Transcriptions can be auto-generated by the ASR service, manually ordered from the ASR service, or manually generated and uploaded to media. Once transcriptions have been applied to a capture or video, you may need to manage them.

This article addresses downloading and editing transcriptions outside of EchoVideo and then uploading them to the media. We recommend using the built-in transcript editor to perform these actions instead. See Using the EchoVideo Transcript Editor for more information. However, downloading, uploading, and reverting to the original can still be done as described in this article.

Admins for all captures and Instructors who own the media can add and manage existing transcriptions. Co-instructors can also add or manage transcriptions on a capture if the co-instructor edits toggle is turned on for any section where the capture is published.

Transcription options are available on the media details page for each piece of media.

Click on a media tile from your Library page (if you are the owner) or select Details from the video icon in the Class List. Then click Accessibility from the Details tab, as identified in the figure below.

The Transcripts entry, as shown in the above figure, opens a menu containing the options described below. Note that these same options are available from the EchoVideo Transcript Editor.

• Order Transcript - Allows you to order or upload a VTT Transcript file for the media if you have one. See also Adding Transcriptions to Captures.
• Upload - Allows you to upload new or updated transcription files (EchoVideo supports the use of VTT files for transcriptions). See also Adding Transcriptions to Captures.
• Edit Transcript - Opens the EchoVideo Transcript Editor.
• Download Original - Download the originally applied transcription (VTT) file. If you use the ASR service for automated transcriptions, this is likely the one returned from that service.
• Revert - Allows you to revert to the originally applied transcription file, removing the edited / current version from access or visibility.
• Delete - Deletes all transcripts and versions from the media.

If you are editing transcriptions outside of the EchoVideo transcription editor, the workflow you would use would probably look something like this:

1. After the capture is published, the machine-generated automatic transcription from the ASR service will be received.
2. Download the original transcription file (described below).
3. Edit the transcription file, either using a transcription editing program or a word processing program.
4. Save the edited file (see the editing article linked in #3 above).
5. Upload the edited .vtt file.

Download Transcriptions

You have two choices for downloading transcription files:

• Download Original: This downloads a copy of the originally applied transcription file. It may be one that was manually uploaded, or it may be an automated transcription provided by the ASR service. If it is the ASR transcription, you will notice entries that read NOTE CONF. These are the confidence rankings for each word, based on how certain the transcription program was that the interpretation of the speech-to-text was correct.
  This file can be edited if necessary and re-uploaded to replace the original with the edited version.
• Export: This downloads the currently applied version of the transcription file. This might be a manually edited version of the original. Or if the capture was edited and republished and was automatically re-transcribed, this might be the updated automatic transcription. Regardless of how it got there or how many versions may have been uploaded between the original and the current version, you can export a .vtt or .txt of the currently applied version of the video's transcriptions.

The figure below is an example of what a downloaded transcription file looks like. It is shown here because the .vtt file looks considerably different than the rendering of the transcriptions in the classroom. For more information on the entries in a downloaded transcription file, see Defining the sections of a WEBVTT file below. For detailed steps and tips for editing a transcription file, see Editing Transcriptions for Accuracy.

Define the Sections of a WEBVTT File

The EchoVideo transcription service only accepts VTT files that comply with the WebVTT specification as outlined by the W3C: https://www.w3.org/TR/webvtt1/. If you are using a transcription program to generate transcription files, it will likely create an acceptable VTT file for you.

If you are downloading the automated transcription, editing it, and then uploading the edited version using a more generic word processing program, there are some important things you need to know. We strongly encourage you to review Editing Transcriptions for Accuracy for more detailed steps and tips.

• You should be able to perform the edits in a simple text editor, such as Notepad, Notepad++, WordPad, or TextEdit.
• You can also perform the edits in a more robust program such as Microsoft Word (allowing you to use a spell checker, or find / replace with special character capabilities, etc.).
• When saving the file, select All Files (*.*) or from the File Type list, and be sure the file has a .vtt extension.
• If All Files is not an option (such as from MS Word) select Text Files (*.txt) as the File Type. You can add the .vtt extension but Word will add .txt after it. This is ok; you can remove it later.
• If your saved file reads something like FileName.vtt.txt because the program added the .txt extension to the file when you saved it, simply go into Windows Explorer or Mac Finder and rename the file, removing the .txtfrom the end.

Concerning making edits in the file:

• Do not remove the WEBVTT title line from the top of the file. This entry is required by the WebVTT specification.
• Do not remove all of the line breaks; leave them in. The specification requires breaks (at least one hard return) between the time cues and the text (cue payload) and then at least one blank line between each cued segment. Meaning, that the time and the text cannot be on the same line, and there must be a blank line before and after each cued segment.
  See the text in the example VTT file shown above. It shows the time cue on a line, the cue payload on the next line, and then two line breaks (hard returns or paragraph marks) to generate a blank line before the next time cue.
• You do not have to, but you can remove the NOTE CONF entries that will be resident in a downloaded automatic transcription file. These indicate the confidence levels of the transcription program in the accuracy of each word transcribed, ensuring it matches what the speaker actually said. You can also leave them right where they are and use them as cited in the next tip, as long as there are also empty line breaks above and below them (see the previous bullet).
  

  Use the NOTE CONF entries and search for confidence scores of 90 and lower. These will help you locate those places in the transcription where the text is most likely in need of editing more easily. See Editing Transcriptions for Accuracy for instructions on how to do this.

The bottom line is that when making edits to an automated transcription file, stick to editing the speech text and leave the rest of the file alone if possible. Using a spell / grammar check to find garbled words and sentences can be helpful, as well as looking for low-confidence markers (as noted in the tip above), to streamline the review and editing process.

Revert to the Original Transcription

The Revert option in the transcript menu is very specific: it reverts the applied transcription file to the original one (the first applied) and removes any edited versions from access. Effectively, it restores the video transcription to its original state.

Because EchoVideo now has a transcript editor, reverting to the original no longer removes access to the other versions of the transcription. Those are available via the Version drop-down list in the editor and can be reapplied to the media there.

To revert a capture or video's transcription to the original

1. From the Accessibility section of the Details tab, click the vertical action menu, as shown above, and select Revert.
2. Click OK on the confirmation message.

The originally applied transcription file is now applied to the video and is the one visible in the classroom.