Sound leaves a mouth, hits a capsule, passes through a preamp, gets converted to data, lands in a recorder and finally reaches an editor. Shows rarely fall apart at the capsule. They fall apart at the joins: a guest who hears themselves echoing back, four people sharing one headphone socket, or a laptop that reboots twenty minutes into a good conversation.

Start by counting people and rooms. One host at a desk needs almost nothing beyond an input and a pair of headphones. Three around a table need three preamps, three headphone feeds and enough cable to reach all of them. A guest joining from another city needs a return path built specifically so their own voice is stripped out of it.