Dictation makes a comeback

Dictation makes a comeback

Dictation and the technology that made it possible—short-hand—died because typing your own thoughts on a computer was far more efficient and cost-effective than dictating to a secretary—even a sexy one like Mad Men’s Christine Hendricks. hendrics secretary

Dictation may make a comeback though; ironically due to the same technology that killed it in the first place: computers.

Voice Recognition has come of age

With Watson’s Jeopardy success almost two years ago already, it makes sense that we should start feeling the impact of voice recognition (VR) in our day-to-day lives.

Even if you’ve never heard of Watson, you’ve seen the apple iPhone ads using Siri, Apple’s personification of VR. Siri executes your spoken commands and responds with a high degree of success; also converting your words into text without the burden of sick days, bad attitude or (Mad Men style) temptation … did I mention Christina already?Christina 2

My Motivation

No, it’s not Christina … although I would prefer her voice to Siri’s.

It’s that I always want to improve my productivity and I’ve never really loved the onscreen iPad/iPhone/Android keyboard.  So when I read about the promise of state-of-the-art voice recognition in the new Apple IOS 7, I was keen to give it a try.

To my great surprise, this technology actually works!

VR is a huge gain to my creativity as I capture more of the ideas that come to me, and a significant gain to my productivity as I can speak faster than I can type.

What you’ll need (my Mac bias)

If you have the latest Apple devices you initiate voice recognition (VR)/dictation by tapping/hitting the little microphone icon on the keyboard. siri micOn a mac there is a shortcut you set in the system preferences e.g. hitting the function key twice.

Don’t despair PC and Android users, I expect that VR technology is as good on non-Apple platforms.  (Please share your Apple and non-Apple experiences in the comments below.)

What to expect

It’s not perfect. VR makes mistakes and mis-identifies words that I’ve spoken. For example the word ‘paper’ sometimes comes out as ‘people’, and the word ‘albeit’ comes out as some variation of words like “I’ll be it.”

Sometimes the words that appear in response to the words that I spoke make absolutely no sense, but on those occasions I discovered something impeded my voice e.g. my hand over the mic or conflicting loud background noise.

But for 90% − 95% of the time the accuracy is amazingly good; as an old boss of mine liked to say “Good enough for government work.”   I can capture my thoughts with enough accuracy that I can come back later and easily correct what VR missed with my keyboard.

 The system sometimes guesses

Whenever the system is not sure of what you’ve spoken it takes a best guess of what it thinks you were trying to say.  There is a pause as it converts your words into text and when the words appear on your screen, you can see the words the computer was not sure about underlined.

Tapping an underlined word presents alternate choices, and if the correct choice is there, tapping it replaces the underlined word.  Often, I found, the correct choice is not there and then you can either use the VR again or simply type in the correct word.

The system (I think) ‘expects’ a North American accent but if you enunciate your words things should work out.

And I can confirm Apple’s claim that the system improves over time and makes fewer and fewer mistakes.

Some tips

This is not yet a perfect technology and it may take some getting used to. Here are some tips and things to be prepared for if you are going to give it a try:

Different neurons

The most interesting thing about using VR is that the process of getting your thoughts on to (virtual) paper is very different to when your thoughts go directly to your fingers as you type.

There seems to be a different brain process going on when writing/typing versus speaking your thoughts—which then appear as sentences on-screen.

If you know of any research to back this up please let me know.

Speak punctuation

The most awkward thing about VR is that you have to speak clearly as well as speak punctuation marks out loud. This takes some getting used to and for me was where the technology had the most problems e.g with open and closed quotation marks.

Pay attention to the feedback

Make sure that you see the little mic icon respond to your voice.

IOS 7 now has a wavy line when Siri is capturing your voice. For PCs I’m sure there is some visual feedback to let you know that it’s capturing and converting your words to text.  Nothing so frustrating as getting off a few great sentences only to find that it wasn’t captured.siri wave

 Only three or four sentences at a time

The technology works by sending your voice into the Cloud where it’s converted and then sent back. This requires memory, bandwidth and processing power, so don’t give it too much speech before you have it converted to text, otherwise you could lose everything you spoke.

In practice (for me) this means no more than four average length sentences at a stretch before I hit the little mic icon on my iPad to check that it really has captured my voice.

 A quiet environment

Not so much because it doesn’t work—the VR still works well with moderate background noise, just not as well as if you are in a quiet environment—but more to not disturb people next to you, or to not feel like a complete nerd as you dictate your thoughts into your device.

The Future of Voice Recognition

My bet is that within the next decade, voice recognition to issue commands or ask questions will become as pervasive it was on the Star Trek TV show and movies.

But VR is dictation ready now

Voice recognition for dictation is functional enough to help people who

  • Write a lot of documents
  • Who are slow typists
  • Who find themselves using mobile devices with small keyboards.
  • Work in fairly quiet environments.

And I definitely check all the above.  It is very satisfying to dictate a short paragraph to your mobile device when you’re in the flow of a thought and have it be captured automatically by the VR.

VR is allowing me to rely more on my iPad as a production device as I don’t need to use the keypad as much.

If you write a lot content, especially on a mobile device like an iPad or iPhone you will want to try voice recognition to dictate your thoughts, emails, documents and articles.

Any thoughts? Contributions/acknowledgments welcome.