Word Online’s Clever Transcription

Speak and Your Words Are Captured

On August 25, Microsoft announced that the Transcribe in Word feature is available to Office 365 users.

Transcribe in Word is only available in the Online (browser) version using the Chrome and Edge browsers. Microsoft has a range of transcription features across Office 365 including the automatic transcript generated for videos uploaded to Stream and the new transcription feature for Teams meetings (Office 365 notification MC220987) that’s due to rolled out in late September. All transcription features available in Microsoft 365 apps are currently only available in English.

Transcription in Word

The idea behind transcription is that you might want to capture text to record what happens during an in-person meeting or interview. Another use case is to transcribe presentations in online conferences (something we’re all attending more of right now). The implementation is simple: open a Word Online document and select the Transcribe option from the dictate menu (Figure 1).

If an existing transcription exists for the document, Word displays it. Otherwise you can start a new transcription (only one recording can exist per document) or upload an audio file for Word to generate a transcript (uploaded files are limited to 200 MB). Figure 2 shows the options. Once you start recording, a control appears to allow you to stop.

Any microphone connected to a workstation can be used for the recording: I achieved acceptable results using a PC’s inbuilt microphone as well as the microphone in a set of Bluetooth earphones.

Saved in OneDrive for Business

When you’re finished recording, Word saves the audio file for the recording to the Transcribed Files folder in your OneDrive for Business Account. Although only one transcription can be associated with a document, many separate transcriptions can exist, each being represented by a file stored in OneDrive. You’re limited to five hours of transcription per month.

Processed by Azure Cognitive Services

After saving the audio file, Word uploads the recording to Azure Cognitive Services, which generates the transcript from the recording. Depending on how long the recording is and how many speakers are detected, this can take a little time, but soon the time-stamped transcript will appear in a pane in the Word document (Figure 3). You can now select individual transcript segments to be inserted in the document or insert the entire transcript.

The text shown in Figure 3 contains portions created by Word’s dictate feature where Word inserts text as you speak (including voice commands for document formatting like insert a new list) and some segments from a transcript. In both cases, the generated text is a reasonable start for an article or other piece, and the transcript segments are an excellent way to capture a statement made by someone.

Updates for Other Word Variants

According to Microsoft, the transcribe feature will come to Word mobile by the end of 2020. The Word desktop apps for Windows and Mac will support voice commands and dictation at the same time. As always with Office desktop apps, their legacy roots mean that it takes more time for new functionality to appear.

Breaking the Habit of a Lifetime

When I started using Word 2.0 to create documents in 1992, I certainly never foresee the day when I could write through voice. Now that I’ve seen how effective the dictate and transcribe options are, I might just do less typing in future.