Problems rendering large MR scene to RPF file type

[Macer]

Hi,

Just wondering if anyone has run into any issues exporting to rpf or rla file types?

I’m creating an animation and I usually use combustion to add motion blur and DOF etc, so rpf files work great, but this is a large model with several building Xref’ed in and hundreds (nearing a thousand) of proxy trees and shrubs with xref’ed mats.

Rendering to tif or jpg is fine, a maximum of 1.8GB memory is being used by max (takes about 2 -3 mins per frame), but if I try and render out rpf or rla, once the geometry has been converted, max just sits there at about 12% cpu with the memory slowly climbing for 30 mins or until it crashes or I quit the process. The memory maxes out 8GB in max and continues with file swapping. This happens whether FG is on or frozen. I realise that rpf files retain a lot more information (I’ve set it to save velocity, z-depth, coverage and colour only) but I didn’t think it would chew through that much memory (while using only 1 core) generating it!

I’ve tried rendering elements out separately (but I’m going to have learn how to combine in post??!) and z-depth is fine but velocity shows similar issues to those above. Any thoughts or suggestions would be greatly appreciated!

I have all the usual large scene set-up options checked (BSP2, scanline disabled, etc). The pc being used is a dual quad xeon @ 3GHz with 8GB memory, Vista 64, and Max Design with SP1 and creativity booster.

Thanks,

Mark

 

[Justin Hunt]

I am going from memory so I could be wrong, I dont think mr supports rpf

 

Do you use the output to combustion option in the render elements? In theory it should layer everything up for you in combustion

 

jhv

[Macer]

Hi Justin, thanks for your response.

 

I've used MR and rpf's several times before on smaller models, so I'm pretty sure the issue is not compatability. In fact it works really well, everything you need in one file.

 

I've played around with render elements since my first post and I ran in to the same problem. I guess its a case of MR having to calculate much more information (ie z-depth, velocity, etc) than it does for a standard flat output (jpg, tif, etc).

 

And there I was thinking MR processed all the information for each render and I was just choosing what to save, but obviously it works the other way around!