Add the 'none' audio backend as an option for qemu-server: it behaves the same as the spice one (id=,driver=), and simply discards audio. Add the option in the Web UI for audio configuration.