iPhone 5: The hurdles to overcome for voice control

iPhone 5: The hurdles to overcome for voice control

Summary: The next iPhone is expected to make voice control a big part of the interface, but there are some big hurdles that must be overcome to get owners to use it.

SHARE:

The next iPhone will be unveiled this week by Apple, and as is usually the case information has trickled out about what we can expect from Cupertino. One of the expected new features of the iPhone 5 is Assistant, a spoken interface that lets users tell the phone what to do by speaking into the phone.

Assistant is reported to be the result of Apple's purchase of Siri, and early reports say it is thoroughly integrated into the iPhone system. The fact is this is nothing new, and Apple may find it as hard to get iPhone owners to adopt the voice control as others have before them.

Voice control, or speech recognition, has been around for a very long time. I have been using it since the premier solution came from IBM, Via Voice. The technology is quite complex, as voice control systems must take any spoken phrase, parse it, and then correctly convert it into text a system can understand. Since no two people speak the same phrases the same way, it is difficult to implement properly on systems with limited processors and memory.

Speech recognition is already used quite a bit in the real world. If you have ever called a company and spoken your menu choices to get sent where you needed to be you have experienced speech reco first-hand. Owners of Android phones/ tablets have enjoyed voice control for a while, with Google search having a very good speech interface.

Many Android phone owners are not aware that Google released Voice Actions a year ago that does a lot of things that the next iPhone's Assistant is supposed to do. This voice control has been integrated into Voice Search, and make it possible to send text messages, and emails by speaking them into the phone. It is integrated into Google Navigation, and the simple phrase "navigate to Reliant Stadium Houston" is all you need to speak to fire up the proper navigation action.

Demo of Voice Actions by Google

Apple will no doubt make Assistant very natural for the user, as that is a trademark of its implementation of a technology like this. I have no doubt we will see ads showing how easy it is to use voice to make the iPhone do things, and user awareness will not be a problem for Apple like it is for other platforms. Even with good user awareness, Apple will have to overcome significant hurdles to get widespread adoption of its speech recognition system.

One technical challenge that such systems must overcome is the handling of background noise. People tend to be in noisy environments when using a gadget like a phone, and this makes it difficult for the recognition system to accurately determine what the speaker is saying. It is essential to isolate and remove background noise from the spoken commands, especially when that includes other folks speaking in the same room as commonly occurs. Background noise must be completely removed from the sessions for the system to have any shot at accurately interpreting what the user speaks into the phone. This is a lot harder than it sounds.

Another challenge that Apple faces with its Assistant feature on the new iPhone is social in nature. Users have been slow to adopt speech reco in the past because they are self-conscious when using them. Many folks are embarrassed to speak commands into a phone. This is human nature, as we as individuals don't like to give the appearance that we are doing things out of the norm, and telling your phone what to do falls in that category. Apple's implementation of speech reco will have to be done in such a way that doesn't make the user feel self-conscious when using it.

If Apple can deal with these two hurdles, it has a shot at making speech recognition a part of the everyday use for iPhone owners. As stated, this is not new technology by any means but if Apple can get iPhone owners to use it regularly it will be a big move forward in mobile.

Related:

Topics: Mobility, Hardware, iPhone, Smartphones, Telcos

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

27 comments
Log in or register to join the discussion
  • Both Google and Apple bought AI firms to merge it with voice recognitios

    Lets see how smart, or not, Apple's version will be comparing to Google's.
    DDERSSS
    • RE: iPhone 5: The hurdles to overcome for voice control

      @DeRSSS

      Or Microsoft's. On WP7.5 have full voice control over the phone and it does rather well. They do this as well through a company they had acquired years back for speech recognition. Wish them luck!
      audidiablo
      • RE: iPhone 5: The hurdles to overcome for voice control

        @audidiablo

        Yep, Apple and Android could learn a thing or two.
        tonymcs@...
  • The real win is for the disabled. Jury is out otherwise.

    The vast majority of consumers will likely never use it, but for some disabled individuals, this will be a huge plus.

    One reason I think it might fail with the majority of consumers is that editing text (to correct the inevitable recognition mistakes) is rather cumbersome on the iPhone. Even on the iPad's larger screen, it is no picnic. Select, copy, and paste are obviously (and literally) afterthoughts which were tacked onto the iOS operating system and using them is not as easy or intuitive as the rest of the OS functions.

    Even in my own case, I've owned voice recognition applications since Via Voice first appeared. Most recently, it was Dragon's software. Each time I bought voice recognition software, I intended to train it and use it extensively. Each time, I ended up abandoning the effort because it was actually no faster than typing for me. The only reason system-wide voice recognition *might* succeed on the iOS devices is because text entry is normally much slower than on a real keyboard. So, if it works with very few mistakes, it might actually be quicker than typing.
    BillDem
    • RE: iPhone 5: The hurdles to overcome for voice control

      @BillDem
      I believe Apple's speech recognition tech will be a success because I have been using "dragon search", an iPhone app on my iPad for over a year and the results have been excellent. I have never had a need to edit a result yet.

