I was interested in Graeme’s post because he took a different approach to the task than I did. Rather than just giving everyone a written copy of his narration like I did, he shared audio files. We got to listen to his narration of his story and the Microsoft Read Aloud version of the transcript. I found that his choice to present the narrations this way was very effective, more so than just having the written transcript. With my written transcript, there was nothing for others to compare it to. There is neither a written record of what I said nor an audio recording. It is interesting that he even compared two different voice to text programs using the same audio recording for each and obtained very similar results.
With the Microsoft Read Aloud, the lack of punctuation in the written record was emphasized. A speech-to-text program cannot record the emphasis and enthusiasm of the story-teller. It didn’t capture any of the humor or urgency present in the story or the tone of or inflections in Graeme’s voice.
Graeme describes the issues with the written text and I had similar issues: missing punctuation, inconsistent verb tenses, and even incorrect words altogether. The tool took the oral language and translated it into what it determined was the most “correct” version of what was being said. In any case, it was definitely not perfect.