When transcribing a two-voice polyphonic texture, what is the primary challenge compared to monophonic transcription?