      Bt the way, I was an early adopter of Via Voice installed on second generation pentium class computer. Unfortunately, my results were less than successful using that tech.
      kenosha77a
    • Might also be popular for in-car hands free usage

      @BillDem
      The voice support in iOS 4 is so poor and so inaccurate that I never bother to use it any more. However, I really tried to use it because it is a good feature if it worked properly. Hopefully with iOS 5 this is a feature that I'll actually be able to successfully use. I will be annoyed though if iPhone 5 ends up being a minor hardware upgrade from iPhone 4 but that this voice control will be exclusive to iPhone 5. I would hate to think that Apple would try to get people to pay hundreds of dollars for a feature that has simply been disabled for older devices.
      toddybottom
  • Ahh... Via Voice

    For it's time it was a pretty awesome piece of software. Even now, most voice recognition does horribly with even a little background noise. Will everyone be sitting on the train, shouting at their phones? I can tell you how frustrating it is to work through those voice menus sometimes. Unless there's a major step forward in noise filtering, I don't see it catching on. MS Kinect has multiple mics that could triangulate where the sound is coming from. You might be able to filter that way?
    crazydanr@...
    • RE: iPhone 5: The hurdles to overcome for voice control

      @crazydanr@... Some phones and tablets already use array microphones to help with the noise. These work pretty well, but in my experience still have trouble when others are speaking in the background.
      JamesKendrick
      • RE: iPhone 5: The hurdles to overcome for voice control

        @JamesKendrick

        My WP7.5 does the same... It does okay with little to moderate noise but in a loud coffee shop or otherwise it starts to have issues with accuracy.

        Regardless it does much better than expected and is a nifty feature.
        audidiablo
  • It is a rare event when Apple releases a major but flakey software item

    FaceTime worked as advertised. I have no reason to doubt Assistant will not be equally successful.
    kenosha77a
    • I think it will probably work well

      @kenosha7777
      Since this technology was purchased by Apple from someone else, I have more confidence that it will work well. Apple's in-house efforts are usually pretty bad. Anyone remember Ping? No? I don't blame you.

      I believe that multi-touch works so well on iOS devices because Apple purchased Fingerworks and did not attempt to build it in-house. Apple developers just aren't very good.
      toddybottom
      • RE: iPhone 5: The hurdles to overcome for voice control

        @toddybottom
        I never saw the need for Ping .. As did many others. The world decided that a second "FaceBook" was not needed. However, Ping did work as advertised, after it's initial launch.

        Are you sure you don't give Apple employees enough credit for system integration efforts. Just because a core technology is provided by an outside agency, the efforts to fully intergrate that technology successfully into their products cannot be discounted.
        kenosha77a
      • FingerWorks was a raw technology and had nothing to do with mobile use, so

        @toddybottom: ... Apple's developers did very good job downscaling it for mobile use. The same is with other technologies.
        DDERSSS
        • Not only 'integrate', but also upgrade the technology

          @kenosha7777: Apple does both things simultaneously.
          DDERSSS
      • RE: iPhone 5: The hurdles to overcome for voice control

        @toddybottom Ping sucks - no doubt about it but wasn't iTunes in house? iOS in house? Look at how those tuned out.
        athynz
  • RE: iPhone 5: The hurdles to overcome for voice control

    Lets see how fast the Apple'aholics re-write history and claim Apple created voice recognition software. My Windows Phone with Mango has full voice control, including speaking my text messages, the implementation is amazing, and in my experience it just works. Since upgrading last week I have not written a single text if I could avoid it.
    clcrockett
    • RE: iPhone 5: The hurdles to overcome for voice control

      @ccrockett@... <br>Not sure where your paranoia is coming from. Apple users usually say the thing they like about apple products is that they implement tech well or to use Steve Job's catch phrase "it just works". I always enjoy watching MS fans use that phrase. I mean do you have to be as unoriginal as the company you support?
      anono
      • RE: iPhone 5: The hurdles to overcome for voice control

        @anono

        Since jobs kicked the bucket i have grown tired of the amount of times that i have seen ill informed apple fans claiming the apple pretty much invented everything
        imaginarynumber
  • Just putting this out there, as i have no idea what will be presented..

    but Apple is not supposed to be presenting speech recognition/speech to text tomorrow.. according to the Siri founder.. and this is two years ago.. Siri is an implementation of real AI.. the existing Siri app is hamstrung in what it can do because of the former limits of existing iPhone hardware, but with an A5 and 1GB of RAM.. natural language and real back and forth conversation with the software will be possible.. and this is why this feature will not run on iPhone 4.. the hardware just not up to the task... so at least in theory.. what we are SUPPOSED to see tomorrow has NOT been seen on a mobile device before.. what we see tomorrow is taunt amount to the Enterprise Computer version 1.0...
    doctorSpoc
  • RE: iPhone 5: The hurdles to overcome for voice control

    Eh, my biggest issue is my voice isn't that great for voice apps. All apps I've encountered so far are pretty miserable.
    CobraA1