When you are using Limecraft for transcribing audio, Limecraft applies speaker segmentation or 'diarisation' by segmenting the transcription in fragments per speaker. Each speech fragment will be marked up with a specific label for each speaker (S1, S2, S3,…) and these labels are consistently applied throughout a single piece of content.
With that said, speaker change detection or diarisation is a process prone to errors. For example, if two speaker interrupt each other without a significant pause, it may be hard for the software to differentiate the speakers.
Hence, in cases it is important for you to have an accurate transcription, we advise you to review the automatically generated result before using it. More information about editing the transcript in this article.