A couple of Workflow Questions

[ben_hamburg]

It does seem like that border is linked to the initial fillup level.
By simply playing around with it, i almost got the surface to be completely flat. i´ll still have to simulate for an hour or so, to see if that flatness at the borders stays that way, so maybe I can decrease the SPF value again (right now it´s at 8 with 20m cells).

[Ivaylo Katev]

no, the border is not linked to the init fill up, it's controlled by the ocean level parameter.

[ben_hamburg]

Yeah, I figured that out after a couple of frames simulating.
I upped the SPF to 20, but the initial sink in or "self arising" wave is still visible. Do i need to up the SPF even more to get ridof it, or is there another way?
Forward Transfer made thinks even worse, don´t know if it is because, the grid is linked to the moving boat or anything else.
or do I need to increase the grid size even more to get rid of it?
Both will come at the prize of bigger simulation times, so probably no option for this project I´m afraid, but clever compositing...

[Ivaylo Katev]

the initial wave is not avoidable, you have just to wait it to pass through. can you post image? on spf 20 self arising waves are not very plausible

[ben_hamburg]

Yeah, the initial wave is not that bad, just the fine border, that seems to stay past the initial wave. But at the borders (I attached a screengrab in a post below) the foam forms straight patterns, because he can´t go through.
maybe I still haven´t simulated enough frames as a preroll and that border would completely disappear and without it the foam could float past the grid borders on the sides without forming those straight lines along its sides. The SPF settings for the scene I already simulated were only at 6, not at 20.
the screenshot I posted above shows the surface of the grid on frame 1, it should be competely flat, but it´s sides are bending up a bit. It´s not very much, maybe a couple of millimiters, but still enough to keep the foam from leaving the grid borders at the side.

On a sidenote:
The problem with the grid not having enough detail to match the displacement of the infinite ocean can be worked around by lowering the detail in the ocean texture.
Right now we´re only simulating with around 20-30m cells due to our lack of time, and that seems to work pretty well.

[Ivaylo Katev]

ah the foam, it is not a good idea to have foam that leaves the grid through the left and right borders, let the foam to leave the box only through the back side. why the back side is good but the left and right side are bad? because the foam's movement is highly related to the water's movement. if you have waves that push the foam outside the grid, the foam switches from the wave's velocity to zero, and this produces visible line. the force that pushes the foam is present only in the grid, this is the water, there is no force outside that continues the waves force, and the foam just stops its outgoing movement in the moment when is no more in the grid. in our video this happens at sec 18, the foam touches the right border and produces line. the back side is not a problem because the intertial forces are directed to the back side, and they continue to push the foam away regardless of the grid. of course small line is visible in the backside too, but is very hard to see it.

[ben_hamburg]

Oh, I see. So, as a rule of thumb and if you are short on simulation time, it is better to make the grid wider for boat trails, rather than larger?
I doubt I´ll find the time after the project, but I´d really like to make a good tutorial about all of this, I´ve learned a great deal via this forum in a very short time...

By the way, there is one more thing bothering me:
Due to the much shorter render times, our workflow is now like this:

1. Render all other geometry and grid surface with good lighting (HDRI Dome light).
2. Change Vray Properties of these to Matte object and set Alpha to -1.
3. Turn of HDRI dome light and environment background.
4. Turn on Special Light rig for foam and splashes.
5. Add Vray atmospheric render element.
6. Hide Foam object and render splashes.
7. Hide splashes element and render foam.

We also had to lower the fog multiplier, because we are loosing the fog coloring through the underwater bubbles.
But now we have every element separetely and can do all our color correction and DoF blurring (by manually blurring the different elements) in post, by turning of the environment we can also get the foam/splashes reflections on the water surface.
At least in theory this gives us very short render times and lots of power in post.

Now for the part, that is bothering me:
We have some scenes setup with multiple grids, and doing the whole process several times, can be quite time consuming.
Is there a way to add several atmospheric render elements and select the different foam/splashes with it?
If not so, it would definitely be on my wish list!

And last but not least:
I know the infinite ocean is kind of a trick solution, because it is not actually interacting with the simulation, and if you are using higher wind speeds, you are bound to get some strange results. But is one of the very few things i found during my stress testing of Phoenix, that might be worth giving some more thought for future developments.
Some of the strange results I saw might also be because we insanced the same grid in the scene. We rendered one in ocean mode and the other one cap mode, all in all they blended together pretty well, just sometimes, when the waves are at their lower peak, the splashes look like they are displaced wrong.

You guys must probably be some of the very few people that were happy to hear, that autodesk didn´t announce the implementation of Naiad into Max...

[ben_hamburg]

Actually, the weird displacement might be a problem I caused myself, because I set the displacement to "Surface" instead of vector. I´ll see after rerendering....

And one more point for my wishlist:
At towards the horizon the ocean surface gets quite noisy. Not sure if this is fixable, I already tried increasing the Vray Image sampler, but that didn´t work either.

[ben_hamburg]

And now another problem occured, I hope you can shed some light on what I´m doing wrong.
To save time, I wanted to reuse the same simulation in different scenes.
SO I imported the Simulation object into my new scene, positioned and rotated it into the desired place and set the input path to the cached sim from my other scene.
But when i hit render, it just rendered black, and the updated viewport mesh showed just one huge displaced mess...

[Ivaylo Katev]

when the rendering is in mesh mode (mesh,ocean, or cap) the displacement is always in vector mode, the other displacement options are valid for the implicit surface and the volumetric modes.
about the noisy horizon - you have to increase both the skyline subdivisions and the image sampler. apparently this will lead to high render times, so i recommend to find some fake solution , why not blending with real image in post?
you can create very easy a blending mask, just blur the alpha of the ocean.
about the reused cache rendering problem - i suppose you have velocity exported right? just few days ago we have a bug fixed, that occurs when you have velocity exported and use some play offset. so, if this is the case, (i hope it is!) just download the latest build and it is fixed

[ben_hamburg]

As for the noise at the horizon:
There is also some flickering during animation.
Right now we´re trying to render an extra pass with Plane and Vraydisplacement, AFAIK that didn´t have the Issue at the horizon, So we can blend those together.

I also checkd and the error indeed only occurs when play at offset is set.
I´ll check the nightly section for newest build then.
But just to make sure I understand the directions right:
If I have a cached simulation, that starts at frame 800 and in my new scene i want it to start at frame 1000:
I would set cache start to 0 and play offset to 200? the nightly tabs are slightly different then in the official manual, so it´s easy to get confused...

Thanks for beeing so patient! Only a couple more days til our deadline and I think we are going to make it in time...

[ben_hamburg]

Plane with vray displacement works perfect for blending the horizon!

Definitely something to look into though for future releases i think...

[Ivaylo Katev]

If I have a cached simulation, that starts at frame 800 and in my new scene i want it to start at frame 1000:
I would set cache start to 0 and play offset to 200?

this will work only if the length option is not used, if you use it, the exact values must be observed, i.e. cashe start=800, play offset=1000

[ben_hamburg]

Lates nightly fixed the play offset bug, thanks!

[ben_hamburg]

So, our deadline got pushed back a little, so we might have time to fix a couple of small things:

We had some mesh penetration on an open boat with water filling up the inside.
Would we need to up the grid cell count or the SPF value to fix that?