The most common use cases are Famicom and NDS, but other platforms/titles with special hardware add-ons may require capturing more or less granular data for proper emulation.
In Famicom's case, binding microphone input to a button by threshold would be a better experience in most cases than emulating via button directly, because voice is expected to vibrate, and therefore button input requires pressing multiple times at a high frequency. With my own test application binding a mic by threshold to a key input, it is clear that microphone input with a rough ballpark threshold is more natural even for a binary input.
Last I heard it needs API changes.
The most common use cases are Famicom and NDS, but other platforms/titles with special hardware add-ons may require capturing more or less granular data for proper emulation.
In Famicom's case, binding microphone input to a button by threshold would be a better experience in most cases than emulating via button directly, because voice is expected to vibrate, and therefore button input requires pressing multiple times at a high frequency. With my own test application binding a mic by threshold to a key input, it is clear that microphone input with a rough ballpark threshold is more natural even for a binary input.