Hello there,
Trying to save out ABC files of meshed particles with velocityPV from bifrost graph. Failing dismally, I'm afraid.
We need it to be an ABC of a polySurface in order to render with VRay. Motion blur is vital to the project
We were previously doing it with nParticles, which just does the velocityPV automatically when using the motionVectorColorSet attribute during export . Can't wrap my head around how to do this in bifrost graph though with bifrostGeoToMaya prior to ABC export.
Is there anyone who can advise me on the right approach here?
(says me, suspecting @syracoj is the man who knows everything)
Solved! Go to Solution.
Solved by syracoj. Go to Solution.
Currently there seems to be a defect where color data created or passed in the graph is not being sent successfully to the DG level via BifrostGeoToMaya. It can be seen on the bifShape but it's not transferring through to the Maya mesh through BifrostGeoToMaya.
Internally, we do have a ticket catalogued to streamline the passing of velocity based color set data to Maya from out of the board but this is still future work.
Oh bugger. Well that accounts for why I was getting nowhere, so on the bright side: I'm not useless (!)
That's caused sand in the works for us badly here, but I really appreciate the explanation. Thank you Syracoj.
Why does heading back to NParticles feel so painful already ? 😜
hmmm...
stay tuned, I may actually have a solution. Need to make sure the generated maya mesh writes out the alembic with said data.
I offer no promises.
the color set data is called point_color. Maybe you need it renamed to a specific name. Getting it out of the bifrost board required it to be renamed to this. Perhaps if there is strict requirements for the name it won't work. Apologies if that is the case.
@syracoj THANK YOU for looking into this. Your results are incredibly promising after the disappointing news. Sorry about the time delayed response - European time here.
The ABC you supplied does work, albeit in strange directions. So we're getting vertex color data at least (and the naming doesn't matter, we can tell the mesh what color set holds the velocity data afterwards).
Sample rendering with and without:
above is the graph I used. Not sure if it's correct. We should probably improve the ways we send out data like this.
One thing to try is to set your frame range in to a negative value (-5) and set the start frame of the particle system to -5 as well. That way the motion color stabilizes by frame 1.
It looks like your motion blur effect could be dialed down. Can't remember but in vray is there a way to adjust motion vector source scale? Or perhaps just dialing down the motion blur length would do the trick.
That's outstanding. Thank you @syracoj
Initial tests work fine here. As you say, the velocity color data needs to be dialed down radically, but it's there and accessible at least. VRay doesn't have a per-object motion blur multiplier that I can see yet, but they're tremendously helpful folks - I'll see what I can find out there.
Thank you again for now !
/RH
So near, but no cigar quite yet.
Initial tests looked like it was working and that it just needed the velocity tuned down. Unfortunately there's still something not quite right here and I can't put my thumb on where it is. The vectors seem all wrong, but strangely "Almost" right. I know that doesn't help at all, but I've attached a couple of scene files here that will hopefully illustrate where it's going wrong.
Just in case the error was in the VRay I also tried it in Redshift (demo) and it's the same there..
Any clue as to where the velocity data is going wrong ?
I know you folks are a bit busy, but it's been a couple of months now - any news on being able to output correct velocity color sets ? Just got into another project that could have sorely used bifrost.
We did a test here where we set the bifrost velocity to be constant along x with a value of 10 on a single particle falling under gravity. We then meshed this using bifrostGeoToMaya and exported an ABC w/ the color set. When we added it to Maya w/ Vray and rendered, the motion blur was consistently correct.
we did output velocities for a multi-particle particle system (using the same graph recipe from above) and the moBlur looks correct. We do see some artifacts when new particles are born but we aren't seeing the phenomenon where the motion blur streaks in the opposite direction. I talked to a dev here and he says this can occur with particle systems at birth though.
As far as per object velocity multiplier control is concerned I found one in vray that can be used to dial down the strength of the moBlur. See attached image.
We are tending to lean toward the idea that the velocities in bifrost are correct. We are not dismissing however that there is a potential problem with the Maya mesh that is generated especially when new particles are birthed. Do you absolutely need to use a blobby surface mesh like this? Could you use sphere instances?
Thank you again so much for looking into this Syracoj. It's reassuring to see the effort put into supporting end users. Massively appreciated.
Unfortunately I'm getting completely different results (see attached image) where the streams vary directionality long after birth, and in numerous/random places. We've tuned the velocity multiplier in a whole variety of ways, and the shutter speeds etc all to rule out wild values as the cuplrit. We've similarly done this with both VRay and Redshift to rule out the renderer as culprit. Unfortunately Arnold isn't an option (neither is Redshift at this point) but it was done just to be sure. Blobby meshes are essential to the type of projects we do very often - two of our clients business are "liquid" based industries (Drainage/Flood prevention and industrial Heat Exchange) where being at the mercy of classic fluid simulators are too precarious.
At this point I'm not sure where or how to continue. OpenVDB/SOuP is an avenue we can take but it's a pity, as Bifrost Graph is so close to being everything we needed with the security of future proofing our studio workflow.
Which version of vray are you on?
Which bifrost plugin version is this?
and
Can you send me this scene file that contains the bifrost graph please?
I went and replicated your scene (I think it's the same, you could still send me your bifrost scene file) but in the scene I have built, the point_color set colors seem wrong.
naturally the ultimate test will be to write out the alembic w/ color set and plug it into vray. So will try that shortly. In this particular example, with collision and emission from 'normal speed(1)' along a given direction I am getting some strange results and perhaps with this workflow there is a problem.
more in a bit.
I can't thank you enough for looking into this. I'm impressed.
I sent the scene via PM btw (easy to miss, I tend to do so all the time)
are you importing the ABC file or are you piping it through a vray_proxy?
I get bad horizontal streaking when I just import the ABC file. But if I add it in the vray_proxy I get 'not horrible' results.
Can't find what you're looking for? Ask the community or share your knowledge.