v22.12.01

Back-end for Transcription Analysis

Summary

This updates the Rehearsal back-end with our first AI Transcription Analysis metrics. Each student response is analyzed for the average pace, the time spent speaking vs pausing, the total word count, and if any questions were asked. An AI Confidence metric is also calculated based on the average confidence of each word transcribed. Future releases will display these metrics to the learner and mentor, on Web, Android, and iOS clients.

Metrics

Each of these metrics will be calculated going forward and stored for future use. 

Pace

The average speaking rate, or pace, varies for the purpose of your speech.  For presentations, the average pace for English speakers in the United States is between 100-150 words per minute for a comfortable pace. Since pace may also vary depending on culture, language and activity, there is no one size fits all. Work with your mentor to establish a comfortable pace for your scenario.

Time Spent Speaking and Pausing

The amount of time spent speaking vs. the time spent pausing between sentences or words can impact how you deliver your message. Pausing can give viewers a moment to process what you are saying, and also emphasize your point. Of course too long of a pause can have a negative effect on your audience.

Number of Questions Asked

Do you ask questions to help engage your audience? Asking open questions prompts your audience to consider their personal opinions and beliefs, and naturally increase interest and engagement.

Word Count

The total number of words spoken in your presentation including words like 'the', 'a', and 'of', and even filler words such as 'uh'.

AI Confidence

Rehearsal's speech recognition evaluates how well it may have transcribed a word and should not be confused with accuracy measurements. Confidence scores provide Rehearsal with feedback on how well our speech-to-text capabilities are working in real-world scenarios.