at www.electricfx.com

[baby]

As soon as I can I try also your technique...





going to check your plugin ;)

[ilasolomon]

baby , please post some pics from a view.
a normal scene, an alpha one, a z-depth + helmet glass &
one without glass (hiden) all in same rez & angle...
let me try for you! :)

[boomji]

i'm posting in the wrong place but i didn't get an answer in the afx forum.
Q:how would i use a uv mapping id(rpf) in afx.how do i get it to work.
i'm thinking, by passing a uv channel in the file any picture file i have in post will wrap around the uv cordinates...have never had the need to use this in production but am just curious.
thanx for listening
b

[ilasolomon]

simple!
just assign an UVW map + a material with map to the obj then
render out to RPF or RLA. (but remember in image setup turn on the UV)
in afx you can apply the 3d channel>extract filter & see the result
in color , but i don't know what do you want to do with that!

[CGTalk Moderation]

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.