Saturday, January 30, 2010

Voice Apps

Text to speech converting applications have been around for sometime now. But speech to text is not so easy. In fact I worked on a small project during my study at ISI on speech recognition and I can safely say it's not simple/easy nor it's fool proof due to many differences in human speech from pronunciation to accent to dialect to ..what not!

Little off track ...My hypothesis of why iPhone became a huge hit as it is now apart from the obvious reason of Apple being a huge brand and it markets its products very well is..
Overtime mobile became much more than just a phone and the size was a constraint under which every developer had to operate due to ease/comfort of carrying. And people started using mobiles as camera/video players so the need for big screen increased and the key pad was eating space in a mobile surface area. I think that's the reason why touch interface became a huge hit and so has iPhone. I believe still there is one problem that remains no matter how many keys you have on a mobile it will never be same as full fledged typical computer keyboard in terms of ease to use.

I recently came across Google Voice based search application. I have been using it for few days now. Though it's not yet perfect in terms of being able to capture exactly what I say but still it does a decent job and to me most importantly it's moving me away from using the tiny key pads on my mobile!

Though it might not be sensible to enable all actions in a mobile to be voice based but it does help sometimes to have voice based commands/operations. Even you can dictate and it can type an SMS. Though there are software which you can buy to convert speech to text(I think Microsoft word has this feature provided you buy speech recognition software/add-in, I think) The beauty is voice apps make more sense in case of mobile as they generally have good built in audio capturing mechanism and with google providing this service for free for search it's fun.

Overtime more and more voice apps would be part of mobiles because there is lot of opportunity in that market.Currently lot of people in rural areas(in India and probably elsewhere) do not know how to speak/write English. Though there are local language text based mobile but still there is a huge population who cannot read/write in their local language and hence do not/cannot use all applications in a mobile(even a SMS). To start with one may send an SMS(not MMS as that could be simple voice recording but costly) based on their local language to start with. Later on these voice apps can help them use many more applications.

In fact you can probably even use speech recognizers to create voice based games or commands in games.

Of course voice based applications can not only help the regular users but also blind people.

Again please feel free to comment below.

1 comment:

Hosted Voice System said...

Hi Partha,

That is a nice post, if you are interested in making a voice app, do drop a line to

Simmortel Voice is leading a new revolution in providing hosted telephony platform for making mobile voice applications.

Abhishek Singh,
Simmortel Voice.