this is a rather non specific and IMO unhelpful statement, there's no problem doing this on a VM and performance should not be an issue.
VMs add a layer of complexity and unknown. Would this be a problem for a system playing 4k video through HDMI? Would it be a problem streaming 8 streams of 24/192 stereo? How about turning 4xDSD into PCM via a direct connected USB DAC?
VMs are potentially an issue any time performance is a factor, or real time processing is required. Both are possible with MC, depending up on how the system will be used.
I think my advice is solid because it avoids all of the uncertainty associated with a VM. For moderate use of MC, a VM is probably fine. With my history and experience in computing, I tend to simplify things to make my life easy.
Perhaps I should stop opinionating about these and other topics. <shrug>
Brian.