Home / video / Transcriptive Automated Video Transcriptions anytime Over 95% Accuracy

 

BSVP On-Site: Transcriptive, Adobe Premiere Pro plugin

Digital Anarchy releases new Transcriptive plugin that helps video editors create transcripts and captions for their Adobe Premiere Pro projects.

Over 95% Accuracy

Transcriptive is a new way of creating captions, subtitles and metadata for SEO or Search. It uses AI and machine learning to generate automatic transcripts with a 96% accuracy rate. It’s the future of transcriptions. It’s an exciting new plugin for Premiere Pro and will change the way you get and work with transcriptions.

The automated transcription plugin uses one of two speech to text programs, IBM Watson or Speechmatics, to turn speech into text.
The text is displayed within a panel in Premiere. It is also editable within the panel.

When you play back the video the transcription plays as well. If you click on a word within the panel, the video will jump to that point in the timeline. Transcriptive also generates timecode in the panel. If the content has multiple speakers, the transcript identifies them individually.

Transcriptive identifies words that it thinks might be wrong and highlights them in red. You can play back the video int 1.5X or 2X faster than real time to skim through the video.

There are two ways of editing. You can be in word processor mode where you would click on the word and make changes. Transcriptive also uses keyboard shortcuts with options such as toggle uppercase/lowercase to make words that weren’t capitalized that should be and visa versa.

As an video editor, you can find portions of the video faster and trim out sections quickly.

You can also export the transcript to Premiere Markers which makes all the text searchable within the timeline. This makes editing more efficient workflow tool.

Multiple Export Options

When you’re done editing the transcript, you can export in different file formats for captions, subtitles, SEO, plain text and more. It accommodates formats for TXT, sequence markers, clip markers, SRT, STL, VTT and XMP. Plain text files can be easily printed and given to a client to review.

Digital Anarchy recommends that the user sign up with the speech to text service directly. Currently Speechmatics is .07 per minute, IBM Watson is .02 per minute. At this time, Watson offers 1000 minutes for free. That would be 16 hours of audio for free with Watson before you would start paying .02 per minute.
The difference is that Speechmatics gives you better results, especially in situations with noisy backgrounds. However, if the audio is clear, the two services are both about 95% accurate.

Transcriptions are created in 3 – 4 times faster than real time. So, a 30 minute segment would be transcribed in about 8 to 10 minutes.

With the automated system, you can create transcriptions anytime, day or night.

For more information on Transcriptive or other plugins from Jim Tierney and his team at Digital Anarchy, visit DigitalAnarchy.com

http://www.youtube.com/user/bsvpTV

http://www.facebook.com/BlueSkyVideoProductions

 
Share this page
 
 

Connect with Us