For anyone struggeling to understand the formula at minute 00:28:14, here's a breakdown: The idea is simple: subtract from the z scale the same amount as we did in x , but in the transform node we deal with scaling factors. So, we start with bbox(0,D_ZSIZE), which is the original z size of the object. Then we want to subtract from the z size, the same amount we did in x, but we can't just use the x channel, we need to take the initial scale of x and subtract the new x size, which is calculated by multiplying the original size with the x channel(scalling factor). And finally we divide it by the z size, which means, we take the new target scale and divide it by the original one, calculating the scaling factor. (Keeping an eye on the brackets) Hope this helps, again, an easier way is to just eyeball it 😄 . Also feel free to suggest a better formula! (bbox(0,D_ZSIZE) - ( bbox(0,D_XSIZE) - bbox(0,D_XSIZE) * ch("sx") ) ) / bbox(0,D_ZSIZE)
@@nevershot3052 How would you do it with a match size? I wanted to make an effect similar to what the inset does. There must be an easier way, but it's always good to know how to work with scaling factors :)
@@cgside Yes, I’m just used to centering on the match size node, although I also use bbox, I haven’t taken the lesson yet, but if match size works, then you can do it faster )
It's just a quick way to get rid of the warning with the default option, the merge option won't create new layers in the stage. USD shenanigans :D Cheers
@@cgside There is no whitewater mat, I must have mistaken it for the cloud material or the white water post-process, but there was also a material in the new AMD Catalog called 'soap bubble'
Despite what everyone would like you to believe, Karma is simply a slow renderer. The jump from CPU to XPU was huge but it still feels antiquated when you compare it to something like Cycles. The constant re-compiling is a killer and I hope this will be addressed in 20.5.
For anyone struggeling to understand the formula at minute 00:28:14, here's a breakdown:
The idea is simple: subtract from the z scale the same amount as we did in x , but in the transform node we deal with scaling factors.
So, we start with bbox(0,D_ZSIZE), which is the original z size of the object.
Then we want to subtract from the z size, the same amount we did in x, but we can't just use the x channel, we need to take the initial scale of x and subtract the new x size, which is calculated by multiplying the original size with the x channel(scalling factor).
And finally we divide it by the z size, which means, we take the new target scale and divide it by the original one, calculating the scaling factor. (Keeping an eye on the brackets)
Hope this helps, again, an easier way is to just eyeball it 😄 .
Also feel free to suggest a better formula!
(bbox(0,D_ZSIZE) - ( bbox(0,D_XSIZE) - bbox(0,D_XSIZE) * ch("sx") ) ) / bbox(0,D_ZSIZE)
Match size won't fit?
@@nevershot3052 How would you do it with a match size? I wanted to make an effect similar to what the inset does. There must be an easier way, but it's always good to know how to work with scaling factors :)
@@cgside Yes, I’m just used to centering on the match size node, although I also use bbox, I haven’t taken the lesson yet, but if match size works, then you can do it faster )
39:00 for here i use clip node after the scatter for points not scattered below the plate. thanks for the tutorial
Despite photorealistic style i add some ambient occlusion and lightening the outer edges on the soapbox but thanks!
great tutorial, could you please do one of those beautiful perfume advertisements videos? thanks anyway 👍
Thank you 😊
Good video , First rule of fight club...
8:35 The new native Houdini's remesh node won't be enough for this? It's in beta though.
Hey, unfortunately, no, it can't keep nice corners. As you said, it's beta. But it will be totally fine to work with the boolean mesh in this case.
Why select Merge SOP Layer Into Existing Active Layer in sop import instead of the default option? What happens if I keep the default options?
It's just a quick way to get rid of the warning with the default option, the merge option won't create new layers in the stage. USD shenanigans :D Cheers
Could it have been a good option to use the new white water material for the foam?
Hey Carlos, I haven't used it yet, but isn't that for flip sims? Cheers
@@cgside There is no whitewater mat, I must have mistaken it for the cloud material or the white water post-process, but there was also a material in the new AMD Catalog called 'soap bubble'
Why does XPU feel so slow on such a simple scene?
Sss and refraction are.expensive .
Despite what everyone would like you to believe, Karma is simply a slow renderer. The jump from CPU to XPU was huge but it still feels antiquated when you compare it to something like Cycles. The constant re-compiling is a killer and I hope this will be addressed in 20.5.