Interresting, I might have to come back to this once, I deal with Animation in Lops. Can't lie, After watching this for the 3rd time...it's starting to actually make sense.
I used the cache method and adjusted the shutter speed in that node. It looks the way I want it to in the viewer when I visualize it with the Karma XPU perspective but there is no blur when I render it. I also tried adjusting the blur from the camera node and there is still no blur in my renders either. Do you know what may be causing this?
Note: Transform motion blur will not work as intended if at all if your transforms are using Houdini Expressions (Like the Labs Turntable SOP does) Transforms need to be keyframed Manually
This constant caching is so annoying when you have to iterate on an animation in Object mode, and have to switch between solaris and standard workarea repeatedly. So far Solaris is slowing me down more than it helps with anything.
oh don't get me started. - too late! Houdini's obj context, despite lacking lighting features like barn doors, is easily the fastest and most powerful lighting tool I've ever used Solaris ditches that in favor of having an edit node pop up every time you touch a light, and if you don't notice it you have to do your edit all over again - plus the idea that lighters should do 10 times as many shots at once, which makes as much sense as modelers making 10 models at once. I tested Solaris in one show then used it a year later in production, and it was a nightmare. Different shots had different LUTs with no way to represent or support them. You'd get different notes on shots that look the same, and once you start doing per-shot lighting, all the advantages of multi-shot are gone, leaving you with orders of magnitude more work. I tried getting USD assets back into the obj context to light them and send them back to Solaris, but there's no tools for that, it would take a real programmer. Really disappointed in SideFX
"If you want motion blur, just add a cache node in after your animation" *entire scene grinds to a halt cooking a 25GB cache file* jeez and it takes20min. to open my scene b/c it's caching this - what a p.i.t.a. It must be something in USD that's breaking the old style motion blur - they've got to think of a workaround
Interresting, I might have to come back to this once, I deal with Animation in Lops. Can't lie, After watching this for the 3rd time...it's starting to actually make sense.
why cache node?
trail does the same thing to generate a V attribute?
just wondering....
for the rest a great explanation... bit new to karma...
I used the cache method and adjusted the shutter speed in that node. It looks the way I want it to in the viewer when I visualize it with the Karma XPU perspective but there is no blur when I render it. I also tried adjusting the blur from the camera node and there is still no blur in my renders either. Do you know what may be causing this?
definitely helps. Thanks!
Extremely important to add motion blur support for volume in XPU engine.
lol ) im couple of hours try to make motion blur on my volume with xpu render)))))))
Note: Transform motion blur will not work as intended if at all if your transforms are using Houdini Expressions (Like the Labs Turntable SOP does)
Transforms need to be keyframed Manually
how do you scale velocity blur?
you can scale geo motion blur and volume blur, but how do you do it with the v attribute?
Multiply v in SOP level should work.
such a useful video :) Thanks!
first time trying karma here. i just give up. i just want to add some motion blur and this is like the 4th video i see wtf
This constant caching is so annoying when you have to iterate on an animation in Object mode, and have to switch between solaris and standard workarea repeatedly. So far Solaris is slowing me down more than it helps with anything.
oh don't get me started. -
too late! Houdini's obj context, despite lacking lighting features like barn doors, is easily the fastest and most powerful lighting tool I've ever used
Solaris ditches that in favor of having an edit node pop up every time you touch a light, and if you don't notice it you have to do your edit all over again - plus the idea that lighters should do 10 times as many shots at once, which makes as much sense as modelers making 10 models at once. I tested Solaris in one show then used it a year later in production, and it was a nightmare. Different shots had different LUTs with no way to represent or support them. You'd get different notes on shots that look the same, and once you start doing per-shot lighting, all the advantages of multi-shot are gone, leaving you with orders of magnitude more work.
I tried getting USD assets back into the obj context to light them and send them back to Solaris, but there's no tools for that, it would take a real programmer. Really disappointed in SideFX
"If you want motion blur, just add a cache node in after your animation"
*entire scene grinds to a halt cooking a 25GB cache file*
jeez and it takes20min. to open my scene b/c it's caching this - what a p.i.t.a.
It must be something in USD that's breaking the old style motion blur - they've got to think of a workaround
cache from $F-1 to $F, or $F-2 to $F, same result.
@@OtakuSquirrel1 You'd probably want to do $F-1 to $F+1 since it needs a frame before and after, otherwise you get half weird looking motion blur.
@@joelchamp1949 ok~