Releasing our V2 Transcription Engine
V2 brings Improved Accuracy, Improved Speaker Identification, and Faster Transcription.
Today, we are excited to announce the release of the second version of our Transcription Engine used to generate Rough Transcripts in DepoDirect. This marks our second transcription release, and is a complete replacement of our previous engine.
Our new engine leverages a much more accurate speech to text engine. This engine provides a 10% accuracy improvement over the previous engine.
Improved Speaker Identification
Our new engine improves on speaker identification by making use of individual and combined audio files. Individual audio files allow us to know who exactly in a meeting was speaking, and assign the appropriate speaker name.
Our new engine is optimized to run in parallel wherever possible. Every deposition requires transcription of multiple files, which we run simultaneously. That means our wait time is only as long as the slowest file to process.
Looking to the Future
We plan to continue iterating on this new engine. We will continue to work on the accuracy, speaker identification, and transcription speed. We also have plans to offer improved formatting, spelling corrections, and so much more.