When you specify a voice combination like [[0,1], [0,1]], all sorts of small things might break, all over. The best solution here is obviously to filter whatever voice combinations we get, to make sure each voice pair is unique. While clients should certainly be responsible for this, there's no reason the framework shouldn't do it too.
When you specify a voice combination like
[[0,1], [0,1]]
, all sorts of small things might break, all over. The best solution here is obviously to filter whatever voice combinations we get, to make sure each voice pair is unique. While clients should certainly be responsible for this, there's no reason the framework shouldn't do it too.