Microsoft to bring its powerful cloud dictation and transcription services to Word
Hints too of "pricing plans" in the works for 2020, suggesting cloud transcription may not be free
4 October 2019 | 0
The dictation service within Microsoft Word has been widely criticised, thought it seems this is about to change.
In 2017, reviewers complained that dictation within Windows was awful, specifically because it worked upon a local model that individual users had to train. Now, Microsoft says it is preparing to bring its far, far superior cloud dictation service to Microsoft Word, beginning in 2020. And it is bringing transcription to Word, too.
Open Word, tap the mic button in the upper right, and you can see what we mean: Unless you have specifically trained the model and speak clearly, Windows can only recognise a majority of your words. Unfortunately, a “majority” is not good enough, as you will often have to dive in and edit, interrupting your workflow.
On the other hand, if you have ever used Cortana, or another cloud service, such as Google Assistant or Amazon Alexa, the recognition accuracy is significantly improved. Any time Microsoft shows off its cloud-powered transcription services, such as at this Build demo, it’s amazing. It is that advanced level of accuracy Microsoft is bringing to Word for the Web, beginning in 2020.
Microsoft says it will use OneDrive and the Azure Speech Services to securely store your audio files, implying that the service will be tied to Office 365 or the corporate version, Microsoft 365. It also won’t be free, apparently: “Audio transcription in Word will be available in early 2020 in Word for the web, with integration into the Word desktop and mobile apps following in the spring. Exact plans and pricing will be announced closer to general availability,” Microsoft says.
You will be able to upload recorded audio and Word will transcribe it, separating it by speaker. Microsoft said in a blog post that this transcription will appear in a Word sidebar, where snippets of it or the whole thing can be brought into the main Word window to be edited. Interestingly, the demonstration GIF Microsoft uploaded shows this happening instantaneously, though other cloud transcription services require a bit of time for processing. It also appears that the transcription will be synced to the recording, as another Microsoft app, OneNote, does.
Depending upon how accurate Microsoft’s transcription is, how quickly it processes, and the price, Word’s transcription services could be a direct strike at Otter.ai, probably the best cloud transcription service available today. Otter offers 600 minutes of transcription for free per month; however, processing takes some time, and the accuracy and the way it assigns dialogue to a particular speaker isn’t always perfect.
Otter, though, has until 2020 to perfect its service and rework, if necessary, its pricing options. Microsoft may be late to cloud-powered dictation and transcription, but the convenience, immediacy and corporate clout of Office will make it a formidable competitor.
IDG News Service