EasyGlass test

#81

Hi, just some infos:

I tried out implementing real absorption and it is supercool - looks even better than the first approach with just linear thickness. That should get rid of many restrictions in messiah! :bounce:

Instead of putting it into the TLHPro:Thickness shader, I decided to include it in EasyGlass directly, since even fully transparent glass has some kind of absorption - you see it most clearly on the outer edges where even normal glass is deep green.

Slowly Easy Glass goes away now from being really easy, since there are some settings needed for it to stay as flexible as I want it to be.
You need to be able to decide if you want thickness measured only for the Front Polygons or for the back also, you may want to use different Ior for the Absorption in Refraction Direction etc.
I personally prefer flexibility and power over one-click solutions, since I often not only want to simulate reality, but also to be able to create effects that reality can’t offer :wink:

This thread also brought me to other new ideas and revealed some problems in messiah that a special material may be able to overcome… Very interesting times indeed.

But first I will concentrate on Glass since this is the most pressing thing IMO.

Sadly, my back hasn’t gone better yet so I can’t do “around the clock” shifts, so please bear with me - finishing this stuff my take a bit of time but I will keep you posted on the progress.

Cheers!

#82

sounds very cool man. this is a good reason to buy the TLHPRO shaders due to the continued work on tweaking it. I feel like its money well spent and looking forward to trying your AON shaders as well :slight_smile:

BTW i rendered a sequence from realflow and the render looks MUCH better then anything out of LW, especially since my shading doesnt get randomly reset and the integration is WAY better. Ill try to get it to a more reasonable file size since right now its 60 megs!

also, i wish there was a way to render motion blur with realflow mesh sequences, there is speak of vectors embedded into the sequences for blur, is that hard to get at? that would complete the connection to render realflow meshes…

#83

Great to hear that you like the tools!

I fear the motionblur is a tricky one: In a replace-object-sequence the objects aren’t actually moving so you will not get normal motionblur.
Realflow therefore embeds motion vectors in the object that basically are morphs between the original state at frame start to the state at frame end.
You can try superblender on the sequence and - like you were told on the LW forum already - animate the morph per frame from 0 to 100 percent.
In messiah that can’t be done in the graph editor AFAIK, but it is easy to do as an expression.
But I doubt that superblender will support that situation, since the mesh changes per frame - and mark no longer actively develops superblender :cry:

So the only secure workaround I can think of would be, to do the sequence with something like 10 times the framerate, simulate and render 250 instead of 25 fps and then do the motionblur afterwards by frameblending the sequence back to 25fps in AE.
Not really practical I fear.

But at least it would overcome the other problem with RF motionblur: Those embedded motion vectors can only be linear, so you can’t get curved motionblur - with those high-fps simulations you would get perfect curved blur but with a high cost for simulation as well as rendering…

But if everything is moving in the frame, you may be able to get away without or with very little antialiasing…

Cheers

#84

Better contact shadows.

#85

It may be off by Modo Standards… Still nice render Wegg…ever closer still.

#86

Just showing what Messiah’s DOF looks like. . . since that was in Mike’s first render.

I’m done.

#87

Great, still a bit noice in the blurred reflections and the displacement ball feels blurred.
How long took that frame to render?

/ Svante

#88

A big difference, what did you do?

And a big thanks to Thomas for tweaking the glass shader even further!

#89

Removed …

#90

very nice. IT would be a great example scene to add to TLHPro :wink:

#91

It will be part of the next release.

Cheers :bowdown:

#92

Removed …

#93

Those whiskey glass renders are excellent! Congrats high fives and kudos!:thumbsup:

As nice as Modo’s renderings look, without the ability to render animations I’m taking a pass
until Lux and Co. get the animation feature set bolted on.

#94

that render looks incredible thomas,

although im seeing a weird shape that doesnt look like it should be there on the bottom of the glass, looks like a sharp bat wing or something.

#95

Thanks.

The shape should be there actually I think. It is a magnified and deformed crossing of the gridlike wood texture. I was iritated at first too, but unless messiahs refraction is completely off (I don’t compute that myself, I just set the values with EasyGlass) it should be correct.

But there are wrong things in the image, like the hard black messiah shadow bug, slightly visible to the left of the glass’ bottom.

Have you done anything lately you can show? Any real realflow scenes?

Cheers,

#96

here is a test render. its 9 megs though

http://www.stooch.net/video/33.mov

not perfect but better then LW (much faster too)

#97

Everything up to 20 Gig is fine with me… :slight_smile:

Well, other than the flickering mesh from RF it looks great. I just love such simulations. :thumbsup: :bounce:

Would be cool to see a RealFlow render where the fluid doesn’t hang in the air but interacts with something so there is more interesting reflections etc. going on.
If you happen to have a sequence lying around that I could test the new absorption with, I would be happy to do so. :wink:

Edit: I just remembered that I already have that partial test sequence from the 33 in my download folder - will give that a try.

Cheers! :bowdown:

#98

yeah im going to do more sims, i really dont like the density and viscousity of the current one. i will share them with you.

i actually wouldnt mind making a sim of that whisky glass :slight_smile: how about dropping some ice cubes in there?

#99

Well that would be cool (Pun intended :slight_smile: ) :thumbsup:

I just don’t know how such a mesh will interact with the glass. In the image above, I modeled the liquid surface directly into the glass, it is no separate mesh to avoid the usual interpenetration that most of the time creates problems.
Have you tried something like that already?

It is possible that a rendertime-boolean shader is needed for such a scenario. You could then simulate the liquid with a slightly negative-bloated glass and then create a perfect collision surface at rendertime…

Well, the stream of ideas for new shaders seems never ending… >LOL<

Cheers!

Edit: Wow, this was my 1,111 posting on CGtalk. Unbelievable.
Let’s have a beer :beer:

#100

Removed …