OK Google voice commands and dictation

Voice commands and dictation have been greatly popularised by Apple’s Siri and are something which many who are averse to trying to type on a small touch screen have been keenly wanting. I’ve previously resisted using voice search, but Android has made some strides in this area recently so I’ve combined having a look at what you can and can’t do with the real world example of bacon in order to provide my impressions.

”OK Google”.

One of the biggest new features is the ability to start Google listening by saying ”OK Google” from anywhere and then being able to say a command or query. This mostly relies on new power saving hardware so currently only works on the new Moto X.

The best option I’ve found is placing a shortcut to ”Voice Search” either on the desktop, or in the notification area (using an app like Notification Toggle). Then you have relatively easy access to Google search which starts listening as soon as you go into it, and once you have given a command, you can continue to give additional commands by saying ”OK Google”.

One of the coolest variations I’ve found on this is that there are several other commands you can try. Several sites mention ”OK Jarvis”, which I couldn’t get to work, but ”OK Dude” did. Sadly ”Ok Bacon” doesn’t yet work, but I’m sure Google are working on this.

Accessibility

Firstly, one of the biggest problems I found was that on my Galaxy Note II, I couldn’t get OK Google to work well with Talkback. ”OK Google” keyword detection was disabled while Talkback was running, and when I did do a search by tapping on the button, I couldn’t get Talkback to interact with the web view of results. On my Nexus 10 with the newer version of Android, I could interact with the web view at least, though in fairness still not quite as smoothly as Siri with VoiceOver on the iPad.

What can you do with it?

So now we’ve got into Google search one way or another, what can we ask it? There are an ever increasing number of commands and variations you can use: Here is one list of OK Google commands.   See that link for detailed instructions but broadly they boil down to:

Searching for information about bacon on Google.
Setting events and reminders to buy bacon
Finding out general information (time, weather, sports scores, flight information, stock price of bacon etc)
Doing maths and conversions (length / currency etc)
Using the phone features (make a phone call or send SMS to order bacon, open an app etc)
Interacting with some apps (send an email, create a note, open a bacon related web page etc)
Maps and navigation instructions

Natural language and context

One of the big strengths is the variation in ways you can phrase your query, for instance:
”What is the weather”, ”Do I need an umbrella / a coat?”, ”Is it sunny / raining?” will all give you the weather (with comments on whether it’s raining or you need an umbrella depending on what you asked).

With many queries though, how you phrase things can make a big difference. For instance, you can say ”Search for bacon” or ”bacon recipes” and it will give you search results. You can also say ”Define bacon” or ”what is bacon?” and it will read out a definition, ask it to ”cook bacon” and it will read out step by step instructions. Interestingly saying ”Cook bacon” gives you instructions from The Art of Manliness, whereas ”Cooking bacon” gives instructions from The Kitchn

One very intuitive feature is that it will use previous questions to work out the context of the current question:

”When did world war 1 start”, reads an answer explaining the precursors to the Great War.
Next asking ”When did it end” gets reworded to ”When did world war 1 end” and gives the appropriate answer.
Asking ”How many were killed” will then also give the answer relating to WW1.

Interacting with your phone

When I asked it to ring ”home”, Google asked me ”Which home?” as I had a contact called ”Home” and one called ”Home visiting doctor”. Very cleverly she listened immediately and when I said ”The first one” she then went to dial my house. ”Take Photo” or ”Record video” opens your default camera app smoothly as well.

When I tried emailing my wife, it dictated my (short) message well, though when I told it to send, instead of sending the message immediately, it opened Aqua Mail with the message ready for me to edit further or send myself.

When I tried to open an app by saying ”Open aqua mail”, it actually crashed Google search – evidently the 700 bacon related apps I have on my phone are too many for it…. however it worked ok on my Nexus 10 with far fewer apps.

I also had trouble creating notes. I have several notepad applications on my phone, however it couldn’t find anything to save a note with. When I tried on my tablet, it happily saved my notes I dictated into Google Keep (which I didn’t have on the phone). One way around this, is to go into your preferred notepad app, and when you create a note and the keyboard appears, you can tap the microphone icon (or set ”Google Voice typing” as your default keyboard) and then you can dictate the note.

Limitations

One of the main things I found OK Google can’t do, which many people would really find useful, is change settings, eg ”Turn on Talkback”, ”increase brightness”, ”magnify” would all be extremely useful.

The voice recognition was able to work out what I was saying most of the time, though not always and sometimes when it got it wrong, that would then flow on to following commands. There is an option to have it analyse everything you ask to facilitate better accuracy so it should learn over time. I also found it cut me off partway through long commands – eg dictating a longer email. I also sometimes said ”OK Google” and then had to try and think how to word what I was about to ask, though because it will accept quite a lot of ’natural’ language, sometimes just saying what you want without thinking of how to word it works really well. If you do pause first it often stops listening just as you start speaking.

Easter Eggs

As with Siri, there are a number of funny questions and commands you can give OK Google, including several good bacon ones:

• “When does the narwhal bacon?”
• “What is the Bacon number of [random actor]?”
• “Make me a sandwich!”
• “Sudo make me a sandwich!”
(I am assuming these last two are BLTs – what else would they be?)

Conclusion

Overall I found it actually worked better than I expected, and being able to say ”OK Google” feels much smoother for issuing subsequent commands than having to tap on the microphone icon like with Siri. I did find it jarring that it will sometimes read aloud what it has found, and other times it doesn’t and I need to concentrate and look at what is on the screen.

While I am still reluctant to randomly talk to my phone in public, I have the icon prominently on my home screen and will keep coming back to see how it advances and I can absolutely see it making accessing the device easier for some people, particularly once Talkback integration is improved.

Is it better than Siri? At this point in time, some things it does better, some not quite as well, particularly if you using Talkback, I would recommend using a device running Android 4.4 for best results. Overall though I think the rivalry between the two companies is good for all of us, as if one introduces a brilliant new feature, the other will find a way to do something similar. As much as fans of one platform or the other will cry ”copycat!”, if the end result is everyone being able to do something more efficiently, then it is the users who win from this.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s