Why do we not see more games and apps accepting speech response?

Joseph Videtto December 8, 2012

Why do we not see more games and apps accepting speech response?
– Are they too complicated to build?
– Are there not reliable open libraries?
– Are the available reliable libraries too expensive or unavailable (e.g. Dragon’s?)
Are there any apps that do accept speech response, besides the few everyone knows about (Siri and Dragon Naturally)
– If one wanted to tinker building such apps, which speech-to-text librarie(s) would be the best options – and what are the pros and cons of using each?

  1. James Bruce
    December 9, 2012 at 11:18 am

    My university dissertation 10 years ago was a speech controlled role play adventure game, so perhaps I could answer here. The technology works, reliably I'd say provided you write a good enough grammar and the vocabulary is limited to a hundred or so words; and there are indeed open libraries you can incorporate - the one I used was from IBM for Java, I believe, though I'm sure there are more nowadays.

    The main problem is that speech control just isn't effective. It's a gimmick, at best, but no real world benefit over a mouse and keyboard. That said, it has found a niche in things like Kinect games, where you don't have a controller to hand. For general computing though, there just isn't a benefit to using it.

    • Joseph Videtto
      December 9, 2012 at 11:51 am

      I would tend to agree with you're first point about the reliability, assuming you limit the grammar and vocabulary. I'd like to give a try to writing my own, and I'm hoping some makeuseof users that have tried out a few of the current speech to text libraries can share based on their experience which are the best to choose today.

      Also - I can imagine many scenarios where it is effective - especially when you're not a fast typer and have a long (e.g. pages) piece to be written, or when you're a worker that's often not in front of a desk, but you have to take notes on your work (say doctors, nurses, teachers). Of course, there is the issue of background noise - I'm not sure this problem can be easily addressed now or any time soon.

      • Douglas Mutay
        December 10, 2012 at 9:48 am

        Agree with you, speech control is actually very effective and even for fast typer, if you can dictate long text you will certainly be faster.

  2. Junil Maharjan
    December 9, 2012 at 5:19 am

    Voice recognition systems are still in their infancy. and there is the issue of accents too and languages as well.

  3. Paul Pruitt
    December 9, 2012 at 12:12 am

    I think people don't like the pain of training the apps. You know Android's voice app works as well at detecting speech for say searching as does Siri. So I think it is catching on in mobile phone as there is no need for training.

    I tried doing a few voice searches on my PC with a dedicated microphone and mostly still got gibberish. Maybe I need a better microphone but the one I have though, cheap is highly rated. So on PCs and laptops without training Dragon Dictate, I guess it just doesn't work yet.

    You can say we are lazy, but a PC and laptop users are so used to putting our thoughts in writing, it is very easy for us and habitual to interface with the computer via a keyboard. It is a well worn path, so why start a new path which will take effort and may not be rewarding when the current one works pretty well.

    With the mobile phone on the other hand, we now increasingly need to interact with it but physical and touch screen virtual keyboard keys are just too small and therefore error prone, so there is a big motivation to try voice commanding the phone. It's nice also that Google and Apple seem to have done their homework right for mobile.