UPDATE 7th September 2020: There have been some major updates to Dictation functionality announced:
This is the new YouTube Video overview:
You can see that the spoken punctuation is inserted when there is a pause or at the end of a sentence. Similarly, words that were transcribed incorrectly initially, are often retrospectively corrected based on context.
One of the features inside of Office 365 that I have been using recently is the Dictate Accessibility feature. In fact, I am composing this blog post using Dictate inside of Microsoft Word right now; everything I have written so far (including the punctuation) has been through dictation.
The mission statement for Microsoft is to empower every person and organization on the planet to achieve more and that includes people with accessibility needs.
The Dictate feature inside of the Office 365 apps makes it very easy for anyone to simply speak to their computer and have a transcription of what they have been speaking displayed in front of them. Unlike other voice to text solutions, there is no need to train the model or the engine doing the dictation because this is a great example of the Microsoft Azure Cognitive Services in action. This is sometimes referred to as:
The Intelligent Cloud powering the Intelligent Edge
You can see this in the image below:
Privacy Considerations
Given Dictate uses the cloud, it does require an internet connection and data is being sent over the internet for processing. This privacy statement is online here:
Important information about Dictate
Dictate is one of the Office Intelligent Services, bringing the power of the cloud to Office apps to help save you time and produce better results.
Your speech utterances will be sent to Microsoft and used only to provide you with this service. When the transcription is done your audio and transcription results are not stored by our service. For more information see, Connected Experiences in Office.
Testing My Typing Speed vs Dictation Speed
I decided to put the Dictate feature to the test and see how well it compared to my typing speeds. I like to think that I am a reasonable typist and normally sit around the 70 to 80 words a minute on a typical day, but can boost up to 90 words a minute on a good day.
In my test, I decided to see how much of the opening paragraphs of The Great Gatsby by F. Scott Fitzgerald (which is one of my favourite novels) I could write in one minute. I would then try to read aloud the same passage over the course of one minute and compare the accuracy and speed of the Dictate feature inside of Office 365 .
You can see the results in the short video below:
I have to admit, I was quite surprised by the speed difference . I achieved 75 words per minute typing which is probably about average for me when trying to copy accurately another piece of text (as opposed to just writing from something I’m thinking about). To see that I could write at 126 words per minute (a 68% increase on an already above average typing speed) using Dictate was a pleasant surprise.
Immediately, you can see the possibilities for students in schools who maybe have difficulty with the physical process of writing, and how Dictate could really unlock their creativity and ideas by allowing them to share them quickly and easily with their classmates.
Considerations When Using Dictate
There are definitely some things that you need to think about when using dictation instead of typing because it changes the thought process and the flow considerably. Even when dictating this blog post I am finding that I am thinking and speaking faster than I would normally type, and so I am having to rethink and re-state some of the comments that I would normally have figured out in my mind whilst typing because it is naturally a slower process.
Here are some bullet points of the things I think are the main things to consider when using dictation that I have found over the last few days of using this extensively:
- Even though you can talk pretty close to naturally you do need to pause to insert punctuation effectively. That is a different thought process to simply speaking aloud, with the punctuation that you require in your sentence needing to be articulated.
- The way dictation works is it will often correct mistakes after you have spoken them so you almost need to become comfortable not looking at what is immediately transcribed, and instead waiting for the end of the sentence where it will often magically fix everything that you have said. This can be a bit disconcerting initially but you start become comfortable with it after a while. Have a look at this video to see what I mean:
- One of the cool features of Dictate is if you have forgotten to hit the stop button when you have been dictating the app automatically turns off the microphone after a few seconds or when it detects that you are doing something else, for example you start switching applications or start typing and editing it will turn off the dictation microphone listening. This is awesome not only from a privacy perspective but also to stop accidentally inserting text into a document!
Final Thoughts:
Overall, I have been super impressed with the accuracy of Dictate inside of Office 365. You definitely still need to reread your text afterwards for accuracy and to remove the odd error that it does make. However because it is part of the Azure Cognitive Services that will inevitably get smarter and more accurate as machine learning is applied.
Given that I was able to write this entire blog post using Dictate with only light editing afterwards shows that this is definitely ready for the classroom!
