i like the way v-ray frame buffer RAW Image file gets its options from the Vray Renderoptions RE and the ability to saves all the render elements within a multilayered EXR
but not all programs can handle multi layered EXRs well and in some scenarios i might not need in post all the render elements in a certain scene so i would benefit ( in terms of performance ) to only loading what i need
is their a way to get the same output but as separate EXR for each render element instead of multilayered EXR ?
using native max’s render output will lose some compression options not supported yet by autodesk like the DWAA for example
and if i used the separate render channels options and saved as EXR Half-Float the Z-Depth channel gets saved in as Half Float too instead of Full Float and i lose the options to configure using the Vray Renderoptions RE
Yes, just render the scene in Standalone (rmb on viewport >.vrscene exporter > Export and Render). I tested it out and checked the .exr’s metadata through exrheader - everything set in the VRayOptionsRE is saved and the VRayZDepth is forced as 32-bit.
ok it works but it also exports a vrscene file i don’t need atm that’s time consuming especially if am sending an animation
any other way ?
or why the separate render channels option not produce a forced ZDepth on float point ?
Not at all: the openEXR standard is built so that only the used channels are loaded from disk.
If you view the beauty, that’s what gets loaded, and so on.
There are minor performance differences if you rendered bufferless to a tiled EXR, but even then, they are minute for a single image (even if big in size.), and only really noticeable with huge sizes, a metric ton of REs, and long sequences.
You can consider exr images as a compact storage option for all your REs, no need to fiddle.
While loading multiple channels from a single file is more efficient, where separate channels in separate files shine is when you can re-render only the required channels, not spending extra time recalculating, antialiasing, and saving channels that don’t change between renders, with good examples being Z-depth and velocity.
We use the normal saver to get the beauty pass into EXR files. Then we use the render dialog Frame Buffer settings to save separate channels for all the other render elements. We disable RGB and alpha in these settings, as we get this in the main beauty files.
Aside from having to set file names in two places this actually works pretty well. There is a checkbox to save each element in its own directory.
Indeed, Data passes have always had their own life in my production experience.
Stuff like normals, point position, Z and so on don’t go nearly through as many revisions as a beauty render with its REs, often require higher bit depth, and they don’t even need lighting or shading most of the time.
Perfect candidates for their own task.
A quick sweep of scripting, and from a beauty a data file is birthed, once, or as many times as needed.
Saving separate REs was, on the other hand, never a thing for me (or those i worked with): exceptionally messy on disk, prone to a number of issues when being handled, as with sequences and a few REs (dozens of multimattes, f.e.) the numbers on disk just exploded, so the risk of leaving behind one or the other RE were higher than with the neat EXR package and an automated shuffle.
I could totally see that. Most of our stuff uses the one team, one shot philosophy. Like AMG one man, one motor… So it is a tight knit group who work on a shot. Easier to keep track of things when the 3D artists are also compositing. If it’s multi departmental setup I could absolutely see that being a nightmare in a bigger shop.
But to the OP the method above should get you the passes in files you desire.
I’m not sure the OP will be able to force the EXR options on the max file saver, as that is used everywhere but for the raw file output, and the all-layers-to-one-exr saving in the VFB.
If he wants DWAA, f.e., he’d be out of luck there.