Polyphonic voice allocator implementation
pool_ stores the state of each voice. lru_ is used to figure out the “age” of each voice - the lower the number, the most recent the voice has been used, the less desirable it is to steal it.
Two pieces of information are stored in the pool_ array. The highest bit is whether the note is active; the lowest 7 bits is the actual MIDI note. To retrieve the note, we mask the 7 lower bits (& 0x7f). To retrieve the activity flag, we mask the highest bit (& 0x80).
Now why do we need to store both pieces of information? When a new note arrive and we need to allocate a channel for it, we want to use, in priority, the least recently used inactive channel (there might still be the “tail” of a decaying envelope). if none is available, the least recently used active channel. So even if a note is off I still need to keep track of it in the array!
The LRU array is filled backwards. So the channel 1 has to be marked as least recently used, and this is the first channel that will be used when the first ever note arrives.