SpeedScriber, Lumberjack, Magic Keywords and Final Cut Pro X: magic combination

I’ve (along with many other people) have been beta testing SpeedScriber, an unreleased app that combines the power of an API for speech to text with a well thought out interface for correcting the machine transcription. Feed the SpeedScriber output to Lumberyard (part of Lumberjack System) and extract Magic Keywords and in a very short period of time (dependent largely on FCP X’s import speed for the XML) and you have an organized, keyworded Event with a fully searchable Transcript in the Notes field.

The results are very encouraging with only about a 3% error rate: mostly correcting punctuation and speaker allocation. The error rate on the words is about 1% for a standard American accent. The area where I’d like to see improvement is in speaker identification. Two speakers in an interview are typically identified as 12-19 ‘speakers’. Fortunately it’s incredibly easy to correct in SpeedScriber.

That transcript is saved out as text, with time stamps every sentence, where a sentence is a very loose concept, but fine for our purposes. That transcript matches the Multicam Clip name in FCP X (or Synchronized Clip, or regular clip) and Lumberyard pairs them up, and extracts Magic Keywords.

Here’s the raw result in FCP X:

And here’s the result (literally) five seconds later after I tidied up the results:

I would say that those keywords are accurate for this interview.