Transcripts of channel-separated stereo audio only contain a single type of error, which is incorrect transcription of an audio region. This is referred to as word error.