{"id":2704,"date":"2017-02-18T05:34:43","date_gmt":"2017-02-18T05:34:43","guid":{"rendered":"http:\/\/truelogic.org\/wordpress\/?p=2704"},"modified":"2017-02-18T05:44:13","modified_gmt":"2017-02-18T05:44:13","slug":"a-speech-translation-mashup-in-less-than-5-hours","status":"publish","type":"post","link":"https:\/\/truelogic.org\/wordpress\/2017\/02\/18\/a-speech-translation-mashup-in-less-than-5-hours\/","title":{"rendered":"A Speech Translation mashup in less than 5 hours"},"content":{"rendered":"

The Web Speech API is a little known framework, which has been developed by Google a couple of years back. It is a very good and accurate javascript based API which is able to take in speech from an audio source like a microphone and convert it into text on the fly. To see it in action, you can check my Youtube video here: https:\/\/www.youtube.com\/watch?v=WNnNJSXHe6s&t=11s<\/p>\n

Recently I did an experiment to take the usage of the Web Speech api further in the form of a single web page which can take in speech input in one language and give a translated speech output in another language.<\/p>\n

The concept is very simple. The processing pipeline is as follows:<\/p>\n

    \n
  1. Take in speech input in a source language and convert it into text<\/li>\n
  2. Take the text and feed it into a translation API which will convert the source text into a target text in a different language<\/li>\n
  3. Take the translated target text and feed it into a Text to speech engine to generate speech in the target language<\/li>\n<\/ol>\n

     <\/p>\n

    Unfortunately Google Chrome is only browser which fully support Web Speech, so this cannot run in any other browser as of the moment.<\/p>\n

     <\/p>\n

    I am using the following libraries\/services:<\/p>\n