No foam where expected ?

[JezUK]

I've created a sea simulation of a 300m long ship, but foam is not visible in places where I'd expect there to be.

Of course, the ship is symmetrical

I've created the sea simulation over 78 seconds with a large grid volume of 200m cells (cell size 300mm, System Units in Meters) - I've also allowed for a long Outside Life of 60 seconds to leave a nice long trail for the Foam and that can be seen clearly in the Perspective shot (you can see clearly the 'trough' created by the grid volume and the foam outside of that).

So, for that view, you can see lots of foam along the side of the ship, behind it and even foam outside of the simulation volume - so far, all good

But then......

......look at the camera view with foam on the left of Camera - there is NO foam at the back, after the outside of the volume grid - you can clearly see it cuts off early at the back off the ship.

Furthermore - when the camera is on the other, camera right, side (i.e. seeing where the ship side is in shadow) - no foam is visible. Or rather, it is clear that foam is 'missing' on that side. Again, in that view too, there is also no foam at the aft of the ship outside of the grid's volume - that foam ends abruptly at the end of the simulation grid edge.

I used the latest production Phoenix FD (as per my signature) - and this simulation took 14 days (on a 7980xe) and we're not going to re-simulate !

Any help please would be much appreciated.

Thank you.

[Gosho.Genchev]

Hey, JezUK ,

we're looking into this. It would help a lot if you could provide us with the scene + 1 cache file.

[JezUK]

Hi Gosho,

Thank you. I will send you the scene plus a link to download one cache file when I return to my desk later today.

Thank you

[Svetlin.Nikolov]

14 days to get this result is unthinkable.. what grid resolution and how many frames are there? How many liquid and foam particles do you have in the end?

I am not sure where to look for the abruptly ending foam at the end of the grid, but if you get that it mean you need more Outside Life (this is now renamed to a more proper name in the nightlies).

Cheers!

[JezUK]

Hi Svetlin,

14 days to get this result is unthinkable..

Hmmmm........

The settings I used were the exact same settings from the Ship tutorial with the only exceptions being that I had to not have such a small cell size as the simulation was claiming it would have taken even longer !!! (and my Outside Life for the foam was 60 seconds, which I could have left it to 20 seconds anyway...)

what grid resolution and how many frames are there? How many liquid and foam particles do you have in the end?

I remember the ship tutorial cell resolution was 165mm and I went with 319mm for the reason given above (so, Tutorial 100m cells versus my 200m cells).

My liquid and foam particles at the end are - 110m Liquid, 713k Foam, 12k Splashes (my ship has a less sharp bow to the tutorial ship).

My ship is 300m long and the ship in the tutorial is 150m long.

Ship tutorial is 1000 frames of sim, mine is 2340 frames.

If this was all linear, then my ship should just take twice the time - but from the predictions given during the initial tests it seemed it wasn't linear at all....... (hence also why I decreased the resolution).

Ideally I'd have preferred to keep the cells the SAME size as the tutorial's (i.e. 165mm) as my sea splashes are a little too granular for my liking....


AN OBSERVATION........

Thanks to your message, I investigated further and noticed that the Performance for my simulation was at around 0.49m vox/s whilst for the tutorial it was 1.34m vox/s ????? (on your tutorial docs a screen grab shows your machine on that tutorial to be 1.49m/s, maybe your machine is slightly better than mine...).

But why such a low performance for my ship ? (this might explain the lengthy solution time - btw, my bad, it was 'only' 12 days, not 14........ !).

(I have a 7980xe, mildly OC'd, with 128Gb ram).

I am not sure where to look for the abruptly ending foam at the end of the grid, but if you get that it mean you need more Outside Life (this is now renamed to a more proper name in the nightlies).

I have attached another screen grab to show you where the foam ends at the end of the grid ) )

Furthermore, my tests I did this morning regarding why no foam on the shadow is of the ship - when I move the sun over to what was the side in shadow, yes, we then see foam !!! (and I checked the previous side where there was sun, but now no sun as it has become the side to be in shadow, and then that has no foam visible so foam seems to be visible only when there is direct v-ray sun !) - I am sending you guys the scene file and a couple of cache files.

Really hope we can sort this out - if I can redo this simulation over a few days rather than a few weeks, I'd rather redo it at the correct cell resolution and get this right....

Many, many thanks.

[JezUK]

Files sent to support@chaosgroup.com via WeTransfer

Thank you.

[Svetlin.Nikolov]

Okay, so from this info, you had 200M voxels / 110M liquid running at 6-7 minutes per frame. This sounds better, and the huge total sim time is due to the very long animation.

Foam and splash are very low in numbers, so the simulation should not be getting slowed down by them as it is right now.

Note that bus speed is as important to the CPU, and saving the caches to a fast drive is also very important. There is a sticky topic in the forums on what the best hardware is, but in general these two factors + CPU are the top 3.

Outside life does not mean outside life, so I've renamed it to Max Outside Age as it was very misleading. So 20 would be too low if the half life is long, and 60 might suffice, but if your foam does not die due to half life, then it must be as long as the entire animation.

Have you noticed the simulation outputting frames more quickly in the beginning than in the end of the sequence? You can check the timestamps of the cache files as well. If so, in Phoenix 3.11 we fixed and issue with geometry interaction that made the simulation gradually slow down after a long sequence, so if you are using a version older than 3.11, it might be this issue that affects your sim's performance.

