Smoke & fire workflow?

[uncia]

Just for fun.
fire.vdb

[uncia]

Some Fire

[grimm]

I suppose these vdb files are not very well generated to give your target effect. In last vdb file, smoke and fire channels are not seperated very well. Smoke densities covers the fire densities. So this causes to a decrease on emission effect.

Yeah, I agree. I split the fire.vdb file's grids out into separate files so I could set them up independently and here is the result:

The strange look is because I scaled the temperature grid a little bit larger and translated it down a bit. It's looking better but I think a better built grid would work even better still.

Jason

[grimm]

Some Fire

That one is looking good!

jason

[uncia]

There is nothing wrong with vdb grids. You can remap values in octane with a help of Multiply node
And you can make that the volume will ignore or accept Environment.
There is no need to split grids.

Some_Fire.ocs

(28.43 KiB) Downloaded 179 times

[grimm]

Some more tests:

I split the grids out of the explosion.vdb file and did this test with it:

It's looking much more like I want but I'm still not too happy with the fire. Which brings up a question, is the emission attenuated by the density or the temperature? So I went back to the original file and brought in two copies so I have the density grid for the fire. Here is the result:

This one is softer because I changed the volume step to 8, the previous image is set to 2. I wanted it to render faster as the first one took over an hour to get to 6000 s/px, the second took 27 min. The look I'm aiming for is almost there and it's looking better and better. It almost look like a oil fire in this image here:

http://web1.cnre.vt.edu/lsg/3104/Main%2 ... /fire2.jpg

Jason

[haze]

Just to give an update, the temperature grid was not actually driving the temperature of the black body emitter, it was erroneously being used as the emitted radiance scaled by the emission texture. The next alpha will have this fixed and you should find it much easier to get the result you are looking for.

[grimm]

Awesome! Thanks haze, I was wondering about that.

Jason

[haze]

I have fixed the issue and the result is attached here.

Please be aware that this result was obtained by hard coding the scaling of the temperature (I had to scale the temperature up by a factor of around 40). I will be temporarily adding some scaling parameters in the volume import preferences so that it is possible to reproduce this effect.

I will be posting a tutorial outlining some workflows, with possibly a screencast to explain it more thoroughly.

As you can imagine, the volumetric rendering is really quite complicated, but we will nail the workflow for you.

[grimm]

Woohoo, thanks haze, it's looking good!

Jason