The main reason you will need to transcode inputs is when you are mixing and spatialising inputs in a Spat Virtual Room. This is because the Virtual Room module requires incoming sources to be in a Channel Based format. Internally, the Room may well be panning in Channel based, Ambisonic or binaural format, but it always needs Channel Based streams as inputs. More about this in the Virtual Room section. Format transcoding may not always need re-spatialising in a Room. There are some contexts where you will not use a Virtual Room in the signal flow,
Here is an example of decoding an Ambisonic signal using an Input Transcoder module. This could also be done in the Master Transcoder section of the graph.
As mentioned in ‘Introduction to Ambisonic’ (section 4.7) a decoding stage is absolutely necessary to render Ambisonic encoded audio to speakers. It can be done like this without the need of a room.