Velocity blending render error

[GrahamMac]

I've finished a simulation of liquid pouring into and splashing out of a Martini glass. The best part of the action all happens within about 1 second so I tried using the Input play speed to slow things down for rendering (using Vray 1.5 sp5 - local rendering). I used 0.25 play speed and the result achieves a quarter speed playback but the motion is very odd with lots of stopping and starting. I noticed the blend method was on default Linear so I tried rendering again using velocity blending but the render crashes around frame six just before any liquid is visible in the render:

UNHANDLED EXCEPTION: Broadcasting
NOTIFY_RENDER_PREEVAL
Last marker is at . \src\globillummap.cpp, line 1416:
GlobalIllumMap::buildlightmap(){1}
etc...

Is this a bug or is there a reason for the error?
Many thanks for any help.

[Ivaylo Katev]

sounds like a bug, however do you have the velocity channel in the caches?

[GrahamMac]

Yes,
Source: velocity checked (set to 1)
Simulator: Output velocity checked
Input velocity checked (but greyed out)

Simulation data files are certainly larger than normal, which confirms the presence of the additional velocity data.

[Ivaylo Katev]

ok i will check this out.

[Ivaylo Katev]

it was a bug really, thanks for the report.

[GrahamMac]

Thanks for letting me know. I've turned back to using the Time scale parameter for now to slow the simulation down, with some success.

However, using the temperature channel with cooling set to zero, I noticed that on very low time scales (around 0.01) the fluid never makes it to the glass. With higher time scales there is no problem for the fluid to reach the glass.
The fluid is emitted from an object placed above the glass from polygons with normals facing down. The fluid travels part of the way down and then starts drifting in various directions and sometimes curls back up a little toward the emission point.

I've found increasing the discharge rate helps force it down into the glass but am curious if there is a better way to keep the direction of the flow intact. Any insight on why this parameter is causing the miss directionality of the fluid would be most interesting to here.

[Ivaylo Katev]

do you have surface tension turned on?

[GrahamMac]

No, surface tension is at zero.
The source emits the fluid from ID 1 polygons of an object placed in the grid with total surface area of around 0.2cm^2. I've chosen this size to ensure the glass doesn't fill up too quickly but also is sufficient to create a thick enough column of fluid to make a nice splash in the glass. For this area I need a discharge value of around 1000 (reduces to zero over time) to make sure the fluid both gets to the glass and has enough momentum to splash up and out of the glass. (settings image attached):

(see video: http://www.vimeo.com/20237262)

I've been trying to make sure the fluid doesn't flow like chocolate or milk and thus had zero viscosity and surface tension. I am now wondering if I should not have them at zero since the fluid currently doesn't create nice sheet formations as it leaves the glass. It separates into too many blobs and tendrils. If I add such tension or viscosity I guess I may have to increase the discharge rate a little to overcome the earlier problem of reaching the glass again.
Any recommendations?

[Ivaylo Katev]

i think the discharge is too high and the jet slops out the liquid

[GrahamMac]

I am trying to achieve a beautiful splash of liquid rising up and out of the glass and am focusing on the look of the liquid in motion. Thus the slopping of liquid out of the glass is intentional.