If the sun is affecting the appearance of the foam, this brings me to my next question - do you have fog material for the ocean mesh's shader and if so, do you have a closing geometry under the ocean? If you don't, particles under water will not render correctly inside the material's fog because V-Ray needs a closed volume where the fog resides. Since the ocean is just an open sheet of geometry, it needs to be capped off, by using e.g. a vrayplane (we updated the ocean toolbar preset with a v-ray plane as well).

I won't have access to my work PC till late Monday as there will be stuff going on at our office building during the weekend, so haven't checked the scene yet, but please try these tips first.

Cheers, man!

[JezUK]

Hi Svetlin,

Sorry it's taken me a while to get back to you on this - a few points;

1) I mentioned that my machine was blasting through the simulation for the Ship tutorial at a rate of 1.34m v/s, yet on my ship simulation was only at .49m v/s - using the same machine, the same hardware / software etc - shouldn't these performance numbers be roughly the same values ? (liquid container settings pretty much the same, just the cell size and volume of grid as well as the length of the simulation was different).

2) I still don't really understand why I wouldn't be seeing any foam outside of the container grid (via a camera view) when it is clearly visible via the perspective view (a test view, not of course a view I will be using for my animation). Sorry, I don't understand your explanation.

3) I've sent you good people the scene file and a couple of cache files, which I know have been downloaded. Any thoughts please ?

4) I notice that if I have VRayAerialPerspective in my scene (switched off for the attached render) then that makes my foam in the render very weak looking and almost like there is no foam.at all. I do have geometry under the sea, i.e. ground and islands etc - but I wouldn't describe it as 'closed' (i.e. it is not like a large swimming pool !) it is more like a large plane. Must it be like physically closed ?

5) For the life of me I cannot make the simulated water be indistinguishable from the ocean mesh (see back of ship) - the container line is very visible (I've outlined it so it is clear what I mean) and I have played a lot with the Ocean Level in the Rendering Rollout for the liquid container, yet it is still blindingly obvious that the container is present there (not noticeable at the front of the ship). Please, any tips for making this less obvious / invisible ? (see attached).

6) Finally any tips for why I'd be getting those turquoise speckles ? (I know it could be anything, but I've tried the render settings and the material settings nothing seems to fix it).

Many thanks.

[Svetlin.Nikolov]

Hey,

Downloaded the scene so the link wouldn't expire, but since I thought you have options for all your questions, went on to do some coding until you replied.

I really need to know which Phoenix version you are using now in order to help you.

The difference between the roughness of surface of the simulation and the ocean extension was discussed in the forums more than once for the last few months and was fixed in the nightlies of Jun 10. This is why it's important to know if you are running a newer version and if this is still an issue.

In order to be completely sure about the performance difference between the two scenes can you run 2-3 frames of the ship example scene, do not close Max and run 2-3 frames of your new scene, then go to the Phoenix log folder and send over the Phoenix log. It would show exactly where the difference is.

Mixing areal perspective with Phoenix smoke or particle shaders is not a good idea, unless you render the smoke in Volumetric Geometry mode or the particles in Render As Geometry mode, because of how atmospheric effects work in Max: https://docs.chaosgroup.com/display/...VRayVolumeGrid If aerial perspective is active, then this is the reason why you don't see the foam correctly. Otherwise, if you have a bottom, then it should be okay vraymtl fog-wise - the volume where the fog of the v-ray mtl resides should just be capped off, but not necessarily closed.

The speckles could be related to the foam, or they could be related to an issue with the ocean surface producing such artifacts that we fixed in the nightlies on 24 Jul.

Will ping you when I know more. Cheers!

[JezUK]

Hi Svetlin,

Thank you for getting back to me so quickly !

Straight off, I'll say that my details are always in my signature and I also put the date on that to show how up to date that information is. So, I'm currently on the production version of Phoenix 3.05, I don't use the nightlies (as I was guessing that is Beta testing territory).

As I'm new to Phoenix, please could you tell me how I should be approaching it from a production point of view, i.e. it is still under development and the nightlies give stable versions with the fixes in them ?

Secondly, I have discovered that for my point 2) ie no foam visible outside of the container, I stand corrected. They are there, BUT they are VERY faint !

Also, I do now get foam on the sides of the ship in shadow (it seems my ground plane fixes that).

But it seems to me, that maybe I need to be using a Nightly version of Phoenix and then retry ? Will that get rid of my 'visible container' issue ?

Or will I need to redo all of my simulation again ? ((((((

Please I'm keeping my fingers crossed you'll say no !!!

Please can you advise which Nightly version would be safest / most stable for me to use.

Many, many thanks.

[Svetlin.Nikolov]

Ohhh, I was really hoping that your signature was outdated. 3.05 is waaay behind on ocean stuff. This Phoenix nightly is stable according to our tests - definitely give it a shot: https://nightlies.chaosgroup.com/#/phoenix3max/20180813

You should not have to simulate again as all of these are just render issues

Cheers!

[JezUK]

Many thanks Svetlin, I'll give that a whirl when I'm back at my desk this afternoon

[JezUK]

Hi Svetlin,

I've tried what you said but the nightlies made it worse - in fact, simply installing the nightly https://nightlies.chaosgroup.com/dow...45_install.zip,

I get renders with no foam at all.

And the issue with the container edge visible is still there (see attached).

[Svetlin.Nikolov]

Roger that, let's see...

[Svetlin.Nikolov]

Hey, so first regarding the missing foam - the particle shaders got disconnected from the foam and splash particles during conversion from the old scene format - there will be a fix for this case in tomorrow's nightlies, and you can work around that by picking the simulator's foam and splash from the 2 particle shaders manually.

Now looking at the ocean issues.

Cheers!