.vdb from Fume rendered in Phoenix. Issues!

[vfxhooker]

I imported a .vdb written from a FumeFX simulation into Phoenix FD... Not v-ray volume grid. Is there a reason that the amount of fire present in the .vdb data is so much less than that in the native FumeFX simulation? The attached image should demonstrate what I mean clearly. The Phoenix render burns up much faster leaving just smoke a lot earlier in the explosion as a result of this difference. I'm under the impression that there's not much I can do with this in Phoenix since it's a simulation, and there are no sources, etc. Just the container.

I was under the impression that the simulation data would read mostly the same, and the main difference would be shading/rendering. This looks like a difference in the way the simulation data is interpreted.

[Svetlin.Nikolov]

Hey,

There shouldn't be a difference between importing caches in Phoenix and the VolumeGrid. The raw cache data is just rendered anew. There might seem to be less fire in the Phoenix render because I've noticed that quite often FFX produces negative fire values. You can check the channel ranges in the channel list under the Simulation rollout or while editing the render curves - they are marked in cyan. In case half the fire is negative, this might be the reason it does not render, so you have to adjust your fire opacity to account for that. This is just a guess of course, but check the channel ranges first. Are you rendering the Fire channel in FFX?

Cheers!

[vfxhooker]

Hi Svetlin,

Thanks for the reply. So I looked, and what I assume would be the fire channel in Phoenix ( Fuel ) shows a range of -0.98 : 0.75, so I'm guessing that you nailed it. I'll play with the fire opacity, etc. and post my results.


Thanks!

[vfxhooker]

So I can't seem to get any new fire to show up in the phoenix render. I can affect the opacity of the fire that you see in the above render, but so far no amount of manipulation of the curves has brought any new fire into the visible range. Any suggestions? Also to answer your question above as I think I understand it... Yes, the above fire on the right side was rendered from FumeFX. I'm exporting Fuel, Oxygen, Smoke, Temperature, and Velocity to the cache file.

[Svetlin.Nikolov]

Hmm, would it be possible to share this single cache so I can see if there isn't an issue on the Phoenix/VVG side?

[vfxhooker]

Sure... I Can do that. Only problem is where. Each frame is massive... 1.8 Gigs per .vdb file. I assume you only need the one frame... If there's something more you'd need, like a scene file, let me know.

I'll see if I can free up some dropbox space in the morning and send you a link.

Thanks!

[flipbook]

WeTransfer, send upto 2gig free

[Svetlin.Nikolov]

We ended up using the Support FTP, but good to know!

So unfortunately this turns out to be one of FFX's hacks. Half the fire trails have positive fire, the other half have negative fire, and the empty space in the rest of the grid has fire=0. I found it using the preview ranges - you can check them in the shot below.

So the only way to exclude the empty space and still render both the positive and negative fire would be to make a curve like this:

[Svetlin.Nikolov]

Alternatively, I could hack it in the code especially for the Use Own Opacity mode, so that you'll be able to input the same curve as in FFX. I suspect this is a failsafe mechanism of theirs aiming to prevent having a box full of fire or smoke, but with Phoenix this might be just the thing you want - some people fill a box with fog using just this trick. I just tested hacking the curve and it worked, showing the trails in their entirety.

[vfxhooker]

Hi Svetlin,

Thanks for the help with this. I was ble to get the fire to show up with the curve you posted, but I really had to overdrive values to get scattering and self illumination to show up. I'm rendering one full explosion now. It'll take a while. It will be interesting to see how it looks in comparison to FFX. Would the data format have any bearing on how Phoenix interprets what is in the container? For instance a .f3d instead of the .vdb? Have other people experienced this same issue with the negative fire values?
I would definitely like to try and run the sim with the original TP particles if possible but as I mentioned in my email to you I have not been able to find any documention for supported workflow and so far the when I choose the TP particles I get errors. I'm wondering if I need to cache the TP particles out to geometry before Phoenix will recognize it? I don't see an option anywhere for a particle source, like FFX has. I know that Pflow has Phoenix operators, but I don't think TP does.

[Svetlin.Nikolov]

Hey,

In tomorrow's nightly there will be the hack that makes zeroes the curve when the fire is zero. Not sure how and why Fume's simulation creates such values, but I guess since I've also seen such values before, probably other people did as well, but never called, so big thanks for the heads up!

There shouldn't be any difference between f3d and vdb containers, unless there is a bug in Fume's export or in Phoenix's import

You should be able to plug a TP system in a Phx source just like any geometry or pflow source - I have a few more task before I get there and I'll check what's going on. Converting TP to actual meshes should help too, but Phoenix should see the particles as well, since the emission from particles is handled differently. There aren't any Phoenix operators for TP yet, though - I hope soon we'll be able to get to implementing those.

Will ping you when I find out more. Cheers!

[vfxhooker]

Awesome. Thanks very much!!