A/V Media: Create Captions and Transcripts

Introduction

This page details the process of producing DAMS-ready captions for a video using the UT Automated Transcription Service and the Amara subtitle editing platform. The Automated Transcription Service uses Amazon Transcribe to attempt a timestamped transcript of a given video, and can output a number of file formats. With Amara, users can edit subtitles directly over the video, to ensure the audio and subtitles are properly synchronized, and that the subtitles are formatted correctly.

The Automated Transcription Service often produces inconsistent results, especially for videos with poor audio quality, Spanish language content, or content with multiple speakers. The automated transcript output by the service will always need to be reviewed and edited, but it may sometimes be faster to create subtitles from scratch in Amara, without loading in the automated output. However, Amara does not support video upload, only linking. This means that even if a user does not want to use the output from the Automated Transcription Service, they should still use it to host any video to be edited in Amara.

1) Use the Automated Transcription Service to generate a preliminary transcript

Log in and upload audio/video file here: https://utdirect.utexas.edu/apps/captioning/services/my_jobs/

All files submitted to the Automated Transcription Service must be under 2GB in size.

Consider transcoding video files to a lower image quality for transcription (the visual information might prove useful for subsequent manual time-coding/linking though, as it can help identify speakers).


Enter a title. The title will determine the filename of the transcript, so you may want to use the local identifier or filename of the video to be transcribed.

Select the language of the media for the transcript. If you would like the transcription service to attempt to identify distinct speakers, click "Advanced options" and select the number of speakers in the video.

When the media file has uploaded, wait a few minutes and refresh the page. (The system will also send an email message several minutes later) If the transcript has loaded, download it as a VTT file.


Each user account has a limit of 50 transcription jobs. Contact Daniel Jacobs to reset this limit when you run out of transcriptions.

2) Ingest the media file and transcript into Amara

Amara is a web-based tool to create and edit captions. Most features can be used with a free account, but for free accounts, captions projects will be visible to the public. UT Libraries' Captioning and Transcription Services maintain a subscription for a UT workspace in Amara, that cannot be viewed by the public. A limited number of licenses/seats are available for creating captions for DAMS assets in the UT workspace. A seat can be made available for the duration of a captioning project. Contact Daniel Jacobs for details.

If you have an Amara account associated with the UT license, log in and access the UT workspace here: https://amara.org/en/teams/university-texas/

Navigate to the "Videos" tab and click the "+ Add videos" button.

Copy and paste the Media File URL (e.g. https://lib-captions.s3-us-west-2.amazonaws.com/Transcribe/prod/store_64668be1-d8fc-44fc-9b34-1ad3e435d0f4.mp4) from the captioning tool page into the Video URL box. Select the language of the video and hit OK.


Amara will not accept MOV file URLs. Upload an MP4 version of a media file to generate the transcript, even if the MOV file will ultimately go in the DAMS.

On the video project page, click the three dots to the right of "Add/Edit subtitles" and select "Upload subtitles". Upload the VTT file downloaded from the automated captioning tool, or from another source.. Select the language of the subtitles and click Upload.


3) Edit the captions in Amara

Click "Add/Edit subtitles", and in the window that opens click "Edit English" (or whichever language your captions are) to open the caption editor.

The caption editor shows the video at the top of the screen and two versions of the captions to be edited on the bottom. The version on the left is a locked copy of the latest version saved in Amara, displayed for your reference and so you can revert changes as needed. If you save and close a captioning project and return later, each saved version will be available to choose from the drop down list at the top of the pane. If you have uploaded more than one language of captions, you can switch between them here as well.

Use the middle pane to enter and edit caption text. Hit tab to play or pause the video, and as the video plays, the caption editing pane and the timeline tool above it will advance in sync with it. To edit the timing of an existing caption, click and drag it in the timeline. Click and drag from the edge of a caption segment to adjust its total length.

The caption editor will alert you as you go to any formatting issues with your captions. Problems will be flagged using a red exclamation point below the caption timestamp, as well as red text explaining any issues with a given caption section if it has been highlighted. The three main rules for formatting are as follows:

  1. A single line of captions shouldn't exceed 42 characters. Break longer lines of captions into a second line.
  2. Captions shouldn't exceed three lines at a time. Split longer caption sections into multiple sections.
  3. The reading rate shouldn't exceed 21 characters per second. Increase the duration of the caption to give it more time on screen, or shorten the text to make it easier to read.

Amara editor keyboard shortcuts

Amara has the following keyboard shortcuts intended to make the editing process faster and easier.

  • Tab: Play or pause the video and timeline
  • Shift + Tab: Skip backwards
  • Enter: Advance to the next caption section
  • Shift + Tab: Insert a line break in the current caption section
  • Control + Tab: Split the current caption section into a new one, beginning at the location of the cursor


The Up and Down arrow keys can also be used to move subtitle sections according to the current playback time. By default, pressing the Up arrow key will move the end of a subtitle to the current time, and pressing the Down key will move the beginning of a subtitle to the current time.

Which subtitle will be moved depends on the location of the current playback marker relative to the subtitles on the timeline:

  • If the marker is in the first half of a subtitle section, hitting the Up key will drag the end of the previous subtitle section to the marker, and push the beginning of the current subtitle to the marker (i.e. making the current section shorter and the previous section longer). Hitting the Down key in the first half of a subtitle section will move the beginning of the current section to the marker.
  • If the marker is in the second half of a subtitle section, hitting the Up key will pull the end of the current subtitle section back to the marker, making it shorter. Hitting the Down key in the second half of a subtitle section will pull the beginning of the next subtitle section to the marker, and pull the end of the current subtitle to the marker (i.e. making the current section shorter and the next section longer).

Click "Using up/down to adjust timings" in the top left of the screen to toggle between the default mode and "legacy syncing behavior", which is less complex but which may be slower to work with.


To save your progress and close the editor, click "Save Draft" in the top right of the window. If you are finished with the subtitles altogether, click "Publish"

4) Download captions

Back in the main Amara page for the video, click the "download" button on the right side of the screen to download a VTT file of the captions you have created.