Malay Speech to Text results assessment

We are currently testing the accuracy of a Malay speech-to-text engine, using content from Suria TV, broadcast in Singapore. The results require listening and reading 57 x 2 min audio files, and providing input to an already created document that captures a Word Error Rate (WER) calculation. You must be able to assess the overall accuracy of the content, divided into Substitutions, deletions, and insertions.