A voice controlled commands for a computer
This project began when I found a library in python capable of turning text into speech. Naturally, I was a bit skeptical of the idea at first, but after testing it out myself, I found it to be very precise and wondered about its applications. I considered a few ideas, but then realized that with my current knowledge of python from previous projects had a lot to do with opening and running files and gathering data. Thus, I decided to see if I could make a voice operated computer assistant like Siri or Alexa.
The implementation of the text to speech aspect of the project was very well documented by the user and went very straight forward with little to no trouble. I decided to implement features to do most of the redundant tasks on my computer. I made voice shortcuts for all of the bookmarks in Chrome and frequently visited programs like Eclipse and Pycharm and general logging on and off. I then used my knowledge of web scraping to create features that could web scrape news headlines and the forecast for the week.
However, despite being very useful for me, I found the application as it stood to be rather lifeless; it had no way of interacting back with the user like Siri, Alexa, or even Cortana. So, I then began researching means of playing back sounds to give my AI a voice. I found a number of options, and a text to voice program online (http://www.oddcast.com/home/demos/tts/tts_example.php ) to use as a voice. After downloading Audacity and learning the basics of sound editing, I had compiled enough responses to automate simple conversations between users.
I even began adding entirely voice based commands like quotes for encouragement or simple conversation and even made the program respond to the name “Margot”. I did for a while, attempt to find a way to sample Morgan Freeman’s voice, because who wouldn’t want Morgan Freeman to open their emails, but found that my current skills in sound editing do not allow for that to happen. However, if I do find time, I intend on watching the Dark Knight and using my minimal skills with Audacity to build a database of Morgan Freeman’s voice as Batman’s butler. I would also like to add in more features and more complex conversations and user interaction that is not completely tailored to my own needs. Currently, the project is opened on my computer and save me the trouble of having to type and press buttons, and occasionally gives me someone to talk to when I am bored.