Yes, an ID Pass or matte ID would provide a long-winded way of doing this, but the downside is that you have to a) render it out as a batch process, rather than a standard render b) it will save it as 2 separate image files instead of one ..and c) would require bringing both into Photoshop, merging into 1 file, setting up a new alpha channel based on the specific selection needed from the ID pass.
My current workflow allows all the above processes ready to be used straight out of the rendered RGBA image, so I was hoping to find something similar in Maya 😞
In terms of how this functionality is used; A large part of my work involves providing high rez 3D renders of posed characters from popular animated TV shows and films to make up the style guides and assets given to, licencees, publishers and print houses to produce all the merchandise associated with the series whether it's wallpaper, bedspreads, books, games, lunchboxes etc. Unfortunately in the print world, they are not conversant in complicated compositing and as such are only interested in having the final artwork with one alpha channel, which is a silhouette of the main character. An ID pass that shows a different colour for every surface, whilst useful to me in other 3D projects, would be totally confusing to licensee clients. The reason it's provided like this and not a 32-bit png for example (where the alpha channel has already been applied) is because it enables the client to have a choice as to whether to use the floor 'as is' (with grounding shadows), or cut it out using alpha channel, or use the alpha channel to cutout a foreground layer whilst using a blending mode on the un-alpha'd background layer to do a basic composite on to other graphics. There are occasions where all 3 of these are used, so having all of them available in the one file, and one that is ready to send to the client as soon as it has rendered (i.e without even needing to bring it in to Photoshop myself), is a real time saver.
In a similar question, how would one go about rendering out a 33-bit png from Maya, where the floor and background are already pre-multiplied as an embedded alpha? So basically having the floor unseen by the camera at render time, but included in all GI and ray-tracing calculations? From what I can see it seems that hiding the floor object will exclude it from all these calculations, thereby making it invisible in reflections, invisible to GI and allowing light from the environment light to be contributing in an upwards direction from below the character, where previously it would have been occluded by the floor layer ..essentially making the render look very different in a very undesirable way. It follows that making the floor transparent would have the same effect. In Lightwave we have a simple 'unseen by camera but not from from GI and ray-tracing' tick-box. Any suggestions?
Cheers for all your replies so far 🙂