Saturday, July 2, 2011

Yaaaay! Voice text messages! Voice *everything*

2 of my favorite iPhone features are voice dialing and Google's voice search.  Android has taken voice recognition to a new level, although the voice dialing has a nearly fatal flaw.



For 2 years, every day I left the office I grabbed my iPhone, and without looking held down the home button and said "Call Jessilem."  Please wipe that silly look off your face, you know you have stranger pet names for your wife than that.  Anyway, my iPhone would dutifully reply "Calling Jessalam, mobile" which was close enough.  Now, once every six months it would say "Calling Justin, mobile" at which point I would have an awkward conversation with an acquaintance trying to pretend I was thinking of him and just wanted to see what was up. :-)

In the wonderful world of Android, every day I leave the office and I grab my Android phone and click the unlock power button, wait a moment, unlock it, click home, wait a moment, touch the voice dialing icon, and say "Call Jessilem on cell."  Then I wait for a moment and look down and realize it takes about a second or two before it actually starts recording.  My Android phone then dutifully displays "Call Dad on cell."  I cancel that and repeat "Call Jessilem on cell." It displays the message, then I click okay, then it dials.

The Android developers got the easy part right: the voice recognition algorithm is just fine.  But they got the whole user interface wrong.  Lets see what they missed.

  • Hands-free
First of all, voice dialing  is a hands-free operation.  Apple got this by triggering voice dialing by holding in the one and only button on the phone.  Picture this: explosions.  aliens.  phaser fire.  Commander Riker grabs his communicator, removes it, presses a button on the back, slides his finger along the symbol of the United Federation of Planets logo, then says "Riker to Captain Picard!  Two to beam up, NOW!"  His communicator then displays a message asking him to confirm it.  Before he can click the button, his head is melted by the Borg he didn't see behind him.
  • Audio queues
On both phones, there is a brief delay before the voice recognition starts.  The iPhone plays a little "doo-doot" sound when it is ready for you to speak.  The droid does not.  If you watch the display it says  "Starting Up" then "Listening" before it starts to record.  The iPhone also gives you audible confirmation that it recognized the command and it audibly tells you what it is about to do.  No looking at the screen, no timing, no guesswork.
  • No confirmation required

The iPhone just makes the call.  There is plenty of time to grab your phone and cancel it.  Audio cancellation would be even better though.  And if the iPhone is unsure it asks for clarification.  "Call Jessilem: home, or mobile?"  (Most voice recognition algorithms know how close they were to a match).  The Android phone requires confirmation even if it got it 100% right, and the confirmation require you to look at the phone.  Of course, you already had to look at the phone to even use the feature.


I assume there are 3rd-party voice recognition apps out there.  I home some of them monitor the power button and wake without requiring screen interaction.  I can imagine other compromises: perhaps another puzzle piece on the unlock screen, one that indicates a desire to make a voice call.  I haven't looked yet.  Does anyone have suggestions for a better app?

FUTURE: I'll talk about how Android voice recognition rocks in another post.

1 comment:

  1. I tell ya, years before iPhone I had MS VoiceCommand on WinMo and it worked like a total charm... you clicked your BT headset and said "Voice Command" and you heard a beep, you said "Call Sammie" and you were done. IF Sammie had more than one phone option a synth voice said "Call Sammie on Mobile?" and you said Yes it went throuh, if you said no it asked "Call Sammie on Home?" and so on. If you only had one phone for a contact then you got the confirm and that was it.

    PLUS you could say "What time is it" or "what is the date?" and the synth voice would just tell you.

    This was free for owners in 2006. I got my Evo 4g since MS trashed the working WinMo for the iZune interface and at first there was no BT option then when it came it was (and still is) like going back to copper wires.

    VLingo is even worse.

    I don't have to unlock the phone as you describe, instead the phone gets unlocked in my pocket whe I dial out. The call goes through but afterwards the phone is still unlocked so I end up butt-dialing. So to use handsfree I have to always be holding the phone. Useless in CA where we are not allowed to hold a phone while in the car.

    Hate to say it but fact is fact... MS WinMo 6+ had this crap (and perfect text handswriring recog) down cold years ago and now we pay more for heavier phones that simply dont get the basics of a TELEPHONE right.

    Ahhhh.. I feel so much better now

    ReplyDelete