VrayMtl: Don't affect Alphachannel checkbox

[Da_elf]

not really. seems natural to me that glass is going to have some sort of effect on the alpha. the more glass you put together the more effect it has. i just want to see an examle of what you want which you said other rendering packages can do. visable examples of what you want will explain better a visual example of what you dont want set by itself

[Dschaga]

ok ... sorry .. da_elf your aswer wasn't serious - mine too, because you are not interrested in solving a problem.
You think i try to harm vray by asking for solutions of a Problem.

You ask me how other render will render this scene? What has it to do with the vray output?

Anyway ..i'm not interrested anymore in showing problems of vray to solve it.

[Da_elf]

i am serious. your saying you want your alpha channel to do something in particular which you said the other rendering packages do. you showed an example of what vray does and that you dont want it to do it that way. im not putting you down. i genuinly want to see the results that you DO want. not the results you DONT want. this way you would be one step closer to people being able to help you get what you DO want out of vray

EDIT..ok ok. i read throiugh the posts again and see that you did not infact say that other renderers could do it. you said its something every renderers should be able to do. im just having a hard time understanding what you want. you basically want that glass should have pure black for alpha channel?

[Dschaga]

Can you see the 90° edges?

They are flipping from bright gray to dark gray, which means that the alphachannel is inconsistent for objects like glas fassades - which means problems with animations from time to time and on hires images we really often have to clean the alpha channel with photoshop.

It also looks like the alphacannel is influenced by internal reflections or refractions (i'm not sure). The results are also independent of having double-sided or Reflect on the back side checked. ..so i think this is a bug or a wrong intepretations of RL.

and about our little infight: *g* - ok?
..let us go on with the little flame here

[Da_elf]

yes. i see what you mean. im wondering if its supposed to be physically correct or of its a bug

[Dschaga]

It can't be physically correct, because this would mean that the density of an object is viewdependent.

[dlparisi]

The reflection are also used by vray to calculate the alpha channel ..and this can't be correct!

got me?

Are you sure this is from the reflection? I can't see why a dark reflection shows up in some of these areas. What exactly is being reflected?

Are you sure you don't have overlapping faces or something like that?

[Dschaga]

There is only this one object - and no, i'm not sure if it's the refraction or the reflection which is shown in the alphachannel.

There can't be a STL error since this is a line which was outlined and extruded.

[Da_elf]

its the internal refraction depth termination or something

[studioDIM]

Refraction

Do i understand you'd prefer an unshaded alpha channel for the transparency?
From the tests i made, the behaviour of glass objects you showed us is given by the refractions.

I'd suggest you to create a separate pass for your glass objects, encapsulating them into a VrayMtlWrapper, checking "Matte object" and changing the alpha contribution to whatever suits you.
That generates unshaded addittive alpha channels for transparent objects, similar to max's own scanline.

Hope it doesn't fuel more flaming :P

Edit: i seem to have missed your inability/unwillingness to use the MtlWrapper, consider the post void

Lele

[Dschaga]

I'm still lightscape user and i know from the LS renderings that there is a slight fresnel like effect on the alphachannel and i also have the feeling that refraction and reflections do not influence the alpha channel.

i seem to have missed your inability/unwillingness to use the MtlWrapper, consider the post void

yep .. we use too many workarounds for limitations of vray, max and windows that i want to optimize the workflow with a working alphachannel.

[studioDIM]

eheh, old skool rulez :P

As a *quick* workaround, you could render the infamous separate pass assigning a falloff map, set to fresnel, IoR override to match your original IoR , in the Vray material opacity map.
That'd give you the desired alphas, respecting your glasses' original IoRs.

Hope it eases the wait

Lele

[Dschaga]

we just do something like an additonal alphamap rendering now ... and ... just to say it anothr time ... we think it's a workaround which costs us time and which makes vray slower than it is or could be with a working alphamap.

btw, we are still on 1.09 and will upgrade to 1.5 directly, so this discussion about a new bugfixed beta is very theoretical to me!

[studioDIM]

Hope it eases the wait

[Da_elf]

ive got a question though. the image you posted with 4 pieces of glass setup the way they were. is that a realworld example? normally you wouldnt see the edges of glass like that and when you do there is always inter-reflections seen anyhow so it is correct kinda
when you look at glass from the right angle you cant see through it. i forget what the physical term is but basically you have gone past a specific angle and no light goes thorugh its all internal reflections. which means there shouldnt be a transparent alpha in that place anyhow