How can I use the Transcript Editor?

Module Description

The Transcript Editor in Aiconix is a powerful tool that allows you to review, edit, and refine the transcripts generated by the DeepVA Speech Recognition module. Whether you need to correct errors or prepare the transcript for export in various formats, this will help you to efficiently use the Transcript Editor.

Accessing the Transcript Editor

• From the Speech Recognition Result:

  After an analysis job is finished, the AI-generated transcript can be found in the Speech Recognition module in the analysis view. This is the raw, unedited version of the transcript, and you can download it as a text or subtitle file. To edit the transcript, click on the text file and microphone icon next to "Transcript." This opens the Transcript Editor, where you can review and make any necessary changes.

• From the Main Menu:

  In the main menu, look for the "Transcripts" module (icon with text and microphone) located in the middle of the left-hand side. Here, you'll find all previously created transcripts. You can also click the "Create Transcript" button to generate a new transcript from a Speech Recognition result.

Editing the Transcript

Once the transcript is open in the editor, you have several tools at your disposal to make corrections and review the text:

• Create New Variant
  A variant of a transcript represents a different version, such as a translation into another language or a different approach to the transcript style. Each variant can have multiple versions, and a new version is created each time you edit and safe the text.
  In the Transcript Editor, variants appear on the right-hand side as cards displaying the following details:
  • ID: Unique number for the variant
  • Name: Custom name you provide
  • Language: Language code (e.g., en for English)
  • Created: Date and time the variant was created
  • Last Modified: Date and time of the last edit

You can easily switch between variants to view or edit them as needed.
In the top bar you will find all necessary buttons for editing the transcript:

• Save Button
  

  After making changes, you can manually store a new version of the transcript. Each saved version is numbered and archived for future reference.

• Approve All Paragraphs 
  Selecting this button with the checkbox symbol will approve all paragraphs in the transcript to be finalized and ready for export.

• Turn on the Edit Mode
  To edit the transcript, click the pencil symbol to enter edit mode. You can then click on any word in the transcript to make corrections via the keyboard. The transcript is interactive—clicking on a word will also move the audio/video player to the corresponding section, allowing you to listen and verify the accuracy of the transcription and correct it, if needed.

  After making changes, click the yellow "Save" button to create a new version of the transcript. You can continue making corrections as needed, and each saved version is automatically archived for easy access later.

• Word-Confidence Coloring (Highlighter Symbol)

  The Word-Confidence Coloring feature helps you proofread transcripts more efficiently by showing how confident the AI was in transcribing each word. You can turn this feature on or off by clicking the highlighter symbol in the top menu bar. When activated, each word is color-coded based on the AI’s confidence level:

  • Red: Low confidence (value below 0.3)
  • Orange: Medium-Low confidence (value between 0.3 and 0.8)
  • Yellow: Medium-High confidence (value between 0.8 and 0.9)
  • Green: High confidence (value above 0.9)

  By hovering over a word, you can see its exact confidence value, allowing you to focus on correcting lower-confidence words first.

Exporting a Transcript

Once your transcript is complete, you can export it in various formats from the Transcript Editor. 

• • • Choose Export Format:
      In the top menu of the editor, click the dropdown menu next to the export icon. You can currently choose from the following formats:

      • .srt (SubRip Subtitle File): Contains transcription text and timestamps for video subtitles.
      • WebVTT (Web Video Text Tracks): A modern subtitle format with extra features for web videos.
      • .docx (Microsoft Word Document): A clean, easy-to-read text document without timestamps.

    • Export the Transcript:
      After selecting your preferred format, click the “Export” button to generate the file.

    • Download the File:
      A new artifact will appear in the overview, which you can download by clicking the "download" button in the right-hand column.

Deleting a Transcript

The Delete Transcript button (represented by a bin symbol) will permanently delete the entire transcript, including all its versions and variants of each paragraph. However, the Speech Recognition result itself will not be deleted. You can always create a new transcript from the original speech recognition result if needed.