Transcribed text, with feeling

In the last year or so, Google’s Speech to Text has gotten really good. There was a time not too long ago when it was terrible, so this is a wonderful and welcome change.

Yet converting speech to text, even when done perfectly, inevitably loses something. Even if the text you end up with is a faithful transcription of your words, it fails to capture your tone and intonation.

I wonder whether there could be a visual representation of computer-transcribed speech that somehow adds back this missing information. Perhaps the visual representation of this affective information would be some combination of text color, word spacing, font size and style, background shade, or other characteristics.

I suppose we could just intermix appropriate emojis at various places within the transcribed text, but that seems somehow unsatisfying. It would be so much more interesting if the text itself could give the sense that it is coming alive with emotion.

2 Responses to “Transcribed text, with feeling”

  1. Al says:

    How about sheet music?

  2. admin says:

    Yes, that’s a great direction! Can we apply that to text?

Leave a Reply