Wine glass filled with wine - Render Artifact

[trancethereal]

Let me first preface I am new and still learning Fusion 360... so I outlined my steps in case my methods are at fault.

Goal: Create and model a realistic wine glass filled with wine. Render the model to provide a realistic effect.

Problem: I am seeing weird artifacts in the render ... the glass looks like it has beveled inside cuts along with an odd circular white pattern on the render. When I render the same bodies independently - they look fine. As I merge the bodies closer, no artifacts ... until the two bodies are correctly centered. If I scale the wine down .999 (most of the artifacts disappear, but still some visible). Scale further to .99, no artifacts. Very weird and not sure why I would need to scale the inside body ... shouldn't a real fluid touch the inside glass? An air gap looks to be an incorrect work around... as it would cause lighting and refraction errors in the render. (inline images follow the above order and include image comments)

How I made the glass (ignore stem for the moment):

1 - Created a sketch - using only the line tool -  I made the base, then up for the height and as the revolve axis, then the top for the radius, then down for a straight slope, then I switched to the spline tool and made the bottom bowl part to connect and form a solid shape. 

2 - Using the revolve command - selected the profile, then the vertical line as the axis. (Bueno!)

How I made the wine (internal contents):

1 - Create an Offset plane ... selected the inside base of the cup... then up 2 inches.

2 - Create an Boundary Fill ... selected the wine cup body and the top of the Offset plan (2 tools), which created 3 cells. I selected the middle cell (which appeared to be the inside contents). BTW - regardless of what cell I selected, the selected result never changed colors ... it remained the same green color making it difficult to figure out which boundry fill I wanted. I made 3 bodies from each cell.. then deleted the two that did not make sense.

How I rendered the two objects:

1 - Wine glass body - appearance ... body = clear glass (from Fusion library)

2 - Wine liquid - appearance ... body = water (then changed attributes on absorption index = 2.5 and color to something dark red)

3 - In Canvas render.

First image is the artifact from the above steps... the other images are various testing to see when the issue reveals itself. 

 

[cekuhnen]

@trancethereal

do you use the new liquid shader ?

[Anonymous]

Hi! and welcome to this awesome world! 

I think you make a nice workflow, but in this case for transparent materials is a little different to make a good render, because its a dielectric material and you should do priority controls to choose what is better for you.

Read the article and if you need more help just make other post here and i will make an screencast for you!

How to create realistic renderings with transparent materials

www.autodesk.com/products/fusion-360/blog/7903-2/

[trancethereal]

Hi Claas,

No - I am unaware of the "liquid shader". I will try to look into it (but if you have a quick cheat sheet or comment on where and how to use, that would be appreciated).

Second - I actually stumbled upon your discussions of a similar issues - (I resigned myself to a air gap solution - while it's not my preference, it's so easy to "implement" via scaling the internal liquid to avoid the artifacts. This is becomes a reluctant choice of speed over accuracy.) I definitely appreciated your thoughts on the subject. 

Third - I read Anuar Mata  comment below and the link provided to Aradhana Vaidya discussion on the subject ... seems to have a logical explanation. I am going to try and report back my findings.

 Thanks,

Chris

[cekuhnen]

@trancethereal @Anonymous

First one has to state that there is no perfect workflow to deal with the following situations:

body and liquid

body and liquid and ice cubes

There are all different and valid approaches. Either via a shader that takes care of priorities or via smartly generating

the render surfaces and texturing it accordingly.

So the priorities shader approach is correct and one option.

I do not agree with how they promote this as the best option and down grade other options.

The reasons often given to not choose a non priorities shader approach I think are questionable.

For example with a complex vessel, all I have to to is create a top surface model and then split the vessel.

I do not need to to any silly tasks such as building the fluid element bigger. This will be very difficult to do when

you deal with something more complex than a glass or bottle model - which in 100% of examples people who defend

the priorities approach visually use.

Further more the comment about render and production model I find also problematic because specifically with

Fusions modeling timeline I can easily at the end of the timeline include the splitting for render purpose or make a dedicated render scene.

So what AD multiple times declares as a problem I rather find a minuscule aspect to keep in mind and it is up to you the user to decide

and adjust your workflow properly. Branch and merging could also have been a nice option here.

So this comes down to your own preference and type of product.

Truly with the ice cube there is a point for priorities - but other wise not really and other methods then based on the model

complexity will be better!

[trancethereal]

Hi Anuar,

Thanks for the link (read the article and it seemed to address the issue very clearly ... and was optimistic this would resolve my issues). However, when I attempted to apply the solutions (dielectric material ranking)... I was unsuccessful. My results showed the liquid to appear to be beyond the interior glass wall (like it was on the outside almost). See pictures.

Steps:

1) Scale liquid body to 1.02 % (this put it just ever so slightly inside the glass wall of the glass).

2) Set dielectric values - Glass (1), Liquid (4) (I have some other bodies being used to make up the model in the base that were given priority 2, 3.. not visible and should not impact this interaction of the render)

Run In Canvas Render ...

Results show the wine fluid touching the outside glass wall rather than touching the interior glass wall (as I thought it was suppose to.) Oddly, I was even expecting worst case solution would have the wine render slightly inside the glass... but reflections and everything else render correctly. Disappointedly, it was worse and renders it all the way to the edge of the glass.

Did I miss something here?

[trancethereal]

Hi Claas,

I am agreeing with much of your sentiments ... as I attempted (unsuccessfully - but giving benefit of doubt) to duplicate the results in the AD example solution work flow. I am puzzled why I was not able to duplicate their results (never mind your valid argument of unnecessary work flow considerations).

So - in my specific example (so that I can apply or attempt what you are trying to explain) - what is your method or workflow? I am not really understanding top surface model / split the vessel approach. 

Using my work flow outlined in the original post - can you insert your work flow steps that I can follow along? Very curious to see your method.

Best,

Chris

[cekuhnen]

@trancethereal @Anonymous

The problem with the trick to do a priorities and just make a bigger fluid body (I hope you use offset not scale!) is that this way

you do not model the surface tension. I attached one image so you can see it. Look how front and back look and how the foam

flows upwards. Also pay attention to how it looks from below.

To be fair renderings like this is something where you also need a proper render tool and Fusion (not judging it) is more geared

towards build-in renderings for products.

A model and shader is half the job. For more control you need to create a good env to place your object into (lights and such).

Fluids are transparent/refractive surface - nothing in the scene and fluids or chrome looks empty.

You also should specify fluid translucency or adsorption right!

[Anonymous]

hahahahaah @cekuhnen please give me one! @trancethereal look my render and tell me you opinion.

 

[trancethereal]

Hi Anuar,

Thank you for the quick update  - Quick clarification ... did you understand my question and concerns of the fluid looking like its on the edge of the glass (as opposed to inside the glass)? 

Your sample images (thanks!) look to have the same issue of the fluid looking like its on the edge. The example Claas showed (wow - amazing) has the same fluid on the edge look It's undeniably realistic and with attention to detail ... that I am questioning my own observation of the fluid on the edge. 

So - is this accurate and how fluids interact .. or something we have to accept. I am wondering if its the former and I am overly literal at this point and not realizing this is how it actually looks.

[trancethereal]

@cekuhnen

Simply amazing results ...  was this done 100% in Fusion? If yes, it should be a showcase of capabilities. Well done and inspiring work (and beyond my abilities at the moment.)

[cekuhnen]

@trancethereal @Anonymous

Ah this is a photo of a beer I was drinking but found useful to make a study of it so you have a real object photo as a reference.

[trancethereal]

Too funny ... well, at least I can recognize what looks "real" and appreciate how "realistic" the details were. 

Thanks again - at least we have an authoritative answer to my question... is it accurate for a liquid to appear as if it on the edge of the outer glass wall (making the inner wall disappear). 

[cekuhnen]

@trancethereal@Anonymous

Well my opinion is not unusual it is kinda common in the CG industry.

Both approach also are good and have the advantages but also disadvantages.

Some major render engines offer today both approach - so I am personally fine with them.

Two things about rendering that can help:

1. never imagine things - always work based on reference photos to compare

2. rendering as a creative tool is about creating a convincing illusion a - it is not really about being 100% physically accurate